Mesenchymal Stem Cells Support Migration ... - Wiley Online Library

Download PDF
8 downloads 697 Views 586KB Size Report
Comment
Mesenchymal Stem Cells Support Migration, Extracellular Matrix. Invasion .... hMSC spheroids for 24 hours, collected, passed through Acrodisc. 0.2-μm HT ...
TISSUE-SPECIFIC STEM CELLS Mesenchymal Stem Cells Support Migration, Extracellular Matrix Invasion, Proliferation, and Survival of Endothelial Cells In Vitro IRINA A. POTAPOVA,a GLENN R. GAUDETTE,b PETER R. BRINK,a RICHARD B. ROBINSON,c,d,e MICHAEL R. ROSEN,c,d,e IRA S. COHEN,a SERGEY V. DORONINa a

Department of Physiology and Biophysics, Institute of Molecular Cardiology, State University of New York at Stony Brook, Stony Brook, New York, USA; bDepartment of Biomedical Engineering, Worcester Polytechnic Institute, Worcester, Massachusetts, USA; cCenter for Molecular Therapeutics and Departments of dPharmacology and ePediatrics, Columbia University, New York, New York, USA Key Words. Mesenchymal stem cells • Endothelial cells • Cell culture • Angiogenic cytokines

ABSTRACT We investigated effects of the paracrine factors secreted by human mesenchymal stem cells (hMSCs) on endothelial cell migration, extracellular matrix invasion, proliferation, and survival in vitro. Human mesenchymal stem cells were cultured as a monolayer or as three-dimensional aggregates in hanging drops (hMSC spheroids). We performed analysis of paracrine factors in medium conditioned by a monolayer of hMSCs and hMSC spheroids. Concentrations of vascular endothelial growth factor (VEGF), basic fibroblast growth factor, angiogenin, procathepsin B, interleukin (IL)-11, and bone morphogenic protein 2 were increased 5–20 times in medium conditioned by hMSC spheroids, whereas concentrations of IL-6, IL-8, and monocyte hemoattractant protein-1 were not increased. Concentrations of VEGF and angiogenin in medium conditioned by hMSC spheroids

showed a weak dependence on the presence of serum, which allows serum-free conditioned medium with elevated concentrations of angiogenic cytokines to be obtained. Medium conditioned by hMSC spheroids was more effective in stimulation of umbilical vein endothelial cell proliferation, migration, and basement membrane invasion than medium conditioned by a monolayer of hMSCs. This medium also promotes endothelial cell survival in vitro. We suggest that culturing of hMSCs as three-dimensional cellular aggregates provides a method to concentrate proangiogenic factors secreted by hMSCs and allows for reduction of serum concentration in conditioned medium. Our data support the hypothesis that hMSCs serve as trophic mediators for endothelial cells. STEM CELLS 2007;25:1761–1768

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

INTRODUCTION Bone marrow-derived mesenchymal stem cells are multipotent adult progenitors capable of differentiating into bone [1], cartilage [2], muscle [3], fat, and marrow stroma [4]. In bone stroma, niche mesenchymal stem cells support lineage progression of hematopoietic progenitor cells [4]. When delivered into ischemic tissues after stroke or myocardial infarction, mesenchymal stem cells improve functions of the central nervous system [5, 6] and the heart [7–9]. Although the mechanisms of these effects have not been elucidated, it has been suggested that the beneficial effects, at least in part, can be attributed to biologically active factors secreted by mesenchymal stem cells. Experiments conducted in vitro have demonstrated that medium conditioned by mesenchymal stem cells supports survival of cardiac myocytes [8] and stimulates migration and proliferation of endothelial cells [10]. Delivery of mesenchymal stem cells or medium conditioned by them stimulates perfusion in ischemic tissue and induces formation of tubular structures by endothelial cells, presumably, via paracrine mechanisms [11, 12]. It was reported that mesenchymal stem cells produce angiogenic cytokines, such as vascular endothelial growth factor (VEGF), basic fibro-

blast growth factor (FGF), hepatocyte growth factor, insulin-like growth factor 1, monocyte hemoattractant protein (MCP)-2, and MCP-3 [5, 10, 13–15]. Neutralization of VEGF and basic FGF or both in medium conditioned by mesenchymal stem cells inhibited but did not completely eliminate stimulation of endothelial cell migration [10, 11]. Thus, it appears that paracrine factors secreted by mesenchymal stem cells may contribute to cell survival as well as angiogenesis, and medium conditioned by mesenchymal stem cells may be used for solution-based therapy to complement and enhance the positive effects of cellular-based therapy. Previous studies of cytokines produced by mesenchymal stem cells have been carried out in vitro in monolayer cell culture [5, 10, 13–15]. Previous work has shown that freefloating mesenchymal stem cells form cellular aggregates of spherical shape [16, 17]. Therefore, we applied a hanging drop technique [18] to obtain human mesenchymal stem cell spheroids. Maintaining mesenchymal stem cells as spheroids reduces the volume of medium required for cell culture. We studied how culturing of human mesenchymal stem cells (hMSCs) in threedimensional (3D) aggregates affects concentrations of cytokines in conditioned medium. We also examined how medium conditioned by hMSCs and hMSC spheroids affects migration,

Correspondence: Sergey Doronin, Ph.D., Department of Physiology and Biophysics, BST-6, Room 124, University of New York at Stony Brook, Stony Brook, New York 11794, USA. Telephone: 631-444-7373; Fax: 631-444-3432; e-mail: [email protected] Received January 11, 2007; accepted for publication March 22, 2007; first published online in STEM CELLS EXPRESS March 29, 2007. ©AlphaMed Press 1066-5099/2007/$30.00/0 doi: 10.1634/stemcells.2007-0022

STEM CELLS 2007;25:1761–1768 www.StemCells.com

MSCs Are Trophic Mediators of Endothelial Cells

1762

extracellular matrix invasion, proliferation, and survival of human umbilical vein endothelial cells (HUVECs) in vitro.

