Protein O-GlcNAcylation Is a Novel Cytoprotective Signal in Cardiac Stem Cells§


  • Ayesha Zafir,

    1. Department of Medicine, Division of Cardiovascular Medicine, Institute of Molecular Cardiology, Diabetes and Obesity Center, University of Louisville, Louisville, Kentucky, USA
    Search for more papers by this author
  • Ryan Readnower,

    1. Department of Medicine, Division of Cardiovascular Medicine, Institute of Molecular Cardiology, Diabetes and Obesity Center, University of Louisville, Louisville, Kentucky, USA
    Search for more papers by this author
  • Bethany W. Long,

    1. Department of Medicine, Division of Cardiovascular Medicine, Institute of Molecular Cardiology, Diabetes and Obesity Center, University of Louisville, Louisville, Kentucky, USA
    Search for more papers by this author
  • James McCracken,

    1. Department of Medicine, Division of Cardiovascular Medicine, Institute of Molecular Cardiology, Diabetes and Obesity Center, University of Louisville, Louisville, Kentucky, USA
    Search for more papers by this author
  • Allison Aird,

    1. Department of Medicine, Division of Cardiovascular Medicine, Institute of Molecular Cardiology, Diabetes and Obesity Center, University of Louisville, Louisville, Kentucky, USA
    Search for more papers by this author
  • Alejandro Alvarez,

    1. Department of Medicine, Division of Cardiovascular Medicine, Institute of Molecular Cardiology, Diabetes and Obesity Center, University of Louisville, Louisville, Kentucky, USA
    Search for more papers by this author
  • Timothy D. Cummins,

    1. Department of Medicine, Division of Cardiovascular Medicine, Institute of Molecular Cardiology, Diabetes and Obesity Center, University of Louisville, Louisville, Kentucky, USA
    Search for more papers by this author
  • Qianhong Li,

    1. Department of Medicine, Division of Cardiovascular Medicine, Institute of Molecular Cardiology, Diabetes and Obesity Center, University of Louisville, Louisville, Kentucky, USA
    Search for more papers by this author
  • Bradford G. Hill,

    1. Department of Medicine, Division of Cardiovascular Medicine, Institute of Molecular Cardiology, Diabetes and Obesity Center, University of Louisville, Louisville, Kentucky, USA
    Search for more papers by this author
  • Aruni Bhatnagar,

    1. Department of Medicine, Division of Cardiovascular Medicine, Institute of Molecular Cardiology, Diabetes and Obesity Center, University of Louisville, Louisville, Kentucky, USA
    Search for more papers by this author
  • Sumanth D. Prabhu,

    1. Department of Medicine, Division of Cardiovascular Medicine, Institute of Molecular Cardiology, Diabetes and Obesity Center, University of Louisville, Louisville, Kentucky, USA
    2. Division of Cardiovascular Disease, University of Alabama at Birmingham, Birmingham, Alabama, USA
    Search for more papers by this author
  • Roberto Bolli,

    1. Department of Medicine, Division of Cardiovascular Medicine, Institute of Molecular Cardiology, Diabetes and Obesity Center, University of Louisville, Louisville, Kentucky, USA
    Search for more papers by this author
  • Steven P. Jones

    Corresponding author
    1. Department of Medicine, Division of Cardiovascular Medicine, Institute of Molecular Cardiology, Diabetes and Obesity Center, University of Louisville, Louisville, Kentucky, USA
    • Institute of Molecular Cardiology, University of Louisville, 580 South Preston Street—321F, Baxter II—321F, Louisville, Kentucky 40202, USA
    Search for more papers by this author
    • Telephone: (502) 852-2460; Fax: (502) 852-8070

  • A.Z.: conception and design, collection and assembly of data, data analysis and interpretation, manuscript writing, and final approval of manuscript; R.R., B.W.L., A.A., and A.A.: collection and/or assembly of data and data analysis and interpretation; J.M. and T.D.C.: collection and/or assembly of data, data analysis and interpretation, and final approval of manuscript; Q.L.: provision of study material; B.G.H.: administrative support, data analysis and interpretation, manuscript writing, and final approval of manuscript; A.B.: financial support, administrative support, manuscript writing, and final approval of manuscript; S.D.P.: financial support and administrative support; R.B.: conception and design, financial support, administrative support, provision of study material, and final approval of manuscript, S.P.J.: conception and design, financial support, administrative support, data analysis and interpretation, manuscript writing, and final approval of manuscript.

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

  • §

    First published online in STEM CELLS EXPRESS January 17, 2013.


