TY - JOUR
AU - Petite, Hervé
AB - Abstract Mesenchymal stem cells (MSCs) hold considerable promise in tissue engineering (TE). However, their poor survival when exogenously administered limits their therapeutic potential. Previous studies from our group demonstrated that lack of glucose (glc) (but not of oxygen) is fatal to human MSCs because it serves as a pro-survival and pro-angiogenic molecule for human MSCs (hMSCs) upon transplantation. However, which energy-providing pathways MSCs use to metabolize glc upon transplantation? Are there alternative energetic nutrients to replace glc? And most importantly, do hMSCs possess significant intracellular glc reserves for ensuring their survival upon transplantation? These remain open questions at the forefront of TE based-therapies. In this study, we established for the first time that the in vivo environment experienced by hMSCs is best reflected by near-anoxia (0.1% O2) rather than hypoxia (1%–5% O2) in vitro. Under these near-anoxia conditions, hMSCs rely almost exclusively on glc through anerobic glycolysis for ATP production and are unable to use either exogenous glutamine, serine, or pyruvate as energy substrates. Most importantly, hMSCs are unable to adapt their metabolism to the lack of exogenous glc, possess a very limited internal stock of glc and virtually no ATP reserves. This lack of downregulation of energy turnover as a function of exogenous glc level results in a rapid depletion of hMSC energy reserves that explains their poor survival rate. These new insights prompt for the development of glc-releasing scaffolds to overcome this roadblock plaguing the field of TE based-therapies. This study demonstrated that human mesenchymal stem cells (hMSCs) located at the core of hydrogels construct implanted in vivo face a near-anoxic microenvironment which quickly becomes ischemic when extracellular nutrients (especially exogenous glucose [glc]) are exhausted. In this ischemic environment, failure of hMSC to adapt their glc consumption to glc shortage as well as their inability to use alternative energy substrates result in a scenario where implanted hMSCs consume their glycolytic reserves through glycolysis in less than 24 hours. Once these glycolytic reserves are exhausted, hMSCs cannot maintain their ATP content and are unable to meet their bioenergetic requirements ultimately leading to an early and massive cell-death within 3 to 7 days post-implantation. Red arrows indicate functions that are downregulated, Green arrows indicate functions that are upregulated. Abbreviations: TCA, tricarboxylic acid cycle; PPP, pentose phosphate pathway; HIF, hypoxia-inducible factor; LDH, lactate dehydrogenase. Open in new tabDownload slide Open in new tabDownload slide Mesenchymal stem cells, Multipotent stromal cells, Metabolism, Hypoxia, Anoxia, Survival, Ischemia, Glycolysis, Bioenergetic, Glycolytic reserves Significance Statement Poor survival of grafted cells at the injury site is the major impediment for developing successful cell-based therapies. In this study, it was shown that hMSCs fail to survive in the ischemic environment that they face in vivo, because these cells are not able to adapt their glucose (glc) consumption and do not possess the necessary glycolytic reserves to maintain their metabolism for more than 3 days. As a result, upon exhaustion of available glc these cells die rapidly and massively. These new insights prompt for the development of strategies aiming at providing extracellular glc to overcome this major roadblock. Introduction Mesenchymal stem cells (often referred to as multipotent stromal cells; MSCs) have elicited great interest in the field of tissue engineering (TE) and regenerative medicine due to their capability to (a) proliferate, (b) secrete various growth factors and cytokines pertinent to new tissue formation, and (c) differentiate into cells such as osteoblasts, chondrocytes, and adipocytes [1-3]. Because of these remarkable properties, TE grafts using MSCs have become the “gold standard” for numerous medical investigations and pertinent applications around the world. Although it is still early to draw definitive conclusions, the results of preclinical animal studies are not as promising as hoped. In fact, in the context of regenerative medicine, exogenously administered MSCs loaded into material scaffolds exhibited poor survival in bone [4, 5], cardiac [6, 7], and kidney [8] applications. A possible explanation for this limited cell survival is that, upon implantation, MSCs encounter a considerable bioenergetic challenge in a harsh ischemic microenvironment characterized by low oxygen tension and nutrient deprivation. In fact, MSCs must rely on energy not only to ensure their homeostasis (which entails non-spontaneous, energy-consuming processes including maintenance of concentration gradients of different ions, cytoskeletal dynamics, DNA repair, basal transcription and translation, protein turnover, and vesicle trafficking [9]), but also to “fuel” their regenerative function. Although a comprehensive accounting of the complete metabolic requirements of MSCs which trigger such responses has never been reported, it is reasonable to hypothesize that an imbalance between the cell energy demand and supply needed to maintain homeostasis and to enact the transcriptional and translational programs which promote effector functions leads to the death of MSCs into implanted TE constructs. Therefore, it is of paramount importance to determine the intracellular energy reserves human MSCs (hMSCs) can rely on upon implantation and to decipher the mechanisms by which these cells align their bioenergetic needs with available energy resources post implantation. Oxygen and nutrients are both critically required for energy-related metabolic pathways in cells and have both the potential to significantly affect cell metabolism and survival. However, until very recently, emphasis has been placed solely on the role of oxygen tension because it modulates several critical cellular processes (e.g., cell adhesion [10, 11], metabolism, proliferation, and differentiation [12, 13]). In addition, since passive diffusion of oxygen from blood capillaries is limited to 100–200 μm radius [14, 15], a hypoxic or even near-anoxic microenvironment surrounding cell-containing constructs of clinical relevant volume is established upon implantation. To date, the mechanism by which MSCs uptake and use oxygen and nutrients under ischemic conditions remains poorly investigated and understood. Recent studies in sheep [16], mice [17], and humans [18] MSCs challenged the pivotal role of oxygen and established that MSCs adapt to near-anoxia conditions in vitro and increased their in vivo survival rate as long as they have a sufficient glucose (glc) supply [18]. This study aims at further elucidating the mechanisms by which glc and its end-products are used by hMSCs to fuel their metabolism upon transplantation. To this aim, we determined the oxygen tension that best reflected the milieu encountered by hMSCs post-implantation and then investigated whether alternative exogenous nutrients can be used to ensure hMSC survival, which energy-providing pathways are active, and how hMSCs allocate their energy-reserves to meet their bioenergetic demands. This paper demonstrates, for the first time, that the 0.1% oxygen level best reflects the in vivo situation encountered by hMSCs located at the core of TE constructs. Under this condition, hMSCs have a very limited internal stock of glc and virtually no ATP reserves. In fact, our data provide evidence of a scenario where implanted hMSCs consume their glycolytic reserves through glycolysis in less than 24 hours. Once these glycolytic reserves are exhausted, hMSCs cannot maintain their ATP content and are unable to meet their bioenergetic requirements; ultimately leading to an early and massive cell-death within 3 days post-implantation. Materials and Methods Chemicals The chemicals used in this study and their respective supplier were as follows: alpha minimum essential medium (αMEM) without glc from Dominique Dutscher (Brumath, France); Dulbecco's minimum essential medium (DMEM), Super script II enzyme, TaqMan primers and Trizol from LifeTechnologies (Saint Aubin, France); human RT2 glc metabolism polymerase chain reaction (PCR) array and the RNeasy MiniKit from Quiagen (Venlo, Netherlands); 2-deoxyglucose (2DG), 3-methyl-adenine, 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT), antimycin A (AnA), citrate buffer, deferoxamine (DFx), dimethyl sulfoxide (DMSO), fibrinogen, glc, glutamine (glut), malonate (Malo), propidium iodide (PI), pyruvate (pyr), serine (ser), sodium oxamate (NaOx), and trypsin-EDTA from Sigma-Aldrich (St Quentin Fallavier, France); antibiotics, fetal bovine serum (FBS), and trypsin (10X) from PAA (Pasching, Austria); CellTiter-Glo Luminescent cell assay from Promega (Charbonnières-les-Bains, France); Glycogen Assay Kit from Biovision (Mountain View, CA); hypoxia-inducible factor (HIF)-1α antibody from Novus Biological (Bio-Techne, Lille, France); β-actin antibody from Abcam (Paris, France) and thrombin from Baxter (Guyancourt, France). Cells and Cell Culture Bone marrow was obtained as discarded tissue during routine bone surgery from five adult donors with the respective patient consent according to Lariboisière hospital (Paris, France) regulations. hMSCs were then isolated and expanded in αMEM supplemented with 10% FBS using a procedure adapted from literature reports [19]. Cells