Endothelial Cell Migration and Invasion Assays

hMSC spheroids were prepared using the hanging drop technique. hMSCs grown as a monolayer were dissociated with trypsin-EDTA solution (Cambrex BioScience), collected by centrifugation, and resuspended in high-glucose Dulbecco’s modified Eagle’s medium (DMEM) (Sigma-Aldrich, St. Louis, http://www.sigmaaldrich.com) supplemented with penicillin-streptomycin (Sigma-Aldrich) and 0%–30% fetal bovine serum (Sigma-Aldrich). Hanging drops were prepared using 40 ␮l of suspension containing 250,000 cells. We used 4 ⫻ 106 hMSCs in a monolayer under normoxic or hypoxic conditions to condition 6 ml of medium. hMSCs were exposed to hypoxia in a BD GasPak EZ Anaerobe Gas Generating Pouch System with an indicator (BD Biosciences, San Diego, http://www.bdbiosciences.com). As certified by the manufacturer, the Anaerobe Gas Generating Pouch System produces an atmosphere containing 10% carbon dioxide and 1% oxygen. Sixteen spheroids (4 ⫻ 106 cells total) were used to condition 640 ␮l of medium. Medium was conditioned by a monolayer of hMSCs or hMSC spheroids for 24 hours, collected, passed through Acrodisc 0.2-␮m HT Tuffryn Membrane low-protein binding filters (Pall Corp., East Hills, New York, http://www.pall.com), and stored at ⫺80°C until use.

An endothelial cell migration assay was performed using a BD BioCoat Endothelial Cell Migration FluoroBlok 24-multiwell insert system (BD Biosciences). The inserts contain a fluorescence-blocking, 3-␮m pore size PET membrane coated with human fibronectin so that the pores of the membrane are not occluded. An endothelial cell invasion assay was performed using a BD BioCoat Endothelial Cell Invasion FluoroBlok 24-multiwell insert system (BD Biosciences). The insert contains a fluorescence-blocking, 3-␮m pore size PET membrane coated with a BD Matrigel matrix that serves as a reconstituted authentic basement membrane. Cells must invade through the matrix barrier to pass through the BD FluoroBlok membrane. The FluoroBlok multiwell insert systems allow quantification of the number of cells that have migrated or invaded through the pores by a bottom-reading fluorometer. To study endothelial cell migration and invasion, HUVECs were serum-starved for 5 hours in EGM-2 basal medium in the presence of 0.1% delipidized bovine serum albumin (BD Biosciences). Cells were detached from the cell culture plate using Hanks’-based Enzyme Free cell dissociation solution (Chemicon, Temecula, CA, http://www.chemicon.com), suspended in serumfree DMEM, and seeded on a BD Falcon FluoroBlok 24-multiwell insert (0.25 ml of HUVECs suspension, 5 ⫻ 104 cells per top chamber). The bottom chambers contained 0.75 ml of serum-free DMEM or serum-free DMEM conditioned by hMSCs, hypoxic hMSCs, or hMSC spheroids. Cells were incubated in the FluoroBlok multiwell insert system for 24 hours at 37°C in a humidified atmosphere of 5% CO2 and labeled with 4 ␮g/ml calcein AM (Molecular Probes Inc., Eugene, OR, http://probes.invitrogen.com) in Hanks’ balanced salt solution (Sigma-Aldrich) for 90 minutes at 37°C and 5% CO2. Fluorescence of cells that passed through the BD FluoroBlok membrane was acquired from the bottom side of the multiwell insert system using a POLARstar OPTIMA microplate reader at excitation/emission wavelengths of 485/520 nm. For each set of experimental conditions, a minimum of four measurements were performed. Three independent experiments were conducted. Photographs of migrated cells were taken using an Olympus IX-51 inverted fluorescent microscope (Tokyo, http://www.olympusglobal.com) at fourfold magnification.

Cytometric Bead Array Analysis

Heme Oxygenase Immunoblot Analysis

Concentrations of angiogenin, FGF, MCP-1, VEGF, interleukin (IL) 6, and IL-8 were measured using a Cytometric Bead Array kit (BD Biosciences). Fifty microliters of premixed beads coated with capture antibodies were incubated with 50 ␮l of standards or samples for 1 hour, 50 ␮l of a mixture of phycoerythrin-conjugated antibodies against specific cytokines was added, and the final mixture was incubated for 2 hours in the dark at room temperature. The beads were washed with the wash buffer supplied with the CBA kit and analyzed by a BD FACSCalibur flow cytometer (Becton, Dickinson and Company, Franklin Lakes, NJ, http://www.bd.com). Data acquisition and processing were performed using the BD FACSComp and FCAP Array software packages from BD Biosciences. Three thousand events were acquired with each sample. Calibration curves were obtained using standards supplied by the manufacturer. Data from four independent experiments were used to estimate the concentrations of the cytokines.

