Small-diameter human vessel wall engineered from bone marrow-derived mesenchymal stem cells

The FASEB Journal • Research Communication

Small-diameter human vessel wall engineered frombone marrow-derived mesenchymal stem cells(hMSCs)

Zhaodi Gong and Laura E. Niklason1

Department of Anesthesiology, Yale University Medical Center, New Haven, Connecticut, USA

ABSTRACT Using biodegradable scaffold and a bio-mimetic perfusion system, our lab has successfullyengineered small-diameter vessel grafts using endothe-lial cells (ECs) and smooth muscle cells (SMCs) ob-tained from vessels in various species. However, trans-lating this technique into humans has presentedtremendous obstacles due to species and age differ-ences. SMCs from elderly persons have limited prolif-erative capacity and a reduction in collagen production,which impair the mechanical strength of engineeredvessels. As an alternative cell source, adult human bonemarrow-derived mesenchymal stem cells (hMSCs) werestudied for their ability to differentiate into SMCs inculture plates as well as in a bioreactor system. In theformer setting, immunofluorescence staining showedthat MSCs, after induction for 14 days, expressedsmooth muscle �-actin (SMA) and calponin, early andmid-SMC phenotypic markers, respectively. In the lat-ter setting, vessel walls were constructed with MSC-derived SMCs. Various factors (i.e., matrix proteins,soluble factors, and cyclic strain) in the engineeringsystem were further investigated for their effects onhMSC cell proliferation and differentiation into SMCs.Based on a screening of multiple factors, the engineer-ing system was optimized by dividing the vessel cultureinto proliferation and differentiation phases. The ves-sel walls engineered under the optimized conditionswere examined histologically and molecularly, andfound to be substantially similar to native vessels. Inconclusion, bone marrow-derived hMSCs can serve as anew cell source of SMCs in vessel engineering. Optimi-zation of the culture conditions to drive SMC differen-tiation and matrix production significantly improvedthe quality of the hMSC-derived engineered vesselwall.—Gong, Z., Niklason, L. E. Small-diameter humanvessel wall engineered from bone marrow-derived mes-enchymal stem cells (hMSCs). FASEB J. 22, 000–000(2008)

Key Words: vessel engineering � smooth muscle cell � bioreactor� cyclic strain � extracellular matrix � soluble factors

In cardiac or peripheral bypass surgery, diseasedarteries are replaced with autologous veins or, lessfrequently, arteries. However, some patients in need of

such operations do not have suitable veins as replace-ment due to the systematic pathological changes intheir vascular system, or previous vein harvest (1).Although synthetic vascular prostheses, such as Dacronand expanded polytetrafluoroethylene (ePTFE), havetheir own merits as conduit in high-flow low-resistanceconditions (such as large peripheral arteries), theirperformance as small diameter vessel graft is far fromsatisfactory (2–5). Biological vascular grafts may be aviable solution, being composed of cellular compo-nents and able to respond physiologically to varioushemodynamic forces and chemical stimuli. Tremen-dous progress has been made in vascular engineeringfield since Weinberg and Bell constructed one of thefirst blood vessels 20 years ago using endothelial cells(ECs) and smooth muscle cells (SMCs) in gelatin (6).Since that time, various approaches to culture vasculargrafts have been developed, including gelatin-based,biodegradable scaffold-based, and sheet-based grafts.

Using polyglycolic acid (PGA) as a scaffold and abiomimetic system, our lab has successfully engineeredvessels with excellent mechanical strength in variousspecies, including bovine, porcine, and canine (7–9).However, translating vascular engineering techniquesfrom animal to human vascular cells has met with somedifficulties. We recently reported the impact of telom-erase (hTERT) gene therapy on culturing human ves-sels (10, 11) Although hTERT expression enabled theculture of engineered human blood vessels, extensionof cellular life span did not appear to alter the intrinsicaging-associated cellular changes of vascular SMC, suchas decline in collagen synthesis and consequent de-creased burst strength.

Based in part on these observations, it seems reason-able to seek alternative cell sources that may be suitablefor vessel engineering. Human stem and progenitorcells have been isolated from a wide range of sources.Their autologous origin, high proliferative capacity,and potential to differentiate into vascular phenotypeshave generated significant attention. Mesenchymalstem cells (MSCs) are variously defined as multipotent

1 Correspondence: Vascular Biology and Transplantation,Yale School of Medicine, P.O. Box 208089, New Haven, CT06520-8089, USA. E-mail: [email protected]

doi: 10.1096/fj.07-087924

10892-6638/08/0022-0001 © FASEB

The FASEB Journal article fj.07-087924. Published online January 18, 2008.

adult stem cells that are present in the bone marrow(12, 13) and other tissues (14–16) and have the abilityto differentiate into multiple cell lineages includingosteoblasts, adipocytes, chondrocytes, myoblasts, andearly progenitors of neural cells (13–15, 17, 18). Theability of MSCs to differentiate into myocytes has beenshown previously. For instance, mesenchymal precur-sor 10T1/2 cells differentiated into skeletal myoblastsafter treatment with 5-azacytidine (19). Galmiche et al.documented that stromal cells from human long-termmarrow cultures are mesenchymal cells that differenti-ate along a vascular smooth muscle differentiationpathway (20). MSCs from the adult rat have beenshown to have a potential to differentiate into SMCswhen exposed to TGF�1 (21). Human MSCs alsoexhibited enhanced differentiation with increased con-tractility in response to TGF�1 (22). In addition togrowth factors, cell-cell contact also plays an importantrole in the creation of an environment conductive toSMC differentiation. Hirschi and colleagues showedthat heterotypic cell-cell interactions mediated EC-induced recruitment of 10T1/2 cells and their differ-entiation to smooth muscle cells (23). In anotherMSC-EC coculture experiment, direct contact with ECsaugmented SMA expression (24).

Beyond these studies of differentiation potential ofMSCs, the application of bone marrow-derived stemcells in vitro and in vivo for vascular engineering is justemerging (25–32). Populations of marrow-derived cellshave been used to regenerate vessels in canine andbovine models, but in some cases these populations arenot fully characterized and the differentiation impactof various factors (such as substrate and mechanicalstimulation) are not well understood (29, 31). Inseveral studies involved in the engineering of heartvalves, the isolation, expansion, and phenotype confir-mation hMSCs were well documented before scaffoldseeding (26–28).

We speculated that various factors that are associatedwith the intact or regenerating vascular wall may beinvolved in directing the differentiation of hMSCstoward an SMC phenotype. Specifically, growth factorsthat are elaborated by platelets and vascular cells aftervessel injury (PDGF, TGF-�1, and bFGF), extracellularproteins found in native vessel wall, and cyclic mechan-ical strain were all examined for their impact on SMCdifferentiation from hMSCs. Intriguingly, many ofthese factors were found to influence extent of differ-entiation toward an SMC fate, implying that local cueswithin injured vessels in vivo may direct hMSCs towarda vessel reparative function.

To test whether human bone marrow-derived MSCscan directly differentiate into SMCs that are functionalfor arterial engineering, vessel walls were engineeredusing human bone marrow-derived MSCs in a biomi-metic culture system in vitro. A vessel engineeringprotocol, having proliferation and differentiationphases, was designed to promote hMSC proliferationand SMC differentiation, respectively. The culturedcells and the engineered vessel wall were examined

histologically for extracellular matrix (ECM) compo-nents, as well as molecularly for the expression of SMCphenotypic markers. When induced to expand andthen differentiate into SMCs in the bioreactor, hMSCsproved to be an excellent starting material for arterialengineering.

MATERIALS AND METHODS

Isolation and cell culture of hMSCs

hMSC cell culture was established as described previously(33) from six different donors (ages 22–45 y, three for eachgender). Briefly, fresh unprocessed human bone marrow(Lonza, Basel, Switzerland) was slowly loaded onto Ficoll-Paque Plus density media (StemCell Technologies, Vancou-ver, BC, Canada) and fractionated at room temperature for30 min at 1200 g with no brake. The mononuclear cell layerat the interface was removed and washed once with Dul-becco’s phosphate buffered saline (PBS; Life Technologies,Inc., Gaithersburg, MD, USA) before being plated in 75 cm(2) flasks in Dulbecco’s Modified Eagle’s Medium (DMEM;Life Technologies, Inc.) containing 10% selected lot of fetalbovine serum (FBS; Hyclone, South Logan, UT, USA) and1% penicillin-streptamycin-glutamate (Life Technologies,Inc.). The screening of the FBS lot was based on the supportof cell proliferation. Cultures were maintained at 37°C in ahumidified atmosphere containing 5% CO2. Medium waschanged 2�/wk. Cultures were passaged to 75 cm2 flasks forprotein analyses or in chamber slides for histochemicalanalysis.

