Metabolic Flexibility Permits Mesenchymal Stem Cell Survival in an Ischemic Environment

Abstract

The application of mesenchymal stem cells (MSCs) for myocardial repair following ischemic injury is of strong interest, but current knowledge regarding the survival and retention of differentiation potency of stem cells under ischemic conditions is limited. The present study investigated the effects of ischemia and its components (hypoxia and glucose depletion) on MSC viability and multipotency. We demonstrate that MSCs have a profoundly greater capacity to survive under conditions of ischemia compared with cardiomyocytes, measured by detecting changes in cellular morphology, caspase activity and phosphatidylserine exposure. MSCs were also resistant to exposure to hypoxia (0.5% O₂), as well as inhibition of mitochondrial respiration with 2,4-dinitrophenol for 72 hours, indicating that in the absence of oxygen, MSCs can survive using anaerobic ATP production. Glucose deprivation (glucose-free medium in combination with 2-deoxyglucose) induced rapid death of MSCs. Depletion of cellular ATP occurred at a lower rate during glucose deprivation than during ischemia, suggesting that glycolysis has specific prosurvival functions, independent of energy production in MSCs. After exposure to hypoxic or ischemic conditions, MSCs retained the ability to differentiate into chondrocytes and adipocytes and, more importantly, retained cardiomyogenic potency. These results suggest that MSCs are characterized by metabolic flexibility, which enables them to survive under conditions of ischemic stress and retain their multipotent phenotype. These results highlight the potential utility of MSCs in the treatment of ischemic disease.

Disclosure of potential conflicts of interest is found at the end of this article.

Cardiac ischemia, Hypoxia, Glucose deprivation, Mesenchymal stem cells, Apoptosis, Metabolic pathways

Introduction

Congestive heart failure (CHF) is a leading cause of morbidity and mortality in the Western world. Myocardial ischemia (MI), one of the main causes of CHF, results in injury and death of cardiomyocytes due to deprivation of oxygen and nutrients. Cardiomyocytes have a high energy requirement for their contractile function, which makes the myocardium very sensitive to MI. Ischemia induces cardiomyocyte death via necrosis in the core of the ischemic region and apoptosis in the infarct border zone (reviewed in [1–3]). Since cardiomyocytes are terminally differentiated postmitotic cells [4], their loss during MI leads to permanent damage. Thus, it is important that cardiomyocyte numbers be maintained either by strategies that prevent cell death/damage or, alternatively, by regenerating the myocardium. Some reports have suggested that cardiac progenitor cells are present in the heart and can facilitate myocardial regeneration [5, 6]. However, their contribution appears to be insufficient, as fibroblasts predominantly repopulate the ischemic zone and a noncontractile scar tissue is formed.

Mesenchymal stem cells (MSCs), a subset of nonhematopoietic stem cells of the bone marrow, are advantageous for tissue repair due to their ease of isolation, high proliferative activity, suitability for allogeneic transplantation, and differentiation potential [7, 8]. MSCs have the ability to differentiate into chondrocytes [9], adipocytes [10], osteocytes [11], and myocytes [12]. A number of studies evaluated the utility of MSCs for cardiac repair. However, there is some lack of clarity regarding the ability of MSCs to differentiate into cardiomyocytes [13, 14]. The demonstration that MSCs injected into noninfarcted murine myocardium differentiated and formed sarcomeric organization suggested that MSC therapy maybe therapeutically effective [14]. Both animal and clinical studies for the treatment of MI have reported positive results using transplanted bone marrow stem cells, with improvements in cardiac function, perfusion, and remodeling [15, 16]. However, complete differentiation of MSCs could not be demonstrated in another study where MSCs were injected into infracted myocardium [15]. It is also possible that MSCs and other cell populations from the bone marrow exert their therapeutic effect by releasing paracrine factors, which enhance cardiac recovery, rather than by direct differentiation into cardiomyocytes [13, 17].

Little is known currently about the mechanisms involved in the response of MSC to an ischemic environment. Fundamental questions remain unanswered regarding the response of MSCs to the different components of ischemia. Ischemia includes both glucose and oxygen deprivation, and these have varied effects depending on the cell type. In the present study we examined the effects of both glucose and oxygen deprivation on MSC viability in comparison with cardiomyocytes and fibroblasts. The objective of this study was to obtain a better understanding of how MSCs may influence postischemic recovery. Our results show that MSCs are more resistant to ischemia than cardiomyocytes and that exposure to ischemia does not impair their differentiation potential. Although cardiomyocytes were sensitive to both hypoxia and glucose deprivation, MSCs could survive hypoxic conditions for at least 72 hours. MSCs were capable of surviving hypoxia because of their ability to rely on glycolysis rather than mitochondrial respiration. This metabolic flexibility may provide the necessary protection in an ischemic environment and allow MSCs to function in a reparative or regenerative capacity.

Materials and Methods

Cell Isolation and Culture Conditions

All procedures involving animal material were performed in accordance with the ethical regulations of the National University of Ireland (Galway, Ireland). All reagents were from Sigma-Aldrich (St. Louis, http://www.sigmaaldrich.com) unless otherwise stated.

MSC Isolation.