Cells were lysed in buffer containing 0.025 M Tris-HCl, pH 7.4, 0.15 M NaCl, 5 mM EDTA, 1% Triton X-100, 0.5% Nonidet P40, and a set of protease inhibitors (Roche Diagnostics, Basel, Switzerland, http://www.roche-applied-science.com) and phosphatase inhibitors (cocktails type 1 and 2; Sigma-Aldrich) for 30 minutes at 4°C. The extract was cleared by centrifugation at 15,000g at 4°C for 30 minutes. Proteins (25 ␮g) were separated in Bis-Tris 12% Criterion gel using XT 3-(N-morpholino)propanesulfonic acid (MOPS) running buffer (Bio-Rad, Hercules, CA, http://www. bio-rad.com). After transfer onto a nitrocellulose membrane (Bio-Rad), heme oxygenase and glyceraldehyde-3-phosphate dehydrogenase (GAPDH) were detected using an anti-heme oxygenase antibody from Abcam (Cambridge, U.K., http://www.abcam.com), anti-GAPDH antibody from Santa Cruz Biotechnology (Santa Cruz, CA, http://www.scbt.com), and affinity-purified peroxidase-labeled F(ab⬘)2 fragments (Kirkegaard & Perry Laboratories, Gaithersburg, Maryland, http://www.kpl.com). Immunoblots were developed with an Amersham Biosciences (Piscataway, NJ, http://www.amersham. com) chemiluminescence kit.

MATERIALS

AND

METHODS

Cell Culture hMSCs and HUVECs were purchased from Cambrex BioScience (Walkersville, MD, http://www.cambrex.com). hMSCs were cultured in Mesenchymal Stem Cell Growth Media (Cambrex BioScience). HUVECs were cultured in EGM-2 BullettKit medium (Cambrex BioScience). Passages 2–5 of hMSCs or HUVECs were used. Cells were maintained at 37°C in a humidified atmosphere of 5% CO2.

Formation of hMSC Spheroids and Preparation of Conditioned Medium

Enzyme-Linked Immunosorbent Assays VEGF, bone morphogenic protein 2 (BMP-2), and procathepsin B were measured using Quantikine enzyme-linked immunosorbent assay (ELISA) kits from R&D Systems (Minneapolis, http://www. rndsystems.com). IL-11 and IL-8 were quantified with DuoSet ELISA Development kits (R&D Systems). The optical densities were measured at 450 nm with a 595 nm reference wavelength using a POLARstar OPTIMA microplate reader (BMG Technologies, Durham, NC, http://www.bmglabtech.com). Calibration curves were obtained using standards supplied by the manufacturer. Concentrations of factors were determined using data from a minimum of three independent experiments, each with at least eight measurements per set of experimental conditions.

DNA Microarray Analysis DNA microarray analysis of hMSCs and 3-day-old hMSC spheroids was performed using an Affymetrix Human Genome U133 Plus 2.0 microarray (Santa Clara, CA, http://www.affymetrix.com). Total RNA was extracted from the cells using an RNeasy Mini kit (Qiagen, Hilden, Germany, http://www1.qiagen.com). Probe hybridizations and data acquisitions were conducted according to the manufacturer’s specifications. t tests of data from three independent hybridizations for each group were performed using the Bioconductor software package (http://www.bioconductor.org).

Potapova, Gaudette, Brink et al.

1763

Measurement of DNA Synthesis with 5-Bromo-2ⴕdeoxyuridine The Cell Proliferation ELISA (Roche) was used to measure the rate of DNA synthesis by HUVECs. Cells were grown in 96-well plates in EGM-2 BullettKit medium. When cells reached a density of 5,000 cells per well, the EGM-2 medium was removed and cells were incubated for 24 hours with DMEM containing 5% fetal bovine serum or DMEM containing 5% fetal bovine serum conditioned by hMSCs, hypoxic hMSCs, or hMSC spheroids. To assay the rate of DNA synthesis, cells were labeled with 10 ␮M 5-bromo2⬘-deoxyuridine (BrdU) for 0, 30, 45, 60, 75, 90, 105, and 120 minutes. Incorporation of BrdU into DNA was detected using an anti-BrdU peroxidase-conjugated antibody and a soluble peroxidase substrate (3,3⬘-5,5⬘-tetramethylbenzidine) according to the manufacturer’s protocol. The absorbance was measured at 450 nm with a 595 nm reference wavelength in a POLARstar OPTIMA microplate reader. In these conditions, the incorporation of BrdU into DNA was a linear function of time. The rate of DNA synthesis was calculated as the slopes of the least squares linear fits to the experimental curves. Four experiments were conducted for each set of experimental conditions.

Measurement of Endothelial Cell Viability Viability of HUVECs was measured using a WST-1-based assay (Roche). WST-1 is metabolized by cellular NAD-dependent dehydrogenase into formazan, the colored product of WST-1. The amount of dye correlates with the number of viable cells. HUVECs were grown in 96-well plates to a density of 10,000 cells per well using EGM-2 BullettKit medium. EGM-2 medium was then replaced with DMEM containing 5% fetal bovine serum or DMEM containing 5% fetal bovine serum and conditioned by hMSCs, hypoxic hMSCs, or hMSC spheroids, and the cells were incubated for 1–3 days. The WST-1 reagent was added to the cell culture medium (10 ␮l per 100 ␮l of cell culture medium), and cells were incubated for 4 hours at 37°C and 5% CO2. The formazan absorbance was measured at 450 nm with the reference wavelength of 595 nm using a POLARstar OPTIMA microplate reader. At least four measurements were performed for each set of experimental conditions in three independent experiments.