were characterized by fluorescence-activated cell sorting analyses and differentiation assays. Briefly, these cells were positive for CD90, CD73, and CD105 and negative for CD45 markers. Their osteogenic, adipogenic, and chondrogenic differentiation potentials were validated (data not shown). Cells from the five donors were then pooled at an equal ratio at passage 1, and cultured under standard cell culture conditions, that is, a humidified 37°C, 5% CO2, and 95% air environment. At 80%–85% confluence, the cells were trypsinized using trypsin-EDTA and passaged. Cells from passage up to numbers 4 and 5 were used for the experiments. Hypoxic Conditions Hypoxic conditions were achieved using a well-characterized, finely controlled, proOx-C-chamber system (C-Chamber, C-374, Biospherix, New York, NY). The oxygen concentration in this chamber was maintained at either 0.1, 1, 5, or 21% with the residual gas mixture composed of 5% CO2 and balance nitrogen for the duration of the experiments. For experiments under hypoxic conditions, the hMSCs were seeded at 12,000 cells per square centimeter in individual wells of 24-well cell-culture plate and further cultured in minimal medium (d-glucose-, l-glutamine-, phenol red-, and sodium pyruvate-free DMEM) containing no serum unless stated otherwise. The culture media were supplemented with either glc (5 g/l), glut (5 g/l), ser (5 g/l), or pyr (5 g/l) for assessing the effect of energy-related metabolites on hMSC survival. The culture media were supplemented with either Malo (10 mM), AnA (5 µM), NaOx (100 mM), or 2DG (10 µM) for assessing the role of glycolysis and tricarboxylic acid (TCA) on hMSC survival. To ensure constant oxygen levels, the hMSC cultures were maintained undisturbed and without supernatant medium changes until the end of the respective experiments. In Vitro MTT Assay The MTT assay was based on the protocol reported for the first time by Mosmann [20]. Briefly, at the prescribed time points, hMSCs were treated with 0.8 mg/ml of MTT dissolved in serum-free αMEM medium for 4 hours. Cells were then washed with phosphate buffer saline (PBS) and incubated with DMSO under gentle shaking for 10 minutes. The respective light absorbance of the resulting solutions was read at 560 nm using a microplate spectrophotometer (SpectraMax 190, Molecular Devices, Sunnyvale, CA). These results were expressed as the absorbance normalized to the total cell protein content: “abs/protein (mg)”. Effects of Oxygen Tension on HIF-1α Bioactivity and Expression hMSCs were exposed to either 0.1, 1, 5, or 21% oxygen for 3 days. The positive control for HIF-1α bioactivity was obtained by adding 500 μM DFx to the supernatant medium of hMSCs cultured under the normoxic (i.e., 21% oxygen) condition. HIF-1α bioactivity was evaluated by transfecting hMSCs with the reporter plasmid pGL3/5 hypoxia responsive element (HRE).CMVmp-Luc which contained five copies of the HRE, a DNA binding sequence for HIF-1α. In addition, antibodies against HIF-1α (in a dilution of 1:750) and antibodies against β-actin (used as a loading control; in a dilution of 1:5,000) were used for immunoblotting. In chemiluminescence, HIF-1α bioactivity and immunoblotting signals were quantified using the IVIS Lumina Bioluminescent imaging system (Xenogen, Caliper Life Science, Tremblay-en-France, France). HIF-1α bioactivity results are expressed as a bioluminescent signal (p/s), HIF-1α expression results are expressed chemiluminescence signal (p/s/cm2/sr). Determination of hMSC Viability Evaluation of cell viability was performed at days 0, 1, 3, 7, and 14 of culture or implantation unless otherwise stated. Cells were recovered after trypsinization and cell viability was assessed using the PI (at a final concentration of 0.5 mg/ml) staining assay. Briefly, the fluorescent PI is incorporated into cells after loss of their cell membrane integrity. Data for the determination of cell viability were collected using the Attune Cytometer (LifeTechnologies, Saint Aubin, France) and were analyzed using Attune software. Real-Time-Polymerase Chain Reaction Analysis mRNA of metabolic genes expressed by hMSCs under the conditions of interest to this study was quantified using quantitative reverse transcriptase PCR (qRT-PCR) and the human RT2 glucose metabolism PCR array (Quiagen, Venlo, Netherlands). Briefly, hMSCs were exposed to either 0.1, 1, 5, or 21% oxygen in the presence of 5 g/l glc for 3 days in vitro. At that time, the hMSCs were rinsed with PBS once, and their RNAs were extracted using Trizol, collected, and kept at −80°C until further analyses. The RNA samples were purified using the RNeasy MiniKit. Reverse transcriptase was performed using 1.5 µg of purified RNA with the superscript II enzyme. The cDNA from hMSCs exposed to 0.1 and 21% oxygen were used for the glucose metabolism RT2 profiler PCR array analysis following the manufacturer's instructions and using the MyiQ Single-Color Real-Time PCR Detection System (Biorad, Marnes-la- Coquette, France). Lactate dehydrogenase (LDH)-A gene expression from hMSCs exposed to either 0.1, 1, 5, or 21% oxygen was quantified using qRT-PCR and specific TaqMan primers following instructions described earlier in this section. These results are expressed using the 2ΔΔCt method, normalized to housekeeping genes, and presented as a fold-change relative to the results obtained under 21% oxygen tension. Determination of ATP Levels, Intracellular Glycolytic Reserves, Exogenous Glucose Contents, and Glucose Consumption hMSCs were cultured in αMEM (0% FBS) supplemented or not with either 0.25, 0.5, 1, or 5 g/l of glc in near-anoxia (0.1% O2) for 7 days and analyses were performed at days 1, 3, and 7 of culture. At these time points, hMSCs were rinsed with PBS once, detached using a scrapper, collected in citrate buffer, and stored at −20°C until further analysis. ATP contents of these hMSC cultures were quantified using the CellTiter-Glo Luminescent cell assay according to manufacturer's instructions. Intracellular glycogen and glc contents were measured using the Glycogen Assay Kit following the manufacturer's instructions. The results were reported as “glycolytic reserves per cell” (glc and glycogen pictogram per cell). In parallel, supernatants from these cell-cultures were collected and stored at −80°C until further analyses. Exogenous glc contents present in the media were monitored using a biomedical ARCHITECT C8000 (Abbott Diagnostic) robot. Data were presented as “exogenous glc level” (mmol per well). Glucose consumption of these cell cultures were determined based on the exogenous glc contents and the number of viable cells over the 7 days of exposure to near-anoxia. Results were reported as “glc consumption” (pmol per day per cell). 3D Hydrogels In Vitro Experiments For assessment of hMSC viability, ATP content and glycolytic reserves in hydrogels, 105 hMSCs were loaded first in 100 µl fibrin hydrogels (fibrinogen at 18 mg/ml; thrombin at 0.72 U/ml) and then placed in individual wells of 12-well cell-culture plate and further cultured in serum-free αMEM medium in the presence or absence of glc (1 g/l) in near-anoxia in vitro for 7 days. At the prescribed time points (d0, d1, d3, and d7), the hydrogels were removed from the hypoxic chambers and cells were recovered for further analyses according to the method described in Enzymatic Digestion of Hydrogels section. Enzymatic Digestion of Hydrogels The cells present in each construct were released from the fibrin hydrogels by digestion with 10X trypsin at 37°C for 15 minutes. The glycolytic reserves, ATP levels, and cell viability were assessed following the protocol described in Determination of ATP Levels, Intracellular Glycolytic Reserves, Exogenous Glucose Contents, and Glucose Consumption section and in Determination of hMSC Viability section, respectively. In Vivo Implantation Ten-week-old male nude mice were obtained from JanvierLabs (Saint-Berthevin, France) and handled in accordance with the new European Directive 2010/63/EU regarding the protection of animals used for scientific purposes. All experimental animal procedures were approved by the Ethics Committee on Animal Research of Lariboisière/Villemin, Paris, France. Preparation of Hydrogel Constructs For both the analyses of HIF-1α and LDH-A expression and the assessment of the TCA activity (via the MTT assay) in vivo, 106 hMSCs were loaded in 100 µl fibrin hydrogels and implanted subcutaneously in nude mice as described previously [21]. For assessment of hMSC viability, ATP contents and glycolytic reserves in vivo, 2 × 105 hMSCs were loaded in a “closed system” consisting in a fibrin hydrogel (200 µl) placed in diffusion chambers (membrane filters with 0.45 µm diameter pores; Millipore, France) as described previously in Moya et al. [22]. The diffusion chambers were then implanted subcutaneously in mice. At days 1, 3, and 7 post-implantation, the diffusion chambers were explanted and cells were recovered for further analyses according to the method described in Enzymatic Digestion of Hydrogels section. Assessment of Oxygen Tension in Hydrogel Constructs In Vivo Oxygen tension in the hydrogel constructs as well as the oxygen tension of the external environment were assessed in vivo using a Presense Needle Type Microsensor (PreSens, Regensburg, Germany) prior implantation (day 0) and at days 1 and 3 post-implantation. Immunochemistry of HIF-1α and LDH-A Prior to implantation (day 0) and at day 3 post-implantation, the hydrogel constructs were explanted