当前位置:文档之家› Embryonic Stem Cell-Derived Endothelial Cells May Lack Complete Functional Maturation in vitro

Embryonic Stem Cell-Derived Endothelial Cells May Lack Complete Functional Maturation in vitro

Fax +41 61 306 12 34E-Mail karger@karger.ch https://www.doczj.com/doc/6f7504545.html, R esearch Paper

J

Vasc Res 2006;43:411–421 D OI :10.1159/000094791 E

mbryonic Stem Cell-Derived Endothelial Cells May Lack Complete

Functional Maturation in vitro

K ara E. McCloskey a ,b Debra A. Smith c Hanjoong Jo c ,d Robert M. Nerem a ,e a T he Parker H. Petit Institute for Bioengineering and Bioscience, Georgia Institute of Technology, A tlanta, Ga. ,

b

S chool of Engineering, University of California, M erced, Calif. , c T he Wallace H. Coulter Department of

Biomedical Engineering, Georgia Institute of Technology and Emory University, A tlanta, Ga. ,

d D ivision of Cardiology, Emory University, A tlanta, Ga. ,

e T he Woodruf

f School of Mechanical Engineering,

Georgia Institute of Technology and Emory University, A tlanta, Ga. , USA and vasculogenic and angiogenic sprouting than MAEC. These results indicate that ESC-derived EC share some key characteristics of ‘mature’ EC, while retaining markers of al-ternate phenotypes including immature endothelium.

C opyright ? 2006 S. Karger AG, Basel I ntroduction V ascular endothelial cells (EC) or endothelial progen-itor cells derived from stem cells could potentially lead to a variety of clinically relevant applications [1] . These cells could be used in therapeutic strategies for the repair and revascularization of ischemic tissue in patients exhibiting vascular defects [2, 3] . Endothelial progenitor cell trans-plantation has been shown to induce new vessel forma-tion in ischemic myocardium and hind limb [3–5] . In addition, it has been shown that EC are able to transdif-ferentiate into cardiomyocytes and may potentially be used to repair cardiac function [6] . Since it is well known that EC inhibit platelet adhesion and clotting, they are needed for lining the lumen of a synthetic or tissue-engi-neered vascular graft or for re-endothelization of injured vessels [7, 8] . Moreover, because EC line the lumen of blood vessels and can release proteins directly into the blood stream, they are ideal candidates to be used as ve-

K ey Words

E mbryonic stem cells ? Endothelial progenitor cells ? Endothelial cells ? Flk1 ? Angiogenesis ? Vasculogenesis ?

Mouse aortic endothelial cells A bstract

S tem cell therapies will only become clinically relevant if the stem cells differentiated in vitro function as their in vivo

counterparts. Here, we employed our previously developed

techniques for deriving endothelial cells ( 1 96% purity) from

mouse embryonic stem cells (ESC) and compared these with

mouse aortic endothelial cells (MAEC) obtained from tho-

racic aortas. I mmunocytochemical analysis of ESC-derived

endothelial cells (EC) demonstrates that both cell types are

positive for the EC markers endothelial nitric oxide synthase

(eNOS), Flk-1, Flt-1, vascular endothelial cadherin (VEcad),

platelet-endothelial cell adhesion molecule-1 (PECAM-1),

and CD34. However, ESC-derived EC express slightly lower

levels of PECAM-1 and VE-cadherin, and significantly lower

levels of acetylated low-density lipoprotein (LDL) uptake

and von Willebrand factor. Although ESC-derived EC do ex-

press VE-cadherin, the VE-cadherin in the ESC-derived EC did

not localize as well at the cell-cell junctions as in the MAEC.

Interestingly, ESC-derived EC express much greater levels of

the endothelial and hematopoietic stem cell marker CD34 R eceived: March 15, 2006

A ccepted after revision: May 31, 2006

P ublished online: July 28, 2006 P rof.Kara E.McCloskey

School of Engineering, University of California, Merced

PO Box 2039

Merced, CA 95344 (USA)

Tel. +1 209 381 7885, Fax +1 209 724 2912, E-Mail kmccloskey@https://www.doczj.com/doc/6f7504545.html, ? 2006 S. Karger AG, Basel A ccessible online at:https://www.doczj.com/doc/6f7504545.html,/jvr

hicles of gene therapy. Lastly, EC may be used for vascu-larizing tissue-engineered materials prior to implanta-tion and for investigating mechanisms of angiogenesis and vasculogenesis.

O ne potential source for these therapeutic endothelial cells is the embryonic stem cell (ESC). The ESC possesses some advantages over the adult stem cell in that the ESC provides an excellent in vitro culture system for studying cellular differentiation events. The ESC is thought to have the capacity for an unlimited number of cell divisions and is much easier to culture and maintain in undiffer-entiated status in vitro [9] . EC have been derived from human ES cells by isolating the differentiating endothe-lium from an embryoid body [10] . Although the embry-oid body system enables investigation of vasculogenesis virtually as it occurs in the embryo [11–14] , the multiple cell-cell contacts and cell lineages make it difficult to study and control the behavior of the maturing EC in de-tail.