RESULTS hMSCs Form Cellular Aggregates when Cultured in Hanging Drops hMSCs form spherical aggregates having a diameter of approximately 1 mm within 24 hours of cultivation in hanging drops. Spheroids do not increase in size from day 1 to day 3 of cultivation. Figure 1A shows a 3-day-old spheroid formed by hMSCs. To determine whether culturing in hanging drops affects gene expression in hMSCs, we compared the gene expression profile of 3-day-old hMSC spheroids with that of a monolayer of hMSCs using the Affymetrix Human Genome U133 Plus 2.0 DNA microarray. Data from the microarray analysis are included in the supplemental online data. We found that in hMSC spheroids 1,731 genes were upregulated and 1,387 genes were downregulated more than twofold (p ⬍ .05). Many of the genes overexpressed in hMSC spheroids could be linked to hypoxia and hypoxia-dependent angiogenic pathways. This group included VEGF, stanniocalcin-1, placental growth factor, angiopoietin 2, transforming growth factor ␤3, macrophage migration inhibitory factor, insulin-like growth factor binding proteins 1 and 5, IL-8, IL-1␣, IL-1␤, acetyl-coenzyme A thiolase, transferrin, aldose reductase, heme oxygenase-1, and cyclooxygenase-2. Genes encoding proangiogenic factors and other important secreted biological regulators, which were upregulated in hMSC-spheroids, are presented in supplemental online Table 1. www.StemCells.com

Figure 1. hMSC SPH. (A): Phase contrast photograph of hMSC SPH on day 3 after the applying hanging drop technique. (B): Western blot analysis of HO-1 expression and control staining for GAPDH in total cell extracts from hMSCs, hypoxic hMSCs, and 1-, 2-, or 3-day-old hMSC SPHs. Abbreviations: hMSC, human MSC; GAPDH, glyceraldehyde-3-phosphate dehydrogenase; HO-1, heme oxygenase 1; Mw, molecular mass; SPH, spheroid.

The heme oxygenase-1 protein expression level was also examined by Western blot. Figure 1B shows that culturing hMSCs under hypoxic conditions or as hMSC spheroids significantly increases the expression of heme oxygenase-1. These data and the data from gene expression analysis suggest that at least some cells in hMSC spheroids are hypoxic.

Concentrations of Cytokines in Medium Conditioned by hMSC Spheroids CBA and ELISA were used to measure the release of cytokines into DMEM containing 30% fetal bovine serum and conditioned by hMSCs, hypoxic hMSCs, or hMSC spheroids. Concentrations of VEGF, basic FGF, angiogenin, IL-6, IL-8, and MCP-1 were measured by CBA (Table 1). Concentrations of VEGF, IL-8, procathepsin B, BMP-2, and IL-11 were measured by ELISA (Table 1). Data obtained by CBA assay correlate with the results from ELISA (Table 1). One-tenth as much medium was used to culture hMSCs in 3D aggregates compared with the same number of cells grown as a monolayer. It was therefore expected that concentrations of cytokines in medium conditioned by hMSC spheroids will be increased. Statistically significant increases (p ⬍ .05) in concentrations of cytokines were observed for VEGF, basic FGF,

MSCs Are Trophic Mediators of Endothelial Cells

1764

Table 1. Concentrations of cytokines in media conditioned by hMSCs or hMSC spheroids as determined by CBA and enzyme-linked immunosorbent assay (ELISA) VEGF (ng/ml)

30% FBS hMSC hMSC SPH1 hMSC SPH2 hMSC SPH3 Hypoxic hMSCs

0.3 ⫾ 0.1

a

basic FGF Angiogenin (pg/ml) (ng/ml)

24 ⫾ 3

ND

Cathepsin B (ng/ml)

ND

BMP-2 (pg/ml)

ND

IL-6 (ng/ml)

ND

IL-8 (ng/ml)

ND

MCP-1 (ng/ml)

0.16 ⫾ 0.03 (0.20 ⫾ 0.02a)

3.8 ⫾ 0.5 (3.6 ⫾ 0.4) 30 ⫾ 3 46 ⫾ 10 28 ⫾ 1a (40 ⫾ 2a)

0.9 ⫾ 0.3 (4.9 ⫾ 0.7) (12 ⫾ 3) 7.8 ⫾ 0.8 4.7 ⫾ 0.3 (2.3 ⫾ 0.2) 4.4 ⫾ 0.7 17 ⫾ 3a (61 ⫾ 2a) (99 ⫾ 8a) 15 ⫾ 2a 1.3 ⫾ 0.1a (1.3 ⫾ 0.3) 5.0 ⫾ 0.8

25 ⫾ 1a (43 ⫾ 2a)

7.2 ⫾ 0.6a (59 ⫾ 1a) (152 ⫾ 17a) 3.8 ⫾ 0.3a 0.6 ⫾ 0.2a (0.5 ⫾0.2)

27 ⫾ 8a (45 ⫾ 2a) 4.7 ⫾ 0.2 (7 ⫾ 1a)

39 ⫾ 22

171 ⫾ 48a 4.2 ⫾ 0.4a (61 ⫾ 2a) (221 ⫾ 6a) 2.7 ⫾ 0.3a 19 ⫾ 3

0.78 ⫾ 0.05 (5.9 ⫾ 1)

(9 ⫾ 2)

6 ⫾ 2 (4 ⫾ 2)

8.1 ⫾ 0.4 6.6 ⫾ 0.6 (5 ⫾ 1)

IL-11 (ng/ml) a

(3 ⫾ 1) (8 ⫾ 3a)

3.9 ⫾ 0.4

(14 ⫾ 4a)

3.9 ⫾ 0.5

(15 ⫾ 3a)

3.7 ⫾ 0.5

(2 ⫾ 1)