Clinical trials demonstrate the regenerative potential of cardiac stem cell (CSC) therapy in the postinfarcted heart. Despite these encouraging preliminary clinical findings, the basic biology of these cells remains largely unexplored. The principal requirement for cell transplantation is to effectively prime them for survival within the unfavorable environment of the infarcted myocardium. In the adult mammalian heart, the β-O-linkage of N-acetylglucosamine (i.e., O-GlcNAc) to proteins is a unique post-translational modification that confers cardioprotection from various otherwise lethal stressors. It is not known whether this signaling system exists in CSCs. In this study, we demonstrate that protein O-GlcNAcylation is an inducible stress response in adult murine Sca-1+/lin CSCs and exerts an essential prosurvival role. Posthypoxic CSCs responded by time-dependently increasing protein O-GlcNAcylation upon reoxygenation. We used pharmacological interventions for loss- and gain-of-function, that is, enzymatic inhibition of O-GlcNAc transferase (OGT) (adds the O-GlcNAc modification to proteins) by TT04, or inhibition of OGA (removes O-GlcNAc) by thiamet-G (ThG). Reduction in the O-GlcNAc signal (via TT04, or OGT gene deletion using Cre-mediated recombination) significantly sensitized CSCs to posthypoxic injury, whereas augmenting O-GlcNAc levels (via ThG) enhanced cell survival. Diminished O-GlcNAc levels render CSCs more susceptible to the onset of posthypoxic apoptotic processes via elevated poly(ADP-ribose) polymerase cleavage due to enhanced caspase-3/7 activation, whereas promoting O-GlcNAcylation can serve as a pre-emptive antiapoptotic signal regulating the survival of CSCs. Thus, we report the primary demonstration of protein O-GlcNAcylation as an important prosurvival signal in CSCs, which could enhance CSC survival prior to in vivo autologous transfer. STEM CELLS 2013;31:765–775


Over the past decade, several reports challenged the conventional wisdom that the adult mammalian heart is a quiescent, postmitotic organ [1–3]. Several lines of evidence now firmly establish the reparative capacity of the myocardium under pathophysiological conditions. This conclusion was based both upon the identification and isolation of resident cardiac stem cells (CSCs) [4, 5] and their therapeutic utility for autologous transplantation and functional improvement in the setting of acute and chronic myocardial infarction [6, 7]. Such promising preclinical findings, although not without controversy, are driving rapid clinical translation [8].

The first wave of clinical application of cardiogenic cell types—skeletal myoblasts [9, 10] and bone marrow-derived cells [11–13]—ushered in a phase of regenerative therapy driven now by mesenchymal stem cells [14, 15], CSCs and cardiosphere-derived cells (CDCs) [8, 16]. In most cases, excellent safety profiles have been demonstrated. Efficacy based on functional assessment (with ejection fraction as the primary endpoint) is less inconsistent [9–13] and has shown promise in the more recent phase I clinical trials [8, 16]. However, a significant impediment in optimizing stem cell therapy, irrespective of cell choice and its route of delivery, remains the issue of marginal cell retention and engraftment due to poor survival in the hostile milieu of the failing myocardium [17, 18]. Enhancing or reinforcing the long-term persistence of CSCs could bolster their therapeutic efficacy.

Without question, understanding the basic biology of CSCs is essential for improving future, more effective iterations of cellular therapeutics in patients. In particular, better understanding of the mechanisms that control cell fate (survival and/or differentiation) could promote more effective cardiac repair [18]. Prosurvival signaling pathways represent potentially viable therapeutic targets in CSCs [7, 19, 20]. Unlike acute cardioprotective interventions where the ischemic insult happens prior to the experimental intervention, CSCs could be manipulated to enhance their survival in the remodeled, infarcted, or otherwise failing heart without limitation of clinical applicability. Here, we test the importance of a potentially novel cytoprotective signal in CSCs.

Numerous stress-inducible pathways have been identified in eukaryotes. In the adult mammalian heart, the glycosylation of proteins by β-O-linkage of N-acetylglucosamine (i.e., O-GlcNAc) is a unique post-translational modification that confers cardioprotection [21–28]. O-GlcNAc transferase (OGT—uridine diphospho-N-acetylglucosamine: peptide β-N-acetylglucosaminyl transferase) is the sole enzyme responsible for O-GlcNAcylating proteins, while O-GlcNAcase (OGA—O-β-N-acetylglucosaminidase) cleaves this post-translational modification. Because protein O-GlcNAcylation represents a metabolically derived, apparently ubiquitous, and recruitable stress response in the cardiovascular system [29], it is rational to posit that such a signaling system might exist in more primitive cardiac cell types, such as CSCs. The objective of this study was to characterize the phenomenon of protein O-GlcNAcylation in CSCs and to assess its potential relevance as a necessary prosurvival signal during hypoxia-reoxygenation injury.

