BASIC AND EXPERIMENTAL RESEARCH

Therapeutic Effect of Human Umbilical CordMultipotent Mesenchymal Stromal Cells in a Rat

Model of StrokeWenbin Liao,1 Jiang Xie,1 Jian Zhong,2,3 Yongjun Liu,1,2 Lei Du,4 Bin Zhou,5,6 Jie Xu,1 Pengxia Liu,1

Shaoguang Yang,1 Jiming Wang,2 Zhibo Han,1 and Zhong Chao Han1,2,7

Background. Human umbilical cord multipotent mesenchymal stromal cells (UC-MSC) have recently been identifiedas ideal candidate stem cells for cell-based therapy. The present study was designed to evaluate therapeutic potentials ofintracerebral administration of UC-MSC in a rat model of stroke.Methods. Rats were subjected to 2-hr middle cerebral artery occlusion and received 2�105 UC-MSC or phosphate-buffered saline as a control. Neurologic function evaluation was conducted weekly after transplantation. Brain injuryvolume and in vivo differentiation of transplanted UC-MSC were detected 2 or 5 weeks after the UC-MSC treatment.In addition, vascular density, vascular endothelial growth factor, and basic fibroblast growth factor expression inipsilateral hemisphere after treatment and in vitro angiogenic potential of UC-MSC were assessed.Results. The transplanted UC-MSC survived for at least 5 weeks in rat brain. Compared with the phosphate-bufferedsaline control, the UC-MSC treatment significantly reduced injury volume and neurologic functional deficits of ratsafter stroke. In ischemic brain, UC-MSC widely incorporated into cerebral vasculature and a subset of them was capableof differentiating into endothelial cells. Furthermore, the UC-MSC treatment substantially increased vascular densityand vascular endothelial growth factor and basic fibroblast growth factor expression in ipsilateral hemisphere of stroke.In vitro induction and tube formation assay further confirmed their angiogenic properties.Conclusions. UC-MSC transplantation could accelerate neurologic functional recovery of rats after stroke, which maybe mediated by their ability to promote angiogenesis.

Keywords: Mesenchymal stromal cells, Human umbilical cord, MCAO, Angiogenesis, Cell-based therapy.

(Transplantation 2009;87: 350–359)

S troke causes 9% of all deaths around the world and is thesecond most common cause of death after ischemic heart

disease (1). Stem cells present tremendous potential in the

treatment of a variety of neurodegenerative diseases (2). Mul-tipotent mesenchymal stromal cells (MSC), which are previ-ously known as mesenchymal stem cells (3), have become apromising candidate for cell-based therapy because of theirmultipotency and immune tolerant feature (4, 5). Accumu-lating evidence revealed that bone marrow MSC (BM-MSC)transplantation could improve neurologic function in ratsafter stroke (6 –9), suggesting the potential use of MSC incerebral ischemia. However, the underlying mechanism ofthe benefit of MSC remains unclear. It has been known thatregulation of cerebral blood flow is critical for the mainte-nance of neural function and hence survival of the organism(10). Thus, augmenting angiogenesis after cerebral ischemiaoffers a promising approach to the recovery of damaged ce-rebral tissue. Strategies such as administration of angiogenicfactors have been proven effective to increase cerebral bloodflow and reduce infarct size after experimental stroke (11, 12).It is known that MSC have the capacity of secretion of variousangiogenic factors and the potential to differentiate into vas-cular cell phenotype (13–15), and recent studies have shownthat MSC may improve neovascularization in ischemic dis-eases, such as critical limb ischemia (CLI) and myocardialinfarction (MI) (16 –18). Therefore, the contribution of MSCinjection to the subsequent neurologic function improve-

This study was supported by 863 projects from Ministry Science and Tech-nology of China (2006AA02A110), National Natural Science Foundationof China (30570357 and 30570905), and Tianjin Municipal Science andTechnology Commission (06YFSZSF01300 and 07JCYBJC11200).

The authors declare no potential conflicts of interest.1 State Key Laboratory of Experimental Hematology, Institute of Hematol-

ogy, CAMS & PUMC, Tianjin, China.2 National Engineering Research Center of Cell Products, AmCellGene Co.

Ltd., Tianjin, China.3 School of Chemical Engineering and Technology of Tianjin University,

Tianjin, China.4 Department of Neurosurgery, Hebei Medical University, Shijiazhuang, China.5 Harvard Stem Cell Institute and Department of Cardiology, Children’s

Hospital Boston, Boston, MA.6 Department of Genetics, Harvard Medical School, Boston, MA.7 Address correspondence to: Zhong Chao Han, Ph.D., M.D., Institute of He-

matology, CAMS & PUMC, 288 Nanjing Road, Tianjin 300020, People’sRepublic of China.