Fluorescence-activated cell sorting (FACS) analysis

Cells were tested for purity by flow cytometry after isolation.FACS analysis was performed on cultured hMSCs from threedifferent donors. Fluorescein isothiocyanate (FITC) -conju-gated mouse anti-human IgGs (CD14, CD45, and CD34 fromAbcam Inc., Cambridge, MA, USA; Santa Cruz Biotechnol-ogy, Santa Cruz, CA, USA; and Miltenyi Biotec Inc., Auburn,CA, USA, respectively) were utilized. Proper isotype IgGs wereserved as controls. For SH2 and SH3 staining, 100 �l super-natants from SH2 and SH3 hybridoma cell (American TypeCulture Collection, Manassas, VA, USA) culture were utilized.After staining, the cells were fixed in 4% paraformaldehyde,and quantitative FACS was performed on a FACStar flowcytometer (BD, Franklin Lakes, NJ, USA).

Confirmation of hMSCs phenotype

MSCs are commonly defined by their ability to differentiateinto osteogenic and adipogenic lineages.

Osteogenic induction

hMSCs (3�104) were seeded in 2 ml mesenchymal stem cellmedium (MSCGM, Lonza) per well of 6-well plates andallowed to adhere to the culture surface for 24 h beforereplacing the MSCGM with the osteogenesis induction me-dium (Lonza). The induced hMSCs were fed every 3–4 daysfor 3 wk by completely replacing the medium with freshosteogenesis induction medium. The noninduced controlhMSCs were fed with MSCGM on the same schedule.

Alkaline phosphatase (ALP) and von Kossa staining (formineralization nodule-formation) was performed to examine

2 Vol. 22 June 2008 GONG AND NIKLASONThe FASEB Journal

osteogenesis in hMSC culture. Mineralized bone noduleswere identified by double labeling for von Kossa stain andALP as described (34).

Adipogenesis induction

hMSCs (2�105) were seeded in 2 ml MSCGM medium perwell of a 6-well plate and fed every 3–4 days until the culturereached confluence. At full confluence, three cycles of induc-tion/maintenance treatment were performed to stimulateoptimal adipogenic differentiation. Each cycle consisted offeeding the hMSCs with adipogenesis induction medium(Lonza) for 3 days followed by 3 days of culture in adipogenicmaintenance medium (Lonza). Noninduced control hMSCswere fed with adipogenic maintenance medium only on thesame schedule. After three complete cycles of induction/maintenance, the extent of adipogenesis was examined underthe microscope by fixation with 10% NBF and stained with 1�g/ml Nile red for 5 min for lipid formation.

Induction of hMSCs into SMCs

hMSCs were grown in basal medium containing low-glucoseDMEM (Life Technologies, Inc.) supplemented with 10%selected lot of premium fetal bovine serum (FBS, Hyclone)and penicillin-streptomycin-glutamine (Life Technologies,Inc.) on glass chamber slides (Nunc, Roskilde, Denmark).SMC induction was performed by adding 1 ng/ml TGF�1 tothe basal culture medium. The control culture was grown inparallel in basal medium. Immnofluorescence staining wasperformed (as described below) for smooth muscle �-actin(SMA) and calponin on day 14. Coronary artery smoothmuscle cells (CASMCs) were cultured as positive controls insmooth muscle growth medium (SmBM) supplemented withSmGM-2 SingleQuots (Lonza) and stained with hMSCs.

Immunofluorescence studies

Immnofluorescence staining for SMC markers was performedas described previously (24). The paraformaldehyde-fixedchamber slides were incubated with monoclonal primaryantibodies (1:100 dilution for both SMA and calponin, and1:50 for SM-MHC, smooth muscle myocin heavy chain; Dako,Copenhagen, Denmark) for 1 h at room temperature. Afterthree washes with PBS, the secondary antibody, a goat anti-mouse IgG conjugated with FITC (1:2000, Santa Cruz Bio-technology) was added and incubated for 30 min at roomtemperature. Nuclei were stained with 4�,6-diamidine-2-phenylindole (DAPI), which was contained in the VectashieldMounting Medium for Fluorescence with DAPI kit (VectorLaboratories, Inc., Burlingame, CA, USA).

Effect of various matrix proteins on hMSCs

In these 2-D culture studies, hMSCs were seeded at 2 �104/ml in DMEM plus 10% FBS medium on 6-well platesuntreated or coated with collagen type I, IV, elastin, fibronec-tin, and laminin (Flexcell International, Hillsborough, NC,USA). After 7 days, the cells were detached and enumeratedwith 3% acetic acid with methylene blue (StemCell Technol-ogies) on a hemacytometer. Cells were lysed with radioimmu-noprecipitation assay (RIPA) buffer with Triton X-100 (Bos-ton BioProducts, Boston, MA, USA) with protease inhibitorcocktail (Sigma, St. Louis, MO, USA). Protein lysates werequantified by Bradford assay, using the Quick Start BradfordDye Reagents (Bio-Rad, Hercules, CA, USA) according to the

manufacturer’s protocol, and stored at �80°C for Westernblot for SMA and calponin.

Effect of various soluble factors on hMSCs

hMSCs were seeded in 6-well plates at 5.6 � 103/cm2 inDMEM plus 5% FBS medium with one of the followingsupplements: 0 (control), 0.01, 0.1, 1, or 10 ng/ml transform-ing growth factor �1 (TGF�1, R&D Systems, Inc., Minneapo-lis, MN, USA); 10 ng/ml PDGF-BB, PDGF-CC, and bFGF(R&D Systems, Inc.); or 50 �g/ml ascorbic acid. Media werechanged on day 4. After 7 days of treatment, cells wereharvested for cell proliferation and differentiation analysis.

Effect of cyclic strain on hMSCs

hMSCs seeded (4�104 cells/ml for cell proliferation and2�104 cells/ml for cell differentiation) in DMEM plus 10%FBS were subjected to cyclic strain using a Flexcell 4000T unit(Flexcell International) in the presence or absence of 10ng/ml PDGF-BB. The stain unit is a computer-regulateddevice that applies cyclic tensile strain to the cell culturethrough regulated vacuum pressure on the bottom cultureplates with a flexible membrane that is untreated or pre-treated with fibronectin and type I collagen (Flexcell Inter-national). The strain causes the flexible plate to stretch acrossa cylindrical loading post to provide equibiaxial strain to thecells. hMSCs were subject to equibiaxial cyclic strain for 5 daysat a frequency of 0.5 Hz, resulting in �8–12% substrateelongation. Unstrained controls were hMSCs prepared in anidentical manner and cultured on unstrained untreated orcollagen I or fibronectin-coated flexible plates for 5 days. Atthe end of the experiment, cell enumeration, protein isola-tion, and Western blot were performed.

Blood vessel wall culture

Basic engineering protocol

As an initial control study to determine the behavior ofhMSCs in bioreactor conditions, human blood vessels (n�6)were engineered from hMSCs using techniques similar tothose previously described (7, 8). To prepare for blood vesselculture, hMSCs were preconditioned in enhanced DMEMmedium (8) supplemented with 10 ng/ml PDGF-BB andbFGF for 48 h before they were seeded on scaffolds in thebioreactor. The enhanced DMEM, supplemented with 20%serum, 10 ng/mL bFGF, and PDGF-BB has been character-ized previously as stimulating SMC growth and collagenmatrix production during engineered vessel growth (8).

Blood vessel bioreactors and PGA mesh scaffolds (Concor-dia Manufacturing LLC, Coventry, RI, USA) were prepared asdescribed previously (7, 8) and seeded with hMSCs (5�106

cells for each vessel). Bioreactors were filled with enhancedDMEM medium and fed with fresh enhanced DMEM for halfof the bioreactor volume 1�/wk. The pulsatile perfusionsystem applied �5% cyclic strain (7) from wk 2–8 of culture.Ascorbic acid was added to the bioreactor 3�/wk. Vesselculture continued for 8 wk, at which time the vessels wereharvested from the bioreactors.