An unambiguous and definitive description of a resident CSC remains elusive, owing to considerable overlap of surface molecular markers with hematopoietic and endothelial progenitor cells [30–33]. In general, adult CSCs are lineage-negative and express, uniquely or in combination, the stem cell antigens stem cell antigen 1 (Sca-1), c-kit, and/or MDR-1; they are self-renewing, clonogenic, and multipotent for the principle cell types of the heart—cardiomyocytes, smooth muscle cells, and endothelial cells [4, 32, 34–36]. Because it is not known whether CSCs express a functional O-GlcNAcylating system, we used Sca-1+/lin mouse CSCs to address this question and the possible functional role of O-GlcNAc levels in modulating CSC survival.


Cell Culture and Flow Cytometric Analysis

CSCs were isolated from adult, male, wild-type mouse (C57BL6) heart cell outgrowth cultures subjected to sequential sorting for c-kit+/lin markers using magnetic immunobeads [37]. The sorted cells were analyzed by flow cytometry at different passages (<P10 for the experiments outlined here) to ascertain the purity of the cultures. Harvested CSCs and cellular controls were blocked with FcBlock [0.005 mg/ml in phosphate buffered saline (PBS) + 1% bovine serum albumin (BSA)] for 10 minutes at 4°C. CSCs and controls were stained with a cocktail of anti-mouse c-kit, Ly6AE (Sca-1), hematopoietic lineage cocktail, CD31, CD34, and CD45 antibodies (BD BioSciences, San Jose, CA; Data were acquired on a LSRII flow cytometer (BD BioSciences, San Jose, CA; and analyzed using BD FACSDiva software (San Jose, CA). Unstained samples of all lines were used in setting discrimination gates. CSCs were cultured in Dulbecco's modified Eagle's medium (DMEM)/F-12 containing leukemia inhibitory factor (1,000 U/ml), basic fibroblast growth factor (20 ng/ml), epidermal growth factor (20 ng/ml), and 10% embryonic stem cell fetal bovine serum, as described [38].

Pharmacological Modulation of O-GlcNAcylation in CSCs

To inhibit OGT and, thereby, reduce O-GlcNAc levels, CSCs were pretreated overnight (∼16 hours) with the OGT enzyme inhibitor TT04 [2H-1, 3-thiazine-6-carboxylic acid, 2-[(4-chlorophenyl) imino] tetrahydro-4-oxo-3-(2-tricyclo[, 7] dec-1-ylethyl-)] (TimTec, Inc., Newark, DE; [22, 39] at a final concentration of 0.001 mmol/l, prior to submitting them to hypoxia-reoxygenation injury. To augment O-GlcNAcylation of cellular proteins, CSCs were treated with thiamet-G (ThG; 0.025 mmol/l) (Cayman Chemical, Ann Arbor, MI; [40], which inhibits OGA.

OGT Gene Deletion

CSCs carrying only loxP-flanked copies of the OGT gene were infected with replication-deficient adenovirus (Vector Biolabs, Philadelphia, PA; carrying the Cre recombinase gene (AdCreGFP) to knock out the OGT gene, or a control adenovirus (AdGFP), at 500 MOI (multiplicity of infection) for 72 hours. Functional expression was ascertained by immunoblot analysis. CSCs were then subjected to hypoxia-reoxygenation for cytotoxicity assays.

In Vitro Hypoxia-Reoxygenation Injury

To simulate ischemia-reperfusion, pretreated CSCs were subjected to 3 or 6 hours of hypoxia in Esumi lethal ischemia medium [22, 23, 28, 41, 42] for glucose and nutrient deprivation (containing 117 mmol/l NaCl, 12 mmol/l KCl, 0.9 mmol/l CaCl2, 0.49 mmol/l MgCl2, 4 mmol/l HEPES, 20 mmol/l sodium lactate, and 5.6 mmol/l L-glucose; pH 6.2) in sealed humidified hypoxic chambers (Billups-Rothenberg Inc., Del Mar, CA; flushed with 5% CO2 and 95% N2, for oxygen deprivation and maintained at 37°C. After the hypoxic period, they were switched to Esumi control medium (containing 137 mmol/l NaCl, 3.8 mmol/l KCl, 0.9 mmol/l CaCl2, 0.49 mmol/l MgCl2, 4 mmol/l HEPES, and 5.6 mmol/l D-glucose; pH 7.4) and allowed to reoxygenate for 0, 1, or 3 hours in the modular incubator. Similarly treated normoxic controls received Esumi control medium for the total period of 6 hours and remained in the incubator during this period.

Protein Expression

Whole cell lysates were prepared using standard protocols for total cellular protein and 10–25 μg (as appropriate) was resolved by SDS-PAGE to immunoblot for detecting protein O-GlcNAcylation [23] or specific protein target, as indicated.