E-mail: tihzchan@public.tpt.tj.cnReceived 1 September 2008. Revision requested 10 October 2008.Accepted 13 October 2008.Copyright © 2009 by Lippincott Williams & WilkinsISSN 0041-1337/09/8703-350DOI: 10.1097/TP.0b013e318195742e

350 Transplantation • Volume 87, Number 3, February 15, 2009

ment in cerebral ischemia may derive from their ability topromote angiogenesis in ischemic brain.

Recently, we have established a method to readily iso-late and expand MSC from human umbilical cord (UC-MSC)and shown that these cells are biologically similar to BM-MSC(19), which is consistent with the observation of several othergroups (20 –22). In comparison with MSC derived from BM,UC-MSC have several advantages, including painless proce-dures, more abundant cell number, and lower risk of viralcontamination. In addition, previous studies have reportedthat the application of UC-MSC could lead to an improved func-tional outcome in neural diseases, such as Parkinson’s diseaseand cerebral global ischemia (23, 24). We have also shown thetherapeutic benefits of UC-MSC transplantation in rat MI andmouse CLI models (25, 26). The present study was designed todetermine the potential of human UC-MSC in improving neu-rologic functional deficits in rats subjected to middle cerebralartery occlusion (MCAO), and whether UC-MSC treatmentcould promote angiogenesis in rat brain after stroke.

MATERIALS AND METHODS

Isolation and Identification of UC-MSCUCs from both sexes were collected from full-term caesar-

ian section deliveries with informed consent of the mothers. Tis-sue collection for research was approved by the institutionalreview board of the Chinese Academy of Medical Science andPeking Union Medical College. UC-MSC were isolated andexpanded as described previously (19). Briefly, the cord wascut with scissor into pieces (1–2 mm3) and digested with0.075% collagenase II (Sigma, St Louis, MO) for 30 min andwith 0.125% trypsin (Gibco, Grand Island, NY) for another30 min with gentle agitation at 37°C. After terminating thetrypsin activity with serum, the digested mixture was thenpassed through a 100-�m filter to obtain cell suspensions.Cells were washed with phosphate-buffered saline (PBS) forthree times and then plated onto noncoated plastic flasks inDulbecco’s minimum essential medium/F12 (1:1) (Gibco) sup-plemented with 10% fetal bovine serum (HyClone, Logan, UT),100 U/mL penicillin-streptomycin (Sigma), 1% glutamine(Sigma), and 10 ng/mL epidermal growth factor (Sigma). Non-adherent cells were removed by changing the medium after 3days. Cultures were maintained at 37°C in a humidified atmo-sphere containing 5% CO2 and passaged at 60% to 80% conflu-ency. Cells of passage four to six were used in this study.

For surface phenotype analysis, cells were stained withfluorescein isothiocyanate- or phycoerythrin-conjugatedmonoclonal antibodies, cells surface markers includedhematopoietic lineage markers (CD34, CD38, CD45,CD14), angiogenic markers (CD31, von Willebrand factor[vWF], Flk-1), adhesion integrins (CD44, CD29, CD49,CD54, CD106, CD166), human leukocyte antigen (HLA)-ABC and HLA-DR, and MSC markers of CD13, CD105(SH2),CD73(SH3), and CD90. Antibodies used were purchased fromBecton Dickinson (BD, San Diego, CA) except vWF antibodywas from Dako (Glostrup, Denmark). Cells were then ana-lyzed by flow cytometry with a FACScan cytometer (BD).

In vitro osteogenic and adipogenic differentiation ofUC-MSC were analyzed with the protocol as described previ-ously (19). Von Kossa and Oil Red O (Sigma) staining were

used to confirm the characteristics of osteocytes and adipo-cytes, respectively.

Tube Formation AssayMatrigel (Sigma) was thawed on ice to prevent prema-

ture polymerization, aliquots of 50 �L were plated into indi-vidual wells of 96-well tissue culture plates and allowed topolymerize at 37°C for 30 min. The cells were washed in serum-containing medium, then resuspended at 106 cells/mL. Cell sus-pension (100 �L) was added to each well supplemented with 10ng/mL vascular endothelial growth factor (VEGF; Sigma),and incubated for 2 to 4 hr at 37°C before being visualizedunder phase contrast microscopy (Olympus, Tokyo, Japan).