MSCs were prepared as reported previously [10]. Briefly, bone marrow was flushed from femurs and tibiae from 8–12-week-old female F-344 rats (Harlan, Oxfordshire, U.K., http://www.harlan.com) with MSC growth medium (44.5% α-minimal essential medium/44.5% Ham's F-12 medium (Gibco, Grand Island, NY, http://www.invitrogen.com) containing 10% fetal bovine serum and 1% antibiotic/antimycotic solution (Gibco). The marrow mononuclear cells were resuspended, centrifuged, plated, and grown in MSC growth medium. The mesenchymal population was isolated on the basis of adherence to plastic, and the nonadherent hematopoietic cells were removed by regular medium changes. When cultures became confluent, the adherent, spindle-shaped MSCs were detached with 0.25% trypsin-EDTA, replated at 5,500 cells per cm², and passaged when they reached 90% confluence. Expression of MSC cell surface markers (CD3, CD19, CD34, CD45, CD106, CD29, CD73, CD90, and CD105) was characterized by flow cytometry and is summarized in supplemental online Table 1.

Fibroblast Isolation.

Dermal fibroblasts were isolated from F-344 neonatal rats (1–4 days old). Briefly, the rat's abdomen was sterilized, and a section of skin was removed, cut into 4-mm strips, and placed on asterisks indented into the surface of the plastic inside 60-mm dishes. Fibroblasts grew out from the strips and into the asterisks. The fibroblasts were grown in MSC growth medium. Cells were detached with 0.25% trypsin-EDTA and passaged every 7–8 days when 90% confluent.

Cardiomyocyte Isolation.

Neonatal cardiomyocytes were prepared and cultured as reported before [18]. Briefly, 1–4-day-old F-344 rats were euthanized and the hearts were excised, homogenized and subjected to overnight trypsin (Langanbach Services, Wicklow, Ireland) digestion at 4°C. Trypsin inhibitor and collagenase (Langanbach Services, Wicklow, Ireland, http://www.langanbach.ie) were added to digest the extracellular matrix prior to differential centrifugation through a discontinuous Percoll gradient. The middle layer of cells was collected, resuspended, and plated at a density of 1 × 10⁵ cells per milliliter in Dulbecco's modified Eagle's medium (DMEM)/Ham's F-12 medium supplemented with 10% newborn fetal calf serum, 100 μM 5-bromo-2-deoxyuridine, 1% insulin transferrin sodium selenite (ITS) liquid supplement medium, 1 mM sodium pyruvate (Gibco), and 1% antibiotic/antimycotic on 0.2% gelatin-precoated flasks (Corning, Lowell, NY, http://www.corning.com). The purity of the preparations was checked by cardiac α-actinin immunostaining and found to be between 85% and 95%. Medium was changed every other day and cells were used between 7 and 14 days of isolation. All cells were cultured in a humidified atmosphere with 5% CO₂ and 95% air at 37°C.

Treatments

To mimic hypoxia, cells were cultured in a hypoxia chamber (Ruskinn Technologies, Leeds, U.K., http://www.ruskinn.com) under 0.5%:94.5%:5.0% O₂:N₂:CO₂ in normal growth medium. Ischemia was induced by applying hypoxia in glucose- and serum-free DMEM (Gibco). To inhibit glycolysis, cells were exposed to 1 mM 2-deoxyglucose (2DG) in glucose-free DMEM supplemented with 1% antibiotic/antimycotic solution for MSCs and fibroblasts, and 1 mM sodium pyruvate, 1% ITS for cardiomyocytes under normoxic conditions. To inhibit oxidative phosphorylation, 0.5, 1, and 2 μM concentrations of the mitochondrial uncoupler 2,4-dinitrophenol (DNP) were added to normal growth medium [19]. As a positive control for apoptosis, cells were treated with 4 μg/ml of Fas ligand (FasL) in combination with 5 μg/ml cyclohexamide (CHX) for 24 hours; for caspase-3-like activity, cells were treated with 500 nM staurosporine (STS) for 4 hours (cardiomyocytes) or 12 hours (MSCs and fibroblasts).

Differentiation Assays

Adipogenesis was induced as described previously [10]. Wells were stained with oil red O to detect lipid vacuoles. Images were acquired using an Olympus BX51 microscope (Olympus, Tokyo, http://www.olympus-global.com) at a final magnification of ×400. Oil red O was then extracted with isopropanol and quantified photometrically using a Wallac 1420 multilabel counter (PerkinElmer Life and Analytical Sciences, Boston, http://www.perkinelmer.com) at 490–520 nm.

Chondrogenesis was induced as described previously [10]. At day 21, pellets were digested with 200 μl of papain (1 μg/ml in 50 mM sodium phosphate, pH 6.5, 2 mM N-acetyl cysteine, 2 mM EDTA) overnight at 65°C. Glycosaminoglycan (GAG) content was measured with 1,9-dimethylmethylene blue (DMMB; 35) by spectrophotometry at 595 nm using chondroitin sulfate as standard. DNA quantitation was carried out using the PicoGreen double standard DNA quantitation kit (Molecular Probes, Eugene, OR, http://probes.invitrogen.com) with phage λ DNA as standard. Chondrogenic differentiation was expressed as μg of GAG per μg of DNA.

To induce cardiomyogenesis, MSCs were seeded at 4 × 10⁴ cells per milliliter and exposed to 10 μM 5-azacytidine for 24 hours. 5-Azacytidine-containing medium was then removed, washed off with phosphate-buffered saline, replaced with MSC growth medium, and the cells were allowed to grow for 14 days, when they were harvested and total RNA was isolated. Induction of cardiomyogenic genes Nkx2.5, α-myosin heavy chain, and cardiac troponin I were detected by reverse transcription-polymerase chain reaction (RT-PCR).

Caspase Enzyme Activity Assay

The ability of caspases to cleave the tetrapeptide aspartate-glutamate-valine-aspartate-aminomethylcoumarine (DEVD-AMC) as a measure of apoptosis was assayed as previously described [20]. Enzyme activity was measured kinetically with a Wallac 1420 multilabel counter using 355 nm excitation and 460 nm emission wavelengths and was expressed as nmol of AMC liberated by 1 mg of total cellular protein per minute.