Cytotoxicity Assay

Cytotoxicity after hypoxia-reoxygenation of CSCs was quantitated by spectrophotometric determination of posthypoxic (or normoxic) lactate dehydrogenase (LDH) activity released in the medium using a commercial kit (Roche Applied Science, Indianapolis, IN; and expressed relative to total LDH content [22, 23, 28, 41].

Cell Viability Assay

CSCs were cultured in 96-well plates (0.02 × 105 cells per well) on day 1 and received ThG or TT04 pretreatment on day 2 (for ∼16 hours). They were then subjected to hypoxia-reoxygenation (day 3) and the number of viable cells was determined colorimetrically at 490 nm after incubation with the tetrazolium compound [3-(4,5-dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H-tetrazolium, inner salt; MTS] [CellTiter 96 AQueous One Solution Reagent (Promega, Madison, WI;]). In similar additional experiments, differential uptake or exclusion of trypan blue was also used to determine the percentage of viable cells.

Confocal Microscopy

For assessment of cell death, CSCs cultured in 35 mm glass-bottomed dishes and subjected to hypoxia-reoxygenation were loaded with the fluorescent DNA-binding dyes DAPI (4',6-Diamidino-2-Phenylindole, Dihydrochloride; 5 μg/ml) and propidium iodide (PI; 5 μg/ml) during the last 30 minutes of reoxygenation. Stained nuclei were visualized with a ×20 objective on a Nikon A1 Confocal Microscope (Nikon Instruments Inc., Melville, NY; using a 405 nm laser (for DAPI) and 561 nm laser (for PI). Data were analyzed using the NIS-Elements software (Nikon Instruments Inc., Melville, NY;


Caspase-3/7 enzyme activity as an index of apoptosis was assayed in total cellular protein by incubation with a caspase-3/7 substrate containing the tetrapeptide sequence DEVD. Aminoluciferin, a substrate of luciferase, is produced in this reaction, generating a luminescent signal (Caspase-Glo 3/7 Assay kit; Promega, Madison, MI; measured by a Modulus luminometer (Turner Biosystems, Sunnyvale, CA; [28]. Bioluminescence was measured in relative luminescent units. Furthermore, to ascertain activation of this pathway, the main cleavage target of caspase-3, that is, poly(ADP-ribose) polymerase-1 (PARP-1) was assessed by immunoblotting as described above.

Assessment of CSC Bioenergetics Using the XF24 Extracellular Flux Analyzer

The bioenergetic response of CSCs was measured using the Seahorse Bioscience XF24 Flux Analyzer (Seahorse Bioscience, North Billerica, MA; Extracellular flux (XF) methodology measures the two major energy producing pathways, oxidative phosphorylation and glycolysis. For mitochondrial respiration, XF analysis measures the oxygen consumption rate (OCR). For glycolytic flux, XF measures the extracellular acidification rate (ECAR). CSCs seeded at an initial density of 20,000 cells per well were pretreated with TT04 (0.001 mmol/l) or thiamet-G (0.025 mmol/l), as indicated above. For XF experiments, the treatment medium was changed to 675 μl assay medium (unbuffered DMEM supplemented with 4 mmol/l glutamine, 5 mmol/l glucose, and 1 mmol/l pyruvate) containing the corresponding pharmacological treatment 1 hour before assay. The XF24 automated protocol consisted of 10 minutes delay following microplate insertion, baseline OCR/ECAR measurements (3 × [3 minutes mix, 2 minutes wait, 3 minutes measure]), followed by injection of Port A (oligomycin, 75 μl) and OCR/ECAR measurement (3 minutes mix, 2 minutes wait, 3 minutes measure), injection of Port B (Carbonyl cyanide 4-(trifluoromethoxy)phenylhydrazone; FCCP, 83.3 μl) and OCR/ECAR measurement (3 minutes mix, 2 minutes wait, 3 minutes measure), and injection of Port C (antimycin A, 92.6 μl) and OCR/ECAR measurement (3 minutes mix, 2 minutes wait, 3 minutes measure). Stocks (1 mmol/l) of oligomycin (75351, Sigma, St. Louis, MO;, FCCP (C2920, Sigma, St. Louis, MO;, and antimycin A (A6874, Sigma, St. Louis, MO; http://www.sigmaaldrich. com) were prepared in 100% DMSO (Dimethyl sulfoxide; 154938, Sigma, St. Louis, MO; Prior to assay, stocks were diluted in assay medium to yield 10 μg/ml oligomycin, 0.01 mmol/l FCCP, and 0.1 mmol/l antimycin A which, after injection, yielded final concentrations of 1 μg/ml oligomycin, 0.001 mmol/l FCCP, and 0.01 mmol/l antimycin A [43–46].

Statistical Analyses

Data are reported as mean ± SEM and were analyzed by unpaired t test or ANOVA with post hoc analysis (Bonferroni's multiple comparison test), as appropriate, using GraphPad Prism 5.0 software (La Jolla, CA; Differences were accepted as significant when p < .05.


