Angiogenic Effects of Human Multipotent Stromal Cell Conditioned Medium Activate the PI3K-Akt Pathway in Hypoxic Endothelial Cells to Inhibit Apoptosis, Increase Survival, and Stimulate Angiogenesis

Authors

  • Shih-Chieh Hung Ph.D., M.D.,

    Corresponding author
    1. Center for Gene Therapy, Tulane University Health Science Center, New Orleans, Louisiana, USA
    2. Stem Cell Laboratory, Veterans General Hospital-Taipei, Taipei, Taiwan
    3. Institute of Clinical Medicine, National Yang-Ming University, Taipei, Taiwan
    • Department of Medical Research and Education, Veteran General Hospital-Taipei, 201, Sec2, Shih-Pai Road, Taipei, Taiwan. Telephone: 886-2-28757396; Fax: 886-2-28757396
    Search for more papers by this author
  • Radhika R. Pochampally,

    1. Center for Gene Therapy, Tulane University Health Science Center, New Orleans, Louisiana, USA
    Search for more papers by this author
  • Sy-Chi Chen,

    1. Stem Cell Laboratory, Veterans General Hospital-Taipei, Taipei, Taiwan
    Search for more papers by this author
  • Shu-Ching Hsu,

    1. Center for Gene Therapy, Tulane University Health Science Center, New Orleans, Louisiana, USA
    Search for more papers by this author
  • Darwin J. Prockop Ph.D., M.D.

    Corresponding author
    1. Center for Gene Therapy, Tulane University Health Science Center, New Orleans, Louisiana, USA
    • Center for Gene Therapy, Tulane University Health Sciences Center, 1430 Tulane Avenue, SL-99, New Orleans, Louisiana 70112, USA. Telephone: (504) 988-7711; Fax: (504) 988-7710
    Search for more papers by this author

Abstract

Recent reports indicated that vascular remodeling and angiogenesis are promoted by conditioned medium from the cells referred to as multipotent stromal cells (MSCs). However, the molecular events triggered by MSC-conditioned medium (CdM) were not defined. We examined the effects of CdM from human MSCs on cultures of primary human aortic endothelial cells (HAECs). The CdM inhibited hypoxia-induced apoptosis and cell death of HAECs. It also promoted tube formation by HAECs in an assay in vitro. Conditioned medium from multipotent stromal cells incubated under hypoxic conditions in serum-free endothelial basal medium for 2 days (CdMHyp) from hypoxic culture of MSCs was more effective than conditioned medium from MSCs incubated under normoxic conditions in serum-free endothelial basal medium for 2 days from normoxic cultures of MSCs, an observation in part explained by its higher content of antiapoptotic and angiogenic factors, such as interleukin (IL)-6, vascular endothelial growth factor (VEGF), and monocyte chemoattractant protein (MCP)-1. The effects of CdMHyp on hypoxic HAECs were partially duplicated by the addition of IL-6 in a dose-dependent manner; however, anti-IL-6, anti-MCP-1, and anti-VEGF blocking antibodies added independently did not attenuate the effects. Also, addition of CdMHyp activated the PI3K-Akt pathway; the levels of p-Akt and several of its downstream targets were increased by CdMHyp, and both the increase in p-Akt and the increase in angiogenesis were blocked by an inhibitor of PI3K-Akt or by expression of a dominant negative gene for PI3K. CdMHyp also increased the levels of p-extracellular signal-regulated kinase (ERK), but there was a minimal effect on p-signal transducer and activator of transcription-3, and an inhibitor of the ERK1/2 pathway had no effect on hypoxia-induced apoptosis of the HAECs. The results are consistent with suggestions that administration of MSCs or factors secreted by MSCs may provide a therapeutic method of decreasing apoptosis and enhancing angiogenesis.

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

Introduction

The plastic adherent cells from bone marrow referred to as mesenchymal stem cells or multipotent stromal cells (MSCs) are capable of self-renewing and have the potential to differentiate into mesenchymal and nonmesenchymal tissues [1]. After systemic administration, they home to injured tissues such as heart [2], brain [3], and skeletal tissues [4] and appear to enhance regeneration of the tissues. MSCs were tested in animal models for a number of diseases and in clinical trials for osteogenesis imperfecta [5], an autosomal recessive mucopolysaccharidosis (Hurler syndrome) [6], and myocardial ischemia [7]. Promising results were reported both in animal models and in clinical trials, but the level of engraftment and differentiation of the cells was generally low. For example, i.v. administration of MSCs stimulated revascularization of ischemic tissues [2, 8], and small numbers of the cells were found to differentiate into both cardiomyocytes [2] and endothelial cells [9]. However, the effects in ischemic limb [10] and myocardial infarction [11, 12] were recently attributed to endocrine or paracrine factors produced by MSCs instead of differentiation of the cells. For these reasons, we have examined the effects of conditioned medium (CdM) from human MSCs on human endothelial cells exposed to hypoxia.