Cytological Stainings

Annexin-V Binding Assay.

Externalization of phosphatidylserine (PS) on the outer leaflet of the plasma membrane of the apoptotic cells was detected by Annexin-V as previously described [21]. Propidium iodide (PI) was added to label necrotic and late apoptotic cells with permeabilized membranes, and the samples were analyzed immediately by flow cytometry (FACSCalibur; BD Biosciences, San Jose, CA, http://www.bdbiosciences.com).

Hematoxylin-Eosin Staining.

Cells were seeded onto 35-mm dishes at a density of 5 × 10⁴ cells per dish. Cells were fixed in 3.7% formaldehyde, the cytosol was stained with eosin Y for 4 minutes and rinsed in tap water, and the nucleus was stained in Harris hematoxylin solution for 5 minutes. Images were acquired using an Olympus BX51 microscope at a final magnification of ×100.

Measurement of Intracellular ATP Concentration

After treatment, cells were harvested by scraping followed by centrifugation, lysed in 50 μl of 0.3% trichloroacetic acid and snap-frozen in liquid nitrogen. For measurement, the lysates were thawed and centrifuged and the pH was neutralized with 250 mM Tris-acetate. ATP concentration of the lysate was then measured with the Enliten ATP detection kit according to the manufacturer's instructions (Promega, Madison, WI, http://www.promega.com) using a Wallac 1420 multilabel counter. The ATP concentration was expressed as nmol of ATP per mg of total protein.

Measurement of Mitochondrial Inner Membrane Potential (ΔΨm)

Membrane depolarization was measured with 100 nM tetramethylrhodamine ethyl ester (Molecular Probes) as described before [22]. Carbonyl cyanide m-chlorophenylhydrazone (CCCP), a proton ionophore, was used a positive control for mitochondrial uncoupling.

RT-PCR

RNA was extracted using the RNeasy Mini kit (Qiagen, Hilden, Germany, http://www1.qiagen.com). Reverse transcription was carried out using 2 μg of total RNA with oligo(dT) primers (Invitrogen, Carlsbad, CA, http://www.invitrogen.com) with SuperScript II reverse transcriptase (Invitrogen). cDNAs for genes of interest were amplified during 26–32 cycles of 30 seconds of denaturing at 94°C, 30 seconds of annealing at 55°C, and 30 seconds of extension at 72°C, followed by a final chain extension for 7 minutes at 72°C with the following primers: Nkx2.5, sense: CAGAACCGCCGCTACAAG, antisense: AGTCCCCGACGCCAAAGT (product size, 325 base pairs [bp]); α-myosin heavy chain, sense: GCAGACCATCAAGGACCT, antisense: GTTGGCCTGTTCCTCCGCC (product size, 310 bp); cardiac troponin I, sense: GCGAAGCAGGAGATGGAG, antisense: TGCCACGCAGGTCATAGA (product size, 250 bp); and glyceraldehyde-3-phosphate dehydrogenase, sense: ACCACAGTCCATGCCATC, antisense: TCCACCACCCTGTTGCTG, used as an internal standard (product size, 450 bp).

Statistical Analysis

Data are presented as mean ± SEM. Differences in adipogenic and chondrogenic potency were assessed using one-way analysis of variance followed by pairwise comparisons using Tukey's post hoc test with a significance level of p < .05, using SPSS 14.0 for Windows (SPSS, Chicago, http://www.spss.com).

Results

MSCs Are More Resistant to Ischemia than Cardiomyocytes

Cardiomyocytes and fibroblasts were isolated from neonatal animals. Neonatal cardiomyocytes were chosen over adult cells as neonatal cells display a stable phenotype in vitro, do not develop calcium hypersensitivity, and beat in culture synchronously. Furthermore, the contraction profile of neonatal cardiomyocytes followed by hypoxia-reoxygenation is compatible with that of in situ hearts during ischemia-reperfusion [23]. MSCs were isolated from adult rats, as Tokalov et al. have shown that the phenotype or differentiation capacity of rat MSCs is not altered by the age of the animal [24]. The cardiomyocyte data presented in this article are derived from a number of donors, whereas in the case of MSCs and fibroblasts, all experiments were repeated at least three times on two separate donors, one male and the other female. No significant differences between the different donors were observed in any experiment.

To determine the extent to which MSCs can tolerate ischemia, cardiomyocytes, MSCs, and fibroblasts were subjected to a combination of serum, glucose, and oxygen deprivation as described in Materials and Methods. Changes in cellular morphology were examined in hematoxylin and eosin (H&E)-stained cultures. In cardiomyocyte cultures, shrunken cells with condensed nuclei, indicative of apoptotic cell death, became detectable after 12 hours of exposure to ischemia, and by 48 hours, most cells were dead (Fig. 1A). On the contrary, the first signs of cell death were observed in MSC cultures subjected to ischemia for 48 hours (Fig. 1A). Fibroblasts were resistant to ischemia in a manner similar to MSCs (Fig. 1A).

Figure 1.