I t has been shown that the 3-D structure is not neces-sary for endothelial maturation from ES cells [15] . Endo-thelial, hematopoietic, and smooth muscle cells have been derived from Flk-1+/E-cadherin– outgrowths from murine ESC cells grown on type-IV collagen coated sur-faces [15] . This 2-D monolayer technique of endothelial differentiation not only allows closer study and control of the in vitro maturation, molecular events, and growth factor requirements of EC derivation [15, 16] , but also uses an induction method that is devoid of the 3-D em-bryo-like self-programmed machinery for vascular dif-ferentiation.

A lthough the 2-D monolayer derivation methods [15–

17] have been very successful in isolating and studying the maturation of EC from murine embryonic stem cells, the long-term maintenance of these murine ESC-derived EC has been limited. Without genetic manipulation, the longest these ESC-derived EC were maintained in culture was 7 days, increasing to two or three passages by cultur-ing cells on OP9 stromal cells [17] .

R ecently, we have developed methodology that ex-pands on the work of Nishikawa’s group [15–17] for the in vitro differentiation and purification (196% pure) of EC populations from mouse ESC [18] . Briefly, 60,000 ESC progress through three different stages of cell induc-tion/expansion and two cell isolation procedures, gener-ating over 300 million EC. These ESC-derived EC display characteristics of the vascular EC in that they express sev-eral endothelial markers [18] , and they form 2-D tube-like structures, as well as complex vessel-like structures in 3-D collagen type-I gels [19] .

I t was not clear, however, whether the ESC-derived cells would function as mature normal EC that had been derived in vivo. Here, we directly compared the endothe-lial marker expression levels and vasculogenic and angio-genic potentials of ESC-derived EC and primary cultured mouse aortic endothelial cells (MAEC).

M ethods

E SC Cultur e

E S-D3 embryonic stem cells (American Type Culture Collec-tion, Manassas, Va., USA) were initially maintained on irradiated or mitomycin C-treated (Sigma) mouse embryonic fibroblast feeder layers in knockout Dulbecco’s modified Eagle medium (KO-DMEM; Gibco) containing 15% ESC qualified fetal bovine serum (Gibco), 5% knockout serum replacement (Gibco), 1,000 units per ml of leukemia inhibitory factor (ESGRO; Chemicon International) and 5 !10–5M?-mercaptoethanol. Cells were then cultured on 0.1% gelatin (no feeders) for one week before switching to differentiation conditions.

I nitiating differentiation, 30,000 cells were transferred to col-lagen type IV-coated dishes (Biocoat, Becton-Dickinson) and cul-tured for 4 days without leukemia inhibitory factor (LIF) in ?-minimal essential medium (Gibco), 15% fetal bovine serum (FBS) (Gibco), and 5 !10–5M?-mercaptoethanol. On day 4, the cells were stained for Flk-1 (FLK11-A, Alpha Diagnostic Internation-al), FACS sorted, and re-plated on collagen type IV in differentia-tion medium supplemented with 50 ng/ml of recombinant human VEGF (VEGF165 , R & D Systems). After culturing for approxi-mately one week, cells exhibiting predominantly two different phenotypes emerged. These included endothelial-like cells with a cobblestone shaped morphology, and more elongated smooth muscle-like cell populations.

M anual Selection of Endothelial Cells from ESC-Derived

Vascular Progenitors

P rior to isolation of the endothelial-like Flk-1 positive out-growths, the cells were first washed with PBS and incubated in Cell Dissociation Solution (Sigma) for 5 min. The cells that exhib-ited endothelial-like morphologies were manually selected using a flame-pulled Pasteur pipet connected to an aspirator assembly (Sigma) fitted with a 0.2-?m syringe filter. Based on cell morphol-ogy alone, 5–10 cells were excised and replated in each well of a collagen type-IV coated 12-well plate. At this stage, the cells were fed endothelial cell EGM-2 medium supplemented with 10 ml FBS, 0.2 ml hydrocortisone, 2 ml hFGF-?, 0.5 ml VEGF, 0.5 ml R3-IGF-1, 0.5 ml ascorbic acid, 0.5 ml hEGF, 0.5 ml GA-1000, 0.5 ml heparin (EGM-2 Bullet Kit, Clonetics), 5 !10–5M?-mer-captoethanol, and an extra 50 ng of recombinant human VEGF (VEGF165 , R & D Systems) per ml of culture medium. These cells were then expanded up to 25 population doublings on either mouse collagen type-IV (Becton D ickinson), 0.1% gelatin (Sig-ma), or fibronectin (Sigma). At each passage, the concentration of VEGF165 was decreased until the only VEGF remaining was the amount provided in the EGM-2 Bullet Kit (amounts are propri-etary and undisclosed).