Characterization of CSCs Derived from Adult Murine Heart

Following isolation (P2) and over progressive passages (<P10), flow cytometric analyses revealed that the CSCs expressed the Sca-1, averaging 79.54% ± 4.5%, but were negative for blood lineage markers, CD31, CD45, were CD34low and did not retain expression of c-kit (Fig. 1) after passage. We did not detect c-kit in cultured murine CSCs, although c-kit expression is stable in human CSCs.

Figure 1.

Characterization of cardiac stem cells (CSCs) isolated from adult murine hearts. (A–C): Flow cytometric analyses reveal abundant expression of stem cell antigen 1 (A), relative absence of lin (B), c-kit (A), and CD31 (C), and low expression of CD34 in CSC cultures (<passage 10). (D): DIC image of CSCs in culture.

Protein O-GlcNAcylation Is a Stress-Responsive Signal in CSCs

CSCs were subjected to 1, 6, or 24 hours of glucose deprivation following which total cell protein was harvested and immunoblotted for analyzing changes in protein O-GlcNAcylation. Glucose-deprived CSCs responded by inducing a dynamic increase in O-GlcNAc levels at 24 hours, and the specificity of the signal was validated using routine control measures (Fig. 2A–2C). On face value it may appear paradoxical that the absence of glucose to provide flux through the hexosamine biosynthetic pathway would still lead to an increase in O-GlcNAc levels; however, this phenomenon occurs in other cell lines, perhaps via glycogen degradation [47]. Increased O-GlcNAcylation in the context of starvation demonstrates an inherently recruitable stress-induced signal that we report here for the first time to be extant in CSCs. Based on our previous work, we hypothesized that this survival response might represent a significant although latent mechanism for promoting CSC survival.

Figure 2.

Stress-induced O-GlcNAcylation of cardiac stem cell (CSC) proteins (n = 5/group). (A): Summary densitometric analysis of O-GlcNAc levels in glucose-deprived CSCs shows O-GlcNAcylation significantly increased at 24 hours (*, p < .001 vs. 0 hours). (B): A representative immunoblot is shown, using the primary antibody CTD110.6. Control measures were adopted to verify the observed O-GlcNAc signal on numerous proteins depicted by several immunopositive bands. The last lane (0+OGA; n = 3/group) is the same sample loaded in lane 1 incubated in vitro with O-GlcNAcase (OGA, which cleaves O-GlcNAc) to confirm antibody specificity. (C): The blot in (B) was coincubated with N-acetylglucosamine that competes for binding with CTD110.6.

CSCs Respond to Hypoxia-Reoxygenation Injury by Altering O-GlcNAc Levels

In order to ascertain whether this might be true under the more lethal stress of hypoxia-reoxygenation, CSCs were subjected to hypoxia for 3 hours and reoxygenated for 0, 1, 3, 6, or 24 hours, following which total proteins were analyzed by Western blotting for effects on O-GlcNAc levels. As evidenced in Figure 3A, 3B, protein O-GlcNAcylation was dynamically altered upon reoxygenation. There was increased cell damage according to LDH activity (Fig. 3C) and caspase-3/7 activation during 1 and 3 hours of reoxygenation (Fig. 3D), and these time points were selected for assessing drug treatments in further experiments. Taken together, it was apparent that CSCs have an innate capacity to respond to hypoxia-reoxygenation injury by activating the O-GlcNAc stress-signaling system. The existence of this basic biological prosurvival adaptation is demonstrated here in CSCs for the first time.

Figure 3.

Effects of hypoxia-reoxygenation (H/R) on cardiac stem cells (CSCs) (n = 5/group). (A, B): CSCs were subjected to 3 hours hypoxia and reoxygenated for various durations. Densitometry of immunoblots demonstrated that O-GlcNAc levels peaked significantly (*, p < .02 vs. normoxic control, NO; n = 4/group) at 3 hours of reoxygenation. This coincided with posthypoxic cellular injury as assessed in (C) significant LDH release (*, p < .02 vs. normoxic control) and (D) significant caspase-3/7 activation in CSCs (*, p < .02 vs. normoxic control). Abbreviation: LDH, lactate dehydrogenase.