and fixed in 4% paraformaldehyde (pH = 7.4) for 36 hours and then embedded in paraffin. A set of non-implanted hydrogel constructs were prepared in parallel at day 0 and processed similarly for histology. Immunodetection of HIF-1α and LDH-A was performed according to a method published by Becquart et al. [21]. In Vivo MTT Assay Before implantation (day 0) and at day 3 post-implantation, 10 µl of a solution containing 0.8 mg/ml of MTT in αMEM was directly injected in the hydrogel constructs. After 2 hours, the hydrogel constructs were explanted, fixed in 4% paraformaldehyde (pH = 7.4) for 36 hours and then embedded in paraffin. Sections, each 5 µm thin, were cut and analyzed using light microscopy. Statistical Analyses Numerical data were expressed as the mean ± SD and were analyzed statistically using the GraphPad Prism Software v6.01 (GraphPad Software, Inc, CA). The quantitative kinetics data were analyzed using two-way analysis of variance followed by Bonferroni's post hoc test. The nonparametric Mann–Whitney U test was used to analyze data from two independent samples. For all analyses, differences at p < .05 were considered statistically significant. Results hMSCs Present in the Core of the Implanted Hydrogel Constructs Exhibited an Upregulation of HIF-1α and LDH-A and a Downregulation of the Mitochondrial Activity The large range of oxygen tension used in vitro to culture hMSCs under hypoxic conditions epitomizes the difficulty in defining the one that best represents the in vivo ischemic milieu encountered by hMSCs loaded in hydrogel constructs post-implantation. To address this issue, oxygen tensions of the subcutaneous tissue and at the core of the hydrogel constructs were monitored prior implantation (at day 0) and post-implantation (at days 1 and 3). HMSCs located at the core of the hydrogel construct were exposed to oxygen tensions of 5.95% ± 0.67% prior implantation and of 0.13% ± 0.06% and 0.14%± 0.04% at day 1 and 3 post-implantation, respectively (Fig. 1A). At the same time, the subcutaneous oxygen tension ranged from 3.30% ± 0.08% to 3.65% ± 0.13% (Fig. 1A). Figure 1 Open in new tabDownload slide Human mesenchymal stem cells experience a hypoxic milieu post-implantation. (A): Oxygen tension measurements of the subcutaneous environment and at the core of the hydrogel constructs (4 mm diameter) either prior (day 0) or after (day 1 and 3) implantation in nude mice (n = 3). Anti-hypoxia-inducible factor-1α and anti-lactate dehydrogenase-A immunostaining of paraffin sections of constructs seeded with 106 cells either (B) prior (at day 0) or (C) after (day 3) subcutaneous implantation and excision in nude mice. The stained cells were visualized both in the outer (a, b) and in the inner area (c, d) of the constructs. 3-(4,5-Dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) staining of paraffin sections of constructs seeded with 106 cells either (D) prior (at day 0) or (E) after (day 3) subcutaneous implantation and excision in nude mice. The MTT solution was locally injected within the construct 2 hours before killing mice and construct excision. Representative of n = 3. Black scale bar = 200 µm. Red scale bar = 200 µm. Abbreviations: HIF, hypoxia-inducible factor; LDH, lactate dehydrogenase; MTT, 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide. Figure 1 Open in new tabDownload slide Human mesenchymal stem cells experience a hypoxic milieu post-implantation. (A): Oxygen tension measurements of the subcutaneous environment and at the core of the hydrogel constructs (4 mm diameter) either prior (day 0) or after (day 1 and 3) implantation in nude mice (n = 3). Anti-hypoxia-inducible factor-1α and anti-lactate dehydrogenase-A immunostaining of paraffin sections of constructs seeded with 106 cells either (B) prior (at day 0) or (C) after (day 3) subcutaneous implantation and excision in nude mice. The stained cells were visualized both in the outer (a, b) and in the inner area (c, d) of the constructs. 3-(4,5-Dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) staining of paraffin sections of constructs seeded with 106 cells either (D) prior (at day 0) or (E) after (day 3) subcutaneous implantation and excision in nude mice. The MTT solution was locally injected within the construct 2 hours before killing mice and construct excision. Representative of n = 3. Black scale bar = 200 µm. Red scale bar = 200 µm. Abbreviations: HIF, hypoxia-inducible factor; LDH, lactate dehydrogenase; MTT, 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide. To further characterize the milieu to which hMSCs are exposed post-implantation, the expression of HIF-1α (hallmark of hypoxia) and of LDH-A (index of an anaerobic metabolism) as well as the mitochondrial activity were assessed either prior or at day 3 post-implantation. Expression of HIF-1α and LDH-A were found to be upregulated after implantation, especially when hMSCs were located at the core of the hydrogel constructs (Fig. 1B, 1C). However, it is noteworthy that, even before implantation, expression of HIF-1α in hMSCs located at the core of the hydrogel construct was higher than the one of cells located at the periphery (Fig. 1Ba vs. 1Bc). This increase in HIF-1α expression is likely the result of the steep gradient of oxygen present in the hydrogel constructs prior implantation, where hMSCs located at the periphery are exposed to higher (>15% O2) oxygen tension; a situation not faced by implanted hMSCs where subcutaneous oxygen tension was inferior to 4% (Fig. 1A). Along the same lines, before implantation, the MTT staining of hMSCs was identical whether cells were located at the core or at the periphery of the hydrogel constructs (Fig. 1Da–1Dd). In contrast, after 3 days post-implantation, the MTT staining of hMSCs present in the core of the explanted hydrogel constructs (Fig. 1Ec, 1Ed) was strongly reduced in comparison with the one obtained at day 0 (Fig. 1Dc, 1Dd) and to one of the cells located in the outer part (Fig. 1Ea, 1Eb). These data indicate a decrease in the mitochondrial activity of hMSCs located at the core of the hydrogel constructs after 3 days of implantation. These results provided evidence that hMSCs seeded in hydrogel constructs expressed the hallmarks of ischemia after 3 days of ectopic implantation in nude mice. In addition, hMSCs experienced different levels of hypoxia depending on whether they were located in the outer or in the inner part of the constructs; the hypoxic stress in the core of the constructs was the most impactful one. The 0.1% Oxygen Tension Condition In Vitro Best Reflected the In Vivo Milieu Experienced by Implanted hMSCs The oxygen tension milieu to which cells are exposed is not a definitive predictor of the energy-related metabolic pathways of cells. In fact, the balance between oxygen demand by cells and its availability is a better predictor for cell metabolism. To establish the oxygen tension that best reflects the in vivo cell metabolism in the core of the hydrogel constructs, hMSCs were exposed to four different oxygen tensions (i.e., 0.1, 1, 5 and 21%) in vitro. At day 3 of culture, the expression and bioactivity of the hypoxic marker HIF-1α, the lactate metabolism and mitochondrial activity were assessed (Fig. 2). Figure 2 Open in new tabDownload slide The 0.1% oxygen tension in cell culture conditions best reflects the hypoxic post-implantation milieu experienced by human mesenchymal stem cells (hMSCs) in vivo. (A): Quantification of the hypoxia-inducible factor (HIF)-1α expression relative to the β-actin expression as a function of oxygen tension using western blot analysis (n = 3). (B): HIF-1α bioactivity assessed by monitoring hypoxia responsive element/Luc expression in hMSCs as a function of oxygen tension (n = 3). (C): Expression of lactate dehydrogenase (LDH)-A in hMSCs cultured under either 0.1, 1, 5, or 21% of oxygen for 3 days; gene expression was normalized first to that of the respective 18S (internal standard), and then to results obtained when the hMSCs were cultured under 21% oxygen. (D): 3-(4,5-Dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide activity of hMSCs exposed to either 0.1, 1, 5, or 21% oxygen for 3 days (n = 3). Results are expressed as means ± SD; *, p ≤ .05 versus the 0.1% oxygen tension cell group (Mann–Whitney test). Abbreviations: Ctrl, control; DFx, deferoxamine; HIF, hypoxia-inducible factor; LDH-A, lactate dehydrogenase-A; MTT, 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide. Figure 2 Open in new tabDownload slide The 0.1% oxygen tension in cell culture conditions best reflects the hypoxic post-implantation milieu experienced by human mesenchymal stem cells (hMSCs) in vivo. (A): Quantification of the hypoxia-inducible factor (HIF)-1α expression relative to the β-actin expression as a function of oxygen tension using western blot analysis (n = 3). (B): HIF-1α bioactivity assessed by monitoring hypoxia responsive element/Luc expression in hMSCs as a function of oxygen tension (n = 3). (C): Expression of lactate dehydrogenase (LDH)-A in hMSCs cultured under either 0.1, 1, 5, or 21% of oxygen for 3 days; gene expression was normalized first to that of the respective 18S (internal standard), and then to results obtained when the hMSCs were cultured under 21% oxygen. (D): 3-(4,5-Dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide activity of hMSCs exposed to either 0.1, 1, 5, or 21% oxygen for 3 days (n = 3). Results are expressed as means ± SD; *, p ≤ .05 versus the 0.1% oxygen tension cell group (Mann–Whitney