M cCloskey/Smith/Jo/Nerem

J Vasc Res 2006;43:411–421 412

I n vitro-Derived Endothelial Cells J Vasc Res 2006;43:411–421413

I mmunostaining E SC and their differentiation products were characterized us-

ing antibodies specific for ESC and endothelial cells. Primary an-

tibodies were detected with fluorochrome conjugated secondary

antibodies and the cells were examined on a Zeiss LSM 510 confo-

cal microscope, FACS Vantage SE cell sorter (Becton Dickinson),

or LSR Flow Cytometer (Becton Dickinson). ESC markers includ-

ed rabbit anti-Oct-4 (H-134; Santa Cruz Biotechnology), and

mouse anti-human SSEA-1 (Santa Cruz Biotechnology). Endo-

thelial markers were rabbit anti-mouse Flk-1 (Research Diagnos-

tics) and rabbit anti-human Flt-1 (Research D iagnostics), goat anti-mouse PECAM-1 (Santa Cruz Biotechnology), FITC conju-

gated rat anti-mouse CD34 (RAM34, a marker for hematopoietic

progenitor and endothelial progenitor cells, PharMingen), rabbit

anti-eNOS (NOS3; Santa Cruz Biotechnology), rabbit anti-hu-

man Von Willebrand factor (Dako) and goat anti-VE-cadherin

(Santa Cruz Biotechnology). Secondary antibodies were PE con-

jugated donkey anti-rabbit IgG (Research D iagnostics), FITC

conjugated donkey anti-goat IgG (Research D iagnostics), and

FITC conjugated donkey anti-rat IgG (Research D iagnostics).

The cells were counterstained with Hoechst 33258 nucleic acid

stain (Molecular Probes).

F luor escent Activated Cell

Sor ting (FACS) F or isolating the Flk-1+ cells after 4 days on collagen type-IV,

the cells were dislodged from the culture dishes with Cell Disso-

ciation Solution (Sigma) for 10–15 min, blocked with 10% normal

donkey serum (Research Diagnostics, Inc.) for 30 min, stained

with anti-mouse Flk-1 for 30 min, and then with secondary anti-

rabbit PE for another 30 min. Fluorescent-labeled cells were then

sorted using a FACS Vantage SE cell sorter (Becton Dickinson)

and replated on collagen type-IV coated dishes as described

above.

I solation of Primary EC

M ouse aortic endothelial cells (MAEC) were isolated from the

thoracic aortas of C57Bl6 mice, cultured in DMEM containing

20% FBS, 100 ?g /ml endothelial cell growth supplement (ECGS,

Sigma), 2.5 U/ml heparin, and 1% penicillin/streptomycin, and

used at passages 5–10 as described by us [20] .

L aminar Shear Str ess

A parallel plate flow device was used to expose the ESC-de-

rived EC to a steady laminar shear stress of 15 dyn/cm 2 for a pe-

riod of 24 h [21] . The shear stress that the cells are exposed to, ?

w ,

is related to the volumetric flow rate,

Q , of the culture medium by:

26w Q

bh ?

? (1)

w here ? is the viscosity of the culture medium. This assumes that

the velocity profile is fully developed, i.e. negligible entrance ef-

fects. The rectangular flow channel possesses a height h = 250 or

800 ?m (depending on the size of the spacer) and width, b =

38 mm.

M atrigel in vitro Tube Formation Assay

M atrigel (BD Biosciences) was added to a few wells of a 24-well

plate in 500-?l volumes and allowed to solidify for 30 min at

37° C . After the Matrigel solidified, 10,000 ESC-derived EC were

suspended in 1 ml of EC medium and added to individual Matri-gel-coated wells. The cells were then incubated at 37°C and 5% CO 2 and observed for tube-like formations with a phase-contrast microscope. C ollagen Gels T he ESC-derived and expanded EC were tested for in vitro for-mation of 3-D vascular structures. The cell-to-gel ratio consisted of one million ESC-derived EC per ml of gel solution. The gel so-lution was made up of 2 mg/ml of rat tail collagen type-I, 0.1 M