Materials and Methods

Cell Culture Conditions

Frozen vials of previously standardized passage 1 MSCs [13] from three normal human volunteers were obtained from the Tulane Center for Preparation and Distribution of Adult Stem Cells (http://www.som.tulane.edu/gene_therapy/distribut.shtml). The cells (approximately 1 million) were plated on a 15-cm diameter plate (Nalgene Nunc International, Rochester, NY, http://www.nalgenunc.com) in complete culture medium (α-minimal essential medium; Gibco-BRL, Gaithersburg, MD, http://www.gibcobrl.com), 17% fetal bovine serum, lot selected for rapid growth of MSCs (Atlanta Biologicals Inc., Norcross, GA, http://www.atlantabio.com), 100 units/ml penicillin, 100 μg/ml streptomycin, and 2 mM l-glutamine (Gibco-BRL). After incubation for 24 hours, the viable adherent cells were lifted by incubation with 0.25% trypsin and 1 mM EDTA at 37°C for approximately 5 minutes. The cells were recovered by centrifugation and replated at 100 cells per cm2, incubated in complete growth medium (CCM) with a change in medium every 2–3 days, and recovered as passage 2 cells after they reached approximately 80% confluency in 8–9 days. For some experiments, the cells were further expanded under the same conditions and used at passage 3 or 4. Human aortic endothelial cells (HAECs) harvested from the aorta of two normal males were obtained from a commercial source (catalog number 304-05a; HAOEC; Cell Applications Inc, Bath, U.K., http://www.cellapplications.com). They were positive for factor VIII and Dil-Ac-LDL uptake. Frozen vials of the cells were plated at 5,000 cells per cm2 and cultured in endothelial growth medium (catalog number 211-500; Cell Applications, Inc.) that contained 2% fetal bovine serum and referred to here as eGM. When the cells reached 80%–90% confluency, they were lifted with trypsin/EDTA, diluted 1:3, and expanded to passages 3–6 under the same conditions.

Culture Under Hypoxic Conditions

Both the MSCs and HAECs were expanded under standard normoxic conditions for cell culture at 37°C in a humidified incubator containing 21% O2, 5% CO2, and 74% N2. To examine hypoxia-induced apoptosis of MSCs, the cells were plated at 2 × 104 cells per well in gelatin-coated 96-well microtiter plates (Nalgene Nunc International) and incubated for 1 day under standard conditions. The incubations were then continued under normoxic conditions or hypoxic conditions for 48 hours in CCM. The hypoxic conditions consisted of culture in an atmosphere-controlled CO2 incubator (Forma Series II) containing a gas mixture composed of 94% N2, 5% CO2, and 1% O2. The nitrogen was provided with a feed from a tank of liquid N2. To examine hypoxia-induced apoptosis of HAECs, the cells were incubated under the same conditions except the culture for the first day was in serum containing eGM and then the cultures were continued either in the same medium or in fetal bovine serum-free endothelial basal medium (eBMsf) (catalog number 210−500; Endothelial Cell Basal Medium; Cell Applications Inc).

Preparation of Conditioned Medium

MSCs were replated at 5,000 cells per cm2 and incubated in CCM for 1 day. The attached cells were washed three times with phosphate-buffered saline (PBS), and the medium was replaced with eBMsf to generate CdM that was serum free and compatible for culture of HAECs. The incubation was then continued under either normoxic conditions, to prepare conditioned medium from MSCs incubated under normoxic conditions in eBMsf for 2 days (CdMNor), or hypoxic conditions, to prepare conditioned medium from MSCs incubated under hypoxic conditions in eBMsf for 2 days (CdMHyp). To prepare conditioned medium from HAECs, the cells were recovered and replated at 5,000 cells per cm2 and incubated in eGM for 1 day. The medium was replaced with eBMsf and incubated under normoxic or hypoxic conditions. After 2 days, the medium was harvested as HAEC-CdM.

Assays for Apoptosis and Cell Survival

Apoptosis was assayed with a cellular dye that detects membrane alterations (phosphatidylserine flip) and stains apoptotic cells (APOPercentage; Accurate Chemical & Scientific Corporation, New York, http://www.accuratechemical.com/accuratechemical). Stained cells were assayed on a colorimetric plate reader (Benchmark Microplate Reader; Bio-Rad, Hercules, CA, http://www.bio-rad.com). Alternatively, cells were harvested by trypsin treatment and pooled with floating cells to analyze the degree of apoptosis in the entire cell population. Cells were then stained with terminal deoxynucleotidyl transferase dUTP nick-end labeling (TUNEL) assay (In Situ Cell Death Detection Kit; Roche Applied Science, Indianapolis, IN, http://roche-applied-science.com) or double-stained with propidium iodide (PI) and fluorescein isothiocyanate-conjugated Annexin V (Sigma, St. Louis, http://www.sigmaaldrich.com) according to the manufacturer's instructions and analyzed by flow cytometry. We defined PI−/Annexin V+ as early apoptotic cells, PI +/Annexin V+ as apoptotic cells, and PI +/Annexin V− as dead cells. Total apoptotic cells were the sum of early apoptotic cells and apoptotic cells. We also stained nuclei with 4′, 6-diamidino-2-phenylindole dihydrochloride hydrate (DAPI; 1:1,000 in PBS; Sigma) by incubation of the cells for 10 minutes in the dark at 0°C–5°C. Apoptotic cells with condensed and fragmented nuclei were assayed by epifluorescence microscopy. For the detection of the active form of caspase 3 and 9, protein lysates were prepared and detected by Western blotting assays with the primary antibodies purchased from Cell Signaling Technology (Beverly, MA, http://www.cellsignal.com) and Chemicon (Temecula, CA, http://www.chemicon.com), respectively. Cell survival was assayed by trypan blue exclusion of cells counted by phase microscopy. Alternatively, cell survival was assayed with a nuclear dye (CyQuant; Molecular Probes, Eugene, OR, http://www.probes.invitrogen.com) and a fluorescence plate reader (FLUORstar OPTIMA; BMG Labtech Inc., Durham, NC, http://www.bmglabtech.com).