Data are shown as concentration ⫾ standard error. Concentrations of factors determined by ELISA are shown in parentheses. The first row shows concentrations of factors measured in Dulbecco’s modified Eagle’s medium in the presence of 30% FBS. Statistically significant changes (p value ⬍ .05) relative to hMSC monolayer under normoxic conditions. Abbreviations: BMP, bone morphogenic protein; FBS, fetal bovine serum; FGF, fibroblast growth factor; hMSC, human MSC; IL, interleukin; MCP, monocyte hemoattractant protein; ND, nondetectable; SPH, spheroid; VEGF, vascular endothelial growth factor. a

angiogenin, procathepsin B, IL-11, and BMP-2 (Table 1). Levels of IL-6 and MCP-1 were not dramatically elevated or even decreased in medium conditioned by hMSC spheroids (Table 1). We found that concentrations of cytokines in medium conditioned by hMSCs under hypoxic conditions were comparable to those in medium conditioned by hMSCs under normoxic conditions (Table 1). Next, we investigated how the presence of serum affects the concentrations of VEGF and angiogenin in conditioned medium. The concentration of VEGF in medium conditioned by a monolayer of hMSCs increased fivefold as serum concentration was raised from 0% to 30% (Fig. 2A). The concentration of angiogenin was increased 2.3-fold as serum concentration was increased from 0% to 25% (Fig. 2B). For hMSC spheroids, the concentration of VEGF increased 1.8-fold as serum was raised from 0% to 5% and remained unchanged with further increases in serum concentration (Fig. 2A). The concentration of angiogenin in medium conditioned by hMSC spheroids showed no dependence on the presence of serum (Fig. 2B). Overall, concentrations of VEGF and angiogenin in serum-free medium conditioned by 3-day-old hMSC spheroids were three times higher (p ⬍ .05) than those in 30% serum containing medium conditioned by a monolayer of hMSCs. Thus, by culturing hMSCs as 3D aggregates, we obtained a concentrated cocktail of angiogenic factors. Importantly, concentrations of angiogenic factors in the medium conditioned by hMSC spheroids showed a limited dependence on the concentration of fetal bovine serum. This feature could provide an advantage for some applications in which it might be necessary to eliminate or at least minimize the presence of contaminating biological products of other origins.

Effects of Medium Conditioned by hMSC Spheroids on Endothelial Cells Media Conditioned by hMSC Spheroids Stimulate HUVEC Migration and Basement Membrane Invasion In Vitro. Migration and invasion of endothelial cells are important for angiogenesis and vessel sprouting and are known to be regulated by angiogenic cytokines. Above, we demonstrated that culturing hMSCs as spheroids resulted in increased concentrations of angiogenic factors in conditioned medium. Therefore, we studied HUVEC migration toward the medium conditioned by hMSCs, hypoxic hMSCs, and hMSC spheroids. Since fetal bovine serum itself causes endothelial cells to migrate, serum-

free DMEM was used to generate the conditioned medium. We found that medium conditioned by hMSCs profoundly stimulated HUVEC chemotaxis compared with nonconditioned DMEM (Fig. 3A, 3B). Effects of medium generated by hMSCs cultured under hypoxic conditions and by 3-day-old hMSC spheroids were comparable to those of hMSCs cultured under normoxic conditions. Medium conditioned by 1- or 2-day-old hMSC spheroids markedly induced endothelial cell migration in comparison with hMSC-conditioned medium (Fig. 3A, 3B). During angiogenesis, activated endothelial cells not only migrate toward proangiogenic factors but also have to invade the vascular basement membrane. We compared HUVEC invasion toward the medium conditioned by hMSCs, hypoxic hMSCs, and hMSC spheroids. Serum-free DMEM was used to generate the conditioned medium. Data are shown in Figure 4A and 4B. We found that medium conditioned by hMSCs stimulated endothelial cell invasion in comparison with nonconditioned DMEM or medium conditioned by hypoxic hMSCs. Substantial activation of HUVEC invasion was observed with medium conditioned by 2- or 3-day-old hMSC spheroids. Maximal invasion was detected with medium conditioned by 1-day-old hMSC spheroids (Fig. 4A, 4B). Our data show that medium conditioned by hMSC spheroids stimulates endothelial cell migration and invasion to a much greater extent than medium conditioned by hMSCs. Medium Conditioned by hMSC Spheroids Stimulates DNA Synthesis in HUVECs. We investigated how DMEM supplemented with 5% fetal bovine serum and conditioned by hMSCs, hypoxic hMSCs, or 2- and 3-day-old hMSC spheroids affects DNA synthesis in HUVECs. The rate of DNA synthesis was determined from the time course of BrdU incorporation into DNA of HUVECs. For a given set of experimental conditions, BrdU incorporation was a linear function of the time. Therefore, the rate of DNA synthesis was estimated from the linear approximations to BrdU incorporation (Fig. 5A). The rate of DNA synthesis by endothelial cells maintained in EGM-2 BullettKit medium was determined for comparison (Fig. 5B). We found that culturing endothelial cells in medium conditioned by hMSCs stimulated DNA synthesis in comparison with nonconditioned DMEM. However, the rate of DNA synthesis remained lower than in the EGM-2 BullettKit medium. Medium conditioned by hypoxic hMSCs was less effective than the medium conditioned by hMSCs. Effects from medium conditioned by 2-day-old hMSC spheroids were comparable to

Potapova, Gaudette, Brink et al.