Optimized culture protocol

Based on the results from matrix protein and soluble factorscreen and cyclic strain experiments (summarized in Fig. 6A),the engineering system was optimized to induce SMC differ-entiation, as shown in Fig. 6B. Seven vessels were engineered

3ADULT BONE MARROW-DERIVED MESENCHYMAL STEM CELLS

based on the optimized protocol and compared with the onesgenerated with the original protocol.

Endothelialization of engineered vessel wall

To show feasibility of EC adhesion to the engineered vessellumen, endothelial seeding was performed on one engi-neered vessel. human umbilical cord-derived endothelial cells(HUVECs; 2.6�106) resuspended in DMEM/10% FBS wereseeded into the lumen of the vessel with a syringe. The endsof the vessel were clamped, and the bioreactor was manuallyrotated every 15 min for 40 min at different positions to alloweven distribution of the HUVEC cells. Then the bioreactorwas returned to the incubator. After 8 h, flow was restarted toallow the perfusion of endothelialized vessel lumens withenhanced DMEM medium at a flow rate that graduallyincreased from �0.02 to 0.1 ml/s (shear stress 0.25–1 dyne/cm2) over 18 h of duration.

Scanning electron microscopy (SEM)

Vascular grafts (endothelized and nonendothelialized) wereharvested, fixed, and prepared as described previously (8).The dried samples were coated with gold using a CressingtonC108 autocoater (Cressington Scientific Instruments, Ltd.,Watford, UK) and examined by SEM (XL30ESEM-FEG; FEICompany, Hillsborough, OR, USA).

Analysis of the engineered vessel walls

Each engineered vessel was cut into three segments, whoseweight was measured and recorded. One segment of eachvessel was fixed immediately in 10% neutral buffered forma-lin (NBF) for 1 h followed by dehydration and embedding inparaffin. Sections (4 �m thick) were cut, deparaffinized, andstained with hematoxyline and eosin (H&E), MassonTrichrome for collagen, Verhoeff-Van Gieson (VVG), andMovat stains for elastin production and immunohistochemi-cal staining. The tissue sections from the engineered humanvessel walls were also stained immunohistochemically forproliferating cell nuclear antigen (PCNA) using the PCNAStaining Kit (Zymed Laboratories Inc., Burlingame, CA, USA)according to the manufacturer’s instructions.

Total protein lysates were isolated from segments of eachengineered vessel with RIPA buffer containing protease in-hibitor cocktail as described above.

Collagen analysis

Collagen analysis was performed according to a protocolpreviously published (10, 35). A 1:10 w/w ratio of hy-droxyproline and collagen was used to calculate the collagencontent of the vessels. The collagen content was calculated asthe percentage of the dry weight. Human umbilical cordartery was included as control.

Western blot analysis

Lysates containing 25 �g of protein from each vessel or cellculture were separated on 10% SDS-PAGE precast gel (Bio-Rad) and transferred to polyvinylidene difluoride (PVDF)membranes (Millipore Corp., Bedford, MA, USA). The mem-branes were incubated for 1 h at room temperature inblocking buffer (5% nonfat dry milk in TBST buffer) andthen incubated with mouse anti-human primary antibodies,

1:200, 1:200, 1:50, and 1:30 dilution in 1% nonfat dry milk(Bio-Rad) in TBST buffer for SMA, calponin, SM-MHC, andsmoothelin (HyCult Biotechnology, Uden, The Netherlands),respectively. Then the blots were probed with horseradishperoxidase (HRP)-conjugated goat anti-mouse IgG (1:2000;Santa Cruz Biotechnology) for 30 min at room temperature.After washing, the blots were developed using the SuperSig-nal West Pico Chemiluminescent detection system (Pierce,Rockford, IL, USA) and exposed to X-ray film to detect theprotein bands. After visualization, membranes were strippedin RestoreTM Western blot Stripping Buffer (Pierce) at roomtemperature for 5–15 min and reprobed for �-actin (1:5000dilution, Sigma), which was used as equal loading control.Quantification of Western blots was performed using Image J(National Institutes of Health, Bethesda, MD, USA). Theresults were presented as relative density after correction with�-actin.

Statistical analysis

All quantitative results were obtained from five samples forcell proliferation analyses and from triplicate samples forWestern blot analyses. Data were expressed as the mean sd.Statistical analysis was performed using Student’s t test. Avalue of P 0.05 was considered to be statistically significant.

RESULTS

Phenotype of isolated hMSC

The surface marker expression of the isolated hMSCsfrom the bone marrow was analyzed by FACS (Fig. 1A).hMSCs were negative for CD14, CD34, and CD45,known markers for monocyte/macrophage, hemato-poietic progenitor cells, and differentiated nonerythro-cyte hematopoietic cells, respectively. However, theywere positive for SH2 and SH3, antibodies that havebeen shown to recognize CD105 (endoglin) and CD73on human MSCs, respectively, but are nonreactive toosteoblasts and osteocytes (36).

hMSC phenotype was further confirmed by osteo-genic and adipogenic induction. Osteogenic inducedhMSCs exhibited changes in cell morphology fromspindle shape to cuboidal shape as they differentiatedand mineralized. Osteogenesis was further examinedafter 3 wk induction by ALP and von Kossa staining,where ALP-positive areas appeared red, while the min-eralized areas were brown to black in color in osteo-genic culture, exhibiting high ALP enzyme activity andmineralization (Fig. 1C). In contrast, no ALP activity ormineralized bone nodules were found in the controlculture (Fig. 1B). Confluent hMSC cultures were alsotreated with adipogenic induction media. After 8 days ofadipogenic induction/maintenance treatment, lipid vacu-oles started to appear in the confluent hMSC culture(data not shown). By day 18, after three cycles of adipo-genic induction and maintenance, lipid-containing adipo-cytes became very prominent, which stained positive forNile red, a lipophilic fluorescence dye (Fig. 1E). Incontrast, no Nile-red positive cells were found in the


hMSC control culture (Fig. 1D). These studies, in combi-nation with the FACS data, confirm the mesenchymalstem cell identity of the starting cellular material.

Induction of hMSCs into SMCs

hMSCs expressed basal levels of SMA and calponinwhen cultured in control conditions (DMEM plus 10%FBS) (Fig. 2a, d). hMSCs were partially induced into anSMC phenotype by exposure to 1 ng/ml TGF�1. Thepartial phenotypic switch from MSC to SMC was dem-

onstrated by immunofluorescence staining for smoothmuscle �-actin (SMA) and calponin, which are earlyand mid-SMC markers, respectively. After 14 days ofexposure to TGF�1, hMSCs were stained positively forSMA (Fig. 2b) and calponin (Fig. 2e) but not forSM-MHC (a late marker of SMC lineage, not shown),indicating a committed yet immature stage of SMCdifferentiation from MSCs after 14 days of exposure.The distribution of SMA and calponin expression inMSCs was similar to that of CASMCs (Fig. 2c, f ), whichserved as positive control.

Figure 1. Confirmation of hMSC phenotype. A) Representative flow cytometric analysis showing reactivity of SH2 and SH3antibody with hMSCs isolated from human bone marrow (green curve). hMSCs were negative with CD14, CD34, and CD45antibody (green curve). Staining with the isotype-matched control antibody is shown in blue; cells-only controls are shown inred. B–E) Osteogenic and adipogenic induction of hMSC culture was performed. Osteogenesis was examined 3 wk aftertreatment in control (B) or induction (C) cultures by ALP and von Kossa staining. After three complete cycles of induction andmaintenance, the extent of adipogenesis was examined by Nile red staining of control (D) and induction (E) cultures.

Figure 2. SMC induction in hMSC culture. hMSCs were grown in basal DMEM medium on glass chamber slides without (a, d)or with (b, e) the addition of 1 ng/ml TGF�1. Immnofluorescence staining (green) was performed for SMA (a–c) and calponin(e, f) on day 14. Nuclei were stained with DAPI (blue). CASMCs were stained for SMA (c) and calponin (f) as positive controls.