In Vitro Induction of UC-MSC into Vascular CellsConfluent UC-MSC were cultured in Dulbecco’s min-

imum essential medium/F12 medium supplemented with 2%fetal calf serum, 100 U/mL penicillin-streptomycin, 50 ng/mLVEGF, and 10 ng/mL basic fibroblast growth factor (bFGF;Sigma) for endothelial cell differentiation. For smooth musclecell induction, 10 ng/mL platelet-derived growth factor-BB(Pepro Tech, Rocky Hill, NJ, USA) instead of VEGF andbFGF was used. Medium was changed every 3 days. Sevendays after induction, immunostaining was performed to de-tect the expression of endothelial specific marker vWF orsmooth muscle cell marker smooth muscle �-actin (ASMA;Chemicon, Tenecula, CA, USA). Diaminobenzidine was usedas a chromogen for light microscopy.

Uptake of DiI-Ac-LDL (Invitrogen, Carlsbad, CA,USA) was also conducted to verify cellular differentiation to-ward endothelial phenotype by incubating the postinducedUC-MSC with 10 �g/mL DiI-Ac-LDL in serum-free mediumat 37°C for 4 hr. VEGF plus bFGF with or without hypoxia(3% O2) were used as induction stimuli.

MCAO Rat Model EstablishmentAdult male Sprague-Dawley rats weighing 230 to 260 g

were used in the experiments. Briefly, rats were initially anes-thetized with 10% chloral hydrate (30 mg/kg). Rectal temper-ature was maintained at 37°C by use of a heat blanket. MCAOwas induced for 2 hr by using a previously described methodof intraluminal vascular occlusion (6, 8).

UC-MSC Labeling and Transplantation ProcedureUC-MSC were labeled with 3 �g/mL CM-DiI (Invitro-

gen) at 37°C for 5 min and then at 4°C for another 15 min.The cells were washed with PBS for three times, and the label-ing efficiency was detected under fluorescence microscope(Olympus). Twenty-four hours after MCAO, rats in cell-treated group (n�20) received 10 �L cell-PBS mixture with atotal number of 2�105 UC-MSC at the following coordi-nates: anterior-posterior (AP)�4.5 mm; and cortex: AP�0mm, middle-lateral (ML)�2.0 mm, dorsal-ventral (DV)�2mm from the bregma based on the atlas by Paxinos andWatson (27). Cells were injected at the speed of 1 �L/min andplaced for another 5 min thereafter. Equal volume of PBS wasadministrated in the control group (n�20), and sham group(rats not subjected to MCAO except exposing the same ves-sels as the MCAO group) received no injection (n�16). Toevaluate the immunogenicity of human UC-MSC and theirangiogenic potential in normoxic condition, equal number of

© 2009 Lippincott Williams & Wilkins 351Liao et al.

cells were also injected into normal animals (n�10). In this set-ting, immune response was detected by CD11b (for macrophageand activated microglia; Chemicon) and myeloperoxidase(MPO, for leukocytes; Abcam, Cambridge, UK) immunostain-ing 3 days after injection, and angiogenic potential of MSC wasanalyzed at the same time point as transplantation in MCAOrats. Rats did not receive any immunosuppression medications.

2,3,5-Triphenyltetrazolium Chloride Stainingand Injury Volume Measurement

Two weeks after transplantation, the rats were killedand their brains were removed carefully and dissected intocoronal 2-mm sections. The fresh brain slices were immersedin a 2% solution of 2,3,5-triphenyltetrazolium chloride(TTC; Sigma) in normal saline at 37°C for 30 min. The cross-sectional area of infarction in each brain slice was examinedwith a dissection microscope and was measured using an im-age analysis software (Image-pro plus software package, Me-dia Cybernetics, Carlsbad, CA). The total infarct volume foreach brain was calculated by summation of the infarcted areaof all brain slices. The ratio of injury volume expressed by thepercentage of infarct volume to volume of contralateral hemi-sphere was used for statistical analysis.

Neurologic EvaluationNeurologic evaluation, including modified neurologic

severity score (mNSS) (7) and Morris water maze test (8), wasperformed before and weekly after MCAO by two individualsblinded to the experimental groups.

mNSS is a composite of motor, sensory, balance, andreflex tests. Neurologic function was graded on a scale of 0 to18 (normal score 0; maximal deficit score 18). In the severityscores of injury, one score point was awarded for the inabilityto perform the test or for the lack of a tested reflex; thus, thehigher score, the more severe is the injury.

Morris water maze test was used to measures learning andmemory. Briefly, the rat was placed in a 1.3-m diameter watertank that is visually separated into four quadrants. In the centerof one quadrant, a platform is hidden 2 cm below the waterline.On the first training day, a single habituation trial was performedby placing each animal on the hidden platform for 60 sec. If theanimal fell or jumped from the platform, it was removed fromthe water and placed back on the platform. The temperature ofthe water maze was 24°C�0.5°C. The latency of quadrant searchand path length of the rat were measured by a video trackingsystem interfaced to a computer.