Figure 2. Cytometric bead array analysis of secreted cytokines. Effect of serum on concentrations of VEGF (A) and angiogenin (B) in medium conditioned by human MSCs (hMSCs) (F) or 3-day-old hMSC spheroids (Œ). Cytokine concentrations are shown as relative percentages compared with medium conditioned by hMSCs cultured in normoxic conditions in the presence of 30% serum. With an exception of the 15% FBS point for hMSCs, all data represent an average of three independent experiments; bars indicate standard deviations. Data for 15% FBS point in case of hMSCs represent average of two independent experiments. Abbreviations: FBS, fetal bovine serum; VEGF, vascular endothelial growth factor.

those induced by conditioned medium from hMSCs. Medium conditioned by 3-day-old hMSC spheroids substantially stimulates DNA synthesis in comparison with hMSCs and provides a rate of DNA synthesis comparable to that obtained with EGM-2 BullettKit medium. Medium Conditioned by hMSC Spheroids Supports HUVECs in Cell Culture. To allow an assessment of cell survival, HUVECs were plated at near confluence and incubated for 3 days with DMEM containing 5% fetal bovine serum, DMEM containing 5% fetal bovine serum and conditioned by hMSCs, hypoxic hMSC, or hMSC spheroids. Endothelial cells incubated with EGM-2 BullettKit medium were used as a control. Pictures of the cells were taken after 24, 48, and 72 hours of incubation www.StemCells.com

1765

Figure 3. Effects of conditioned media on endothelial cell migration. Human umbilical vein endothelial cells (HUVECs) in serum-free DMEM were placed in the upper chamber of an Endothelial Cell Migration FluoroBlok 24-multiwell insert system, allowed to migrate toward conditioned media for 24 hours, and labeled with calcein AM. Lower chambers contained nonconditioned serum-free DMEM or serum-free DMEM conditioned by hMSCs, hypoxic hMSCs, or 1-, 2-, or 3-day-old hMSC SPHs (SPH1, SPH2, and SPH3, respectively). Media were conditioned for 24 hours. (A): Photographs show fluorescence signals associated with the bottom of the insert membrane. (B): Quantitative analysis of HUVEC migration toward conditioned media. Data are presented as relative percentages, compared with medium conditioned by hMSCs cultured in normoxic conditions. Data presented are the mean of four inserts ⫾ SD. ⴱ, p ⬍ .05 compared with hMSCs; #, p ⬍ .05 compared with hMSC SPH1. Abbreviations: DMEM, Dulbecco’s modified Eagle’s medium; hMSC, human MSC; SPH, spheroid.

with the conditioned medium (Fig. 6A), and a colorimetric WST-1 assay was performed to quantify effects of conditioned medium on HUVEC survival (Fig. 6B). HUVEC viability was not changed when cells were incubated in EGM-2 BullettKit medium for 3 days. In contrast, the number of viable HUVECs was substantially reduced when cells were incubated for 72 hours with nonconditioned medium or medium conditioned by hMSCs or hypoxic hMSCs. No decline in viable cell number was detected after 72 hours of incubation of HUVECs with medium conditioned by 1-, 2-, or 3-day-old hMSC spheroids (Fig. 6A, 6B). This shows that medium conditioned by hMSC spheroids supports endothelial survival in cell culture.

MSCs Are Trophic Mediators of Endothelial Cells

1766

Figure 4. Effects of conditioned media on endothelial cell invasion. Human umbilical vein endothelial cells (HUVECs) in serum-free DMEM were placed in the upper chamber of an Endothelial Cell Invasion FluoroBlok 24-multiwell insert system, allowed to invade for 24 hours, and labeled with calcein AM. Lower chambers contained nonconditioned serum-free DMEM or serum-free DMEM conditioned by hMSCs, hypoxic hMSCs, or 1-, 2-, or 3-day-old hMSC SPHs (SPH1, SPH2, and SPH3, respectively). Media were conditioned for 24 hours. (A): The photographs show fluorescence signals associated with the bottom of the insert membrane. (B): Quantitative analysis of HUVEC invasion toward conditioned media. Data are presented as relative percentages compared with medium conditioned by hMSCs cultured in normoxic conditions. Data presented are the mean of four inserts ⫾ SD. ⴱ, p ⬍ .05 compared with hMSCs; #, p ⬍ .05 compared with hMSC SPH1. Abbreviations: DMEM, Dulbecco’s modified Eagle’s medium; hMSC, human MSC; SHP, spheroid; SPH, spheroid.

DISCUSSION Formation of 3D aggregates by hMSCs alters the microenvironment that cells experience. As a first step toward understanding how changes in cell culture conditions affect hMSCs, we compared the gene expression profile of hMSC spheroids with that of a monolayer of hMSCs. DNA microarray analysis revealed that gene expression was substantially altered in hMSC spheroids. We found that transcription of several hypoxia-regulated genes, including VEGF, cathepsins, and FGFs, was upregulated in hMSC spheroids. It has been reported that dense cellular

Figure 5. Effects of conditioned media on BrdU incorporation in DNA of endothelial cells. Human umbilical vein endothelial cells (HUVECs) were incubated for 24 hours with EGM-2 BullettKit medium, nonconditioned DMEM supplemented with 5% fetal bovine serum, or DMEM supplemented with 5% fetal bovine serum and conditioned by hMSCs, hypoxic hMSCs, or 2- or 3-day-old hMSC SPHs (SPH2 and SPH3, respectively). Media were conditioned for 24 hours. BrdU was added, and the cells were incubated with 10 ␮M BrdU for 0, 30, 45, 60, 75, 90, 105, and 120 minutes. (A): Representative time courses of BrdU incorporation after 24 hours of HUVEC incubation with nonconditioned medium (F), hMSC-conditioned medium (f), or 3-day-old hMSC SPHconditioned medium (Œ). (B): Rates of BrdU incorporation in HUVECs. The rates were calculated as the slopes of the linear approximations of the kinetics of BrdU incorporation. Columns summarize the result of four independent experiments; bars indicate standard deviation. ⴱ, p ⬍ .05 compared with hMSCs; #, p ⬍ .05 compared with hMSC SPH3. Abbreviations: BrdU, 5-bromo-2⬘-deoxyuridine; FBS, fetal bovine serum; hMSC, human MSC; SPH, spheroid.