Protein O-GlcNAcylation can be Pharmacologically Manipulated in CSCs and Inhibition of OGT Sensitizes CSCs to Posthypoxic Injury

To evaluate whether the endogenous response of O-GlcNAcylation imparted any effects on cell survival, we took a loss-of-function approach using the putative OGT inhibitor, TT04, at a concentration of 0.001 mmol/l (Fig. 4A). This concentration was selected based on absence of cytotoxic effects the drug might have alone (data not shown). CSCs treated overnight (∼16 hours) with TT04 had decreased levels of O-GlcNAc-modified proteins, demonstrating that this was a viable route to examine the effects of OGT loss-of-function on hypoxia-reoxygenation injury in CSCs. Reduction in O-GlcNAc levels by pretreatment (∼16 hours) of CSCs with the OGT inhibitor TT04 prior to submitting them to 3 hours of hypoxia and 3 hours reoxygenation exacerbated the extent of cell death. This was demonstrated by augmented LDH release (Fig. 4B) and by enhancement of caspase-3/7 activities (Fig. 4C), both of which were significantly greater than that observed with hypoxia-reoxygenation alone. We also assessed posthypoxic cell viability and found it to be dramatically limited via OGT inhibition, as demonstrated by the MTS assay (Fig. 4D) and the trypan blue exclusion assay (Fig. 4E). The percentage of cells positive for PI was also significantly higher with TT04 treatment during hypoxia-reoxygenation (Fig. 4F).

Figure 4.

Role of O-GlcNAc transferase (OGT) inhibition on posthypoxic cardiac stem cell (CSC) survival (n = 5/group). (A): Summary densitometric analysis of immunoblots (representative shown) demonstrating that TT04 (an OGT inhibitor) can reduce O-GlcNAc levels in CSCs by treatment with 0.001 mmol/l for ∼16 hours (*, p < .01 vs. 0 mmol/l TT04). (B): OGT inhibition significantly exacerbated posthypoxic injury in CSCs, assessed by LDH activity (*, p < .01 vs. normoxic control [NO]; #, p < .001 vs. H/R). (C): Reduction in O-GlcNAc levels by TT04 also significantly increased caspase-3/7 activation following H/R (*, p < .01 vs. NO; #, p < .001 vs. H/R). (D): OGT inhibition significantly exacerbated post-H/R CSC death, as determined by the MTS assay (*, p < .01 vs. NO; #, p < .001 vs. H/R), and (E) as evidenced by trypan blue exclusion (*, p < .01 vs. NO; #, p < .001 vs. H/R; n = 4/group). (F): Propidium iodide-indicated cell death was significantly higher following H/R due to OGT inhibition (*, p < .01 vs. NO; #, p < .001 vs. H/R). Abbreviation: H/R, hypoxia-reoxygenation.

CSCs expressing cre recombinase showed significant loss of OGT protein (Fig. 5A) and reductions in O-GlcNAc levels, in the absence of cytotoxic effects (Fig. 5B). The genetic loss of OGT recapitulated the pharmacologic studies of OGT inhibition, thereby assuaging any concerns regarding efficacy and specificity of the OGT inhibitor. To assess the effect of OGT gene deletion on posthypoxic survival, CSCs were subjected to hypoxia-reoxygenation 72 hours after treatment with AdCreGFP (500 MOI). A significant elevation of posthypoxic LDH release was observed with AdCreGFP treatment (Fig. 5C). Posthypoxic cell viability was significantly reduced by AdCreGFP (Fig. 5D). It was evident that limiting O-GlcNAcylation exacerbated cell death in CSCs. Thus, the endogenous O-GlcNAcylation in CSCs is important for cell survival. Nevertheless, the endogenous cytoprotective capacity could be further enhanced, as tested below.

Figure 5.

Sensitization of cardiac stem cells (CSCs) to posthypoxic damage by OGT gene deletion (n = 5/group). CSCs carrying the loxP-flanked OGT gene were infected with AdCreGFP (500 MOI) for 72 hours, resulting in (A) significant loss of OGT expression (*, p < .001 vs. AdGFP), (B) significant reduction in protein O-GlcNAcylation (*, p < .001 vs. AdGFP), (C) significant aggravation of posthypoxic cellular injury, assessed by LDH activity (*, p < .001 vs. AdGFP), and (D) significant loss of cell viability, according to MTS assay (*, p < .001 vs. AdGFP; n = 3/group). Abbreviations: LDH, lactate dehydrogenase; OGT, O-GlcNAc transferase.

Augmented Protein O-GlcNAcylation Protects CSCs from Posthypoxic Injury

Gain-of-function for O-GlcNAcylation was achieved by treating CSCs overnight (∼16 hours) with thiamet-G (ThG: 0.025 mmol/l), which is a potent inhibitor of OGA (the enzyme that removes the O-GlcNAc modification from proteins). This resulted in elevated levels of O-GlcNAc as assessed by Western blot (Fig. 6A) that were increased over and above vehicle-treated CSCs following reoxygenation (Fig. 6B). To ascertain whether CSCs could be rescued from posthypoxic damage via reversing the loss of O-GlcNAc during hypoxia-reoxygenation, CSCs were pretreated with ThG and it was found that elevated O-GlcNAc levels could improve cell survival with regard to reduction of LDH released into the medium (Fig. 6C). A significant reduction in caspase-3/7 activities was observed with ThG (Fig. 6D). The protective effects of enhanced O-GlcNAc signaling in CSCs were also demonstrated by a significant increase in posthypoxic cell viability, according to the MTS assay (Fig. 6E) and PI positivity (Fig. 6F).