Immunohistochemical StainingAfter anesthetized, rat brains were fixed by transcardial

perfusion with saline, followed by perfusion and immersionin 4% paraformadehyde, and dehydrated with 20% sucrose inPBS overnight. A series of adjacent 10 �m thick coronal cryostatsections were cut from the brain tissue (AP��1.0 to AP�1.0from the bregma). Antibodies against neuronal specific enolase(NSE; Chemicon), microtubule associated protein 2 (MAP2;Chemicon), glial fibrillary acidic protein (GFAP; Dako), vWF,ASMA, VEGF (Santa Cruz Biotechnology, Santa Cruz, CA),bFGF (Santa Cruz), CD11b (Chemicon), and MPO (Abcam)immunostaining were used, respectively. 4�,6-Diamidino-2-phenylindole (DAPI; Invitrogen) was used to identify nuclear.

Vascular Density and VEGF, bFGF ExpressionAnalysis

To determine the effect of the UC-MSC transplantationon vascular density, rats were killed 7 days after treatment forbrain tissue section, and vWF immunostaining was used to vi-sualize vascular regions. Morphometric measurements wereperformed on five equidistant slices (AP��1.0 to AP�1.0 fromthe bregma) from three animals per group. Per slice, five randomareas of ischemic boundary zone were photographed at a magni-fication of 20� and used for measurement. Vascular areas werecalculated using Image-Pro Plus software package, and vasculardensitywasexpressedbythepercentageofvascularareas.VEGFandbFGF expression quantitation were analyzed by the same methodand expressed by the area of positive staining per square millimeterin ischemia border. The measurements were conducted by two in-dependent examiners who were blinded to treatment assignment.

Statistical AnalysisData are presented as mean�SD. One-way analysis of

variance was used to make comparisons of parameters amonggroups. If the omnibus tests among groups were significantlydifferent, post hoc tests between groups using unpaired t testswere used. P less than 0.05 was considered significant. The SPSS(Chicago,IL,USA)softwarepackagewasusedforthestatisticaltests.

RESULTS

Isolation and Identification of UC-MSCMSC could be isolated from UC with 100% harvesting

efficiency after digested with enzyme cocktail. The cells formed amonolayer of bipolar spindle-like cells with a whirlpool-likearray within 2 weeks (Fig. 1A), and in vitro expansion analysis

FIGURE 1. In vitro differentiation of UC-MSC into osteocytes and adipocytes. Fibroblast-like UC-MSC grow in the growthmedium (A). Positive stain of Von Kossa (B) and Oil red O (C) reflect the osteogenesis and adipogenesis of UC-MSC,respectively. Scale bar�100 �m.

352 Transplantation • Volume 87, Number 3, February 15, 2009

revealed that their proliferation characteristics did not changeeven after 30 passages. Flow cytometric analysis revealed thatUC-MSC were positive for mesenchymal antigens, includingCD13, CD105(SH2), CD73(SH3), CD90, and adhesion mol-ecules CD29, CD 44, CD54, CD49, CD106, and CD166; butnegative for hematopoietic antigens or endothelial cellmarkers (Table 1). Furthermore, these cells express HLA-ABC but not HLA-DR. When cultured with osteogenic oradipogenic conditional medium, UC-MSC could differenti-ate into osteocytes and adipocytes, as detected by positivestaining of von Kossa and Oil Red O, respectively (Fig. 1B, C).

Neurologic Function Improvement AfterUC-MSC Transplantation

Human UC-MSC were injected into the brain of nor-mal rats to evaluate the host immune reaction against thetransplanted human cells. The result showed that injectedUC-MSC did not elicit any obvious immune response aroundthe injection site 3 days after injection (Fig. 2A-a, b), onlysparse leukocytes or marcophages/microglia acumulatedaround the injection site for which small number of dead cellsor cell debris may be responnsible.

Two-hour MCAO and reperfusion in rats caused aninjury that led to neurologic functional deficits as measuredby mNSS. These rats presented with high scores in sensori-motor tests during the early phase after injury. Recoverybegan on day 3 after injury (Fig. 2A-c) and persisted at allsubsequent evaluation points in both PBS-treated and UC-MSC-treated groups. Statistically, the mNSS scores for theUC-MSC-treated group were significantly decreased at day10, 17, 24 and day 31 compared with the PBS-treated group(Fig. 2A-c). This result was also supported by the Morris wa-ter maze assay (Fig. 2A-d), a test of learning and memoryability. The latency of Morris water maze assay in UC-MSC-

treated group significantly decreased compared with controlgroup at 14, 21 days after treatment, suggesting UC-MSCtransplantation fostered neurologic functional recovery afterstroke.