test). Abbreviations: Ctrl, control; DFx, deferoxamine; HIF, hypoxia-inducible factor; LDH-A, lactate dehydrogenase-A; MTT, 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide. First, the HIF-1α expression assessed by Western blotting (Fig. 2A) and the HIF-1α bioactivity assessed by the HRE/l uc expression (Fig. 2B) were both affected by the level of oxygen under which the hMSCs were cultured. Exposure of hMSCs to 0.1% oxygen levels, resulted in a 8.9-, 8.7-, and 36.5-fold increase in HIF-1α expression and in a 3.1-, 18.7-, and 32.2-fold increase in HIF-1α bioactivity compared with those of cells maintained at 1, 5, and 21% oxygen, respectively (Fig. 2A, 2B). These data highlighted the specific effect of near-anoxia (0.1% O2) on hMSCs. Second, LDH-A mRNA expression was evaluated in hMSCs cultured either at 0.1, 1, 5, or 21% oxygen. A significant (p < .05) increase in LDH-A mRNA expression was observed in hMSCs cultured at 0.1% oxygen (Fig. 2C) when compared with either 1, 5, or 21% oxygen. These data suggest that through activation of HIF-1α, LDH-A expression was upregulated and that anerobic conversion of glc into lactate appeared to be the preferred pathway for hMSCs under near-anoxia. Third, mitochondrial activity (determined using the MTT assay) was similar in hMSCs cultured at 1, 5 and 21% oxygen but significantly (p < .05) decreased when the hMSCs were cultured at 0.1% oxygen (Fig. 2D). Overall, the results of these experiments provided evidence of an oxygen-dose effect on the hMSC response to oxygen tension: a 36.5-fold increase in HIF-1α expression, a 32.2-fold induction HIF-1α bioactivity, a drastic increase in LDH-A gene expression and, most importantly, a significant (p < .05) decrease in mitochondrial activity was observed when hMSCs were cultured the at 0.1% oxygen. For these reasons, we conclude that 0.1% oxygen tension best reflected the in vivo milieu to which the hMSCs located at the core of the hydrogel constructs were exposed. The 0.1% oxygen condition in vitro was used throughout the rest of the study. The Presence of Extracellular Glucose Is Essential for hMSC Long-Term Survival Under Near-Anoxia In Vitro After establishing an in vitro model reflecting the in vivo milieu experienced by hMSCs post-implantation, the main cellular metabolism and its related pathways (summarized in Fig. 3A) used by hMSCs when cultured under near-anoxia (0.1% O2) were investigated. Figure 3 Open in new tabDownload slide Extracellular glucose (glc) is necessary to ensure long-term survival of human mesenchymal stem cells (hMSCs) exposed to near-anoxia. (A): Schematic overview of the energy-related pathways (in blue) in cells along with pertinent inhibitors tested in this study (in red; 2-deoxy glc, sodium oxamate, antimycin A, and malonate). (B): Viability of hMSCs cultured under near-anoxia (0.1% O2) in a minimal medium supplemented or not with either glutamine (glut), serine (ser), pyruvate (pyr), or glc at a final concentration of 5 g/l for up to 14 days (n = 3). Results are expressed as means ± SD. *, p ≤ .05 versus the minimal medium cell group (two-way analysis of variance followed by Bonferroni's post hoc test). Abbreviations: AnA, antimycin A; 2DG, 2-deoxyglucose; Malo, Malonate; MCTs, Monocarboxylate transporters; NaOx, sodium oxamate; TCA, tricarboxylic acid. Figure 3 Open in new tabDownload slide Extracellular glucose (glc) is necessary to ensure long-term survival of human mesenchymal stem cells (hMSCs) exposed to near-anoxia. (A): Schematic overview of the energy-related pathways (in blue) in cells along with pertinent inhibitors tested in this study (in red; 2-deoxy glc, sodium oxamate, antimycin A, and malonate). (B): Viability of hMSCs cultured under near-anoxia (0.1% O2) in a minimal medium supplemented or not with either glutamine (glut), serine (ser), pyruvate (pyr), or glc at a final concentration of 5 g/l for up to 14 days (n = 3). Results are expressed as means ± SD. *, p ≤ .05 versus the minimal medium cell group (two-way analysis of variance followed by Bonferroni's post hoc test). Abbreviations: AnA, antimycin A; 2DG, 2-deoxyglucose; Malo, Malonate; MCTs, Monocarboxylate transporters; NaOx, sodium oxamate; TCA, tricarboxylic acid. The potential benefits of exogenous glc and other energy-related metabolites (specifically, glut, ser and pyr; all present in the αMEM media) were studied. The amino acids glut and ser can be metabolized through glutaminolysis and serinolysis processes, respectively (Fig. 3A). The pyr plays a pivotal role in hMSC metabolism as both the end-product of glycolysis and substrate of either acetyl-CoA or lactate (Fig. 3A). In this study, we used a minimal medium supplemented with either glc or one of the aforementioned energy-related metabolites. HMSCs, cultured either in the absence of nutrients or in the presence of glut or ser, exhibit similar survival rate with only less than 40% and 10% of hMSCs still viable after 3 and 7 days of exposure to near-anoxia, respectively (Fig. 3B). In contrast, in the presence of either glc or pyr, more than 60% of hMSCs remained viable at day 7 (Fig. 3B). However, cell viability in the presence of pyr then drastically decreased at day 14 (7%) while, in the presence of glc, the cell survival rate remained above 60% after 14 days of exposure to near-anoxia (Fig. 3B). Taken together, these data demonstrate that glc and, to a lesser extent, pyr appear to be the only exogenous metabolites used by hMSCs to sustain near-anoxia. Glycolysis Is the Preferred Energy-Providing Pathway Used by hMSCs to Ensure Their Survival in near Anoxia In human cells, glc can be obtained either from the extracellular milieu via glc transporters from the GLUT family or from intracellular glycolytic reserves including glycogen. Glucose is then catabolized either (a) via glycolysis into two molecules of pyr generating two molecules of ATP or (b) via the pentose phosphate pathway to generate NADP(H) and specialized products (Fig. 3A) .Then, pyr is converted either into lactate by lactate dehydrogenase or into acetyl-CoA to enter into the TCA cycle to produce ATP solely in the presence of oxygen (Fig. 3A). In this study, the gene expression profiles of metabolic enzymes involved in either the TCA cycle, the pentose phosphate, the glycogenolysis, or the glycolysis pathways in hMSCs cultured under near-anoxia for 3 days were assessed. A downregulation of the gene expression of the enzymes involved in the TCA (Fig. 4A), the pentose phosphate (Fig. 4B), and the glycogenolysis (Fig. 4C) pathways was observed. In contrast, under the same conditions, the expression of the enzymes involved in the glycolysis pathway was upregulated (Fig. 4D) indicating that glycolysis is the favorite pathway for energy supply when hMSCs are cultured in near-anoxia. Figure 4 Open in new tabDownload slide Glycolysis is the preferred energy-providing pathway used by human mesenchymal stem cells (hMSCs) to maintain their ATP content and ensure their survival under near-anoxia. Tricarboxylic acid (TCA) cycle (A), pentose phosphate (B), glycogen metabolism (C), and glycolysis (D) pathway-related gene expressions of hMSCs cultured under near anoxia (0.1% O2) for 3 days. Gene expression was normalized with respect to expression of the house-keeping genes expression of beta-2-microglobulin, ribosomal protein L13a, and beta-actin. Expression levels are presented in the 2(-ΔΔCT) format relative to those obtained when the hMSCs were maintained under 21% oxygen tension. Values are presented as means ± SD from three independent experiments. (E): Viability of hMSCs cultured under near-anoxia in a minimal medium (0% fetal bovine serum, 0 g/l glucose) in the absence or presence of either two TCA inhibitors (malonate or antimycin A) or of two glycolysis inhibitors (sodium oxamate or 2-deoxyglucose) for up to 14 days. Results are expressed as means ± SD (n = 6); *, p ≤ .05 versus the minimal medium cell group for each time point (two-way analysis of variance [ANOVA] followed by Bonferroni's post hoc test). (F): Time course of ATP content in hMSCs per well previously cultured in a minimal medium under near-anoxia for 24 hours and then exposed to 2DG for up to 3 hours. Results are expressed as means ± SD (n = 3); $, p ≤ .05 versus the near anoxia cell group (two-way ANOVA followed by Bonferroni's post hoc test). Abbreviations: ACLY, ATP citrate lyase; ACO, aconitase; AGL, amylo-alpha-1, 6-glucosidase, 4-alpha-glucanotransferase; ALDOB, ALDOC, aldolase-B and -C; CS, citrate synthase; 2DG, 2-deoxyglucose; DLAT, dihydrolipoamide S-acetyltransferase; DLD, dihydrolipoamide dehydrogenase; DLST, dihydrolipoamide S-succinyltransferase; ENO2, enolase 2; FH, fumarate hydratase; GBE1, glucan (1,4-alpha-) branching enzyme 1; G6PD, glucose-6-phosphate dehydrogenase; GPI, glucose-6-phosphate isomerase; GSK3A, GSK3B, glycogen synthase kinase 3 alpha and beta; GSY, glycogen synthase1 and 2 (muscle GYS1 and liver GSY2); HK3, hexokinase 3; H6PD, hexose-6-phosphate dehydrogenase; IDH, isocitrate dehydrogenase (soluble IDH3G); MDH2, malate dehydrogenase (soluble MDH1, MDH1B); OGDH, oxoglutarate (alpha-ketoglutarate) dehydrogenase; PC, pyruvate carboxyalse; PCK1, PCK2, phosphoenol pyruvate carboxykinase 1 and 2; PDHB, pyruvate dehydrogenase-B; PGAM, phosphoglycerate mutase 2 (muscle PGAM2); PGK2, phosphoglycerate kinase 2; PGLS, 6-phosphogluconolactonase; PGM1, PGM2, PGM3, phosphoglucomutase 1, 2, and 3; PHKA, phosphorylase kinase alpha, beta, gamma 1, and 2 (muscle PHKA, PHKB, muscle PHKG1, testis