structures develop hypoxia at distances beyond the diffusion capacity of oxygen (typically 150 –250 ␮m) [19]. The diameter of hMSC spheroids is approximately 1 mm, and it is likely that at least some cells on the interior of the spheroids are hypoxic. To corroborate findings of DNA microarray analysis, we assayed expression of heme oxygenase-1 as one of the protein markers of hypoxic cells. Heme oxygenase-1 was upregulated in hMSC spheroids at both the mRNA and protein levels. We found that the level of heme oxygenase-1 expression in hMSC spheroids was comparable to that found in hMSCs under hy-

Potapova, Gaudette, Brink et al.

Figure 6. Effects of conditioned media on human umbilical vein endothelial cell (HUVEC) viability. HUVECs were incubated for 24, 48, and 72 hours with EGM-2 BullettKit medium, nonconditionedDMEM supplemented with 5% fetal bovine serum, or DMEM supplemented with 5% fetal bovine serum and conditioned by hMSCs, hypoxic hMSCs or 1-, 2-, or 3-day-old hMSC SPHs. Media were conditioned for 24 hours. (A): Representative phase contrast photographs of the cells incubated with the conditioned media for 24, 48, and 72 hours. (B): WST-1 assays for HUVECs incubated with the conditioned media for 72 hours. Data presented are the mean of eight experiments ⫾ SD. ⴱ, p ⬍ .05 compared with hMSCs; #, p ⬍ .05 compared with hMSC SPH1. Abbreviations: FBS, fetal bovine serum; h, hours; hMSC, human MSC; SPH, spheroid.

www.StemCells.com

1767

poxic conditions. We hypothesized that hypoxia could be one of the reasons for stimulation of the transcription of angiogenic factors. This suggestion is consistent with reported observations that hypoxia leads to increases in the production of VEGF and basic FGF by hMSCs [10]. However, we found that exposure of hMSC to hypoxia for 24 hours did not substantially alter the levels of the studied cytokines in conditioned medium. Only ELISAs have demonstrated that VEGF secretion by a monolayer of hMSCs under hypoxic conditions was statistically significantly increased 1.8 times (p ⬍ .05). This increase correlates with previously reported observations for ELISA-based measurements of VEGF in medium conditioned by hMSC under hypoxic conditions [10]. Enrichment of conditioned medium with angiogenic factors by hMSC spheroids was achieved primarily because of a reduction in the ratio between the volume of cell culture medium and the number of cultured cells. The measured increase in the concentrations of VEGF, angiogenin, basic FGF (for 3-day-old hMSC spheroids), procathepsin B, and BMP-2 in medium conditioned by hMSC spheroids correlates with the expected concentration effects due to the reduction in medium volume. Concentrations of IL-6, IL-8, and MCP-1 were not increased in medium conditioned by hMSC spheroids. Overall, culturing hMSCs in 3D aggregates resulted in medium with elevated concentrations of angiogenic factors. Diffusion of cytokines in the extracellular environment is expected to be affected by formation of dense 3D cellular aggregates. This may lead to high local concentrations of factors inside hMSC spheroids. We found that secretion of VEGF and angiogenin by a monolayer of hMSCs depends on the concentration of fetal bovine serum in the cell culture medium. Serum had no effect on the concentration of angiogenin in medium conditioned by hMSC spheroids. Addition of 5% fetal bovine serum stimulated secretion of VEGF by hMSC spheroids, but further increases in the concentration of serum had no effect. We hypothesize that production of paracrine factors by hMSC spheroids is sustained via an autoregulatory loop due to the high local concentrations of secreted factors inside hMSC spheroids. Therefore, production of cytokines by hMSC spheroids is less dependent on the concentration of fetal bovine serum. This is an important feature of hMSC spheroids that allows for reduction or even elimination of serum from the cell culture medium while maintaining the production of proangiogenic factors. It has been reported that medium conditioned by a monolayer of mesenchymal stem cells under normoxic or hypoxic conditions stimulates endothelial cell migration [10, 11]. Medium conditioned by hMSC spheroids even in the absence of serum retained the ability to stimulate HUVEC migration in vitro and was superior in that capacity to medium conditioned by a monolayer of normoxic or hypoxic hMSCs. Successful angiogenesis requires activation of endothelial cell invasion through the vascular basement membrane. Medium conditioned by hMSCs or hMSC spheroids stimulated endothelial cell invasion through a reconstituted basement membrane in vitro. Stimulation of endothelial cell invasion by the medium conditioned by hMSC spheroids in the absence of serum was greater than that shown by the medium conditioned by a monolayer of hMSCs. The ability of medium conditioned by hMSCs to stimulate mitotic propagation of endothelial cells is uncertain: one report suggests stimulation [10], whereas another has failed to find this effect [11]. Our data suggest that exposure of hMSCs to hypoxia reduces their ability to activate DNA synthesis in endothelial cells. Medium conditioned by hMSCs cultured as a monolayer or as hMSC spheroids stimulates DNA synthesis in endothelial