Effect of various matrix proteins on hMSCs

To determine the impact of various ECM proteins thatare present in the vascular wall on hMSC differentia-tion, we cultured hMSCs on Flexcell 6-well cultureplates that had been precoated with ECM molecules(and not subjected to cyclic strain). None of theexamined matrix proteins (collagen type I, collagentype IV, laminin, elastin, and fibronectin) had a signif-icant effect on hMSC proliferation as compared tountreated surfaces. The presence of 10 ng/ml PDGF-BB, which is an important mitogen and chemoattrac-tant for mesenchymal cells (37–38), increased hMSCproliferation similarly across all the different matrixproteins studied (Fig. 3A).

As assessed by immunoblotting, no difference wasfound in SMA expression between each matrix proteinand the untreated surface, with or without PDGF-BB(Fig. 3B, n�5). ECMs including collagen I, IV, elastin,and fibronectin significantly increased calponin expres-sion compared to untreated surfaces. This is an interest-ing finding, given that these four matrix proteins arepresent in native vasculature at different locations. Incontrast, laminin exerted the opposite effect on calponincompared to control (Fig. 3C, D). The inhibitory effect ofPDGF-BB on SMA was confirmed in hMSCs cultured onuncoated or elastin and fibronectin-coated plates (Fig.

3B), while PDGF-BB decreased calponin on almost allcoating matrix proteins except elastin (Fig. 3C).

Effect of various soluble factors on hMSCs

The enhanced culture medium that is used to growengineered arteries (7, 8) contains soluble factors suchas PDGF-BB, bFGF, vitamin C, and a high percentage ofserum. Serum, in turn, contains a variety of growth factors,including those liberated from platelets, that affect mesen-chymal cell behavior (e.g., TGF-�1 and PDGF). Seven daysof exposure to various soluble factor revealed that TGF�1induced a dose-dependant inhibition of hMSC cell prolif-eration at concentrations between 0.1 and 10 ng/ml. Incontrast, other factors, including PDGF-BB, PDGF-CC,bFGF, and vitamin C, significantly increased hMSC cellnumber over a 7-day period (Fig. 4A).

TGF�1, although not affecting SMA protein expres-sion (Fig. 4B), significantly increased calponin expres-sion in hMSCs, in a dose-dependent fashion, from 0.1to 10 ng/ml (Fig. 4C, D). Of note, expression levels ofcalponin in hMSCs exposed to 10 ng/ml of TGF-�1were similar to levels expressed by CASMCs with similarprotein loading. The differentiation effect of TGF�1 isconsistent with results reported by other groups. How-ever, PDGF-BB, PDGF-CC, and bFGF all significantly

SMA

Calponin

Figure 3. Effect of various matrices on hMSC cell proliferation and differentiation. hMSCs were seeded at 2 � 104/ml in DMEMplus 10% FBS medium on 6-well plates untreated or coated with collagen type I, IV, elastin, fibronectin, and laminin in thepresence or absence of 10 ng/ml PDGF-BB. A) After 7 days, cell numbers were counted with methylene blue on ahemacytometer for cell proliferation (**P0.05, PDGF-BB treatment vs. no BB; no difference between each matrix anduntreated with or without PDGF-BB; n�5). B, C) Western blots were performed on the protein lysates from each treatmentgroup on SMA (B) and calponin (C, *P0.05 vs. untreated; **P0.05, PDGF-BB vs. no BB; n�3). D) Representative Westernblot on SMA, calponin, and �-actin.


reduced both SMA (Fig. 4B, D, E) and calponin (Fig. 4C,D) levels. Since both PDGF-BB and bFGF are included inthe enhanced DMEM medium that is used for basic vesselculture, it might be anticipated that these factors wouldinhibit hMSC differentiation into SMC.

Effect of cyclic strain on hMSCs

Native vessels, and engineered vessels during culture,are exposed to cyclic mechanical strain. Cyclic straininhibited cell proliferation after 5 days applicationwhen hMSC were cultured on three different kinds ofmatrices—untreated, pronectin (fibronectin), and col-lagen type I (Fig. 5A). In the absence of matrix coating,cyclic strain alone significantly decreased SMA expres-sion. In the presence of PDGF-BB, cyclic strain dis-

played opposite effect on SMA expression: an increaseon collagen I matrix and a decrease on fibronectin(Fig. 5B). In the absence of PDGF-BB, calponin levelwas significantly decreased on collagen I but increasedon fibronectin by cyclic strain (Fig. 6C). The inhibitionof PDGF-BB on SMC protein expression was furtherconfirmed regardless of cyclic strain and differentmatrix coating, suggesting that PDGF-BB should beavoided during SMC differentiation despite its potentmitogenic effect on hMSCs. The observation that fi-bronectin increased calponin expression in the ab-sence of cyclic strain (Fig. 5C comparison of calponin inS�B� in nontreated vs. fibronectin-coated, P0.05) isconcordant with the results from Fig. 3C.

In summary, while cyclic strain inhibited hMSC pro-liferation independent of ECM (untreated, collagen

E

Figure 4. Effect of various factors on hMSC cell prolifer-ation and differentiation. hMSCs were seeded at 5.6 �103/cm2 in DMEM plus 5% FBS medium with one of thefollowing supplements: 0 (control), 0.01, 0.1, 1, or 10ng/ml TGF�1; 10 ng/ml PDGF-BB, PDGF-CC, bFGF; or50 �g/ml ascorbic acid. A) After 7 days, the cells wereenumerated with 3% acetic acid with methylene blue ona hemacytometer (**P0.05 vs. control; n�5). B, C)Western blots were performed on the protein lysates fromeach treatment group on SMA (B) and calponin (C;**P0.05 compared to control, n�3). D) RepresentativeWestern blot on SMA, calponin, and �-actin. E) Immu-nofluorescence staining for SMA in hMSC culture aftertreatment with TGF�1, PDGF-BB, PDGF-CC, bFGF, andvitamin C in addition to control for 7 days. Concentrationof each factor is labeled in the figure.


type I, or fibronectin), its influence on hMSC differen-tiation was significantly tied to the presence of ECM aswell as PDGF-BB.

Optimization of the engineering system

Based on the above results, we developed an “opti-mized” protocol for engineering vessels from hMSCsthat was divided into two phases. Under the optimizedconditions, instead of culturing the vessel walls in thesame enhanced DMEM medium for 8 wk, the cultureperiod was divided into 4 wk of proliferation phase andanother 4 wk of differentiation phase. Before cellseeding, PGA scaffold was precoated with 10 �g/cm2

fibronectin (BD Biosciencies, San Jose, CA, USA),which in combination with cyclic strain stimulatedhMSC differentiation into SMC (Fig. 5C). The basiccomponents of the enhanced DMEM medium were

essentially the same as the original protocol, exceptbFGF was removed from the culture medium due to itspotent inhibition on SMC differentiation. Anotherpotent hMSC mitogen, PDGF-BB, was retained duringthe first 4 wk of culture. After 4 wk, PDGF-BB wassubstituted with 1 ng/ml TGF�1, which had been seento significantly enhance SMC differentiation fromhMSCs (Fig. 4C). The pulsatile cyclic strain in thebioreactor was not initiated until wk 4, when thedifferentiation phase started, due to its observed inhi-bition of hMSC cell proliferation.

Histological comparison of vessels walls engineeredbefore and after optimization

Engineered vessels were cultured using previously de-scribed protocols and using the “optimized” protocoldescribed above. Thirteen SMC vessel walls were engi-

Figure 5. Effect of cyclic strain on hMSC cell proliferation and differentiation. hMSCs were cultured in 6-well plates noncoatedor coated with collagen I or fibronectin. In each surface treatment group, there were four culture conditions: control (S�B�),PDGF-BB only (10 ng/ml, S�B�), cyclic strain only (S�B�), and PDGF plus cyclic strain (S�B�). A–C) Cells were subject tothe detailed condition for 5 days and harvested for cell number (A) and Western blot on SMA (B) and calponin (C). D)Representative Western blot film (n�3).


neered in total from six different donors: half accord-ing to the original protocol (n�6), and the rest usingthe optimized protocol (n�7). After 8 wk of bioreactorculture, the histological appearance (H&E stain) of thevessel walls cultured according to the original protocolhad some similarity to that of native vessels (Fig. 7a, c).Masson’s trichrome stain of the engineered vessel re-vealed some production of collagen (blue stain, Fig. 7b,d). Remnants of the PGA polymer near the lumen ofthe vessel wall were stained dark blue, indicating incom-plete degradation of PGA after 8 wk of culture, whichhas been observed previously (7). Movat and VVGstains did not reveal the deposition of elastin (data notshown).