sodium hydroxide, 10% FBS, and concentrated MCDB 131 (Sig-

ma). The ingredients were all mixed on ice and pipetted into 6- or 12-well plates and immediately placed in the incubator. After one week, vascular formations were observed in the collagen gels. R esults F irst, we determined whether ESC-derived EC display similar cell morphology during growth, at confluence under static no-flow conditions and in response to lami-nar shear stress exposure. Compared with MAEC, ESC-derived EC exhibited similar morphologies while grow-ing at subconfluence ( f ig. 1 a and d). In contrast, as ESC-derived EC become confluent ( f ig. 1 b ) they appeared to be slightly more flattened and more closely packed to-gether at their greatest density compared with MAEC ( f ig. 1 e ). Another characteristic response of MAEC is cell-shape alignment response to the imposed laminar flow direction [20, 22] . Morphologies of ESC-derived EC ( f ig. 1 c ), after exposure to 24 h of shear stress at 15 dynes/cm 2, also closely matched the morphologies of MAEC ( f ig. 1 f ). Both cell populations underwent cell shape re-alignment parallel to the flow direction within 24 h. C ell diameters of the two cell populations were very close ( f ig. 2 ). The mean diameter of the ESC-derived EC (d = 15.7 ?m , f ig. 2 a ) measured approximately 1 ?m less than the mean diameter of the MAEC (d = 17.1 ?m , f ig. 2 b ). Although the difference between the cell populations is small, the difference is statistically significant (p ! 0.01). M any cells express specific markers associated with cell phenotype. These markers are very helpful in identi-fying cells. Markers of both EC and ESC were employed to verify cell phenotype. Mouse ESC, including our ES-D3, express very high levels of the transcription factor Oct-4 and cell surface marker SSEA-1 [18, 23] . ESC mark-ers Oct-4 and SSEA-1 were evaluated, and determined not to be expressed on either ESC-derived EC or MAEC ( f ig. 3 ). These results verified that the ESC-derived EC progenitor cells had differentiated significantly and would no longer be categorized as undifferentiated ESC. Similarly, EC markers eNOS, Flk1, and Flt1 were found

M cCloskey /Smith /Jo /Nerem

J Vasc Res 2006;43:411–421414on both ESC-derived EC and MAEC with very similar

levels of expression ( f ig. 4 a –c, g–i); however, the EC and

hematopoietic progenitor cell marker CD 34 was ex-

pressed on ESC-derived EC, but not expressed at all on

MAEC ( f ig. 4 d and j). Two additional EC markers, VE-

cadherin ( f ig. 4 e and k) and PECAM1 ( f ig. 4 f and l), were

expressed on both ESC-derived EC and MAEC, but to a

lesser extent on ESC-derived EC compared with MAEC.

C onfocal laser scanning microscopy may be used in addition to flow cytometry for evaluation of cell marker

expression. Although data from confocal laser scanning

microscopy is not as easily quantifiable, it is beneficial for

assessment of molecule expression at localized areas

within the cells. In EC including MAEC, eNOS is well

known to be expressed in the plasma membrane, caveolae and Golgi. The first images in f igure 5 depict eNOS ex-pression in both cell populations. ESC-EC showed a sim-ilar eNOS expression pattern ( f ig. 5 a ) to that of MAEC. VE-cadherin is also known to be expressed at the cell-cell junctions of EC ( f ig. 5 e ), but we did not see a significant junctional staining pattern in ESC-derived EC ( f ig. 5 b ). In fact, the cell-cell staining pattern of the MAEC was also limited compared with typical human EC staining patterns. This is due to the inferior antibodies available for mouse VE-cadherin compared with human VE-cad-herin. The most telling EC marker examined might be the acetylated LDL uptake of the cells. ESC-derived EC ( f ig. 5 c

) took up negligible amounts of LDL, while MAEC F

ig . 1. Phase microphotographs of ESC-derived EC (left column) and MAEC (right

column). The first row depicts their mor-

phology at subconfluence, or low density.

At subconfluence, both ESC-derived EC

(a ) and MAEC (d ) exhibit similar spindle-

shaped morphologies indicative of migra-

tory cells. At confluence, the ESC-derived

EC (b ) appear to be flatter and packed

closer together than MAEC (e ). ESC-de-

rived EC (c ) and MAEC (f ) show similar

alignment to the direction of imposed

laminar shear stress for 24 h.

I n vitro-Derived Endothelial Cells J Vasc Res 2006;43:411–421

415

C e l l s (n )ESC-EC

D mean = 15.85 ?m 180

160

140

120

100

80

60

40

20

a

46

81020

4060C e l l l s (n )MAEC D mean = 17.11 ?m

12010080604020046810204060b F

ig . 2. The cell size of MAEC is slightly greater than that of ESC-derived EC. Cell size distributions of ESC-de-rived EC (a )

and MAEC (b ) were determined by Coulter Counter and shown as histograms. According to Stu-dent’s t test, the difference in the diameters of the 2 cell types is statistically significant (p ! 0.01). ESC-EC 685134170

100

101102103104MAEC

Oct-4SSEA-16851

34

17

0100101102103104

6851

34

17

100101102103104

685134170

100101102103104a b d c Ab

F

ig . 3. ESC cell markers Oct-4 and SSEA-1 are not expressed in ESC-derived EC and MAEC. ESC-derived EC were negative for both Oct-4 (a ) and SSEA-1 (b ) ESC markers. Likewise, MAEC were negative for both Oct-4 (c ) and SSEA-1 (d ) ESC markers.