UC-MSC Treatment Reduces Injury Volume ofStroke

Macroscopic appearance of whole brain tissue showedreduced lesion area on the cortex surface of the UC-MSC-treated rats (Fig. 2A-f) compared with the control group (Fig.2A-e). The extent of the infarct lesions was examined withTTC staining (8). Normal brain stains red with this method,but the cerebral infarct regions in the brains of the MCAOmodel rats showed no or reduced staining, clearly delineatingareas of infarction. TTC staining 2 weeks after treatmentshowed that rats in the UC-MSC-treated group (Fig. 2A-h)had smaller injury volume of ischemic brain than those in thePBS-treated group (Fig. 2A-g). In the control group, the ratioof injury volume to contralateral hemisphere was 35.2%�7.7%,whereas in UC-MSC-treated group, the ratio decreased to18.7%�2.9% (Fig. 2A-i, P�0.05), suggesting a neuroprotectiveeffect of UC-MSC.

In Vivo Differentiation of UC-MSC TowardNeuron/Gliacyte

Five weeks after transplantation, CM-DiI-labeledMSCs were still observed in rat brain with a preferential dis-tribution in ischemic boundary regions. By immunohisto-chemistry, a small number of donor cells were found positivefor neural specific markers NSE, MAP2, or GFAP (Fig. 2B).The percentage of CM-Dil-labeled UC-MSC expressing NSE,MAP2, and GFAP were �1%, �3%, and �2%, respectively.

UC-MSC Treatment Increases Vascular Densityand Angiogenic Factors Expression in IschemicRat Brain

vWF immunostaining revealed that the UC-MSC treat-ment significantly increased the vascular density in ischemicboundary zone 7 days after the treatment compared with thePBS-treated group (Fig. 3A). Above 90% blood vesselsaround ischemic region contained transplanted UC-MSC,which incorporated into cerebral vasculature or located invessel wall (Fig. 3B). Immunostaining assay demonstrateda part of these incorporated donor cells expressed endo-thelial cell-specific protein vWF (Fig. 3C), whereas donorcells positive for smooth muscle cell marker AMSA werenot detected.

VEGF and bFGF are well-known angiogenic factors.Treatment of UC-MSC substantially increased VEGF andbFGF expression around ischemic region (Fig. 3D, E), and apart of them were secreted by the transplanted UC-MSC (Fig.3F). Thus, the UC-MSC treatment increased the vesseldensity through direct incorporation into host vasculature orup-regulation of angiogenic factors, which leads to improvedcirculation after ischemia.

To evaluate the influence of hypoxia on donor cellsmigration and their capability of promoting angiogenesis, wealso intracerebrally injected UC-MSC into animals withoutMCAO. Seven days after transplantation, most of the UC-MSC remained in the injection regions in normoxic brainwith few migrated cells adjacent to the injection regions (Fig.

TABLE 1. Phenotype characteristics of UC-MSC

Surface markersPositive rate

(%)Surfacemarkers

Positive rate(%)

Hematopoieticmarkers

Cell adhesionmolecules

CD34 � CD44 ����

CD38 � CD29 ����

CD45 � CD49 ����

CD14 � CD54 ���

Mesencymalmarkers

CD106CD166

�

CD73(SH3) ���� ���

CD105(SH2,endoglin)

����MHC

CD90(Thy-1) ���� HLA-ABC ���

CD13 ���� HLA-DR �

Endothelialmarkers

CD31(PECAM) �

CD133 �

vWF �

Surface markers of UC-MSC were detected by flow cytometric analysis.�, �1%; �, 1%–25%; ��, 25%–50%; ���, 50%–70%; ����, 75%.

© 2009 Lippincott Williams & Wilkins 353Liao et al.

4C, D). In contrast, in hypoxic brains, many more cells mi-grated in distance from the injection site toward the ischemicregion (Fig. 4A, B). We also compared the vascular densityand number of incorporated cells in vasculature around theinjection regions in normoxic and hypoxic condition. Results

showed that vascular density was not different between thetwo groups (Fig. 4G, left; P0.05), but the number of incor-porated donor cells in vessels was significantly higher in theMCAO group than in the normoxic group (Fig. 4G, right;P�0.05).