With optimization, both cellularity and collagen pro-duction were substantially improved in the engineeredvessel wall based on H&E (Fig. 7e, g) and Masson’sTrichrome stains (Fig. 7f, h), respectively. PCNA stainshowed that more cells were proliferating after optimi-zation, which distributed evenly through the vessel wall,in contrast to few proliferative cells only in proximity tothe lumen of the vessel wall before optimization (Fig.7i, j). The result that more cells are proliferating at theend of culture using the “optimized” conditions issomewhat surprising, given that conditions were cho-sen to favor differentiation during the latter part ofculture. Local paracrine factors secreted by differenti-ating SMC may have contributed to the observed in-

crease in replication and cellularity under “optimized”conditions.

Collagen production was quantified by hydroxypro-line assay. Without optimization, the average collagencontent of engineered vessels was 5.1 2.5% of totaldry weight (n�6). After optimization, collagen signifi-cantly increased to 22.1 7.8% of dry weight (P0.05,Fig. 8A, n�4), which is roughly half of the collagencontent of native vessels (7). But based on our mechan-ical testing of burst pressure on two vessels engineeredfrom each protocol, the vessel walls engineered withthe optimized protocol had higher burst pressures (7–8psi) compared to those engineered with the originalprotocol (�3–4 psi). Both the burst pressure andhandling of the vessel walls engineered with the two-phase protocol were superior to the ones with originalprotocol.

All of the engineered vessels expressed SMA andcalponin proteins by immunoblotting (Fig. 8B). WhileSMA may be expressed by myofibroblasts and pericytes,calponin expression generally connotes true smoothmuscle cell differentiation. After optimization, the ex-pression levels of SMA and calponin were substantiallyincreased to levels comparable with the CASMC con-trol. SM-MHC, late SMC marker, and smoothelin, acytoskeletal protein that is only found in contractilesmooth muscle cells (39), were not detected in any ofthe engineered vessel walls (data not shown), indicat-

Figure 6. A) Effects of various soluble factors,matrix proteins, and cyclic strain on hMSC cellproliferation and differentiation. B) Optimiza-tion of the engineering system to grow a vesselwall using hMSCs.


ing an incomplete differentiation into an SMC pheno-type. Though collagen production was robust in engi-neered vessels, Western blots for tropoelastin did notreveal this elastin monomer (data not shown), andmature elastin was not detected by staining. This resultis in contrast to results from differentiated human SMCused for vascular engineering, wherein tropoelastinproduction is detectable throughout culture, despitelack of mature, insoluble elastin (10).

Scanning electron microscopy

Without endothelialization, sheets of tightly packedcollagenous ECM produced by hMSCs were evident onthe inner lumen of the engineered vessel (Fig. 9A). Incontrast, the lumen of the endothelialized vessel wallappeared very smooth, with flattened ECs covering thewhole lumen and little evidence of the underlying ECMand hMSCs (Fig. 9B). This result shows the feasibility ofproducing a complete engineered vessel from hMSCsand luminal ECs.

DISCUSSION

A long-term goal of this study was to explore thefeasibility of using hMSCs from adult marrow as a newcell source to generate a vascular smooth muscle cellwall in engineered arteries for adult humans. Consis-tent with previous reports, we showed that factors suchas TGF�1 induced a partial conversion of hMSC to anSMC phenotype. We also analyzed the impact of othergrowth factors, substrate molecules and mechanicalstimulation on the phenotypic differentiation of hMSC.

Using information gleaned from the study of individualfactors, we devised an “optimized” protocol wherebyhMSC proliferation and differentiation could be con-trolled in the vessel engineering bioreactor. Under theoptimized protocol, we observed significantly morecellularity and collagen deposition than those obtainedunder standard conditions. Importantly, many of thefactors that drove both hMSC proliferation and differ-entiation into an SMC phenotype are factors that arefound in the native or regenerating vessel wall.

Interestingly, evidence of partial induction of hMSCsinto SMC was obtained after 2 wks’ culture in enhancedDMEM medium that has been used to grow vessels inmultiple species. By immunofluorescence and Westernblotting, we noticed some basal level expression of SMAand calponin in hMSC cultured in control condition(DMEM plus 10% FBS). This finding is in agreementwith previous report that bone marrow stromal cellshave a phenotypic similarity to a subset of vascularsmooth muscle cells (20). Because of the high TGF�1level in our selected serum lot (�2 ng/ml in 10% FBSmedium), which significantly increased SMA and cal-ponin expression, we believe TGF�1 might be thecritical factor that led to the basal expression of SMCmarkers. Since it is inevitable to have TGF�1 in any lotof FBS, it is not surprising that SMA and calponin areexpressed in other MSC cultures (40, 41).

The fact that we did not observe SM-MHC, late SMCmarker, after 2 wks’ exposure to TGF�1 suggests thatthe MSC-derived SMC were still at an early differentia-tion stage. Furthermore, PGA scaffold hydrolysis hasbeen found to induce dedifferentiation of SMC in theproximity of polymer remnants in engineered vessels,which may have contributed to the incomplete differ-

Figure 7. Representative histological stainings on engi-neered vessel walls before (a–d, i) and after (e–h, j) theoptimization of engineering conditions: H&E (a, c, e, g),Masson’s Trichrome (b, d, f, h), PCNA (i, j). Scale bar isshown in each picture.


entiation observed in engineered vessels under bothoriginal and optimized conditions (8).

In investigating factors that may drive hMSC differ-entiation toward a vascular SMC phenotype, we focusedon those soluble, matrix and physical factors that maybe associated with the intact or regenerating vessel wall.Specifically, growth factors elaborated by platelets andvascular cells after vessel injury (PDGF, TGF-�1, andbFGF), extracellular proteins found in native vesselwall, and cyclic mechanical strain were all examined.Intriguingly, many of these factors were found toinfluence the extent of differentiation toward an SMCfate, implying that local cues within injured vessels invivo may direct hMSC toward a vessel reparative func-tion.

Inadequate collagen synthesis by the partly differen-tiated SMC (5% by dry weight) under the originalprotocol contrasted with the average vessel collagencontent under the optimized protocol (�20% by dryweight). This increase in collagen synthesis is consistentwith a more complete conversion to SMC phenotypeunder the optimized protocol. Although all the engi-neered vessel walls expressed SMA and calponin, these

appeared to be more highly expressed in vessels cul-tured under optimized conditions, as compared tooriginal conditions. Indeed, though myosin heavychain and smoothelin were not detectable in hMSC-derived vessels cultured for 8 wks, levels of alpha actinand calponin were comparable to those of CASMC.

Whereas one principle function of SMC is contrac-tion, this cell also has an important synthetic function.It is the major source of the ECM components of theblood vessel wall (42–44), which contributes impor-tantly to the mechanical strength of the vessel. Thecontractile and synthetic functions of SMCs appear tobe inversely correlated and may be described by theterm phenotypic modulation (45). The extent of SMCphenotypic modulation and plasticity appears to bedependent on many factors, among which are solublefactors, ECMs, and mechanical forces (46). The ECMthat surrounds cells is a highly organized and dynamicstructure that contributes to the control of cellularfunction and is involved in the maintenance of SMCs’state of proliferation and differentiation (47). Amongthe five ECMs we’ve examined, although none of themaffected hMSC proliferation and SMA expression, all ofthe ECMs (including fibronectin, collagen I and IV,and elastin) significantly increased calponin expressionexcept laminin, which, in contrast, significantly de-creased calponin level.

SMCs on various ECMs, such as collagen, Matrigel (abasement membrane-rich matrix material) and fi-

Figure 8. Collagen and Western blot analysis on engineeredvessels. A) Collagen assay on the engineered vessel wallsbefore (n�6) and after optimization (n�4). Human umbili-cal artery served as positive control. B) Western blot onprotein lysate from six engineered human vessels for SMAand calponin. L216 v1 and v2, L219v1 and v2, and L237 v1and v2 were the pairs of vessels grown in the same bioreactorfrom the same donor-derived MSCs. Lysates from CASMCswere used as positive control. �-Actin served as equal loadingcontrol.