M cCloskey /Smith /Jo /Nerem

J Vasc Res 2006;43:411–421416( f ig. 5 f ) showed a robust fluorescent signal demonstrat-

ing the robust acetylated LDL uptake. Also, the ESC-de-

rived EC did not express detectable levels of von Wille-

brand factor (vWF), whereas MAEC did express this fac-

tor VIII-related antigen (data not shown).

A lthough the ESC-derived EC exhibited limited ex-pression of some markers of mature EC, they did exhibit increased vasculogenic and angiogenic behaviors. When both ESC-derived EC and MAEC were seeded on Matri-

gel, they rapidly formed tube-like networks on the Matri-

ESC-EC Ab 40

30

20C o u n t s 10

10010110210310468

5134170

100

10110210310468

513417

100

10110210310468

5134170100101102103104685134170100101102103104685134170100101102103104eNOS Flt-1Flk-1CD34VE-caderin PECAM-1a b c d e f 6851

34

17

0100101102103104MAEC C o u n t s 403020

100100101102103104685134170100101102103104

6851

34

1701001011021031046851

34

170100101102103104685134170100101102103104g h i j k l F ig . 4. Characterization of EC markers in ESC-derived EC in comparison to MAEC. Confluent ESC-derived EC (a –f ) and MAEC (g –l ) were stained with antibodies

specific to eNOS, Flt1, Flk1, CD 34, VE-

cadherin, and PECAM1. The primary an-

tibodies were then detected by PE- or

FITC-conjugated secondary antibodies,

and the immunopositive cells determined

by FACS. As controls, secondary antibody

alone groups (shown in black lines in each

histogram) were included. The EC mark-

ers eNOS (a , g ), VEGF receptors, Flt1 (b ,

h ) and Flk1 (c , i ), were expressed similarly

in ESC-derived EC (a ) and MAEC (g ). The

CD 34 hematopoietic progenitor cell and

EC marker was expressed in ESC-derived

EC (d ), but expressed in much lower

amounts in MAEC (j ). VE-cadherin was

expressed in both ESC-derived EC (e )

and MAEC (k ), but expressed at greater levels

in MAEC. PECAM1 was expressed in both

ESC-derived EC (f ) and MAEC (l ), but to

a greater extent in MAEC.

I n vitro-Derived Endothelial Cells J Vasc Res 2006;43:411–42141750 m a b c d

e

f

F

ig . 5. Characterization of eNOS and VE-cadherin expressions and uptake of acetylated LDL in ESC-derived EC and MAEC. The eNOS expression patterns of ESC-derived EC (a ) and MAEC (d ) were similar. VE-cadherin expression is often localized at the cell-cell junctions as seen in MAEC (e ), but this localization was not fre-quently observed in ESC-derived EC (b ). The characteristic high uptake of acetylated LDL uptake observed in MAEC (f ) was significantly lower in ESC-derived EC (c ).

M cCloskey /Smith /Jo /Nerem

J Vasc Res 2006;43:411–421418gel surface within 24 h; however, the ESC-derived EC

( f ig. 6 a ) exhibited significantly more sprouting than the

MAEC ( f ig. 6 d ). After 2 days on the Matrigel, MAEC

sprouting was no longer observed and replaced by indi-

vidual cells becoming rounder and clustering into aggre-

gates ( f ig. 6 e ). In contrast, the ESC-derived EC migrated towards one another and continued sprouting from a central cell aggregate ( f ig. 6 b ). Likewise, when MAEC were seeded as single cells in collagen gels, they did not exhibit significant migration or network formation 25 m 25 m 50 m 50 m 100 m 25 m a b c d

e

f

F

ig . 6. Comparisons of angiogenic responses of ESC-derived EC vs. MAEC. ESC-derived EC (a , b ) and MAEC (d , e ) were grown on Matrigel for 1 day (a , d ) and 2 days (b , e ). As shown by the phase photomicrographs, ESC-derived EC exhibit significantly more tubule formation than MAEC. The bottom images are photographs of

ESC-derived EC (c )

and MAEC (f ) grown in collagen gels for 7 days. Note that ESC-derived EC assembled into multicellular vessel-like structures in the collagen gel, whereas MAEC did not.

( f ig. 6 f), whereas ESC-derived EC formed complex mul-ticellular networks ( f ig. 6 c). These multicellular vessel-like networks were counted and normalized per ml of collagen gel ( f ig. 7 ). We also compared the effect of cell confluence on vessel forming capabilities in the collagen gel study. As shown in f igure 7 , we found that sub-conflu-ent ESC-EC formed significantly more vessel-like net-works on the confluent cells.