Assays of Cytokines

Samples of CdMNor or CdMHyp were centrifuged at 1,500g for 10 minutes to remove cell debris and concentrated 40× by tangential flow dialysis (Bio-Rad). Membranes from a human protein cytokine array kit (Human Cytokine Antibody Array; RayBiotech, Norcross, GA, http://www.raybiotech.com) were blocked with a blocking buffer and incubated with 1 ml of the concentrated CdM at room temperature for 2 hours. The membranes were then washed three times with wash buffer I and twice with wash buffer II, both at room temperature and 5 minutes for each time of wash, and assayed by chemiluminescence. To verify the results, the same samples and concentrated HAEC-CdM were assayed with enzyme-linked immunosorbent assay (ELISA) kits for interleukin (IL)-6 (RayBiotech), monocyte chemoattractant protein (MCP)-1 (RayBiotech), and vascular endothelial growth factor (VEGF) (Endogen, Woburn, MA, http://www.endogen.com).

Real-Time Reverse Transcription-Polymerase Chain Reaction

Total RNA was isolated (RNAqueous; Ambion, Austin, TX, http://www.ambion.com) from MSCs and HAECs and reverse transcribed to cDNA by using M-MLV reverse transcriptase (Invitrogen, Carlsbad, CA, http://www.invitrogen.com). The assays were performed using an automated instrument (ABI PRISM 7700; Applied Biosystems, Bedford, MA, http://www.appliedbiosystems.com) and a commercial kit (TaqMan 2-Step; Applied Biosystems). A 50-μl volume reaction was used containing 200 ng of the sample cDNA; TaqMan buffer; 200 mmol/l each of deoxy-ATP, deoxycytidine triphosphate, and deoxyguanosine triphosphate; 400 mmol/l deoxyuridine triphosphate; 5.5 mmol/l magnesium chloride; 0.025 U/ml polymerase (AmpliTaq Gold DNA Polymerase; Applied Biosystems); 0.01 U/l N-glycosylase (AmpErase; Applied Biosystems); 200 nmol/l forward primers and reverse primers; and 100 nmol/l fluorescent probe. The polymerase chain reaction (PCR) cycling conditions included an initial phase of 2 minutes at 50°C followed by 10 minutes at 95°C for the N-glycosylase reaction, 40 cycles of 15 minutes at 95°C, and 1 minute at 60°C. A commercial supplier (Biosource International, San Jose, CA, http://www.biosource.com) was used for primers for IL-6 and controls (GHC0064) with IL-6 fluorescence resonance energy transfer (FRET) probe (GHC0065); primers for β-actin and control (GHO0124) with β-actin FRET probe (GHO0125); and primers for VEGF and control (GHG0114) with VEGF FRET probe (GHG0115). MCP-1 primers were designed using designated software (TaqMan probes; Primer Express; Applied Biosystems): forward, 5′-c ccagtcacctgctgttata-3′; reverse, 5′-tgctggtgattcttctata-3′; and MCP-1 FRET probe, aggaagatctcagtgcagaggct.

Cell Proliferation Assays

HAECs were plated at 2 × 104 cells per well in gelatin-coated 96-well microtiter plates (Nalgene Nunc International) and incubated in eGM. After 24 hours, eGM was replaced with eBMsf or CdMHyp with or without PI3K inhibitor (LY294002; Cell Signaling Technology), and cells were exposed to hypoxia for another 16 hours. Cell number was assayed with a nuclear dye (CyQuant).

Tube Formation Assay

Ninety-six-well microtiter plates were coated with 100 μl per well of Matrigel or Growth Factor Reduced Matrigel (GFR-Matrigel; BD Biosciences, San Jose, CA, http://www.bdbiosciences.com). HAECs were suspended in basal medium or CdM with and without IL-6-, MCP-1-, or VEGF-neutralizing antibodies and plated at 4 × 104 per cm2 in triplicate. Cells were incubated under hypoxic conditions. After 24–48 hours, the formation of tube-like structures was examined microscopically. Photographs were taken of each well, and the lengths of >100 tubes were measured and assayed with a software program (Photoshop; Adobe, San Jose, CA, http://www.adobe.com).