PHKG2); PKLR, pyruvate kinase, liver and RBC; PRPS1, PRPS1L1, PRPS2, phosphoribosyl pyrophosphate synthetase 1, 1-like 1 and 2; PYGL, phosphorylase, glycogen liver; RBE, RPIA, ribulose-5-phosphate-3-epimerase and isomerase A; RBKS, ribokinase; SDH, SDHD, succinate dehydrogenase complex D; SUCLA2, SUCLG, succinate-CoA ligase A2, -G; TALDO1, transaldolase 1; TKT, transketolase; UGP2, UDP-glucose pyrophosphorylase 2; TPI1, triosephosphate isomerase 1. Figure 4 Open in new tabDownload slide Glycolysis is the preferred energy-providing pathway used by human mesenchymal stem cells (hMSCs) to maintain their ATP content and ensure their survival under near-anoxia. Tricarboxylic acid (TCA) cycle (A), pentose phosphate (B), glycogen metabolism (C), and glycolysis (D) pathway-related gene expressions of hMSCs cultured under near anoxia (0.1% O2) for 3 days. Gene expression was normalized with respect to expression of the house-keeping genes expression of beta-2-microglobulin, ribosomal protein L13a, and beta-actin. Expression levels are presented in the 2(-ΔΔCT) format relative to those obtained when the hMSCs were maintained under 21% oxygen tension. Values are presented as means ± SD from three independent experiments. (E): Viability of hMSCs cultured under near-anoxia in a minimal medium (0% fetal bovine serum, 0 g/l glucose) in the absence or presence of either two TCA inhibitors (malonate or antimycin A) or of two glycolysis inhibitors (sodium oxamate or 2-deoxyglucose) for up to 14 days. Results are expressed as means ± SD (n = 6); *, p ≤ .05 versus the minimal medium cell group for each time point (two-way analysis of variance [ANOVA] followed by Bonferroni's post hoc test). (F): Time course of ATP content in hMSCs per well previously cultured in a minimal medium under near-anoxia for 24 hours and then exposed to 2DG for up to 3 hours. Results are expressed as means ± SD (n = 3); $, p ≤ .05 versus the near anoxia cell group (two-way ANOVA followed by Bonferroni's post hoc test). Abbreviations: ACLY, ATP citrate lyase; ACO, aconitase; AGL, amylo-alpha-1, 6-glucosidase, 4-alpha-glucanotransferase; ALDOB, ALDOC, aldolase-B and -C; CS, citrate synthase; 2DG, 2-deoxyglucose; DLAT, dihydrolipoamide S-acetyltransferase; DLD, dihydrolipoamide dehydrogenase; DLST, dihydrolipoamide S-succinyltransferase; ENO2, enolase 2; FH, fumarate hydratase; GBE1, glucan (1,4-alpha-) branching enzyme 1; G6PD, glucose-6-phosphate dehydrogenase; GPI, glucose-6-phosphate isomerase; GSK3A, GSK3B, glycogen synthase kinase 3 alpha and beta; GSY, glycogen synthase1 and 2 (muscle GYS1 and liver GSY2); HK3, hexokinase 3; H6PD, hexose-6-phosphate dehydrogenase; IDH, isocitrate dehydrogenase (soluble IDH3G); MDH2, malate dehydrogenase (soluble MDH1, MDH1B); OGDH, oxoglutarate (alpha-ketoglutarate) dehydrogenase; PC, pyruvate carboxyalse; PCK1, PCK2, phosphoenol pyruvate carboxykinase 1 and 2; PDHB, pyruvate dehydrogenase-B; PGAM, phosphoglycerate mutase 2 (muscle PGAM2); PGK2, phosphoglycerate kinase 2; PGLS, 6-phosphogluconolactonase; PGM1, PGM2, PGM3, phosphoglucomutase 1, 2, and 3; PHKA, phosphorylase kinase alpha, beta, gamma 1, and 2 (muscle PHKA, PHKB, muscle PHKG1, testis PHKG2); PKLR, pyruvate kinase, liver and RBC; PRPS1, PRPS1L1, PRPS2, phosphoribosyl pyrophosphate synthetase 1, 1-like 1 and 2; PYGL, phosphorylase, glycogen liver; RBE, RPIA, ribulose-5-phosphate-3-epimerase and isomerase A; RBKS, ribokinase; SDH, SDHD, succinate dehydrogenase complex D; SUCLA2, SUCLG, succinate-CoA ligase A2, -G; TALDO1, transaldolase 1; TKT, transketolase; UGP2, UDP-glucose pyrophosphorylase 2; TPI1, triosephosphate isomerase 1. We next sought to confirm that the glycolysis but not the TCA pathway was involved in hMSC survival by blocking these energy-related metabolic pathways using specific inhibitors displayed in red in Figure 3A. To this aim, hMSCs were exposed to near-anoxia in the presence or absence of two TCA inhibitors, specifically either Malo (a competitive inhibitor of the succinate which binds to the TCA enzyme succinate dehydrogenase) or AnA (an indirect TCA inhibitor which block the electron transport chain, thus impeding ATP production) or of two glycolysis inhibitors, specifically either NaOx (a structural analog of pyr which inhibits l(+)-lactate dehydrogenase) or 2-deoxy-d-glucose (2DG; a competitive inhibitor for the production of glucose-6-phosphate from glc at the phosphoglucoisomerase level). In the presence of either Malo or AnA, the hMSC viability was similar to that observed in the minimal medium after 1, 3, and 7 days of culture (Fig. 4E). In contrast, the addition of either 2DG or NaOx induced a drastic decrease in cell viability; in fact, there were no more viable cells after 3 days of culture (Fig. 4E) suggesting a critical role of glycolysis for ensuring hMSCs survival in near-anoxia. Cells need a constant supply of energy in the form of ATP to keep them alive. To further confirm that glycolysis is the main energy provider-pathway to hMSCs in near anoxia, the glycolysis pathway was blocked using 2DG in hMSCs previously cultured under near anoxia for 24 hours and the ATP content in cells was monitored over 3 hours. Addition of 2DG led to a complete depletion of ATP content within 5 minutes (Fig. 4F). Taken together, these data provided strong evidence that hMSCs heavily rely on glycolysis as their predominant energy-related pathway to maintain their ATP content and to ensure their survival under near-anoxia. Glucose Availability Does Not Dictate hMSC Glucose Consumption in Near-Anoxia Since hMSCs heavily rely on glc through glycolysis to fuel their metabolism under near-anoxia, we next investigated the potential regulation of hMSC glc consumption by glc availability. To this aim, we evaluated glc consumption by hMSCs cultured in the presence of either 0.25, 0.5, 1, or 5 g/l of glc for 7 days under near-anoxia. Under these conditions, glc consumption by hMSCs cultured in either low (0.25–0.5 g/l) or high (5 g/l) glc levels were not different throughout the duration of the near-anoxic episode despite a 10- to 20-fold increase in glc availability (Fig. 5). In contrast, under a physiological glc level (1 g/l), glc consumption by the hMSCs was decreased (–50%) when compared with the other cell groups (0.25, 0.5, and 5 g/l glc; Fig. 5). Taken together, these results suggest that glc consumption by hMSCs is not regulated by glc availability although the existence of an optimum at 1 g/l. Figure 5 Open in new tabDownload slide Glucose availability does not impact human mesenchymal stem cell (hMSC) glucose (glc) consumption in near anoxia. Evaluation of glc consumption by hMSCs cultured in alpha minimum essential medium (0% fetal bovine serum) supplemented with either 0.25, 0.5, 1, or 5g/l glc in near-anoxia (0.1% O2) for 7 days. Results are expressed as “pmol/day/cell” in means ± SD (n = 3); *, p ≤ .05 versus the 1 g/l cell group (one-way analysis of variance followed by Bonferroni's post hoc test). Figure 5 Open in new tabDownload slide Glucose availability does not impact human mesenchymal stem cell (hMSC) glucose (glc) consumption in near anoxia. Evaluation of glc consumption by hMSCs cultured in alpha minimum essential medium (0% fetal bovine serum) supplemented with either 0.25, 0.5, 1, or 5g/l glc in near-anoxia (0.1% O2) for 7 days. Results are expressed as “pmol/day/cell” in means ± SD (n = 3); *, p ≤ .05 versus the 1 g/l cell group (one-way analysis of variance followed by Bonferroni's post hoc test). HMSCs Do Not Possess the Necessary Glycolytic Reserves to Survive Without Exogenous Glucose Under Near-Anoxia in In Vitro and In Vivo Conditions The bioenergetic self-sufficiency of hMSCs is a major concern for their use in TE because hMSCs may have access to a limited exogenous glc supply upon implantation. To address this issue, we monitored the available exogenous glc levels in 2D in vitro hMSC-cultures. In addition, we evaluated the glycolytic reserves (specifically, intracellular glc and glycogen contents), and their effect on the ATP levels and cell viability of hMSCs either cultured in vitro (in 2D and 3D cultures) or implanted in vivo in nude mice. For the 2D in vitro cultures, hMSCs were cultured either in the absence of glc, or in a low (0.25 g/l) or physiological (1g/l) level of glc under near anoxia conditions in vitro. In the case of hMSCs in the 1 g/l cell-group, only 15% of the available exogenous glc was consumed throughout the 7-days of exposure to near-anoxia (Fig. 6A). A steady but slow decrease (–25% and −54% after 3- and 7-days of exposure, respectively) without complete exhaustion of the glycolytic reserves of hMSCs was observed (Fig. 6B) along with maintenance of sufficient ATP levels (–53% and −59% after 3- and 7-days, respectively; Fig. 6C) that permitted the hMSC survival in near-anoxia (Fig. 6D). In contrast, hMSCs exposed to low glc (0.25 g/l) exhibited a rapid and extensive exhaustion of the available exogenous glc (-48% at day 3 of culture; Fig. 6A) paralleled with exhaustion of their glycolytic reserves (–77% at day 3 of culture; Fig. 6B) and ATP levels (–49% at day 3; Fig.6C). Under this condition, the hMSCs failed to maintain a sufficient ATP level to ensure their survival under near-anoxia for 7 days (Fig. 6C, 6D). When cultured in the absence of glc, hMSCs depleted their glycolytic reserves in less than 24 hours (Fig. 6B), their ATP content within 3 days of culture (Fig. 6C), and massively died (–92%) within the first 3 days of exposure to near anoxia (Fig. 6D). Figure 6 Open