MSCs are more resistant to ischemia than cardiomyocytes. Cells were subjected to hypoxia in the absence of serum and glucose for various amounts of time. (A): Images of H&E-stained cells. The arrows indicate apoptotic bodies in cardiomyocytes, MSCs, and fibroblasts (overall magnification, ×100). (B–D): Representative fluorescence-activated cell sorting quantitation of early apoptotic (Annexin-V⁺/propidium iodide [PI]⁻; black bar) and late apoptotic/necrotic (Annexin-V⁺/PI⁺; white bar) cells after exposure to different periods of ischemia. (B): Cardiomyocytes. (C): MSCs. (D): Fibroblasts. CHX was used in combination with FasL to induce apoptotic cell death as a positive control. The presented graphs are representatives of three independent experiments. (E–G): Caspase-3-like activity in ischemic cardiomyocytes (E), MSCs (F), and fibroblasts (G), measured by DEVDase assay. Enzyme activity is expressed as fold activation compared with controls. STS, an inducer of apoptotic cell death, was used as a positive control for caspase-3-like activity. Data are shown as representative samples or as mean ± SEM; n = 3. The asterisk (*) denotes a statistically significant difference (p < .05) compared with the control. Abbreviations: CHX, cyclohexamide; FasL, Fas ligand; h, hours; MSC, mesenchymal stem cell; STS, staurosporine.

To quantitate cell death induced by ischemia, the percentage of Annexin-V⁺/PI⁻ and Annexin-V⁺/PI⁺ cells was determined using flow cytometry as described in Materials and Methods (Fig. 1B–1D). Annexin-V⁺/PI⁻ cells were identified as early apoptotic cells, and Annexin-V⁺/PI⁺ cells were identified as late apoptotic/necrotic cells (early/late). Apoptotic cardiomyocytes were detectable as early as 4 hours after induction of ischemia. The percentage of dying cardiomyocytes increased gradually up to 24 hours and thereafter decreased because of disintegration of the dead cells by secondary necrosis (Fig. 1B). MSCs displayed no significant Annexin-V positivity until 48 hours of exposure to ischemia, when 16.2%/2.4% early/late apoptotic cells were detected; these percentages rose to 25.8%/3.2% at 72 hours (Fig. 1C). Fibroblasts exhibited a 16.7%/3.8% early/late cell death ratio at 4 hours of ischemia; this ratio was maintained up to 36 hours and thereafter increased to 44.0%/13.2% by 60 hours. Subsequently, the level of cell death declined, probably because of clearance of dead bodies by secondary necrosis (Fig. 1D). The classic apoptotic stimulus CHX, in combination with the apoptosis-inducing cytokine FasL, was used as a positive control to induce apoptosis. Both MSCs and fibroblasts displayed a high percentage of Annexin-V⁺ cells in response to FasL/CHX, confirming that apoptotic MSCs and fibroblasts externalize PS, and thus the lack of Annexin-V positivity is a true indicator of the absence of cell death (Fig. 1C, 1D).

In addition to the morphology, the apoptotic mode of cell death was confirmed by measuring caspase activity (Fig. 1E–1G). Caspase activity (DEVD-cleavage activity) was detectable in cardiomyocyte cultures at 4–36 hours of ischemia, with the activity slowly decreasing to basal levels at later times (coinciding with disintegration of dead cells; Fig. 1E). MSC cultures showed no caspase activity until 48 hours, when a 3.1 ± 0.1-fold increase was detected that gradually rose to 7.9 ± 1.6 at 72 hours (Fig. 1F). Similarly, fibroblasts showed no significant caspase activity until 48 hours, when caspase activity rose by 13.5 ± 1.8-fold, after which it decreased again (Fig. 1G). STS, a well-characterized caspase activator, was used as a positive control. STS treatment resulted in a 32.6 ± 1.8 and 18.2 ± 3.7-fold increase in caspase activity in MSCs and fibroblasts, respectively, confirming that caspases are activated in both cell types upon induction of apoptosis (Fig. 1F, 1G). Taken together, these data show that MSCs are dramatically resistant to ischemia compared with cardiomyocytes.

Effect of Ischemia, Inhibition of Glycolysis, and Reoxygenation on Cell Survival

Initially, cardiomyocytes respond to ischemia by increasing the rate of glycolysis to maintain ATP levels. However, the reduced coronary blood flow causes accumulation of lactate that inhibits glycolysis through negative feedback [25, 26]. To model these in vivo conditions in vitro, ischemia (glucose and oxygen deprivation as above) was induced in the presence of 2DG, an inhibitor of hexokinase.

Although addition of 2DG did not have any significant effect on ischemia-induced cardiomyocyte death, it exacerbated the effect of ischemia on MSCs (Fig. 2). H&E staining showed increased number of dying MSCs in the presence of 2DG (Fig. 2A), which was confirmed by Annexin-V/PI staining. This showed a maximum of 64.2%/5.5% early/late apoptotic cells in the presence of 2DG, compared with a maximum of 25.3%/4.3% in the absence of 2DG (Figs. 2C, 1C, respectively). Caspase activation also occurred 12 hours earlier in MSCs when 2DG was added, peaking at a much higher induction level (58.7 ± 6.4-fold increase; Fig. 2F). Fibroblasts responded to addition of 2DG in a manner similar to MSCs, displaying earlier and enhanced caspase activation and cell death (Fig. 2A, 2D, 2G). These results indicate that glycolysis is an important energy-producing pathway in MSCs and fibroblasts that contributes to prolonged survival in an ischemic environment, although MSCs still displayed significant resistance to ischemia + 2DG compared with cardiomyocytes.

Figure 2.