Figure 6.

Evaluation of the effects of augmented O-GlcNAcylation on posthypoxic survival of cardiac stem cells (CSCs) (n = 5/group, unless indicated otherwise). (A): Thiamet-G, a potent OGA inhibitor, was used to significantly augment O-GlcNAc levels in CSCs (0.025 mmol/l, ∼16 hours treatment; *, p < .01 vs. 0 mmol/l). (B): O-GlcNAc levels remained significantly higher in ThG-treated CSCs following reoxygenation for 3 hours (*, p < .001 vs. normoxic control [NO]; #, p < .01 vs. H/R). (C): Increased cell survival was evident following hypoxia-reoxygenation injury by augmenting O-GlcNAcylation in CSCs, as assessed by significantly lower LDH release (*, p < .001 vs. NO; #, p < .001 vs. H/R), (D) significant reduction in caspase-3/7 activation (*, p < .001 vs. NO; #, p < .001 vs. H/R; n = 3/group), (E) significant recovery in posthypoxic cell viability (*, p < .001 vs. NO; #, p < .001 vs. H/R), and (F) significant protection from H/R-induced cell death determined by lowered PI positivity (*, p < .001 vs. NO; #, p < .001 vs. H/R). Abbreviations: H/R, hypoxia-reoxygenation; LDH, lactate dehydrogenase; PI, propidium iodide.

Inhibition of OGT Potentiates Apoptotic Pathways Following Hypoxia-Reoxygenation Via Augmenting the Cleavage of PARP-1

Because caspase-3/7 activation observed in CSCs during hypoxia-reoxygenation appeared to be a critical mechanism for the ensuing cell death processes exacerbated by the inhibition of OGT, we were interested to verify whether this might further lead to the enhanced cleavage of the main downstream target of caspase-3, that is, PARP-1. According to immunoblot analysis, the cleavage product of PARP-1 appeared when CSCs were reoxygenated, which was exacerbated by treatment with the OGT inhibitor, TT04 (Fig. 7A). In short, the posthypoxic activation of caspase-3/7 induced by lowered protein O-GlcNAcylation resulted in the downstream cleavage of full length PARP-1, sensitizing CSCs to apoptotic cell death. By increasing O-GlcNAcylation (via thiamet-G), PARP-1 cleavage was reduced, thereby demonstrating a robust prosurvival mechanism in CSCs (Fig. 7B).

Figure 7.

Potential mechanism of cell death regulated by O-GlcNAc in cardiac stem cells (CSCs) (n = 4–5/group). (A): Potentiation of apoptotic pathways is demonstrated by O-GlcNAc transferase inhibition following H/R. Posthypoxic recovery (3 hours reoxygenation) showed that reduction in O-GlcNAc levels significantly elevated PARP-1 cleavage (expressed as the ratio cleaved PARP/uncleaved PARP) (*, p < .05 vs. normoxic control (NO); #, p < .01 vs. H/R), whereas (B) enhancing O-GlcNAcylation (via ThG-mediated OGA inhibition) mitigated posthypoxic PARP1 cleavage (*, p < .05 vs. NO; #, p < .01 vs. H/R). Abbreviations: H/R, hypoxia-reoxygenation; PARP, poly(ADP-ribose) polymerase.

Bioenergetic Profile of CSCs Does Not Change by Modulating O-GlcNAc Levels

To directly assess the effects of pharmacological manipulation of O-GlcNAcylation on CSC bioenergetics, XF analysis was performed on CSCs treated with vehicle, thiamet-G, or TT04. Bioenergetic profiles were generated by determining basal OCR, ATP-linked OCR (basal OCR—oligomycin OCR), proton leak (oligomycin OCR—antimycin A OCR), maximal OCR (FCCP OCR—antimycin A OCR), and nonmitochondrial OCR (antimycin A OCR). A decrease in O-GlcNAc signal (by TT04) did not result in alterations in OCR or ECAR values compared with vehicle (data not shown). Likewise, increased O-GlcNAcylation (ThG group) did not alter OCR or ECAR values compared with vehicle (data not shown). These data exclude one potential mechanism of cytoprotection.

Early Akt Activation Occurs in Response to Hypoxia-Reoxygenation Injury in CSCs

To assess whether Akt phosphorylation occurs during the early period of reoxygenation (at 30 minutes posthypoxia) in CSCs, pAkt/total Akt ratios were determined by immunoblot (data not shown). Hypoxia-reoxygenation induced significant activation of Akt; this was not blocked by ThG pretreatment; however, Akt phosphorylation was significantly enhanced by TT04 during reoxygenation. Thus, activation of Akt does not appear to be a mechanism of ThG-mediated cytoprotection.