Figure 9. SEM images of the luminal surface of nonendothe-lialized (A) and endothelialized (B) vessels engineered fromhMSCs using the optimized protocol. Scale bars � 50 �m.


bronectin, evoked changes in SMC morphology consis-tent with a change in the differentiated phenotype(48–50). ECM proteins such as fibronectin and colla-gen I were found to promote proliferation of humanairway SMCs and suppress contractile protein expres-sion. In contrast, basement membrane elements suchas laminin inhibited proliferation and supported amore contractile phenotype (51). The discrepancy be-tween other results and our findings may reflect thedifference in cell source (differentiated SMCs vs. bonemarrow-derived hMSCs) as well as the limitations of thein vitro culture system and the specific experimentalconditions and matrices examined. Further explora-tion and confirmation is clearly needed in this complexarea.

Components of our enhanced DMEM medium,PDGF-BB, bFGF, and vitamin C, were examined fortheir effect on hMSC proliferation and differentiation.In addition, PDGF-CC, a newly discovered member ofthe VEGF/PDGF superfamily, which has been shown toinduce the differentiation of bone marrow cells intosmooth muscle cells and stimulate their growth duringvessel sprouting (52), was examined. TGF�1, a potentmultifunctional cytokine that coordinately up-regulatesa variety of SMC differentiation marker genes in cul-tured SMCs and is released from platelets after vascularinjury (53, 54), was also included in the study. Surpris-ingly, among all the factors examined, vitamin C, acofactor for collagen synthesis, stimulated hMSC pro-liferation the most. PDGF-BB is a key mediator of SMCphenotypic switching. It has been shown to potentlysuppress expression of SMC marker genes as well as toincrease the rate of proliferation and migration incultured SMCs (55–58). Whereas both PDGF-BB andbFGF elicited a potent proliferative response in hMSCs,they significantly decreased SMA and calponin proteinexpression. bFGF reduced SMA and calponin levels tonearly undetectable levels by immunoblotting. TGF�1inhibited proliferation of hMSCs in a dose-dependentmanner (59). However, calponin expression was signif-icantly elevated by TGF�1 at concentration between 0.1and 10 ng/ml, consistent with previous reports onmultipotent adult potential cells (MAPCs) cultured inserum-free condition (60). PDGF-CC exerted similareffect on hMSC proliferation and differentiation asPDGF-BB.

Cyclic strain has been shown to enhance extracellu-lar matrix remodeling and synthesis, as well as cellularproliferation of differentiated SMCs (61–66). For ex-ample, Birukov et al. (61) demonstrated potentiation ofSMC proliferation in serum-activated cultures. In con-trast, Chapman et al. (67) reported that physiologicalcyclic stretch causes cell cycle arrest in cultured vascularSMCs, which was verified by others (68, 69). In thepresent study, while cyclic strain inhibited hMSC pro-liferation independent of ECM (untreated, collagentype I or fibronectin), its influence on hMSC differen-tiation was significantly tied to the presence of ECM aswell as PDGF-BB. For example, in the absence of matrixcoating, cyclic strain alone significantly decreased SMA

expression. In the presence of PDGF-BB, cyclic straindisplayed the opposite effect on SMA expression: anincrease on collagen I matrix and a decrease on fi-bronectin (Fig. 5B).

Between the two major SMC markers, SMA andcalponin, SMA is less specific due to its transientexpression in early stages of cardiac and skeletal myo-cytes (44, 70), as well as in myofibroblasts in healingwounds (71) and tumors (72). In addition, treatment ofECs and myofibroblasts with TGF� can also induceSMA expression (73, 74). Thus, SMA alone does notprovide definitive evidence for SMC lineage. In con-tract, in adult organisms, the expression of calponinappears to be restricted almost exclusively to vascularsmooth muscle (45). The different specificity of thesetwo SMC markers might explain the different changesof the two proteins in response to each factor.

In summary, human vessel walls have been success-fully constructed with hMSCs, providing a new cellsource to generate small-diameter vessel grafts. This isthe first report of utilizing hMSCs for engineered vesselculture in a bioreactor setting. These vessel walls ex-hibit some similarities in terms of morphology, histol-ogy, and protein synthesis to native counterparts, whichare mainly composed of SMCs and their ECM. Furtherinvestigation is necessary to improve the mechanicalstrength of vessels derived from MSCs, and this willlikely come from improvements in SMC differentiationand matrix synthesis. Based on our initial success on theendothelialization of these hMSC-derived vessel walls,endothelialization with EPC-derived ECs, and in vivoimplantation may be the next steps in achieving thegoal of a marrow-derived vascular replacement.

The authors are grateful for Dr. Clay Quint and Mr.Thomas Petersen for their help with the FACS analysis, Dr.Clay Quint for the endothelialization and SEM imaging, andDr. Thomas Hitchcock for his assistance with the histologicalimages. This work was funded by U.S. National Institutes ofHealth grants RO1HL083895 and HL063766 (both toL.E.N.).

REFERENCES

1. Niklason, L. E. (1999) Replacement arteries made to order.Science 286, 1493–1494

2. Veith, F. J., Gupta, S. K., Ascer, E., White-Flores, S., Samson,R. H., Scher, L. A., Towne, J. B., Bernhard, V. M., Bonier, P.,Flinn, W. R., Astelford, P., Yao, J. S. T., and Bergan, J. J. (1986)Six-year prospective multicenter randomized comparison ofautologous saphenous vein and expanded polytetrafluoroethyl-ene grafts in infrainguinal arterial reconstructions. J. Vasc. Surg.3, 104–114

3. Chard, R. B., Johnson, D. C., Num, G. R., and Cartmill, T. B.(1987) Aorta-coronary bypass grafting with polytetrafluoroeth-ylene conduits. Early and late outcome in eight patients. J. Tho-rac. Cardiovasc. Surg. 94, 132–134

4. Sapsford, R. N., Oakley, G. D., and Talbor, S. (1981) Early andlate patency of expanded polytetrafluoroethylene vascular graftsin aorta-coronary bypass. J. Thorac. Cardiovasc. Surg. 81, 860–864

5. Steinthorsson, G., and Sumpio, B. (1999) Clinical and biologicalrelevance of vein cuff anastomosis. Acta Chir. Belg. 99, 282–288

6. Weinberg, C. B., and Bell, E. (1986) A blood vessel modelconstructed from collagen and cultured vascular cells. Science231, 397–400


7. Niklason, L. E., Gao, J., Abbott, W. M., Hirschi, K. K., Houser, S.,Marini, R., and Langer, R. (1999) Functional arteries grown invitro. Science 284, 489–493

8. Niklason, L. E., Abbott, W., Gao, J., Klagges, B., Hirschi, K. K.,Ulubayram, K., Conroy, N., Jones, R., Vasanawala, A., Sanzgiri,S., and Langer, R. L. (2001) Morphologic and mechanicalcharacteristics of bovine engineered arteries. J. Vasc. Surg. 33,628–638

9. Niklason, L. E., and Seruya, M. (2002) Small-diameter vasculargrafts. In Methods of Tissue Engineering (Atala, A., and Lanza, R.,eds) pp. 905–913, Academic Press, San Diego, CA, USA

10. McKee, J. A., Banik, S. S. R., Boyer, M. J., Hamad, N. M., Lawson,J. H., Niklason, L. E., and Counter, C. M. (2003) Human arteriesengineered in vitro. EMBO Rep. 4, 633–638

11. Poh, M., Boyer, M., Solan, A., Dahl, S. L. M., Dawn, P., Banik,S. S. R., McKee, J. A., Klinger, R. Y., Counter, C. M., andNiklason, L. E. (2005) Blood vessels engineered from humancells. Lancet 365, 2122–2124

12. Friedenstein, A. J., Gorskaja, J. F., and Kulagina, N. N. (1976)Fibroblast precursors in normal and irradiated mouse hemato-poietic organs. Exp. Hematol. 4, 267–274

13. Pittenger, M. F., Mackay, A. M., Beck, S. C., Jaiswal, R. K.,Douglas, R., Mosca, J. D., Moorman, M. A., Simonetti, D. W.,Craig, S., and Marshak, D. R. (1999) Multilineage potential ofadult human mesenchymal stem cells. Science 284, 143–147