D ifferences in the interaction of the EC and ESC-EC suspended within the collagen gels were also observed. The MAEC compacted the collagen gels by almost 50% over two weeks, whereas the ESC-derived EC did not compact the collagen gels by more than 20% ( f ig. 8 ). This assay showed that there are significant differences be-tween the functional relationship of the ESC-derived EC and its three-dimensional collagen scaffolding and that of the MAEC and its collagen scaffolding. The compac-tion of the gel indicates a migratory behavior of the cells in which the cells pull and tug on the collagen fibers. The highly angiogenic behavior and limited compaction of the ESC-derived EC in the collagen indicate that these cells are degrading the collagen by making pathways for migration, rather than pulling and tugging on the colla-gen fibers during migration.

D iscussion

H ere we compared the characteristics of two EC types: EC derived from mouse ESC, and cultured MAEC iso-lated from mouse thoracic aortas (macrovessel). The most notable differences between these two cell popula-tions are that ESC-derived EC showed limited acetylated LDL uptake and vWF expression, but an increased ex-pression of CD34 and increased vasculogenic and angio-genic sprouting compared to MAEC. Another significant difference between the two cell populations is the lower VE-cadherin localization at the endothelial cell-cell junc-tions of the ESC-derived EC.

I n the analysis of the differences in EC markers, it is important to recognize the inherent heterogeneity of EC. Endothelial cells exhibit specialized properties specific for the different organs in which they reside [24–26] , and when similar cell surface markers are identified, it is not unusual that the same protein be expressed to different degrees in different vascular regions [24–26] . Therefore, differences between endothelial cell populations derived from stem cells in vitro compared with in vivo-derived EC might be representing subpopulations of endothelial cells from microvessels, macrovessels, tumor EC, or or-gan-specific endothelium, such as endothelium of the kidney, lungs, or blood-brain barrier rather than an indi-cation of incomplete functional maturation.

F ig. 7. Effects of confluency on vessel-like structure formation in ESC-derived EC and MAEC. Average number of vessel-like struc-tures per ml of collagen gel for ESC-derived EC and MAEC as de-scribed in figure 6 was quantified by a computational image anal-ysis of photomicrographs. Note that no vessel-like structures were observed with MAEC (n = 3), while ESC-derived EC formed ro-bust vessel-like structures. Some cells were cultured to confluence (C) before seeding in collagen gels, but we noticed that if the cells were not allowed to reach confluence (NC), then the cells would form vessel-like structures more readily. According to Student’s t test, the difference between the nonconfluent ESC-EC and the nonconfluent MAEC is statistically significant (p !

0.01, n = 3). F ig. 8. Differential effects of ESC-derived EC and MAEC on col-lagen gel compaction. The diameters of the collagen gels seeded with ESC-derived EC and MAEC over a two-week period were determined by computational image analysis of photomicro-graphs. Shown are the mean gel diameters and standard deviation (n = 2) of the two cell populations.

I n vitro-Derived Endothelial Cells J Vasc Res 2006;43:411–421419

T he most extensive study of the immunohistochemi-cal expression of various vascular beds in normal human tissues included expression of CD31 (PECAM1), CD34, and vWF in the peripheral tissues and microvasculature of the kidneys, lungs, spleen, liver, heart, lymph nodes, bone marrow, skin, and large vessels including the aorta, inferior cava vein, renal artery, femoral artery and vein and pulmonary artery and vein [26] . These results indi-cated that the expression patterns in some of the EC pop-ulations were heterogeneous. The most heterogeneous ones included the EC of the kidney, lung, spleen and liver. The fenestrated endothelium of the kidney glomeruli strongly expressed CD34, but did not stain positive for vWF. Likewise, the alveolar wall of the lung stained for CD34, but was also negative for vWF, although vWF in-creased gradually with the vessel caliber in the lung. A second study reports on the expression of vWF in arteri-oles, capillaries and venules of various tissues, but notes a lower vWF expression in umbilical artery and veins and a lack of vWF expression in the lymphatics, liver sinu-soids, and the glomerular capillaries [27] . The heteroge-neity of vWF expression, depending upon different vas-cular beds such as capillaries and tumor vessels, was con-firmed by another study as well [24] .