Effects of Neutralizing Antibodies to IL-6, MCP-1, or VEGF

CdM was combined with neutralizing antibodies from a commercial source (R&D Systems, Hornby, Ontario, Canada, http://www.rndsystems.com) against IL-6 (MAB206), MCP-1 (MAB279), and VEGF (MAB293). The final concentrations of the antibodies were 10 μg/ml and therefore were 25- to 1,000-fold greater than the concentrations for half-maximal inhibition of 10 ng/ml of the recombinant proteins (R&D Systems). The mixtures were then added to cultures of HAECs for incubation for 1–2 days.

Western Blotting Assays

HAECs were incubated in hypoxia for 1 day in eBMsf with or without IL-6 or in CdMHyp with or without an inhibitor of PI3K (Cell Signaling Technology) or the inhibitor of mitogen-activated protein kinase kinase (MAPKK), also known as MAPK/extracellular signal-regulated kinase (ERK) kinase or MEK kinase (PD98059; Cell Signaling Technology). The cells were lysed in buffer (M-PER or NE-PER; Pierce Biotechnology, Rockford, IL, http://www.piercenet.com) supplemented with protease inhibitor cocktail (Halt; Pierce Biotechnology), and protein concentration was determined (Micro BCA Kit; Pierce Biotechnology). The cell lysate (20 μg of protein) was fractionated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (NuPAGE; 4%–12% Bis-Tris gels; Invitrogen). The sample was transferred to a filter (Immobilon P; Millipore, Bedford, MA, http://www.millipore.com) by electroblotting (XCell II Blot Module; Invitrogen). The filter was blocked for 1 hour with Tris-buffered saline (TBS) containing 5% nonfat dry milk and 0.05% Tween 20 and then incubated overnight at 4°C with the primary antibodies. The primary antibodies purchased from Cell Signaling Technology were against p-Akt (1:1,000; Ser473), Akt (1:1,000), p-signal transducer and activator of transcription (STAT)-3 (1:1,000; Ser727), p-ERK (1:2,000; Thr202/Tyr204), ERK (1:1,000), p-glycogen synthase kinase (GSK)-3β (1:1,000; Ser9), p-forkhead transcription factors (FKHR) (1:1,000; Ser256), p-acute-lymphocytic-leukemia-1 fused gene from chromosome X (AFX) (1:1,000; Ser193), and p-endothelial nitric oxide synthase (eNOS) (1:1,000; Ser1177). The filter was washed three times for 10 minutes each with TBS containing 0.1% Tween 20. Bound primary antibodies were detected by incubating for 1 hour with horseradish peroxidase-conjugated goat anti-mouse (1:3,000) or anti-rabbit IgG (1:2,000; BD Biosciences). The filter was washed and developed using a chemiluminescence assay (West-Femto Detection Kit; Pierce Biotechnology).

Adenoviral Vectors and Transfection

Adenoviral vectors expressing a dominant negative mutant of the p85 subunit of PI3K (AdGFP-p85) and constitutively activated mutant of PI3K (AdGFP-p110) were kindly provided by Dr. Patrice Delafontaine (Department of Medicine, Tulane University Health Sciences Center, New Orleans). An adenoviral vector encoding eGFP gene (green fluorescent protein) was used as a control vector. The HAECs were transduced at a multiplicity of infection of 100 at 48 hours prior to the experiment. The transduction efficiency was over 90% as assayed by green fluorescent protein-positive cells by epifluorescence microscopy.

Statistics

Data and results are reported as means ± SD. Statistical comparisons between two groups were performed with unpaired Student's t test. Pearson product moment correlation coefficients were obtained to estimate linear correlations among variables.

Results

Effect of CdM on Hypoxia-Induced Apoptosis and Cell Death

To test whether hypoxia induces apoptosis, MSCs and HAECs were plated at 2 × 104 cells per well in gelatin-coated 96-well plates and cultured in CCM and eGM, respectively. The next day, the cells were incubated in air or hypoxia for 48 hours in the same medium with a decrease in serum concentration. The decrease of serum concentration increased the level of apoptosis both in MSCs and HAECs. With MSCs, exposure to hypoxia induced a significant decrease in apoptosis at all tested serum concentrations (Fig. 1A). With HAECs, exposure to hypoxia further increased the apoptosis both in serum-containing and serum-free media (Fig. 1A). We also found that MSCs survived and proliferated in hypoxia for long-term culture; however, most of the HAECs and tumor cell lines did not survive under hypoxic conditions. Since the inhibitor of protein synthesis, cycloheximide at 20 μM or PI3K inhibitor LY294002 at 20 μM, did not alter the susceptibility of MSCs to hypoxia-induced apoptosis (supplemental online Fig. 1), the mechanisms by which MSCs protect themselves from hypoxia-induced apoptosis do not need new protein synthesis and are not mediated via the PI3K-Akt pathway.