14. Noth, U., Osyczka, A. M., Tuli, R., Hickok, N. J., Danielson,K. G., and Tuan, R. S. (2002) Multilineage mesenchymaldifferentiation potential of human trabecular bone-derivedcells. J. Orthop. Res. 20, 1060–1069

15. Zuk, P. A., Zhu, M., Ashjian, P., De Ugarte, D. A., Huang, J. I.,Mizuno, H., Alfonso, Z. C., Fraser, J. K., Benhaim, P., andHedrick, M. H. (2002) Human adipose tissue is a source ofmultipotent stem cells. Mol. Biol. Cell 13, 4279–4295

16. Jiang, Y., Vaesse, B., Lenvik, T., Blackstad, M., Reyes, M., andVerfaillie, C. M. (2002) Multipotent progenitor cells can beisolated from postnatal murine bone marrow, muscle, andbrain. Exp. Hematol. 30, 896–904

17. Zhao, L. R., Duan, W. M., Reyes, M., Keene, C. D., Verfaillie,C. M., and Low, W. C. (2002) Human bone marrow stem cellsexhibit neural phenotypes and ameliorate neurological deficitsafter grafting into the ischemic brain of rats. Exp. Neurol. 174,11–20

18. Jiang, Y., Jahagirdar, B. N., Reinhardt, R. L., Schwartz, R. E.,Keene, C. D., Ortiz-Gonzalez, X. R., Reyes, M., Lenvik, T., Lund,T., Blackstad, M., Du, J., Aldrich, S., Lisberg, A., Low, W. C.,Largaespada, D. A., and Verfaillie, C. M. (2002) Pluripotency ofmesenchymal stem cells derived from adult marrow. Nature 418,41–49

19. Davis, R. L., Weintraub, H., and Lassar, A. B. (1987) Expressionof a single transfected cDNA converts fibroblasts to myoblasts.Cell 51, 987–1000

20. Galmiche, M. C., Koteliansky, V. E., Briere, J., Herve, P., andCharbord, P. (1993) Stromal cells from human long-termmarrow cultures are mesenchymal cells that differentiate follow-ing a vascular smooth muscle differentiation pathway. Blood 82,66–76

21. Seruya, M., Shah, A., Pedrotty, D., du Laney, T., Melgiri, R.,McKee, J. A., Young, H. E., and Niklason, L. E. (2004) Clonalpopulation of adult stem cells: life span and differentiationpotential. Cell Transplant. 13, 93–101

22. Kinner, B., Zaleskas, J. M., and Spector, M. (2002) Regulation ofsmooth muscle actin expression and contraction in adult hu-man mesenchymal stem cells. Exp. Cell. Res. 278, 72–83

23. Hirschi, K. K., Rohovsky, S. A., and D’Amore, P. A. (1998)PDGF, TGF�, and heterotypic cell-cell interactions mediateendothelial cell-induced recruitment of 10T1/2 cells and theirdifferentiation to a smooth muscle fate. J. Cell Biol. 141, 805–814

24. Ball, S. G., Shuttleworth, A. C., and Kielty, C. M. (2004) Directcell contact influences bone marrow mesenchymal stem cellfate. Int. J. Biochem. Cell Biol. 36, 714–727

25. Kaushal, S., Amiel, G. E., Guleserian, K. J., Sharpira, O. M.,Perry, T., Sutherland, F. W., Rabkin, E., Moran, A. M., Schoen,F. J., Atala, A., Soker, S., Bischoff, J., and Mayer, J. E. (2001)Functional small-diameter neovessels created using endothelialprogenitor cells expanded ex vivo. Nat. Med. 7, 1035–1040

26. Kadner, A., Heorstrup, S. P., Zund, G., Eid, K., Maurus, C.,Melinitchouk, S., Grunenfelder, J., and Turina, M. I. (2002) A

new source for cardiovascular tissue engineering: Human bonemarrow stromal cells. Eur. J. Cardiothorac. Surg. 21, 1055–1060

27. Hoerstrup, S. P., Kadner, A., Melnitchouk, S., Trojan, A., Eid,K., Tracy, J., Sodian, R., Visjager, J. F., Kolb, S. A., Grunenfelder,J., Zund, G., and Turina, M. I. (2002) Tissue engineering offunctional trileaflet heart valves from human marrow stromalcells. Circulation 106, I143–150

28. Perry, T. E., Kaushal, S., Sutherland, F. W., Guleserian, K. J.,Bischoff, J., Sacks, M., and Mayer, J. E. (2003) Bone marrow asa cell source for tissue engineering heart valves. Ann. Thorac.Surg. 75, 761–767

29. Matsumura, G., Miyagawa-Tomita, S., Shin’oka, T., Ikada, Y.,and Kurosawa, H. (2003) First evidence that bone marrow cellscontribute to the construction of tissue-engineered vascularautografts in vivo. Circulation 108, 1729–1734

30. Wu, X., Rabkin-Aikawa, E., Guleserian, K. J., Perry, T. E.,Masuda, Y., Sutherland, F. W., Schoen, F. J., Mayer, J. E., andBischoff, J. (2004) Tissue-engineered microvessels on three-dimensional biodegradable scaffolds using human endothelialprogenitor cells. Am. J. Physiol. Heart. Circ. Physiol. 287, H480–487

31. Cho, S. W., Lim, S. H., Kim, I. K., Hong, Y. S., Kim, S. S., Yoo,K. J., Park, H. Y., Jang, Y., Cahng, B. C., Choi, C. Y., Hwang,K. C., and Kim, B. S. (2005) Small-diameter blood vesselsengineered with bone marrow-derived cells. Ann. Surg. 241,506–515

32. Liu, J. Y., Swartz, D. D., Peng, H. F., Gugino, S. F., Russell, J. A.,and Andreadis, S. T. (2007) Functional tissue-engineered bloodvessels from bone marrow progenitor cells. Cardiovasc. Res. 75,618–628

33. Gong, Z., and Wezeman, F. H. (2004) Inhibitory effect ofalcohol on osteogenic differentiation in human bone marrow-derived mesenchymal stem cells. Alcohol Clin. Exp. Res. 28,468–479

34. Aubin, J. E. (1999) Osteoprogenitor cell frequency in rat bonemarrow stromal populations: role for heterotypic cell-cell inter-actions in osteoblast differentiation. J. Cell. Biochem. 72, 396–410

35. Woessner, J. F. (1961) The determination of hydroxyproline intissue and protein samples containing small proportions of thisimino acid. Arch. Biochem. Biophys. 93, 440–447

36. Haynesworth, S. E., Baber, M. A., and Caplan, A. I. (1992) Cellsurface antigens on human marrow-derived mesenchymal cellsare detected by monoclonal antibodies. Bone 13, 69–80

37. Berk, B. C. (2001) Vascular smooth muscle growth: Autocrinegrowth mechanisms. Physiol. Rev. 81, 999–1030

38. Betsholtz, C. (2004) Insight into the physiological functions ofPDGF through genetic studies in mice. Cytokine Growth FactorRev. 15, 215–228

39. Van der Loop, F. T., Schaart, G., Timmer, E. D., Ramaekers,F. C., and van Eys, G. J. (1996) Smoothelin, a novel cytoskeletalprotein specific for smooth muscle cells. J. Cell. Biol. 134,401–411

40. Li, J., Sensebe, L., Herve, P., and Charbord, P. (1995) Non-transformed colony-derived stromal cell lines from normalhuman marrows. II. Phenotypic characterization and differen-tiation pathway. Exp. Hematol. 23, 133–141.