T he variable expression of CD34 between ESC-de-rived EC and MAEC is not very surprising, because not all endothelial cells are positive for the CD34 antigen. CD34 is expressed on hematopoietic progenitor cells as well as endothelium of most capillaries; however, CD34 is not typically expressed on endothelium of larger vessels [28] , and although many freshly isolated endothelial cells express the CD34 molecule, its expression is sometimes diminished after several passages in culture [28, 29] . Our result that the ESC-derived EC expressed higher levels of CD34 than that of MAEC is consistent with this notion. Interestingly, CD34 expression has been shown to be en-riched on villi and sprouting processes of EC [29, 30] . The expression of VE-cadherin has been shown to inversely correlate with the invasiveness of EC [31] . In view of our finding that the ESC-derived EC showed enhanced vas-culogenic and angiogenic properties, enhanced CD34 ex-pression, and lower VE-cadherin at the cell junctions compared to MAEC, one might speculate that our ESC-derived EC may be similar to a subpopulation of EC re-lated to angiogenic ‘tip cells’ [32–34] .

I f one were to infer conclusions about the subpheno-type of the ESC-derived EC by comparing the vWF and CD34 expression alone, one might alternatively conclude that our ESC-derived EC might be capillary EC or tumor EC; however, the low LDL uptake of these cells is also an important consideration. A group comparing endothelial precursor cells (EPC) with mature endothelial cells and tumor endothelium found that the EPC were weakly pos-itive for VE-cadherin and CD34, and that they expressed variable levels of PECAM1 [35] . The EPC population, however, did not express other markers associated with mature endothelial cells, including acetylated LDL up-take and thrombomodulin. Additionally, EPC were able to invade into clusters of malignant cells, whereas H UVECS did not invade the cell clusters. Because LDL uptake and thrombomodulin expression are related to important functional properties of mature EC, the EPC may not be fully mature functional EC, but instead may represent a slightly immature EC population. These char-acteristics are reminiscent of ESC-derived EC reported in our current study, raising a possibility that the ESC-de-rived EC are immature EPC-like EC that need further differentiation for functional completeness.

A lthough these observational results are somewhat in-conclusive, they clearly outline the differences that do ex-ist between MAEC and in vitro-derived EC. Because these differences may have a significant impact on the functional capabilities of stem cell-derived cells in stem cell therapies, it is important to identify these differences and their potential impact on cell therapy approaches. For example, the potential immature nature of the ESC-derived EC might be advantageous for cell therapy, be-cause we could expect these cells to mature further into sub-phenotypes under appropriate in vivo signals. These ESC-derived EC might also be advantageous in that they are significantly more angiogenic and could be expected to revascularize ischemic tissues more quickly compared with mature EC. Alternatively, these in vitro-derived cells may be missing an important signal for complete functional differentiation, and further optimization of in vitro differentiation techniques may be required. The in-corporation of ESC-derived EC into angiogenic animal models or animal models of ischemia is needed in order to further elucidate the advantages or limitations of ESC-derived EC.

A cknowledgments

T his work was supported by the National Heart, Lung, and Blood Institute under a National Research Service Award Num-ber F32 HL071461 and the Georgia Tech/Emory Center (GTEC) for the Engineering of Living Tissues, an Engineering Research Center funded by the National Science Foundation under Award Number EEC-9731643.

M cCloskey/Smith/Jo/Nerem

J Vasc Res 2006;43:411–421 420

R eferences

1 Dzau VJ, et al: Therapeutic potential of en-

dothelial progenitor cells in cardiovascular diseases. Hypertension 2005;46:7–18.

2 Soker S, Machado M, Atala A: Systems for

therapeutic angiogenesis in tissue engineer-

ing. World J Urol 2000,18:10–18.

3 Kalka C, et al: Transplantation of ex vivo ex-

panded endothelial progenitor cells for ther-

apeutic neovascularization. Proc Natl Acad Sci USA 2000;97:3422–3427.

4 Kawamoto A, et al: Therapeutic potential of

ex vivo expanded endothelial progenitor cells for myocardial ischemia. Circulation 2001;103:634–637.

5 Kocher AA, et al: Neovascularization of

ischemic myocardium by human bone-mar-

row-derived angioblasts prevents cardio-

myocyte apoptosis, reduces remodeling and improves cardiac function. Nature Med 2001;7:430–436.

6 Condorelli G, et al: Cardiomyocytes induce

endothelial cells to trans-differentiate into cardiac muscle: implications for myocardi-

um regeneration. Proc Natl Acad Sci USA 2001;98:10733–10738.

7 Griese DP, et al: Isolation and transplanta-

tion of autologous circulating endothelial cells into denuded vessels and prosthetic grafts: implications for cell-based vascular therapy. Circulation 2003;108:2710–2715.

8 Kaushal S, et al: Functional small-diameter

neovessels created using endothelial progen-

itor cells expanded ex vivo. Nature Med 2001;7:1035–1040.

9 Odorico JS, Kaufman DS, Thomson JA: Mul-

tilineage differentiation from human em-

bryonic stem cell lines. Stem Cells 2001;19:

193–204.

10 Levenberg S, et al: Endothelial cells derived

from human embryonic stem cells. Proc Natl

Acad Sci USA 2002;99:4391–4396.