in new tabDownload slide The glycolytic reserves of human mesenchymal stem cells (hMSCs) are insufficient to maintain cell viability under near anoxia in vitro. Time-courses of (A) exogenous glucose (glc) levels in the supernatant media, (B) glycolytic reserves (including both intracellular glycogen and glc contents) per cell, (C) ATP content per cell and (D) cell viability (expressed as a percentage of the viable cell number obtained at day 0) in hMSCs cultured in alpha minimum essential medium (0% fetal bovine serum) supplemented or not with either 0.25 or 1 g/l of glc in near-anoxia (0.1% O2) for 7 days. Results are expressed as means ± SD (n = 3); *, p ≤ .05 versus the near anoxia cell group; $, p ≤ .05 versus the near anoxia + glc (1 g/l) cell group (two-way analysis of variance followed by Bonferroni's post hoc test). Figure 6 Open in new tabDownload slide The glycolytic reserves of human mesenchymal stem cells (hMSCs) are insufficient to maintain cell viability under near anoxia in vitro. Time-courses of (A) exogenous glucose (glc) levels in the supernatant media, (B) glycolytic reserves (including both intracellular glycogen and glc contents) per cell, (C) ATP content per cell and (D) cell viability (expressed as a percentage of the viable cell number obtained at day 0) in hMSCs cultured in alpha minimum essential medium (0% fetal bovine serum) supplemented or not with either 0.25 or 1 g/l of glc in near-anoxia (0.1% O2) for 7 days. Results are expressed as means ± SD (n = 3); *, p ≤ .05 versus the near anoxia cell group; $, p ≤ .05 versus the near anoxia + glc (1 g/l) cell group (two-way analysis of variance followed by Bonferroni's post hoc test). Similarly to what was observed with the 2D cultures, hMSCs loaded into fibrin hydrogels (3D culture) and cultured in the presence of glc (1g/l), did not exhaust their glycolytic reserves (Fig. 7A), and maintained a low, but sufficient, ATP level (Fig. 7B). These energy related conditions allowed hMSCs to survive at survival rate of 77% under near-anoxia for 7 days (Fig. 7C). In contrast, in the absence of extracellular glc, hMSCs loaded in fibrin hydrogels, rapidly consumed (–83%) their glycolytic reserves within the first 24 hours of culture (Fig. 7A), failed to maintain their ATP levels for more than 3 days of culture (Fig. 7B) and died massively (-91%) within the first 3 days of exposure to near-anoxia (Fig. 7C). Figure 7 Open in new tabDownload slide Quick exhaustion of human mesenchymal stem cells (hMSCs) energy reserves when loaded within hydrogels explains their rapid death post-implantation. Time-courses of (A) glycolytic reserves, (B) ATP content per hydrogel, and (C) cell viability (expressed as a percentage of the viable cell number obtained in the hydrogel at day 0) of hMSCs loaded within fibrin hydrogels (105 cells in 100 µl hydrogels) and cultured in alpha minimum essential medium (0% fetal bovine serum) supplemented or not with 1 g/l of glucose in near-anoxia (0.1% O2) for 7 days. Results are expressed as means ± SD (n = 3); $, p ≤ .05 versus the near anoxia cell group (two-way analysis of variance followed by Bonferroni's post hoc test). Time-courses of (D) glycolytic reserves per implant, (E) ATP content per implant, and (F) cell viability (expressed as a percentage of the viable cell number present in the hydrogel at day 0) of hMSCs loaded within fibrin hydrogel (2.105 cells in 200 µl hydrogel) in diffusion chambers and subcutaneously implanted in nude mice for up to 7 days. Results are expressed as means ± SD (n = 3); *, p ≤ .05 versus implanted hMSCs d0 (Mann-Whitney). Abbreviation: hMSC, human mesenchymal stem cell. Figure 7 Open in new tabDownload slide Quick exhaustion of human mesenchymal stem cells (hMSCs) energy reserves when loaded within hydrogels explains their rapid death post-implantation. Time-courses of (A) glycolytic reserves, (B) ATP content per hydrogel, and (C) cell viability (expressed as a percentage of the viable cell number obtained in the hydrogel at day 0) of hMSCs loaded within fibrin hydrogels (105 cells in 100 µl hydrogels) and cultured in alpha minimum essential medium (0% fetal bovine serum) supplemented or not with 1 g/l of glucose in near-anoxia (0.1% O2) for 7 days. Results are expressed as means ± SD (n = 3); $, p ≤ .05 versus the near anoxia cell group (two-way analysis of variance followed by Bonferroni's post hoc test). Time-courses of (D) glycolytic reserves per implant, (E) ATP content per implant, and (F) cell viability (expressed as a percentage of the viable cell number present in the hydrogel at day 0) of hMSCs loaded within fibrin hydrogel (2.105 cells in 200 µl hydrogel) in diffusion chambers and subcutaneously implanted in nude mice for up to 7 days. Results are expressed as means ± SD (n = 3); *, p ≤ .05 versus implanted hMSCs d0 (Mann-Whitney). Abbreviation: hMSC, human mesenchymal stem cell. For the in vivo experiments, hMSCs were loaded in fibrin hydrogels and loaded into Millipore diffusion chambers (to avoid infiltration of host cells in hydrogel constructs), and subcutaneously implanted in nude mice for up to 7 days. In a trend similar to that observed in the in vitro data, a rapid decrease (–66%) of hMSC glycolytic reserves was observed during the first 24 hours post-implantation and paralleled a rapid decrease in the ATP content of these hMSCs: specifically, 40% of the ATP stock was lost within 24 hours, and 63% within 3 days; no more ATP was detected at day 7 post-implantation (Fig. 7E). As a result, the implanted hMSCs died massively (–94%) during the first 3 days of implantation (Fig. 7F). Taken together, these data demonstrate that, as long as a glc supply is available, hMSCs are able to survive in an oxygen-deprived environment (such as those the hMSCs encountered in our in vitro, near-anoxic models or in vivo when they were loaded in the diffusion chambers and implanted subcutaneously in nude mice) by maintaining their glycolytic reserves and ATP contents above a threshold value that ensures cell-survival. In the absence of a glc supply, however, hMSCs experienced an almost immediate exhaustion of their glycolytic reserves and, consequently, depletion of intracellular ATP that ultimately caused their death. Discussion Upon implantation, survival and maintenance of functional hMSCs in the 3D environment of TE constructs before their vascularization is the most important challenge for developing TE constructs of clinically-relevant volume for regenerative medicine applications. The traditional view maintains that the delay in the neovascularisation of TE constructs post-implantation results in insufficient oxygenation which causes ischemia within those implants and, consequently, leads to cell dysfunction including premature cell death. Our research group challenged this paradigm and established that severe oxygen depletion per se is not responsible for the massive hMSC death observed post-implantation. In fact, hMSCs can withstand exposure to severe, continuous, near-anoxia (i.e., 0.1% oxygen), provided that glc is available for the duration of the experiments [18]. In this study, we established that 0.1% oxygen best reflected the in vivo post-implantation milieu. Under this condition, and with the exception of glycolysis (which was the main bioenergetic pathway that promoted ATP resources), other energy-related pathways (including the TCA, the pentose phosphate, and the glycogenolysis pathways) were downregulated. Most importantly, we demonstrated for the first time that rapid exhaustion of intracellular energy reserves (specifically, glc and glycogen) evidenced very low energy reserves within hMSCs and explained the poor hMSC survival observed in near anoxia in vitro and in vivo after subcutaneous implantation of hMSC-containing constructs in a mouse model. Cell Culture Under 0.1% Oxygen Conditions Best Reflects the In Vivo Post-Implantation Milieu Mammalian tissues function in environments where oxygen tension ranges from 2% to 10% [23-25]; for instance, the oxygen tension in the MSC niche in the bone marrow ranges from 2% to 8% (for review, Mohyeldin et al. and; Spencer et al. [26, 27]). Despite its critical role in the ability of hMSCs to support the energy-demanding processes of stem cell homeostasis, survival and ultimately functionality post-transplantation, the oxygen tension to which hMSCs loaded into TE constructs are exposed during the early post-implantation phase in vivo remains unknown. In this study, we determined, for the first time, that hMSCs loaded in TE constructs of limited size (100 µl—4 mm diameter) experienced oxygen tension of 0.13% ± 0.06% 24 hours post-implantation. This oxygen tension is lower than the one measured in vivo in soft-tissue ischemia in mice [28] and in human tumors [29] where it ranges from 0.3% to 4.2%. A possible explanation for this extremely low oxygen level is that blood supply is totally nonexistent within the hMSC-containing constructs during the immediate post-implantation phase. Whether these findings can be generalized to all types of cell-containing constructs await further measurements of oxygen levels in other TE constructs. On one hand, one must resist the temptation to draw broad conclusions regarding the oxygen tension encountered by hMSCs in the context of TE applications as oxygen delivery to a tissue construct is a function of diffusion, which in turn is controlled by material scaffold porosity and permeability (for review Karande et al. [30]). On the other hand, the data of this study are