FIGURE 2. Neurologic function improvement after UC-MSC treatment (A) and in vivo neural differentiation of UC-MSC(B). (A) No obvious immune response of UC-MSC in rat brain (n�5) as showed by CD11b (a) and MPO (b) immunostaining.Neurologic evaluation shows functional recovery of rats after UC-MSC transplantation (c, d: n�6 in UC-MSC-treated group,n�5 in PBS group, n�6 in sham group). Representative macroscopic appearance of brain cortex (e, f: areas with black dotmargin show the ischemia area) and TTC staining (g, h) in UC-MSC-treated group (f, h) and control group (e, g). Quantitativeanalysis of the ratio of injury volume reveals the difference is significant (i: n�3 per group; P�0.05). Scale bar: a�50 �m;e�2 mm. (B) CM-DiI (red) labeled UC-MSC could be observed in ischemic brain 5 weeks posttransplantation (n�6), arrowsshow UC-MSC colocalize with NSE, MAP2, or GFAP. DAPI (blue) staining was used to identify nuclear. Scale bar�50 �m.

354 Transplantation • Volume 87, Number 3, February 15, 2009

In Vitro Tube Formation Assay and Induction ofUC-MSC into Vascular Cells

Tube formation assay revealed that UC-MSC werecapable of forming capillary-like structure on Matrigel(Fig. 5A), whereas fibroblasts were negative for this test(Fig. 5B). In presence of 50 ng/mL VEGF and 10 ng/mL

bFGF, UC-MSC could express vWF (Fig. 5C), whereas 10ng/mL platelet-derived growth factor-BB directed UC-MSC into smooth muscle cell phenotype as UC-MSC wereexpressing AMSA (Fig. 5E).

Uptake of DiI-labeled ac-LDL is considered to be one ofthe typical functions for mature endothelial cells (28). After

FIGURE 3. UC-MSC increase vascular density and angiogenic factors expression. Representative images show thevascular density around ischemic areas in UC-MSC-treated group and control group (A, n�3 in each group). UC-MSCwidely incorporate into endothelium or locate in vessel wall in ipsilateral hemisphere of stroke (B), and a subset of UC-MSC(red) colocalize with vWF (green) in rat brain (C, top), whereas UC-MSC do not express ASMA in vivo (C, bottom). Repre-sentative samples of VEGF and bFGF immunostaining at 7 days after transplantation in UC-MSC-treated group and controlgroup (D, n�4 in each group). (E) Quantitative analysis of VEGF and bFGF density in ischemia boundary regions. Apopulation of transplanted cells are positive for VEGF and bFGF (F). DAPI (blue) staining was used to identify nuclear. Allscale bars represent 50 �m in this figure.

© 2009 Lippincott Williams & Wilkins 355Liao et al.

culture in endothelial differentiation medium containingVEGF and bFGF for 7 days, UC-MSC were positive for ac-LDL uptake (Fig. 5G, H). Moreover, hypoxia further en-hanced their capacity of ac-LDL uptake (Fig. 5I, J), whereasUC-MSC cultured in normal growth medium were negativefor ac-LDL uptake (Fig. 5K, L), suggesting that the hypoxia-induced milieu is preferred for driving MSC into endotheliallineage.

DISCUSSIONUC tissue has recently been regarded as a rich source of

stem/progenitor cells (21, 22, 29 –31). The UC tissue is com-posite of amniotic epithelium, two arteries and one vein, andWharton’s Jelly, and at least six distinct zones are recognizedbased on the structural and functional studies (32). Until to-day, all of the components of UC have been reported to con-tain stem/progenitor cells and these cells shared similarcharacteristics with BM-MSC based on morphology, pheno-typic profile, and in vitro differentiation analysis (20 –22, 29,30), which was proposed by Mesenchymal and Tissue StemCell Committee of the International Society for Cellular

Therapy to define human MSC (33). Recently, we isolatedMSC from UC by digesting the complete tissue and found thismethod increased the success rate and the absolute number ofcolony-forming unit-fibroblast (19). Also, cells obtained us-ing our method fulfilled the minimal criteria of MSC by In-ternational Society for Cellular Therapy. Therefore, digestionof the whole tissue may represent a more promising approachto obtain MSC from UC as compared with site-specific isola-tion methods, especially for the potential translation intoclinic of cell-based therapy in which the cells should be insignificant numbers and in the form that require little prepa-ration. A recent study also reported the potential use of UC-derived cells by whole tissue digestion in neural diseases (34).They found that these cells showed typical MSC characteris-tics and genetic stability in vitro and more effectively, com-pared with BM-MSC, rescued photoreceptors and visualfunctions in a rodent model of retinal disease. Of note is theheterogenicity of these UC-derived MSC, which requires fur-ther clarification. Actually, it is a common challenge toidentify MSC in regard to the lack of unique marker ofthem.