Figure Figure 1..

Conditioned medium inhibits hypoxia-induced apoptosis and increases survival of HAECs. (A): MSCs were cultured for 2 days under normoxic or hypoxic conditions in 16%, 5%, or 0% FBS-supplemented α-minimal essential medium. HAECs were grown under normoxic or hypoxic conditions in endothelial cell growth medium with or without serum. Apoptosis assayed by APOPercentage assay method. Values are mean ± SD of n = 3. * p < .05, ** p < .01 versus control under normoxic conditions. (B–E): HAECs were incubated under hypoxic conditions for 16 hours and in eBMsf, CdMNor, CdMHyp, or in a 1:1 mixture of CdMNor/Hyp. (B): Apoptosis assayed by APOPercentage assay method. (C): Survival of HAECs assayed by trypan blue exclusion. Values are mean ± SD of n = 3; * p < .05, ** p < .01 versus cells in eBMsf. (D): Apoptosis assayed by TUNEL assay. (E): The detection of activated caspase 3 and 9 and actin by Western blotting assays. Abbreviations: BM, basal medium; CdMHyp, conditioned medium prepared under hypoxic conditions; CdMNor, conditioned medium prepared under normoxic conditions; CdMNor/Hyp, conditioned medium prepared under normoxic or hypoxic conditions; eBMsf, serum-free endothelial basal medium; eGM, endothelial growth medium containing 2% fetal bovine serum; FBS, fetal bovine serum; HAEC, human aortic endothelial cell; Hyp, hypoxic; MSC, multipotent stromal cell; Nor, normoxic; OD, optical density; TUNEL, terminal deoxynucleotidyl transferase dUTP nick-end labeling.

Because MSCs survived from hypoxia, we hypothesized that molecules or proteins secreted by MSCs could inhibit hypoxia-induced cell apoptosis or death. We next investigated whether CdM from MSCs cultured in normoxic (CdMNor) or hypoxic (CdMHyp) conditions rescued HAECs from hypoxia-induced apoptosis. With exposure of HAECs to 16 hours of hypoxia, all of CdMHyp, CdMNor, and a 1:1 mixture of CdMHyp and CdMNor decreased the level of apoptosis, but only CdMHyp and the mixture of CdMHyp and CdMNor had a significant difference as compared with the control basal medium (Fig. 1B). However, the level of apoptosis at 48 hours of hypoxia exposure was decreased by both CdMHyp and CdMNor (data not shown). Also, both CdMHyp and CdMNor improved cell survival after 16 hours of hypoxia (Fig. 1C). The effect of CdMHyp on the protection of HAEC from hypoxia-induced apoptosis was also demonstrated by the TUNEL assay (Fig. 1D) and determined by the detection of activated caspase 3 and 9 (Fig. 1E). Since HAEC-CdMHyp (CdM prepared from hypoxic HAECs) did not significantly decrease the level of apoptosis after 16 hours of hypoxia exposure, this protective effect specifically belonged to MSCs (supplemental online Fig. 2).

CdM Promotes Angiogenesis by HAECs

In the presence of GFR-Matrigel, hypoxic HAECs did not demonstrate tube formation (Fig. 2A). In contrast, the replacement of eBMsf with CdMHyp or CdMNor stimulated tube formation (Fig. 2A). Since the tube formation is incomplete in CdMNor and complete in CdMHyp, the angiogenic effect was therefore greater with CdMHyp than with CdMNor. In the presence of Matrigel containing growth factors, the hypoxic HAECs developed tubules when incubated either in eBMsf or in CdM (Fig. 2B). The tube length was 172 ± 92 μm, 207 ± 82 μm, and 265 ± 109 μm in eBMsf, in CdMNor, and in CdMHyp, respectively (n > 40). The tube length was slightly greater in CdM than in eBMsf and was significantly greater in CdMHyp than in CdMNor (p < .01, Student's t test).

Figure Figure 2..

Conditioned medium promotes tube formation by hypoxic human aortic endothelial cells in Growth Factor Reduced Matrigel (GFR-Matrigel) assay. (A): HAECs suspended in eBMsf, CdMNor, or CdMHyp were seeded on GFR-Matrigel in 96-well plates at 4 × 104 cells per well and cultured in hypoxic conditions for 48 hours. The cells were photographed by phase microscopy at ×4 magnification. (B): Same as (A) but cells seeded in standard Matrigel. Abbreviations: CdMHyp, conditioned medium prepared under hypoxic conditions; CdMNor, conditioned medium prepared under normoxic conditions; eBMsf, serum-free endothelial basal medium.