11 Choi K, et al: A common precursor for hema-

topoietic and endothelial cells. Development

1998;125(4):725–732.

12 Risau W, et al: Vasculogenesis and angiogen-

esis in embryonic-stem-cell-derived embry-

oid bodies. Development 1988;102:471–478. 13 Vittet D, et al: Embryonic stem cells differ-

entiate in vitro to endothelial cells through

successive maturation steps. Blood 1996;88:

3424–3431.

14 Wang R, Clark R, Bautch VL: Embryonic

stem cell-derived cystic embryoid bodies

form vascular channels: an in vitro model of

blood vessel development. D evelopment

1992;114:303–316.

15 Nishikawa SI, et al: Progressive lineage anal-

ysis by cell sorting and culture identifies

FLK1+VE-cadherin+ cells at a diverging

point of endothelial and hemopoietic line-

ages. Development 1998;125:1747–1757.

16 Yamashita J, et al: Flk1-positive cells derived

from embryonic stem cells serve as vascular

progenitors. Nature 2000;408(6808):92–96.

17 Nishikawa SI, et al: Cell biology of vascular

endothelial cells. Ann NY Acad Sci 2001;947:

35–40; discussion 41.

18 McCloskey KE, et al: Purified and proliferat-

ing endothelial cells derived and expanded

in vitro from embryonic stem cells. Endothe-

lium 2003;10:329–336.

19 McCloskey KE, Gilroy ME, Nerem RM: Use

of embryonic stem cell-derived endothelial

cells as a cell source to generate vessel struc-

tures in vitro. Tissue Eng 2005;11:497–505.

20 Hwang J, et al: Oscillatory shear stress stim-

ulates endothelial production of O2- from

p47phox-dependent NAD(P)H oxidases,

leading to monocyte adhesion. J Biol Chem

2003;278:47291–47298.

21 Levesque MJ, Nerem RM: The elongation

and orientation of cultured endothelial cells

in response to shear stress. J Biomech Eng

1985;107:341–347.

22 Sorescu GP, et al: Bone morphogenic protein

4 produced in endothelial cells by oscillatory

shear stress induces monocyte adhesion by

stimulating reactive oxygen species produc-

tion from a nox1-based NAD PH oxidase.

Circ Res 2004;95:773–779.

23 Saito S, Liu B, Yokoyama K: Animal embry-

onic stem (ES) cells: self-renewal, pluripo-

tency, transgenesis and nuclear transfer.

Hum Cell 2004;17:107–115.

24 Belloni PN, Tressler RJ: Microvascular endo-

thelial cell heterogeneity: interactions with

leukocytes and tumor cells. Cancer Metasta-

sis Rev 1990;8:353–389.

25 Garlanda C, Dejana E: Heterogeneity of en-

dothelial cells. Specific markers. Arterio-

scler Thromb Vasc Biol 1997;17:1193–1202.

26 Pusztaszeri M, et al: Immunohistochemical

expression of endothelial markers CD31,

CD34, von Willebrand Factor, and Fli-1 in

Normal Human Tissues. J Histochem Cyto-

chem 2006;54(4):385-395.

27 Turner RR, et al: Endothelial cell phenotypic

diversity. In situ demonstration of immuno-

logic and enzymatic heterogeneity that cor-

relates with specific morphologic subtypes.

Am J Clin Pathol 1987;87:569–575.

28 Fina L, et al: Expression of the CD34 gene in

vascular endothelial cells. Blood 1990,75:

2417–2426.

29 Delia D, et al: CD34 expression is regulated

reciprocally with adhesion molecules in vas-

cular endothelial cells in vitro. Blood 1993;

81:1001–1008.

30 Schlingemann RO, et al: Leukocyte antigen

CD34 is expressed by a subset of cultured en-

dothelial cells and on endothelial abluminal

microprocesses in the tumor stroma. Lab In-

vest 1990;62:690–696.

31 Martin-Padura I, et al: Expression of VE

(vascular endothelial)-cadherin and other

endothelial-specific markers in haemangio-

mas. J Pathol 1995;175:51–57.

32 Eichmann A, et al: Guidance of vascular and

neural network formation. Curr Opin Neu-

robiol 2005;15:108–115.

33 Gerhardt H, Betsholtz C: How do endothe-

lial cells orientate? EXS 2005;94:3–15.

34 Gerhardt H, et al: Neuropilin-1 is required

for endothelial tip cell guidance in the devel-

oping central nervous system. D ev D yn

2004;231:503–509.

35 Bagley RG, et al: Endothelial precursor cells

as a model of tumor endothelium: character-

ization and comparison with mature endo-

thelial cells. Cancer Res 2003;63:5866–

5873.

I n vitro-Derived Endothelial Cells J Vasc Res 2006;43:411–421421

相关主题
相关文档 最新文档