FIGURE 4. Brain hypoxia after MCAO en-hances migration and vascular incorporationof transplanted UC-MSC. UC-MSC (red) inhypoxic brain (A) show greater migration ascompared with those in normoxic brain (C).(B, D) Magnification of boxed areas of (A)and (C), respectively. (E, F) Representativesamples of vascular incorporation and vWFexpression of UC-MSC around injection re-gion in hypoxic (E) or normoxic (F) brain.(G) Quantitative analysis of vascular densityand the number of incorporated cells in twogroups (n�3 in each group). Scale bar: (A,C)�200 �m; (B, D–F)�50 �m.

356 Transplantation • Volume 87, Number 3, February 15, 2009

In the present study, we showed that UC-MSC couldsurvive and improve neurologic function in rats after stroke.The critical question of host immune response in cell-basedtherapy closely relates with the survival of implanted donorcells and their therapeutic effects in transplantation. Thoughemerging evidence has suggested that MSC are nonimmu-nogenic or hypoimmunogenic because of their distinct im-munophenotypic profile associated with absence of HLA-IImolecule and low expression of costimulatory factors (4, 5,19, 35), it should be cautious in regard to allogeneic and xe-nogeneic transplantation because immune rejection of hu-man BM-derived cells has been reported when transplantedin rat after MI (36). The observation that UC-MSC did notelicit obvious immune response when injected into rat probablyassociated with the relatively immunoprivileged environment inthe brain (37, 38), and this may explain the long-term survival ofthese cells (at least 5 weeks) in rat brain.

Previous studies have reported the benefits of MSCtreatment in rats after stroke. However, the underlying mech-anism of MSC therapy for ischemia animals is still unclear.Chen et al. (6) found that BM-MSC survived, migrated, dif-ferentiated into neural phenotypes, and significantly acceler-ated functional recovery in rats after stroke. Our work alsofound that exogenous UC-MSC could express neural specificmarkers in ischemic brain, which confirmed our previousresults that UC-MSC could be induced into neural phenotypein vitro (19). However, it is hard to attribute functional im-provement to cell replacement because transdifferentiation isnot a common event in our study, which is consistent withother studies (6, 39). Moreover, whether these cells function-ally coupled with host cells remains unknown.

We then focused on the angiogenic potential of UC-MSC in this study because accumulating data revealed thatMSC were capable of promoting neovascularization in isch-emic diseases (16 –18), and we have previously reported thatUC-MSC enhanced angiogenesis in both rat MI and mouseCLI model (25, 26). UC-MSC treatment increased vasculardensity after ischemia; thus, may improve the cerebral circu-

lation and maintain neural function because patients withstroke with a greater cerebral blood vessel density seem tomake better progress and survive longer than patients withlower vascular density (40). Actually, BM-MSC have beendemonstrated to have potential of promoting angiogenesis inMCAO animals (41– 43). Angiogenic factor up-regulation byMSC was believed to be a mediator for amplified neovascu-larization after stroke (42, 43), which is consistent with ourobservation that UC-MSC increased VEGF and bFGF expres-sion in ischemic brain. We also found that VEGF and bFGFwere partially produced by transplanted UC-MSC. Indeed,both BM- and UC-derived MSC could secrete angiogenic fac-tors, such as VEGF, bFGF, and placenta growth factor (PIGF)(13, 15, 16), all of which are known to be effective in reducingcerebral ischemic injury (11, 12). MSC of different sourcesmay have different expression profiles of angiogenic factors(15, 23), which may determine their distinct angiogenic ef-fects in stroke.

Direct integration of UC-MSC into host vasculaturemay be another mechanism to increase angiogenesis after ce-rebral ischemia. Approximately 90% blood vessels in periis-chemia regions contained exogenous UC-MSC, either inendothelium or blood vessel wall, and a part of them differ-entiated into endothelial cells, which is consistent with ourprevious study of transplantation of UC-MSC into MI ratmodel (26). The angiogenic potential of UC-MSC was con-firmed by in vitro assay as they could form capillary-likestructure and differentiate into vascular cells. Angiogenesisrequires sprounting of new capillaries from preexsisting ves-sels in which endothelial cells retain their ability to proliferate(44). The incorporation of UC-MSC into vasculature andtransformation to endothelial cells may increase the cellnumber of “endothelial cells pool,” thus augmenting sprout-ing angiogenesis. In addition, smooth muscle cells, pericytes,and extracellular matrix are crucial for vascular remodelingand maturation (44). Therefore, our observation that partialUC-MSC located in blood vessel wall suggested these cellsmight also participate into vascular remodeling. Although we

FIGURE 5. In vitro tube formation assayand differentiation of UC-MSC into vascularcells. UC-MSC form capillary structure onMatrigel (A), human fibroblasts are negativeof tube formation (B). UC-MSC differentiateinto vascular cell lineage in conditioned me-dium as showed by immunopositive for vWF(C) or ASMA (E). (D, F) Controls in which pri-mary antibody is omitted. VEGF and bFGFsupplement enable UC-MSC to uptake DiI-ac-LDL (red) (G), and this effect is enhancedby hypoxia (I), whereas UC-MSC in growthmedium fail to uptake DiI-ac-LDL (K). (H, J, L)Merged pictures of (G, I, K) with nuclearstaining (DAPI, blue). Scale bar�50 �m.