Inhibition of glycolysis with 2DG increases sensitivity of MSCs to ischemia. Cells were cultured with 1 mM 2DG in the absence of serum and glucose in hypoxia (ischemia + 2DG) for various amounts of time. (A): Images of H&E-stained cells. The arrows indicate apoptotic bodies in cardiomyocytes, MSCs, and fibroblasts (overall magnification, ×100). (B–D): Representative fluorescence-activated cell sorting quantitation of early apoptotic (Annexin-V⁺/propidium iodide [PI]⁻; black bar) and late apoptotic/necrotic (Annexin-V⁺/PI⁺; white bar) cells after exposure to different periods of ischemia + 2DG. (B): Cardiomyocytes. (C): MSCs. (D): Fibroblasts. CHX was used in combination with FasL to induce apoptotic cell death as a positive control. The graphs presented are representatives of three independent experiments. (E–G): Caspase-3-like activity in ischemic cardiomyocytes (E), MSCs (F), and fibroblasts (G), measured by DEVDase assay. Enzyme activity is expressed as fold activation compared with controls. STS, an inducer of apoptotic cell death, was used as a positive control for caspase-3-like activity. Data are shown as representative samples or as mean ±SEM; n = 3. The asterisk (*) denotes a statistically significant difference (p < .05) compared with the control. Abbreviations: 2DG, 2-deoxyglucose; CHX, cyclohexamide; FasL, Fas ligand; h, hours; MSC, mesenchymal stem cell; STS, staurosporine.

Reperfusion, occurring after prolonged ischemia alters the extracellular milieu in the heart in a manner that aggravates cellular damage by inducing calcium influx and production of reactive oxygen species [26, 27]. The response of MSCs to these alterations was also tested and compared with that of cardiomyocytes. Although reoxygenation enhanced cardiomyocyte death, it did not exacerbate the effect of ischemia in MSCs, as no increase in DEVDase activity or appearance of apoptotic/necrotic cell morphology could be observed (supplemental online Fig. 1).

How ischemic stress affected the functionality of MSCs was studied next. MSCs were exposed to 6–12 hours of ischemia in the presence of 2DG, after which adipogenic, chondrogenic and cardiomyogenic differentiation potency was assessed. Ischemia + 2DG treatment did not diminish the ability of MSCs to produce lipid vacuoles, as seen with oil red O staining after culturing the cells in adipogenic medium (Fig. 3A). Colorimetric quantification of oil red O uptake showed slightly reduced values in the ischemia + 2DG-treated samples compared with the control (p < .05; Fig. 3B), which was probably due to the lower cell numbers in the ischemia-treated samples.

Figure 3.

Mesenchymal stem cells (MSCs) retain their differentiation potential after exposure to Isch. MSCs were subjected to Isch + 2DG for 6–12 h. Cells were then induced to differentiate along the adipogenic, chondrogenic, and cardiomyogenic pathways under normoxic conditions. (A): Representative images of oil red O-stained MSCs cultured in adipogenic medium. Overall magnification was ×400. (B): Quantitation of oil red O uptake by photometry. The graph shows the absorbance values at 490 nm. (C): GAG production by Con and Isch + 2DG-treated MSCs after induction of chondrogenesis. The graph shows the averaged GAG concentration per 1 μg of DNA. The asterisk (*) denotes a statistically significant difference (p < .05) compared with the Con. (D): Effect of Isch + 2DG on the induction of cTnI, Nkx2.5, and αMHC in MSCs exposed to 5-Aza treatment. Induction of cardiomyogenic marker mRNAs was detected by reverse transcription-polymerase chain reaction. RNA from H9c2 rat embryonic cardiomyocytes was used as a positive control. GAPDH was used to confirm equal loading. The images presented are representatives of three independent experiments. Abbreviations: 5-Aza, 5-azacytidine; Con, control; cTnI, cardiac troponin I; 2DG, 2-deoxyglucose; GAG, glycosaminoglycan; GAPDH, glyceraldehyde-3-phosphate dehydrogenase; h, hours; Isch, ischemia; αMHC, α-myosin heavy chain.

Chondrogenic differentiation was determined by quantifying the level of sulfated GAGs secreted by MSCs cultured in chondrogenic medium using the colorimetric DMMB assay. GAG production by MSCs treated with ischemia + 2DG for 6 or 12 hours was not significantly different from that of control MSCs (p > .05; Fig. 3C). Proteoglycan secretion by MSCs was also studied in situ by toluidine blue staining, which confirmed the results of the DMMB assay (supplemental online Fig. 2).

Finally, cardiomyogenic potency induced by 5-azacytidine was not hindered by ischemia either. MSCs exposed to ischemia for 6 or 12 hours were equally capable of inducing the expression of cardiomyogenic markers Nkx2.5 and troponin I, as well as α-myosin heavy chain, detected by RT-PCR analysis in total RNA samples (Fig. 3D) [28].

MSCs Are Able to Tolerate Hypoxic Conditions but Depend on Active Glycolysis

During ischemia, several factors may contribute to cellular injury, including lack of glucose, growth/survival factors, and reduced oxygen concentration [29]. To compare the effect of oxygen deprivation on MSCs, fibroblasts, and cardiomyocytes, cells were subjected to hypoxic conditions for various times (4–72 hours). H&E-stained cardiomyocyte cultures showed cells with apoptotic morphology after 12 hours of hypoxia (Fig. 4A). At 24 hours of hypoxia, numerous apoptotic cells with dark and fragmented nuclei were present (Fig. 4A). By 48 and 72 hours, most cells had undergone secondary necrosis and had lifted off the plates. MSCs and fibroblasts, on the other hand, displayed no visible signs of cell death across the 72-hour time course examined (Fig. 4A). Analysis of PS externalization by Annexin-V confirmed the morphological data (data not shown).

Figure 4.