Cytokines Present in CdMHyp and CdMNor

To define the active components in CdM, samples were assayed using an antibody array containing 120 cytokines. Both CdMNor and CdMHyp contained IL-6; several angiogenesis factors, such as VEGF, MCP-1, and angiogenin; and several additional proteins such as the tissue metalloprotease inhibitors tissue inhibitor of metalloproteinase-1 and -2, urokinase plasminogen activator receptor, and neutrophil activating protein-2 (Fig. 3A). They also contained traced concentrations of insulin-like growth factor binding protein-2, -3, -4, and -6. The results suggested that the levels were higher in CdMHyp than in CdMNor. ELISAs on CdM prepared from two different donors of MSCs confirmed the presence of IL-6 and VEGF and demonstrated higher levels in CdMHyp than in CdMNor (Fig. 3B, 3D). The ELISAs also confirmed the presence of MCP-1 in CdM, but levels were not higher in CdMHyp (Fig. 3C). In contrast, HAEC-CdM did not contain measurable levels of IL-6 or VEGF. MCP-1 was present and increased by incubation of the HAECs in hypoxia (Fig. 3C). Real-time reverse transcription-PCR assays on the cells indicated the mRNA levels were consistent with the secreted levels of the cytokines (Fig. 3E).

Figure Figure 3..

Release of cytokines from MSCs in normoxic and hypoxic conditions. (A): Human protein cytokine array. CdMNor and CdMHyp were blotted onto two membranes, which are arrayed in duplicate with 120 human cytokines. The rectangles highlight cytokines higher in the CdMHyp than in CdMNor. (B–D): Enzyme-linked immunosorbent assays for IL-6, MCP-1, and VEGF from conditioned medium from MSCs and HAECs. ** p < .01 versus MSC-CdM or HAEC-CdM prepared under normoxic conditions or hypoxic conditions. (E): Real-time RT-PCR for mRNAs for IL-6, MCP-1, and VEGF. Abbreviations: Acrp30, adipocyte complement-related protein of 30 kDa; CdM, conditioned medium; CdMHyp, conditioned medium prepared under hypoxic conditions; CdMNor, conditioned medium prepared under normoxic conditions; eBMsf, serum-free endothelial basal medium; HAEC, human aortic endothelial cell; Hyp, hypoxic; IGFBP, insulin-like growth factor binding protein; IL, interleukin; MCP, monocyte chemoattractant protein; MSC, multipotent stromal cell; NAP, neutrophil activating protein; Nor, normoxic; RT-PCR, reverse transcription-polymerase chain reaction; TIMP, tissue inhibitor of metalloproteinase; uPAR, urokinase plasminogen activator receptor; VEGF, vascular endothelial growth factor.

Effects of IL-6 on HAEC Cell Death, Apoptosis, and Angiogenesis

Because IL-6 was reported to have antiapoptotic effects in cells such as multiple myeloma cells [14], we tested the hypothesis that the effects of CdM from MSCs could be explained by its content of IL-6. HAECs were cultured in eBMsf under hypoxia with or without recombinant IL-6. IL-6 inhibited apoptosis (supplemental online Fig. 3A) and increased cell survival (supplemental online Fig. 3B) in a dose-dependent manner. Taking into consideration the concentration of IL-6 in CdMHyp, the protective effect of recombinant IL-6 is similar to CdMHyp. However, blocking antibodies against IL-6, MCP-1, or VEGF did not reverse the effects of CdMHyp on apoptosis (supplemental online Fig. 3A) or cell survival (supplemental online Fig. 3B). Similarly, tube formation by hypoxic HAECs in GFR-Matrigel was promoted by the addition of up to 10 ng/ml IL-6 (supplemental online Fig. 3C). However, the effect of CdMHyp was only partially inhibited by antibodies against IL-6 and not influenced by antibodies against MCP-1 or VEGF (supplemental online Fig. 3C). These results suggested that IL-6 is only one of the active factors in CdMHyp.

CdM Induces Akt Phosphorylation and Increased Activation of Downstream Targets of Akt

We next tested the hypothesis that one or more of the active factors in CdM may activate the PI3K-Akt signaling pathway that plays a crucial role in survival, proliferation, microvascular permeability, and angiogenesis in HAECs and other endothelial cells [15, 16]. Incubation of HAECs in CdMHyp induced phosphorylation of Akt (Ser473), ERK, and STAT-3 (Fig. 4A). The effects occurred as early as 5 minutes after the addition of CdMHyp, and they were not reproduced by the addition of IL-6 to HAECs cultured in eBMsf. The results further supported the conclusion that the effects of CdMHyp on HAECs were not entirely explained by the presence of IL-6. To further support the observation that CdMHyp activates the PI3K-Akt pathway, we next examined whether CdMHyp also induced phosphorylation of downstream targets in the signaling cascade. HAECs incubated in hypoxia and eBMsf had either undetectable or low constitutive levels of phosphorylated GSK-3β, FKHR, and AFX (Fig. 4B). Incubation in CdMHyp induced phosphorylation of GSK-3β, FKHR, and AFX in a time-dependent manner. Importantly, the induced phosphorylation of Akt and its downstream targets was inhibited by pretreatment for 1 hour with the PI3K inhibitor LY294002 (10 μM). In contrast, the addition of the PI3K inhibitor LY294002 did not inhibit the phosphorylation of ERK, suggesting that the PI3K-Akt pathway was not targeting to the MEK/ERK pathway. Hypoxic HAECs cultured in eBMsf had a high level of phosphorylated eNOS, another downstream target of Akt (Fig. 4B). CdMHyp slightly increased the levels of p-eNOS, and the effect was partially blocked by LY294002.