Enhanced Protein O-GlcNAcylation Mitigates Oxidative Damage to CSCs

To assess whether reduction of oxidative stress might be a contributory mechanism involved in the protective effect of O-GlcNAcylation, CSCs were subjected to 150 minutes of hydrogen peroxide stress. Exposure to peroxide significantly enhanced cytotoxicity as determined by LDH release (Supporting Information Fig. S1). It was evident that LDH release could be significantly suppressed by augmenting protein O-GlcNAcylation in CSCs using ThG.


This work establishes the existence of a stress-responsive glycosignaling molecular entity, that is, O-GlcNAc, in adult CSCs. Because protein O-GlcNAcylation is associated with cytoprotection in differentiated cells, we examined whether manipulation of O-GlcNAc levels in CSCs would affect their survival under hypoxia-reoxygenation. We observed that injured CSCs respond to reoxygenation by inducing protein O-GlcNAcylation in a time-dependent manner. Blockade of OGT reduced O-GlcNAc levels and sensitized CSCs to reoxygenation injury, while OGA inhibition enhanced protein O-GlcNAcylation and markedly reduced posthypoxic death.

O-GlcNAcylation of proteins is a stress-activated signal constituting a global cell survival response to diverse noxious stimuli [48–51]. Alterations in cardiac O-GlcNAc levels were found to occur under acutely stressful pathological conditions, using models of in vivo myocardial ischemia/reperfusion injury [26, 52], ischemic preconditioning [21], and trauma-hemorrhagic shock [27, 53] as well as upon in vitro exposure of cardiomyocytes to oxidative, hypoxic, or endoplasmic reticulum (ER) stress [21, 22, 28]. Increasing glucosamine availability attenuates posthypoxic injury in isolated perfused hearts [54, 55] and neonatal rat cardiomyocytes [25], whereas inhibiting the rate-limiting enzyme glutamine:fructose amidotransferase (GFAT) negated the same [25]. We have previously validated the cytoprotective action of OGT [22] in the response of cardiomyocytes to acute oxidative/hypoxia-associated damage using a pharmacological inhibitor of OGT (TT04), genetic loss-of-function (translational silencing with RNAi or deletion through cre-lox recombination), and, conversely, adenoviral overexpression to enhance global O-GlcNAc levels [22]; similar findings were reported by other laboratories [26]. Such methods to manipulate OGA activity/expression demonstrated it to be a sensitizing stimulus for cellular injury in vitro [21, 23, 25] as well as during acute in vivo myocardial ischemia-reperfusion [21].

In terms of potential mechanistic insights, apoptosis is implicated in cardiovascular dysfunction caused by acute myocardial infarction and heart failure [56, 57]. In this study, we identified caspase activation as one (of likely multiple) potential target in cell death pathways. Caspase activation became even more pronounced when protein O-GlcNAcylation in CSCs was reduced. As a potential consequence of caspase activation, we observed enhanced cleavage of poly(ADP-ribose)polymerase-1 (PARP-1), a caspase substrate and mediator of cell death. Thus, O-GlcNAc may partially regulate cell death by targeting molecular events such as caspase-mediated cleavage of PARP-1. In addition to attenuating apoptosis, our data indicate that O-GlcNAc could protect CSCs by limiting oxidative stress. Future studies will identify the specific targets of O-GlcNAc modification related to this process.

A principal requirement of any cell-based therapy for repairing the damaged myocardium with functionally competent cells may include elevated endurance capabilities [18]. To withstand pathological stress (e.g., hypoxia, oxidants, and inflammation) in the postinfarct myocardium, adoptively transferred cells must be appropriately “primed” to minimize loss due to cell death. As demonstrated here, targeting O-GlcNAcylation represents a promising approach for simple, but highly protective, modification of CSCs to promote retention of transplanted CSCs. Of course, it is likely that survival of CSCs in situ is subject to further regulation by various paracrine signals. Transplantation of O-GlcNAc-enhanced CSCs in an animal model of myocardial infarction is required to clearly demonstrate the full translational capacity of enhanced O-GlcNAcylation in CSCs; such studies are the focus of ongoing work in our laboratory.


In summary, this study demonstrates that the O-GlcNAcylation system is present in adult CSCs and is a determinant of cell viability. Because protein O-GlcNAcylation represents a robust prosurvival signal during hypoxia-reoxygenation injury in CSCs, our findings clearly signify a potential mechanism for fine-tuning cellular therapeutics for myocardial infarction.


This work was supported by grants from the NIH (R01 HL083320, R01 HL094419, P20 RR024489, P01 HL078825, and R37 HL055757).


The authors indicate no potential conflicts of interest.