MSCs are resistant to hypoxia. Cells were subjected to hypoxia for various amounts of time. (A): Images of H&E-stained cells. The arrows indicate apoptotic bodies in cardiomyocyte and fibroblast cultures (overall magnification, ×100). (B–D): Caspase-3-like activity in hypoxic cardiomyocytes (B), MSCs (C), and fibroblasts (D) measured by DEVDase assay. Enzyme activity is expressed as fold activation compared with Cons. STS, an inducer of apoptotic cell death, was used as a positive control for caspase-3-like activity. Data are shown as mean ± SEM; n = 3. (E, F): MSCs were subjected to hypoxia for 48 h and then induced to differentiate into adipocytes and chondrocytes under normoxic conditions. (E): Quantitation of oil red O uptake by photometry. The graph shows the absorbance values at 490 nm. (F): GAG production by Con and hypoxia-treated MSCs after induction of chondrogenesis. The graph shows the averaged GAG concentration per 1 μg of DNA. Data are shown as representative samples or as mean ± SEM; n = 3. The asterisk (*) denotes a statistically significant difference (p < .05) compared with the Con. (G): Effect of hypoxia on the induction of cTnI, Nkx2.5, and αMHC in MSCs exposed to 5-Aza treatment. Induction of cardiomyogenic marker mRNAs was detected by reverse transcription-polymerase chain reaction. RNA from H9c2 rat embryonic cardiomyocytes was used as a positive control. GAPDH was used to confirm equal loading. The images presented are representatives of three independent experiments. Abbreviations: 5-Aza, 5-azacytidine; Con, control; cTnI, cardiac troponin I; GAG, glycosaminoglycan; GAPDH, glyceraldehyde-3-phosphate dehydrogenase; h, hours; αMHC, α-myosin heavy chain; MSC, mesenchymal stem cell; STS, staurosporine.

Cardiomyocyte cultures subjected to hypoxia displayed a 2.8 ± 0.3-fold increase in DEVDase activity, which was maintained throughout the 72-hour time course (Fig. 4B). MSCs showed no increase in DEVDase activity in response to hypoxia up to the 72 hours examined (Fig. 4C). Fibroblasts exhibited a small increase in DEVDase activity in response to hypoxia (Fig. 4D) after 24 hours (although statistically nonsignificant), which was maintained up to 72 hours, correlating with the phased, low level of annexin-V positivity (data not shown).

MSCs not only survived hypoxia but also retained their multipotency. Forty-eight hours of hypoxic treatment did not reduce the ability of MSCs to differentiate into either adipocytes, measured by oil red O staining of lipid vacuoles, or chondrocytes, measured by detecting sulfated GAGs (Fig. 4E, 4F). In fact, colorimetric quantification of oil red O uptake showed a slightly increased value in the hypoxia-treated sample compared with the control (p < .05; Fig. 4E). Finally, RT-PCR analysis showed that induction of markers of cardiomyogenic differentiation was not reduced by exposure to hypoxia for 48 hours (Fig. 4G).

To examine how glucose deprivation affects MSC survival, we examined the ability of all three cell types to survive glycolytic inhibition under normoxic conditions. To inhibit glycolysis, cells were treated with 2DG, an inhibitor of hexokinase, and exposed to glucose deprivation in serum-free medium for various periods of time. Onset of cell death was measured by cellular morphology and detection of caspase activity. H&E-stained cardiomyocytes showed a substantial amount of cell death at 12 hours of glucose deprivation, with very few live cells remaining after 24 hours (Fig. 5A). MSCs and fibroblasts were more resistant than cardiomyocytes but also displayed numerous apoptotic-like condensed cells after 24–36 hours (Fig. 5A). The apoptotic morphology induced by glucose deprivation was verified by caspase activation (Fig. 5B–5D). In cardiomyocytes, glucose deprivation induced early caspase activation detectable after 4 hours, comparable to that induced by ischemia (Fig. 5B). MSCs and fibroblasts, on the other hand, responded to glucose deprivation with much stronger caspase activation, comparable to that induced by ischemia in the presence of 2DG (Fig. 5C, 5D). These data indicate that active glycolysis is essential for MSC survival.

Figure 5.

Effect of glucose deprivation on cardiomyocyte, MSC, and fibroblast cell viability. Cells were cultured with 1 mM 2DG in the absence of serum and glucose for various times under normoxic conditions. (A): Images of H&E-stained cells. The arrows indicate apoptotic bodies in cardiomyocyte, MSC, and fibroblast cultures (overall magnification, ×100). (B–D): Caspase-3-like activity induced by glucose deprivation in cardiomyocytes (B), MSCs (C), and fibroblasts (D) measured by DEVDase assay. Enzyme activity is expressed as fold activation compared with controls. STS, an inducer of apoptotic cell death, was used as a positive control for caspase-3-like activity. Data are shown as representative samples (A) or as mean ± SEM; n = 3. The asterisk (*) denotes a statistically significant difference (p < .05) compared with the control. Abbreviations: 2DG, 2-deoxyglucose; h, hours; MSC, mesenchymal stem cell; STS, staurosporine.

MSC Survival Is Independent of Oxidative Phosphorylation

As shown above, MSCs are completely resistant to hypoxia and more resistant to ischemia than the other cell types examined. We hypothesized that MSCs appear to adapt to a low oxygen concentration (0.5%), probably by sourcing their energy requirements from glycolysis, suggesting that they can survive in the absence of oxidative phosphorylation. To test this, mitochondrial respiration was uncoupled with DNP [30]. Treatment with DNP resulted in the loss of mitochondrial transmembrane potential (ΔΨm) in both cardiomyocytes and MSCs. The loss of ΔΨm occurred as early as 2 hours after treatment in both cell types (data not shown), and the mitochondria remained depolarized up to 72 hours (Fig. 6A). CCCP or a higher concentration of DNP (up to 2 mM) induced ΔΨm loss to the same extent as 0.5 mM DNP, confirming that 0.5 mM DNP was sufficient to completely diminish the proton gradient (data not shown). Treatment with 0.5 mM DNP induced death in cardiomyocytes, reflected by the increasing percentage of Annexin-V-positive cells (Fig. 6B, 6C). At 48 and 72 hours, a PI-only positive population appeared, representing the nuclear fragments remaining from the disintegrated dead cells (Fig. 6B, 6C). On the contrary, MSCs displayed no increase in cell death after DNP treatment up to the 72 hours examined (Fig. 6B, 6C).