Figure Figure 4..

Conditioned medium activates Akt and its downstream targets via PI3K-Akt signaling in human aortic endothelial cells (HAECs). (A): HAECs were incubated under hypoxia in serum-free endothelial basal medium (eBMsf) for 18 hours followed by incubation in CdMHyp for 5 and 15 minutes or in eBMsf with IL-6 (10 or 100 ng/ml) for 15 minutes. Cells were lysed, electrophoresed, and immunoblotted. (B): HAECs were incubated under hypoxic conditions first in eBMsf for 18 hours, then in eBMsf with either vehicle (0.01% DMSO) or LY294002 (10 μM) in vehicle for 1 hour, and finally in CdMHyp for 0–30 minutes. Cells were lysed and immunoblotted. Abbreviations: AFX, acute-lymphocytic-leukemia-1 fused gene from chromosome X; CdMHyp, conditioned medium prepared under hypoxic conditions; DMSO, dimethyl sulfoxide; eNOS, endothelial nitric oxide synthase; ERK, extracellular signal-regulated kinase; FKHR, forkhead transcription factors; GSK, glycogen synthase kinase; IL, interleukin; min, minutes; STAT, signal transducer and activator of transcription.

Inhibition of PI3K Blocks the Effects of CdMHyp on Apoptosis, Survival, and Angiogenesis by HAECs

To study the biologic sequelae of Akt activation in HAECs, we next examined the effects of the PI3K inhibitor LY294002 on hypoxia-induced apoptosis, cell survival, and tube formation. Incubation of HAECs in eBMsf under hypoxic conditions induced apoptosis and cell death of the cells as assayed by PI and Annexin V or the nuclear dye DAPI that detects apoptotic cells by fragmented and condensed nuclei (not shown). CdMHyp decreased apoptosis and cell death (Fig. 5A, 5B) and increased survival of the HAECs (Fig. 5C). Therefore, the results were consistent with assays carried out with a membrane stain for apoptosis and trypan blue exclusion for cell survival (Fig. 1B, 1C). Increasing concentrations of the inhibitor LY294002 blocked the effects of CdMHyp on inhibiting the sum of apoptosis and cell death (Pearson correlation, r = .779, p < .01) (Fig. 5B) and increasing cell survival (Fig. 5C). In contrast, no effect on apoptosis was observed with the MAPKK inhibitor PD98059 (50 μM), which blocks ERK signaling (not shown). Similarly, LY294002 blocked the effects of CdMHyp promoting tube formation by HAECs in GFR-Matrigel (Fig. 5D). The inhibitor LY294002 had no effect on tube formation when the experiments were repeated in the presence of 5% fetal calf serum (data not shown), suggesting that activation of the PI3K-Akt pathway was not important in the CdMHyp-induced effects on HAECs that were not serum deprived.

Figure Figure 5..

LY294002 inhibits CdMHyp induced survival and angiogenesis. Human aortic endothelial cells (HAECs) were incubated for 16 hours under hypoxia in eBMsf or in CdMHyp without or with the addition of the PI3K inhibitor LY294002 (2, 10, 50 μM). (A): Apoptosis and cell death were evaluated quantitatively by flow cytometry after PI/Annexin V staining (medium/LY294002 at μM). Data are gated for apoptotic cells (PI± and Annexin+, red) and dead cells (PI+ and Annexin V−, blue). (B): Percentage of total apoptotic cells and dead cells in each condition. (C): Survival of HAECs by counting the total cell number with CyQuant cell assay. Data represent mean ± SD of three independent experiments. ** p < .01 versus cells in eBMsf; # p < .05, ## p < .01 versus cells in CdMHyp. (D): Tube formation by HAECs under hypoxia for 24 hours in Growth Factor Reduced-Matrigel and CdMHyp in the presence or absence of LY294002 (2, 10, or 50 μM). The addition of LY294002 at all tested concentrations blocked tube formation. Abbreviations: CdMHyp, conditioned medium prepared under hypoxic conditions; eBMsf, serum-free endothelial basal medium; Hyp, hypoxic; LY, LY294002; Nor, normoxic; PI, propidium iodide.

Overexpression and Downregulation of PI3K

Transduction with an adenoviral vector carrying a constitutively activated construct of PI3K (AdGFP-p110) activated Akt and its downstream targets in HAECs cultured in eBMsf (Fig. 6A). Transduction with a dominant negative construct of PI3K (AdGFP-p85) blocked the CdMHyp-mediated activation of Akt and its downstream targets (Fig. 6A). The p110 construct also blocked hypoxia-induced apoptosis (Fig. 6B) and promoted tube formation by hypoxic HAECs incubated in eBMsf (Fig. 6C). In contrast, the p85 construct blocked the ability of CdMHyp to inhibit hypoxia-induced apoptosis (Fig. 6B) and to stimulate tube formation in hypoxic HAECs (Fig. 6C).