consistent with the 100–200 µm range-limits for oxygen diffusion in mammalian tissues and it is the authors’ belief that the oxygen tensions measured in this study best reflect the in vivo situation experienced by hMSCs located at the core of TE cell-containing constructs with dimensions bigger than the aforementioned diffusion range-limit of oxygen. Because it is challenging to determine the accurate oxygen tension that a cell might experience in vivo, we attempted to further validate that the 0.1% oxygen tension best reflected the in vivo situation by assessing the expression of select hallmarks of hypoxia both in vivo in hMSCs dispersed within a fibrin gel and implanted ectopically in nude mice for 3 days and in vitro after exposure of hMSCs to oxygen tension ranging from 21% to 0.1%. After implantation, the hMSCs located at the core of the hydrogel constructs expressed the hypoxic markers HIF-1α and LDH-A (Fig. 1C), and most importantly, exhibited a highly reduced mitochondrial activity (Fig. 1E). In the in vitro experiments, 0.1% oxygen was the only oxygen tension tested which both induced expression of hypoxic markers and massively downregulated the mitochondrial activity. These results indicated that the 0.1% oxygen tension best reflected the in vivo situation encountered by hMSCs located at the core of TE cell-containing constructs with dimensions bigger than the diffusion range-limit of oxygen. Long-Term hMSC Survival Depends on the Use of Glucose via Glycolysis Cells are capable of metabolizing various carbon substrates, including glc, fatty acids, ketone bodies, and amino acids. In this study, hMSCs exposed to long-term (14 days) near-anoxia (0.1% O2) conditions survived only in the presence of exogenous glc (but not of pyr, glut or ser); these results confirmed, and extended, previous studies which reported that the presence of glc enhanced rat [31], sheep [16], and human MSC [18] viability when these cells were exposed to low (0.1%–1%) oxygen tension. These results are also in line with the observations according to which hypoxia-preconditioning of hMSCs led to decreased glc consumption, possibly promoting cell retention in vivo, since glc remained available to the cells for longer period of time [32]. Most importantly, this study provided evidence for the first time that glycolysis (but not pentose phosphate, TCA, and glycogenolysis) is the unique metabolic pathway involved in hMSC survival under near-anoxia (Fig. 4D–4F). Interestingly, hMSCs exposed to near anoxia for 7 days of culture survived in the presence of pyr in a glc-deprived milieu. In addition, inhibition of pyr conversion into lactate using NaOx (a l(+)-lactate dehydrogenase inhibitor) strongly reduced hMSC survival (–82%) when cultured under near anoxia in a glc- and pyr-deprived milieu (Fig. 4E). A possible explanation for the beneficial role of pyr in hMSC survival is that, under near anoxia, pyr is used to produce cofactors, such as NAD(H), which catalyse ATP production through glycolysis [9]; this process is limited by the availability of cofactors such as NAD(H) [33]. By improving ATP production, pyr could temporarily enhance hMSC viability in the near-anoxia milieu. Nevertheless, definitive conclusions regarding the roles of pyr, glut and ser await studies using a panel of specific inhibitors of these pathways in pertinent animal models. In addition to glycolysis, autophagy can furnish ATP and energy-related substrate to the cells. Although it was out of the scope of this study to explore in detail the role of autophagy in hMSCs, we observed that inhibition of autophagy had a deleterious impact on hMSC survival under near-anoxia independently of the presence of exogenous glc (Supporting Information Fig. S1) confirming that maintenance of a high autophagic activity is crucial for hMSC survival under near anoxia [22]. HMSCs Fail to Adapt Their Glucose Consumption to Exogenous Glucose Availability and Have Virtually No Internal Energy Reserves This study provided evidence that hMSCs failed to survive in vitro longer than 3 days of culture under conditions simulating the in vivo post-implantation milieu; these results corroborated previous studies by our research group [16, 21] and by other researchers [34]. Since glc availability to cells is limited after implantation of TE cell-containing constructs, we further investigated the causes of hMSC death by monitoring their energy resources, particularly the ATP content as well as the intracellular glc and glycogen stocks using in vitro and in vivo models (Figs. 5, 6B, 6C, 7B–7E, respectively). In vitro, using the 2D cell culture models, we observed (a) that the hMSCs are unable to modulate glc consumption in response to glc availability (Fig. 5) and (b) that the successive exhaustion of glycolytic reserves and of the ATP content is responsible for the observed hMSC poor survival in near-anoxia (Fig. 6). In other words, because hMSC autonomy (in terms of energy-related metabolites) is limited in a glc-deprived environment, an exogenous glc supply is required to maintain hMSC viability under near anoxia over 3 days. When loaded in 3D hydrogel constructs, the hMSCs displayed similar behavior than the one observed in the 2D cell culture models with a concomitant exhaustion of the glycolytic reserves (Fig. 7A) and ATP content (Fig. 7B) resulting in massive cell death (Fig. 7C). It is worth to mention that, the kinetics of exhaustion of glycolytic reserves in 2D and 3D cultures present a slight discrepancy albeit a similar overall pattern. In fact, in the absence of exogenous glc, in 2D, a full exhaustion of glycolytic reserves is observed within a day (Fig. 6B) whereas in 3D, glycolytic reserves are still present after one day, albeit massively reduced (-83%; Fig. 7A). Nevertheless, these depletions of the glycolytic reserves resulted in a drastic loss of cell viability in both culture conditions at day 3 and day 7 (Figs. 6D, 7C). These in vitro results were corroborated by the in vivo studies which demonstrated similar decrease in cell viability (Fig. 7F) and ATP content (Fig 7E) within 3 days post-implantation. In addition, a three-fold decrease of glycogen and glc stocks per hydrogel constructs was observed in vivo within the first day of implantation (Fig. 7F); these energy stocks were not sufficient to ensure cell viability at day 3 and 7. The critical role of glycolysis in hMSC survival was further substantiated by the facts that the addition of 2DG inside the hydrogel construct led to massive cell death (Supporting Information Fig. S2); and that supplementation of hydrogel constructs with glc increased hMSCs survival post-implantation in an ectopic nude mice model (Deschepper et al. [18]). A possible explanation for the incomplete exhaustion of glc and glycogen stocks in the hydrogels implanted in vivo is that passive diffusion of glc from the surrounding tissues may have occurred. Further studies are required to definitively establish this explanation. Limitations of This Study A limitation of the in vitro models and as well as of the in vivo closed system used in this study is that they exclude the impact of the in vivo inflammatory response of the anatomical site and extends of the tissue injury on hMSC survival. Nevertheless, because hMSCs died within 3 days in the closed system (which does not stimulate the cellular aspects of the immune response), it is tempting to speculate a prime role of the rapid depletion for energy reserves in the observed massive cell death post-implantation. Perspectives for TE Strategies The findings of this study have implications for the design of future protocols of preconditioning hMSCs to enhance their in vivo survival [35, 36]. In fact, the new information regarding the metabolic adaptation of hMSCs under near-anoxia conditions may be exploited in developing delivery systems to slowly release exogenous nutrients (glc should be the primary choice) to transplanted hMSCs. It is becoming increasingly clear that, when metabolic perturbations are excessively severe or of protracted duration (such as the conditions encountered by hMSCs post-implantation), metabolic checkpoints may initiate either apoptotic or necrotic cell death pathways [37]. Improved understanding of the metabolic checkpoints involved in hMSC survival/death post-implantation may facilitate development of novel pharmacological approaches that selectively either block or stimulate cell death by inducing specific metabolic changes. The results of this study provide evidence that the metabolic checkpoints involved in monitoring ATP levels might be relevant targets for novel TE strategies. Conclusion In this study, we determined, for the first time, that the 0.1% oxygen culture condition best reflected the in vivo situation encountered by hMSCs located at the core of TE hydrogel constructs with dimensions bigger than 200 μm (the diffusion range limit of oxygen). This new scientific information has practical implications because cell energy metabolism is modulated by the oxygen tension (specifically, the TCA pathway is drastically down-regulated in 0.1% but not at 1% O2 environments) and has been implicated in cell fate decision (for review, see Green and Llambi, 2015 [38]). Most importantly, this study established that cell energy was produced via glycolysis only and that, under near-anoxia (0.1% O2) conditions, the cellular energy reserves were exhausted resulting in hMSC death within the first three days post-implantation. In this respect, development of strategies aiming at providing extracellular glc supplies is an attractive, but yet untapped, alternative to enhance both hMSC viability and functions pertinent to neo-tissue formation. Acknowledgments We thank Prof. R. Bizios and E. Potier for valuable comments regarding this article. This work was supported by ANR VIASTEM, VITABONE, and ANR IPSOAT. Author Contributions A.M. and J.P.: conception and design, collection and assembly of data, data analysis and interpretation, manuscript writing; M.D.: conception and design, data analysis and interpretation; N.L., K.O., C.D., and M.B.: provision of study material; D.L.