MSCs Are Trophic Mediators of Endothelial Cells

1768

cells. Medium conditioned by 3-day-old hMSC spheroids had the greatest effect. We found that HUVECs did not survive for 3 days in DMEM supplemented with 5% fetal bovine serum or in the same medium conditioned by a monolayer of normoxic or hypoxic hMSCs. In contrast, the same medium conditioned by hMSC spheroids effectively supported endothelial cells in cell culture for 3 days with an efficacy close to that offered by the commercial endothelial cell culture medium (EGM-2). These data demonstrate that medium conditioned by hMSC spheroids contains the complete set of factors required for endothelial cell survival. In conclusion, we suggest that culturing hMSCs as 3D cellular aggregates is a simple and effective method to concentrate angiogenic cytokines secreted by hMSCs. For the first time, we have shown that medium conditioned by hMSCs stimulates basement membrane invasion by HUVECs in vitro. In addition, the cocktail of factors produced by hMSC spheroids stimulates mitotic propagation, migration, and invasion of endothelial cells to a much higher extent than medium conditioned by a monolayer of hMSCs. Moreover, hMSC spheroid-condi-

REFERENCES 1

Haynesworth SE, Goshima J, Goldberg VM et al. Characterization of cells with osteogenic potential from human marrow. Bone 1992;13:81–88. 2 Yoo JU, Barthel TS, Nishimura K et al. The chondrogenic potential of human bone-marrow-derived mesenchymal progenitor cells. J Bone Joint Surg Am 1998;80:1745–1757. 3 Wakitani S, Saito T, Caplan AI. Myogenic cells derived from rat bone marrow mesenchymal stem cells exposed to 5-azacytidine. Muscle Nerve 1995;18:1417–1426. 4 Majumdar MK, Thiede MA, Mosca JD et al. Phenotypic and functional comparison of cultures of marrow-derived mesenchymal stem cells (MSCs) and stromal cells. J Cell Physiol 1998;176:57– 66. 5 Chen J, Li Y, Katakowski M et al. Intravenous bone marrow stromal cell therapy reduces apoptosis and promotes endogenous cell proliferation after stroke in female rat. J Neurosci Res 2003;73:778 –786. 6 Li Y, Chen J, Zhang CL et al. Gliosis and brain remodeling after treatment of stroke in rats with marrow stromal cells. Glia 2005;49: 407– 417. 7 Dai W, Hale SL, Martin BJ et al. Allogeneic mesenchymal stem cell transplantation in postinfarcted rat myocardium: Short- and long-term effects. Circulation 2005;112:214 –223. 8 Gnecchi M, He H, Noiseux N et al. Evidence supporting paracrine hypothesis for Akt-modified mesenchymal stem cell-mediated cardiac protection and functional improvement. FASEB J 2006;20:661– 669. 9 Orlic D, Kajstura J, Chimenti S et al. Bone marrow cells regenerate infarcted myocardium. Nature 2001;410:701–705. 10 Kinnaird T, Stabile E, Burnett MS et al. Marrow-derived stromal cells express genes encoding a broad spectrum of arteriogenic cytokines and promote in vitro and in vivo arteriogenesis through paracrine mechanisms. Circ Res 2004;94:678 – 685.

tioned medium contains all the necessary factors to promote endothelial cell survival in cell culture. Our results support the hypothesis that hMSCs can serve as a powerful trophic mediator of endothelial cells.

ACKNOWLEDGMENTS We appreciate Dr. Shelly Cohen’s help in statistical interpretations of the data. We also thank Tamara M. Potapova for technical assistance with hMSC culture and collection of conditioned media. This work was supported by an American Heart Association Scientist Development Grant to S.V.D. and NIH Grants HL67101 and HL28958.

DISCLOSURE

POTENTIAL CONFLICTS OF INTEREST

OF

The authors indicate no potential conflicts of interest.

11 Gruber R, Kandler B, Holzmann P et al. Bone marrow stromal cells can provide a local environment that favors migration and formation of tubular structures of endothelial cells. Tissue Eng 2005;11:896 –903. 12 Kinnaird T, Stabile E, Burnett MS et al. Local delivery of marrowderived stromal cells augments collateral perfusion through paracrine mechanisms. Circulation 2004;109:1543–1549. 13 Kamihata H, Matsubara H, Nishiue T et al. Implantation of bone marrow mononuclear cells into ischemic myocardium enhances collateral perfusion and regional function via side supply of angioblasts, angiogenic ligands, and cytokines. Circulation 2001;104:1046 –1052. 14 Nagaya N, Fujii T, Iwase T et al. Intravenous administration of mesenchymal stem cells improves cardiac function in rats with acute myocardial infarction through angiogenesis and myogenesis. Am J Physiol Heart Circ Physiol 2004;287:H2670 –H2676. 15 Tang YL, Zhao Q, Qin X et al. Paracrine action enhances the effects of autologous mesenchymal stem cell transplantation on vascular regeneration in rat model of myocardial infarction. Ann Thorac Surg 2005;80: 229 –236. 16 Kale S, Biermann S, Edwards C et al. Three-dimensional cellular development is essential for ex vivo formation of human bone. Nat Biotechnol 2000;18:954 –958. 17 Winter A, Breit S, Parsch D et al. Cartilage-like gene expression in differentiated human stem cell spheroids: A comparison of bone marrowderived and adipose tissue-derived stromal cells. Arthritis Rheum 2003; 48:418 – 429. 18 Kelm JM, Timmins NE, Brown CJ et al. Method for generation of homogeneous multicellular tumor spheroids applicable to a wide variety of cell types. Biotechnol Bioeng 2003;83:173–180. 19 Nichols MG, Foster TH. Oxygen diffusion and reaction kinetics in the photodynamic therapy of multicell tumour spheroids. Phys Med Biol 1994;39:2161–2181.

See www.StemCells.com for supplemental material available online.