Figure Figure 6..

Overexpression and downregulation of PI3K-Akt signaling. (A): Western blot assays of human aortic endothelial cells (HAECs) at day 3 after adenoviral transduction. HAECs transfected with the activated PI3K gene (AdGFP-p110) and cultured in hypoxia and eBMsf showed enhanced phosphorylation of Akt and its downstream targets FKHR and GSK-3β but not ERK. HAECs transfected with the dominant negative gene (AdGFP-p85) and cultured in CdMHyp had decreased levels of Akt and its downstream targets FKHR and GSK-3β but not ERK. (B): HAECs were cultured in eBMsf under hypoxic conditions for 16 hours and apoptosis was assayed by DAPI staining. Images of cells expressing GFP fluorescence (green) and nuclei stained with DAPI (blue); ×600. HAECs transfected with AdGFP-p110 and cultured in eBMsf showed no apoptosis. HAECs transduced with AdGFP-p85 and cultured in CdMHyp showed apoptotic cells with nuclear condensation. (C): HAECs transduced with AdGFP-p110 and cultured in eBMsf and hypoxia had an increase in tube formation in Growth Factor Reduced-Matrigel as compared with cells transduced with AdGFP. HAECs transduced with AdGFP-p85 had a decrease in tube formation as compared with cells infected with AdGFP. Photomicrographs of identical fields by phase and GFP fluorescence are shown. Abbreviations: AFX, acute-lymphocytic-leukemia-1 fused gene from chromosome X; CdMHyp, conditioned medium prepared under hypoxic conditions; DAPI, 4,6-diamidino-2-phenylindole; eBMsf, serum-free endothelial basal medium; ERK, extracellular signal-regulated kinase; FKHR, forkhead transcription factors; GFP, green fluorescent protein; GSK, glycogen synthase kinase; O/E, overexpression.

Discussion

MSCs were previously shown to enhance collateral remodeling of vessels, angiogenesis, and regeneration of tissues in several systems in vivo, including assays in chick embryo chorioallantoic [17] and in rodent models for ischemia of hind limb or heart [10, [11]–12]. Several reports suggested the effects were mediated by the large number of cytokines and growth factors that MSCs secrete [18, [19]–20], but the signaling pathways activated in the target cells were not defined. The results here demonstrated that CdM from MSCs inhibited apoptosis, increased cell survival, and enhance tube formation by hypoxic endothelial cells. The CdMHyp from hypoxic MSCs was more effective than CdMNor from normoxic MSCs and contained higher levels of several factors previously shown to have similar effects such as IL-6 and VEGF. Moreover, the results demonstrated that the effects were largely explained by activation of the PI3K-Akt pathway. Incubation of hypoxic HAECs with CdMHyp increased phosphorylation of Akt. It also increased phosphorylation of GSK-3β, FKHR, and AFX, downstream targets of Akt that are known to regulate proliferation, differentiation, cell survival, and angiogenesis [21, [22]–23]. However, neutralizing antibodies to IL-6, VEGF, and MCP-1 did not block the effects of CdMHyp. Also, IL-6 partially mimicked the effects of CdMHyp, but it did not activate the PI3K-Akt pathway. Therefore, the results indicated that the effects of CdMHyp on hypoxic endothelial cells are largely mediated through activation of the PI3K-Akt and that activation is produced by the combined effects of several components produced by MSCs.

Incubation of hypoxic endothelial cells with CdMHyp also increased the phosphorylation of ERK and briefly produced a small increase in the phosphorylation of STAT-3. However, inhibition of PI3K with LY294002 did not decrease phosphorylation of ERK, and inhibition of ERK activity with the MAPKK inhibitor PD98059 had no effect on the suppression of apoptosis in hypoxic endothelial cells. Therefore, it is unlikely that activation of the ERK pathway made a major contribution to the observed effects.

An incidental observation was that the effects of CdMHyp on endothelial cells were not explained by activation of eNOS, another downstream target of Akt that was previously shown to mediate stromal cell-derived factor-1α enhanced ischemic vasculogenesis and angiogenesis [24]. The results support previous suggestions [10, 11] that MSCs or the components of CdM can potentially provide therapeutic benefits in pathological conditions that involve endothelial cells, such as wound repair, angiogenesis in ischemic tissues, microvascular permeability, vascular protection, and hemostasis.

Disclosure of Potential Conflicts of Interest

The authors indicate no potential conflicts of interest.

Acknowledgements

This work was supported in part by Grants from the NIH (HL73755 and HL075161), the NSC of Taiwan (NSC95-2314-B-075-047-MY3), HCA the Healthcare Company, and the Louisiana Gene Therapy Research Consortium. We thank Dr. Patrice Delafontaine (Division of Cardiology, School of Medicine, Tulane University Health Sciences Center, New Orleans) for providing the adenoviral vectors.

Ancillary