-A.: data analysis and interpretation; H.P.: conception and design, data analysis and interpretation, financial support, manuscript writing, final approval of manuscript. Disclosure of Potential Conflicts of Interest The authors indicated no potential conflicts of interest. References 1 Caplan A. Mesenchymal stem cells . J Orthop Res 1991 ; 9 : 641 – 650 . Google Scholar Crossref Search ADS WorldCat 2 Owen M , Friedenstein AJ. Stromal stem cells: Marrow-derived osteogenic precursors . Ciba Found Symp 1988 ; 136 : 42 – 60 . Google Scholar OpenURL Placeholder Text WorldCat 3 Friedman MS , Long MW, Hankenson KD. Osteogenic differentiation of human mesenchymal stem cells is regulated by bone morphogenetic protein-6 . J Cell Biochem 2006 ; 98 : 538 – 554 . Google Scholar Crossref Search ADS WorldCat 4 Logeart-Avramoglou D , Oudina K, Bourguignon M et al. . In vitro and in vivo bioluminescent quantification of viable stem cells in engineered constructs . Tissue Eng Part C Methods 2009 ; 15 : 447 – 458 . Google Scholar OpenURL Placeholder Text WorldCat 5 Manassero M , Paquet J, Deschepper M et al. . Comparison of survival and osteogenic ability of human mesenchymal stem cells in orthotopic and ectopic sites in mice . Tissue Eng Part A 2016 ; 22 : 534 – 544 . Google Scholar Crossref Search ADS WorldCat 6 Terrovitis JV , Smith RR, Marbán E. Assessment and optimization of cell engraftment after transplantation into the heart . Circ Res 2010 ; 106 : 479 – 494 . Google Scholar Crossref Search ADS WorldCat 7 Haider HK , Ashraf M. Strategies to promote donor cell survival: Combining preconditioning approach with stem cell transplantation . J Mol Cell Cardiol 2008 ; 45 : 554 – 566 . Google Scholar Crossref Search ADS WorldCat 8 Togel F , Yang Y, Zhang P et al. . Bioluminescence imaging to monitor the in vivo distribution of administered mesenchymal stem cells in acute kidney injury . Am J Physiol Renal Physiol 2008 ; 295 : F315 – F321 . Google Scholar Crossref Search ADS WorldCat 9 Lunt SY , Vander Heiden MG. Aerobic glycolysis: Meeting the metabolic requirements of cell proliferation . Annu Rev Cell Dev Biol 2011 ; 27 : 441 – 464 . Google Scholar Crossref Search ADS WorldCat 10 Lash GE , Fitzpatrick TE, Graham CH. Effect of hypoxia on cellular adhesion to vitronectin and fibronectin . Biochem Biophys Res Commun 2001 ; 287 : 622 – 629 . Google Scholar Crossref Search ADS WorldCat 11 Grayson WL , Zhao F, Izadpanah R et al. Effects of hypoxia on human mesenchymal stem cell expansion and plasticity in 3D constructs. J Cell Physiol 2006 ; 207 : 331 – 339 . 12 Potier E , Ferreira E, Andriamanalijaona R et al. . Hypoxia affects mesenchymal stromal cell osteogenic differentiation and angiogenic factor expression . Bone 2007 ; 40 : 1078 – 1087 . Google Scholar Crossref Search ADS WorldCat 13 Fehrer C , Brunauer R, Laschober GT et al. . Reduced oxygen tension attenuates differentiation capacity of human mesenchymal stem cells and prolongs their lifespan . Aging Cell 2007 ; 6 : 745 – 757 . Google Scholar Crossref Search ADS WorldCat 14 Thomlinson RH , Gray LH. The histological structure of some human lung cancers and the possible implications for radiotherapy . Br J Cancer 1955 ; 9 : 539 – 549 . Google Scholar Crossref Search ADS WorldCat 15 Carmeliet P. Mechanisms of angiogenesis and arteriogenesis . Nat Med 2000 ; 6 : 389 – 395 . Google Scholar Crossref Search ADS WorldCat 16 Deschepper M , Oudina K, David B et al. . Survival and function of mesenchymal stem cells (MSCs) depend on glucose to overcome exposure to long-term, severe and continuous hypoxia . J Cell Mol Med 2011 ; 15 : 1505 – 1514 . Google Scholar Crossref Search ADS WorldCat 17 De Meester C , Timmermans AD, Balteau M et al. . Role of AMP-activated protein kinase in regulating hypoxic survival and proliferation of mesenchymal stem cells . Cardiovasc Res 2014 ; 101 : 20 – 29 . Google Scholar Crossref Search ADS WorldCat 18 Deschepper M , Manassero M, Oudina K et al. . Proangiogenic and prosurvival functions of glucose in human mesenchymal stem cells upon transplantation . Stem Cells 2013 ; 31 : 526 – 535 . Google Scholar Crossref Search ADS WorldCat 19 Gnecchi M , Melo LG. Bone marrow-derived mesenchymal stem cells: Isolation, expansion, characterization, viral transduction, and production of conditioned medium . Stem Cells Regen Med Methods Protoc 2009 ; 482 : 281 – 294 . Google Scholar OpenURL Placeholder Text WorldCat 20 Mosmann T. Rapid colorimetric assay for cellular growth and survival: Application to proliferation and cytotoxicity assays . J Immunol Methods 1983 ; 65 : 55 – 63 . Google Scholar Crossref Search ADS WorldCat 21 Becquart P , Cambon-Binder A, Monfoulet LE et al. . Ischemia is the prime but not the only cause of human multipotent stromal cell death in tissue-engineered constructs in vivo . Tissue Eng Part A 2012 ; 18 : 2084 – 2094 . Google Scholar Crossref Search ADS WorldCat 22 Moya A , Larochette N, Paquet J et al. . Quiescence preconditioned human multipotent stromal cells adopt a metabolic profile favorable for enhanced survival under ischemia . Stem Cells 2017 ; 35 : 181 – 196 . Google Scholar Crossref Search ADS WorldCat 23 Okazaki K , Maltepe E. Oxygen, epigenetics and stem cell fate . Regen Med 2006 ; 1 : 71 – 83 . Google Scholar Crossref Search ADS WorldCat 24 Brahimi-Horn MC , Chiche J, Pouysségur J. Hypoxia and cancer . J Mol Med 2007 ; 85 : 1301 – 1307 . Google Scholar Crossref Search ADS WorldCat 25 Brahimi-Horn MC , Pouysségur J. Oxygen, a source of life and stress . FEBS Lett 2007 ; 581 : 3582 – 3591 . Google Scholar Crossref Search ADS WorldCat 26 Mohyeldin A , Garzón-Muvdi T, Quiñones-Hinojosa A. Oxygen in stem cell biology: A critical component of the stem cell niche . Cell Stem Cell 2010 ; 7 : 150 – 161 . Google Scholar Crossref Search ADS WorldCat 27 Spencer JA , Ferraro F, Roussakis E et al. . Direct measurement of local oxygen concentration in the bone marrow of live animals . Nature 2014 ; 508 : 269 – 273 . Google Scholar Crossref Search ADS WorldCat 28 Ceradini DJ , Kulkarni AR, Callaghan MJ et al. . Progenitor cell trafficking is regulated by hypoxic gradients through HIF-1 induction of SDF-1 . Nat Med 2004 ; 10 : 858 – 864 . Google Scholar Crossref Search ADS WorldCat 29 McKeown SR. Defining normoxia, physoxia and hypoxia in tumours-implications for treatment response . Br J Radiol 2014 ; 87 : 20130676 . Google Scholar Crossref Search ADS WorldCat 30 Karande TS , Ong JL, Agrawal CM. Diffusion in musculoskeletal tissue engineering scaffolds: Design issues related to porosity, permeability, architecture, and nutrient mixing . Ann Biomed Eng 2004 ; 32 : 1728 – 1743 . Google Scholar Crossref Search ADS WorldCat 31 Mylotte L. A , Duffy AM, Murphy M et al. . Metabolic flexibility permits mesenchymal stem cell survival in an ischemic environment . Stem Cells 2008 ; 26 : 1325 – 1336 . Google Scholar Crossref Search ADS WorldCat 32 Beegle J , Lakatos K, Kalomoiris S et al. . Hypoxic preconditioning of mesenchymal stromal cells induces metabolic changes, enhances survival, and promotes cell retention in vivo . Stem Cells 2015 ; 33 : 1818 – 1828 . Google Scholar Crossref Search ADS WorldCat 33 Maldonado EN , Lemasters JJ. ATP/ADP ratio, the missed connection between mitochondria and the Warburg effect . Mitochondrion 2014 ; 19 : 78 – 84 . Google Scholar Crossref Search ADS WorldCat 34 Muschler GF , Nakamoto C, Griffith LG. Engineering principles of clinical cell-based tissue engineering . J Bone Joint Surg 2004 ; 86 : 1541 – 1558 . Google Scholar Crossref Search ADS WorldCat 35 Estrada JC , Albo C, Benguría A et al. . Culture of human mesenchymal stem cells at low oxygen tension improves growth and genetic stability by activating glycolysis . Cell Death Differ 2012 ; 19 : 743 – 755 . Google Scholar Crossref Search ADS WorldCat 36 Rosová I , Dao M, Capoccia B et al. . Hypoxic preconditioning results in increased motility and improved therapeutic potential of human mesenchymal stem cells . Stem Cells 2008 ; 26 : 2173 – 2182 . Google Scholar Crossref Search ADS WorldCat 37 Green DR , Galluzzi L, Kroemer G. Metabolic control of cell death . Science 2014 ; 345 : 1250256–1250256 . Google Scholar Crossref Search ADS WorldCat 38 Green DR , Llambi F. Cell death signaling . Cold Spring Harb Perspect Biol 2015 ; 7 : a006080 . Google Scholar Crossref Search ADS WorldCat © 2017 AlphaMed Press This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/open_access/funder_policies/chorus/standard_publication_model)
TI - Human Mesenchymal Stem Cell Failure to Adapt to Glucose Shortage and Rapidly Use Intracellular Energy Reserves Through Glycolysis Explains Poor Cell Survival After Implantation
JF - Stem Cells
DO - 10.1002/stem.2763
DA - 2018-03-01
UR - https://www.deepdyve.com/lp/oxford-university-press/human-mesenchymal-stem-cell-failure-to-adapt-to-glucose-shortage-and-1A8oM0KXtc
SP - 363
EP - 376
VL - 36
IS - 3
DP - DeepDyve
ER -