Figure 6.

Effect of 2,4-dinitrophenol on cardiomyocyte and MSC survival. Cells were cultured in the presence of 0.5 μM DNP for various times under normoxic conditions. (A): Loss of mitochondrial inner membrane potential (ΔΨm) measured with tetramethylrhodamine ethyl ester staining in cardiomyocytes (upper panel) and MSCs (lower panel) (B): Representative fluorescence-activated cell sorting analysis of cell death induced by 0.5 mM DNP in cardiomyocytes (upper panel) and MSCs (lower panel). (C): Quantitation of cell death induced by 0.5 mM DNP in cardiomyocytes and MSCs with Annexin-V assay. The graph shows the percentage of live, Annexin-V⁻/PI⁻ cells after increasing times of DNP (0.5 mM) treatment. Data shown are representatives of three independent experiments. Abbreviations: DNP, 2,4-dinitrophenol; h, hours; MSC, mesenchymal stem cell; PI, propidium iodide.

Glucose Deprivation Significantly Reduces the ATP Concentration in MSCs

The high sensitivity of MSCs to glucose deprivation but not to inhibition of mitochondrial ATP production suggested that MSCs source their ATP primarily from glycolysis. Thus, ATP concentration in MSCs exposed to hypoxia, ischemia, ischemia + 2DG, or glucose deprivation was determined. Hypoxia gradually reduced the ATP concentration from 66% ± 17% after 12 hours to 39% ± 4% at 48 hours (Fig. 7). In addition to hypoxia, depletion of glucose from the culture medium due to consumption could have also contributed to the low ATP level at 48 hours. Ischemia and ischemia + 2DG diminished ATP synthesis, and the ATP concentration dropped to 6% ± 3% after 12 hours treatment (Fig. 7). Glucose deprivation resulted in an intermediate drop in ATP concentration, leveling at 20% ± 4% between 12 and 48 hours, confirming that active glycolysis is required for 80% of the total cellular ATP synthesized in MSCs (Fig. 7).

Figure 7.

Effect of ischemia and its components on mesenchymal stem cell (MSC) ATP concentration. MSCs were exposed to hypoxia, glucose deprivation, ischemia, or ischemia + 2DG for 12–48 h. After treatment, cells were lysed immediately, and intracellular ATP concentration was determined. The graph shows the average ATP concentration of three independent experiments as fraction of the ATP concentration in untreated cells ± SEM; n = 3. Abbreviations: 2DG, 2-deoxyglucose; h, hours.

Discussion

Despite multiple positive preclinical and clinical studies using cell-based therapies for MI, greater clarity on the mechanisms underlying these beneficial effects is required. Therefore, this study examined the effects of different components of ischemia, as well as ischemia followed by reperfusion, on MSC survival and functionality in comparison with fibroblasts and, more importantly, to cardiomyocytes. To our knowledge, no complete comparative study has been carried out to date examining the effect of the components of ischemia on MSC survival and multipotentiality in comparison with cardiomyocytes.

The findings of this study illustrate that MSCs are much more resistant to ischemia than cardiomyocytes, suggesting that MSCs can cope with the acidic conditions developing during prolonged ischemia. MSCs were also fully resistant to 6 hours of ischemia followed by reperfusion. This finding has major importance, as it is well known that the reactive oxygen species generated by reperfusion are detrimental to cardiomyocytes. In agreement with the results of Zhu et al. [29], it was found that ischemia triggers apoptosis in MSCs, but it occurred only after exposure for 48 hours, compared with 4 hours for cardiomyocytes. To mimic inhibition of glycolysis occurring during in vivo ischemia, the glycolytic inhibitor 2DG was included in the in vitro model [25]. Again, MSCs displayed considerable resistance compared with cardiomyocytes. At the same time, inhibition of glycolysis during ischemia exacerbated cell death in MSCs, suggesting that glycolysis is a significant contributor to MSC survival in ischemic conditions.

Exposure to ischemia in the presence of 2DG did not reduce the differentiation potency of MSCs, showing that in ischemic stress conditions, MSCs can retain multipotency. In addition to differentiation or transdifferentiation, MSCs may aid cardiac tissue recovery by secretion of cytokines, paracrine growth factors, and/or antiapoptotic factors. When the identity of these factors becomes known, the effect of ischemic exposure on their secretion by MSCs must be examined to conclude that full MSC functionality is maintained. In case of severe acute myocardial infarction, and especially in cases with no revascularization option, it is possible that MSCs can home to and survive in the damaged area by secreting vascular endothelial growth factor and other growth factors. MSCs may reduce scar tissue formation and/or enhance angiogenesis, leading to recovery of a contractile tissue. Because of the robustness of the MSCs, combination of partial/stepwise reperfusion with cell therapy may offer a new avenue in MI therapy. Overall, these data show that MSCs may possess great potency to assist myocardium regeneration and question the need for overexpression of high-risk prosurvival or antiapoptotic genes in MSCs to retain their viability/functionality in ischemic conditions [31–33].