41. Kinner, B., Zaleskas, J. M., and Spector, M. (2002) Regulation ofsmooth muscle actin expression and contraction in adult hu-man mesenchymal stem cells. Exp. Cell. Res. 278, 72–83

42. Chamley-Campbell, J. H., Campbell, G. R., and Ross, R. (1979)The smooth muscle cell in culture. Physiol. Rev. 59, 1–61

43. Chamley-Campbell, J. H., and Campbell, G. R. (1981) Whatcontrols smooth muscle phenotype? Atherosclerosis 40, 347–357

44. Ruzicka, D. L., and Schwartz, R. J. (1988) Sequential activationof alpha-actin genes during avian cardiogenesis: vascularsmooth muscle alpha-actin gene transcripts mark the onset ofcardiomyocyte differentiation. J. Cell Biol. 107, 2575–2586

45. Owens, G. K. (1995) Regulation of differentiation of vascularsmooth muscle cells. Physiol. Rev. 75, 487–517

46. Carey, D. J. (1991) Control of growth and differentiation ofvascular cells by extracellular matrix proteins. Annu. Rev. Physiol.53, 161–177

47. Ingber, D. E., Dike, L., Hansen, L., Karp, S., Liley, H., Maniotis,A., McNamee, H., Mooney, D., Plopper, G., Sims, J., and Wang,N. (1994) Cellular tensegrity: exploring how mechanicalchanges in the cytoskeleton regulate cell growth, migration, and


tissue pattern during morphogenesis. Int. Rev. Cytol. 150, 173–224

48. Clyman, R. I., McDonald, K. A., and Kramer, R. H. (1990)Integrin receptors on aortic smooth muscle cells mediate adhe-sion to fibronectin, laminin, and collagen. Circ. Res. 67, 175–186

49. Hedin, U., Bottger, B. A., Luthman, J., Johansson, S., andThyberg, J. (1989) A substrate of the cell-attachment sequenceof fibronectin (Arg-Gly-Asp-Ser) is sufficient to promote transi-tion of arterial smooth muscle cells from a conractile to asynthetic phenotype. Dev. Biol. 133, 489–501

50. Pauly, R. R., Passaniti, A., Crow, M., Kinsella, J. L., Papadopou-los, N., Monticone, R., Lakatta, E. G., and Martin, G. R. (1992)Experimental models that mimic the differentiation and dedif-ferentiation of vascular cells. Circulation 86(Suppl. III), 68–73

51. Hirst, S. J., Twort, C. H. C., and Lee, T. H. (2000) Differentialeffects of extracellular matrix proteins on human airway smoothmuscle cell proliferation and phenotype. Am. J. Respir. Cell Mol.Biol. 23, 335–344

52. Li, X., Tjwa, M., Moons, L., Fons, P., Noel, A., Ny, A., Zhou,J. M., Lennartsson, J., Li, H., Luttun, A., Ponten, A., Devy, L.,Bouche, A., Oh, H., Manderveld, A., Blacher, S., Communi, D.,Savi, P., Bono, F., Dewerchin, M., Foidart, J. M., Autiero, M.,Herbert, J. M., Collen, D., Heldin, C. H., Eriksson, U., andCarmeliet, P. (2005) Revascularization of ischemic tissues byPDGF-CC via effects on endothelial cells and their progenitors.J. Clin. Invest. 115, 118–127

53. Bjorkerud, S. (1991) Effects of transforming growth factor-beta1 on human arterial smooth muscle cells in vitro. Arterioscler.Thromb. 11, 892–902

54. Hautmann, M. B., Madsen, C. S., and Owens, G. K. (1997)Transforming growth factor � (TGF�) control element drivesTGF�-induced stimulation of smooth muscle �-actin gene ex-pression in concert with two CArG-elements. J. Biol. Chem. 272,10948–10956

55. Dandre, F., Owens, G. K. (2004) Platelet-derived growth fac-tor-BB and Ets-1 transcription factor negatively regulate tran-scription of multiple smooth muscle cell differentiation markergenes. Am. J. Physiol. Heart. Circ. Physiol. 286, H2042–H2051

56. Heldin, C. H., Westermark, B. (1999) Mechanism of action andin vivo role of platelet-derived growth factor. Physiol. Rev. 79,1283–1316

57. Raines, E. W. (2004) PDGF and cardiovascular disease. CytokineGrowth Factor Rev. 15, 237–254

58. Yoshida, T., Gan, Q., Shang, Y., Owens, G. K. (2007) Platelet-derived growth factor-BB represses smooth muscle cell markergenes via changes in binding of MKL factors and histonedeacetylases to their promoters. Am. J. Physiol. Cell. Physiol. 292,C886–C895

59. Majack, R. A. (1987) Beta-type transforming growth factorspecifies organizational behavior in vascular smooth muscle cellcultures. J. Cell Biol. 105, 465–471

60. Ross, J. J., Hong, Z., Willenbring, B., Zeng, L., Isenberg, B., Lee,E. H., Reyes, M., Keirstead, S. A., Weir, E. K., Tranquillo, R. T.,and Verfaillie, C. M. (2006) Cytokine-induced differentiation ofmultipotent adult progenitor cells into functional smooth mus-cle cells. J. Clin. Invest. 116, 3139–3149

61. Birukov, K. G., Shriinsky, V. P., Stepanova, O. V., Tkachuk, V. A.,Hahn, A. W., Resink, T. J., and Smirnov, V. N. (1995) Stretch

affects phenotype and proliferation of vascular smooth musclecells. Mol. Cell. Biochem. 14, 131–139

62. Kim, B. S., Nikolovski, J., Bonadio, J., and Mooney, D. J. (1999)Cyclic mechanical strain regulates the development of engi-neered smooth muscle tissue. Nat. Biotechnol. 17, 979–983

63. Li, Q., Muragaki, Y., Ueno, H., and Ooshima, A. (1997) Stretch-induced proliferation of cultured vascular smooth muscle cellsand a possible involvement of local rennin-angiotensin systemand platelet-derived growth factor (PDGF). Hypertens. Res. 20,217–223

64. Mills, I., Cohen, C. R., Kamal, K., Li, G., Shin, T., Du, W., andSumpio, B. E. (1997) Strain activation of bovine aortic smoothmuscle cell proliferation and alignment: Study of strain depen-dency and the role of protein kinase A and C signalingpathways. J. Cell. Physiol. 170, 228–234

65. O’Callaghan, C. J., and Williams, B. (2000) Mechanical strain-induced extracellular matrix production by human vascularsmooth muscle cells: role of TGF-beta (1). Hypertension 36,319–324

66. Seliktar, D., Nerem, R. M., and Galis, Z. S. (2001) The role ofmatrix metalloproteinase-2 in the remodeling of cell-seededvascular constructs subjected to cyclic strain. Ann. Biomed. Eng.29, 923–934

67. Chapman, G. B., Durante, W., Hellums, J. D., and Schafer, A. I.(2000) Physiological cyclic stretch causes cell cycle arrest incultured vascular smooth muscle cells. Am. J. Physiol. Heart. Circ.Physiol. 278, H748–754

68. Park, J., Chu, J., Cheng, C., Chen, F., Chen, D. J., and Li, S.(2004) Differential effects of equiaxial and uniaxial strains onmesenchymal stem cells. Biotechnol. Bioeng. 14, 359–368

69. Kurpinski, K., Park, J., Thakar, R., and Li, S. (2006) Regulationof vascular smooth muscle cells and mesenchymal stem cells bymechanical strain. Mol. Cell. Biomech. 3, 11–24

70. Sawtell, N. M., and Lessard, J. L. (1989) Cellular distribution ofsmooth muscle actins during mammalian embryogenesis: ex-pression of the alpha-vascular but not the gamma-enteric iso-form in differentiating striated myocytes. J. Cell Biol. 109,2929–2937

71. Darby, I., Skalli, O., and Gabbiani, G. (1990) Alpha-smoothmuscle actin is transiently expressed by myofibroblasts duringexperimental wound healing. Lab. Invest. 63, 21–29

72. Cintorino, M., Belllizzi De Marco, E., Leoncini, P., Tripodi,S. A., Xu, L. J., Sappino, A. P., Schmitt Graff, A., and Gabbiani,G. (1991) Expression of alpha-smooth-muscle actin in stromalcells of the uterine cervix during epithelial neoplastic changes.Int. J. Cancer 47, 843–846

73. Arciniegas, E., Sutton, A. B., Allen, T. D., and Schor, A. M.(1992) Transforming growth factor beta 1 promotes the differ-entiation of endothelial cells into smooth muscle-like cells invitro. J. Cell Sci. 103, 521–529

74. Marotti, K. R., Castile, C. K., Boyle, T. P., Lin, A. H., Murray,R. W., and Melchior, G. W. (1993) Severe atherosclerosis intransgenic mice expressing simian cholesteryl ester transferprotein. Nat. Lond. 364, 73–75

Received for publication August 3, 2007.Accepted for publication December 6, 2007.


Documents

Small-diameter human vessel wall engineered from bone marrow-derived mesenchymal stem cells