To further investigate the components of ischemia to which MSCs were most sensitive, the effect of hypoxia and glucose deprivation was studied. MSCs cultured in hypoxia showed no signs of apoptosis or necrosis, suggesting that hypoxia had little contribution to MSC apoptosis observed after ischemic treatment. Besides viability, hypoxia can affect cell adhesion, metabolism, proliferation and growth factor secretion [34], and thus its effect on MSC multipotency was studied. MSCs exposed to 48 hours of hypoxia maintained the ability to differentiate into chondrocytes and adipocytes, as well as into cardiomyocyte-like cells. Although presently there is no generally accepted protocol to fully differentiate MSCs to cardiomyocytes, treatment with 5-azacytidine leads to induction of a number of cardiomyogenic marker genes and thus was suitable to assess the effect of hypoxic and ischemic stress on the induction of cardiomyogenesis [28]. These results are in agreement with the studies of Ren et al. [35] and Grayson et al. [34, 36], who showed that human MSCs retained their ability to differentiate into osteoblasts and adipocytes after exposure to hypoxia. Lennon et al. [37] showed that cultured rat MSCs functioned optimally in an atmosphere of reduced oxygen concentration (5%), which more closely approximates the in vivo oxygen tension of the bone marrow [38] (4%–7%). The current study shows that MSC viability and functionality remain unaffected when the oxygen concentration is as low as 0.5%. Qu et al. [39] have shown that in continuous hypoxia and hypoxia/reperfusion, cocultures of MSCs exert antiapoptotic effects on cardiomyocytes.

MSCs displayed the highest sensitivity to glucose deprivation, indicating that they may require active glycolysis for survival. Terminal oxidation has the capacity to produce larger quantities of ATP than glycolysis. During glucose deprivation, ketogenic amino acid-, fatty acid-, or ketone body-derived acetyl-CoA is available for the citric acid cycle to supply NADH+H⁺/FADH₂ to fuel the terminal oxidation. Thus, either glucose deprivation-induced apoptosis in MSCs is not a consequence of ATP depletion or MSCs rely on glucose-based ATP production to survive.

Inhibition of mitochondrial respiration with DNP [19] showed that MSCs do not require oxidative phosphorylation to survive. MSCs displayed no cell death upon mitochondrial uncoupling, whereas cardiomyocytes could not survive without aerobic respiration. This explains how MSCs could survive hypoxia, as mitochondrial energy production does not appear to be essential for their viability. When ATP levels were examined, hypoxia induced a slow, gradual decline in ATP concentration. In hypoxia, glycolysis is the only source of ATP, leading to increased glucose consumption by the cells and depletion of glucose from the culture medium, which probably also contributed to the declining ATP levels. This reduced ATP concentration (38% ± 6% of control at 48 hours) was sufficient to retain MSC viability. Ischemia induced an almost complete depletion of ATP after 12 hours, and addition of 2DG did not have a significant additive effect. MSCs were able to cope with minimal ATP concentration, as 48 hours of exposure to ischemia was necessary to induce 15% cell death. Glucose deprivation under normoxic conditions resulted in an intermediate reduction of the ATP concentration, averaging approximately 22% ± 6% of the control over the 48-hour time course examined. These findings indicate that MSCs are metabolically flexible and can survive solely on anaerobic respiration in a hypoxic/ischemic environment. Glucose-dependent metabolic pathways (primarily glycolysis and possibly mitochondrial degradation of glucose-derived pyruvate) produced approximately 80% of the total cellular ATP in MSCs. However, the high sensitivity of MSCs to glucose deprivation cannot be explained solely by the loss of ATP. Ischemia induced a more pronounced drop in ATP concentration but failed to induce cell death in the same time frame. In addition to ATP production, there is evidence that glycolysis is central in maintaining cell viability. Akt, a potent antiapoptotic kinase, has been shown to inhibit cell death in a manner that depends on glucose hydrolysis through glycolysis [40]. This dependence of Akt on glucose to promote cell survival may be due to Akt-dependent stimulation of hexokinase (HK) and the action of the produced glucose-6-phosphate as a signaling molecule. Control of glucose metabolism by Akt may be a critical component of cell survival, because maintenance of glucose metabolism on its own was sufficient to block Bax activation, cytochrome c release from mitochondria and growth factor withdrawal-induced cell death [41]. Akt has also been shown to localize HK to the mitochondria, where it keeps the mitochondrial permeability transition pore (an ion channel causing cytochrome c release by inducing mitochondrial depolarization) in a closed state [41]. Glucose deprivation impairs the ability of Akt to maintain mitochondrion-HK association. Thus, glucose deprivation and the resulting inhibition of glycolysis lead not only to ATP depletion but also to loss of essential prosurvival signals, the combination of which rapidly leads to MSC apoptosis.

Summary

Our results show that MSCs have a profoundly greater capacity to survive ischemia than cardiomyocytes and that exposure to ischemia does not impair their differentiation potency. Although cardiomyocytes were sensitive to hypoxia and glucose deprivation, MSCs tolerated these conditions considerably longer because their metabolism could become independent of mitochondrial energy production. Of the different components of ischemia, MSCs showed the highest sensitivity to inhibition of glycolysis, and its inhibition led to accelerated cell death. In summary, MSCs showed a metabolic flexibility that may provide the necessary protection in an ischemic environment and allow MSCs to function in a reparative or regenerative capacity. These important features of MSCs may have considerable significance in improving the efficiency of stem cell therapy in cardiac disease.

Acknowledgements

We thank Garry Duffy, Caroline Curtin, Miriam Kearns, Aoife O'Reilly, Linda Howard, and Georgina Shaw for technical assistance and Cynthia Coleman for critical reading of the manuscript. This work was supported in part by grants from Science Foundation Ireland and Medtronic Vascular Division, Santa Rosa, CA.

Disclosure of Potential Conflicts of Interest

The authors indicate no potential conflicts of interest.