© 2009 Lippincott Williams & Wilkins 357Liao et al.

did not observe in vivo smooth muscle cell marker AMSAexpression of these UC-MSC [which may attribute to our useof later passages (P4 –P6) cells for treatment, see Ref. (45)],they may function as pericytes because a recent report dem-onstrated a close relationship of MSC with pericytes (46). Inthis setting, Lindolfo et al. (46) analyzed the evidence ofperivascular location for MSC and proposed that MSC stabi-lize blood vessels and contribute to tissue homeostasis underphysiological conditions and assume a more active role in therepair of focal tissue injury. Supporting evidence was thatSarugaser et al. (31) reported UC perivascular cells as a sourceof mesenchymal progenitors and a following study (20) fur-ther revealed these cells expressed high level of CD146, a cellmarker of pericytes. These data implied that transplantedUC-MSC probably functioned as supporting cells for bloodvessel maturation. In addition, MSC are capable of producingplentiful adhesion molecules, such as vascular cell adhesionmolecule, cadherins and integrin (19, 23, 31), which play acrucial role in angiogenesis (44, 47).

Hypoxia is probably responsible for cell migration andthe overwhelming integration of UC-MSC into vasculature inrats after stroke. UC-MSC in normoxic brain tissue predom-inantly stayed where they were injected and only a smallnumber of migrated cells could be observed adjacent to theinjection site. In contrast, UC-MSC transplanted in ischemicbrain selectively migrated toward ischemic penumbra andfrequently coupled into blood vessels. In vitro assay also re-vealed that hypoxia was favorable for induction of UC-MSCinto endothelial cell. These observations were consistent withprevious studies that hypoxic condition could enhance mi-gration and angiogenic potential of MSC (48 –50). Short-term exposure of MSC to low oxygen was shown to be able toincrease their expression of migration related molecules andengraftment in vivo (48), and hypoxic preconditioning couldlead to increased expression of angiogenic factors and en-hance the capacity of MSC to repair infarcted myocardium(49, 50). Therefore, cerebral ischemia created a hypoxic mi-croenvironment that may promote UC-MSC migration andprovoke their angiogenic potentials through direct differen-tiation into vascular cells or secretion of angiogenic factors.

In conclusion, we demonstrated that UC-MSC couldreduce neurologic defects after stroke in rats, and this benefitmay attribute to their ability to promote angiogenesis in isch-emia brain. Given the advantages of UC over BM as a sourceof MSC, our data suggest that UC-MSC treatment may be-come an ideal strategy of stem cell-based therapy for centralnervous system injury and disease.

REFERENCES1. Murray CJ, Lopez AD. Mortality by cause for eight regions of the world:

Global Burden of Disease Study. Lancet 1997; 349: 1269.2. Yu D, Silva GA. Stem cell sources and therapeutic approaches for cen-

tral nervous system and neural retinal disorders. Neurosurg Focus 2008;24: E11.

3. Horwitz EM, Le Blanc K, Dominici M, et al. Clarification of the no-menclature for MSC: The International Society for Cellular Therapyposition statement. Cytotherapy 2005; 7: 393.

4. Porada CD, Zanjani ED, Almeida-Porad G. Adult mesenchymal stemcells: A pluripotent population with multiple applications. Curr StemCell Res Ther 2006; 1: 365.

5. Noel D, Djouad F, Bouffi C, et al. Multipotent mesenchymal stromalcells and immune tolerance. Leuk Lymphoma 2007; 48: 1283.

6. Chen J, Li Y, Wang L, et al. Therapeutic benefit of intracerebral trans-

plantation of bone marrow stromal cells after cerebral ischemia in rats.J Neurol Sci 2001; 189: 49.

7. Kang SK, Lee DH, Bae YC, et al. Improvement of neurological deficitsby intracerebral transplantation of human adipose tissue-derived stro-mal cells after cerebral ischemia in rats. Exp Neurol 2003; 183: 355.

8. Iihoshi S, Honmou O, Houkin K, et al. A therapeutic window for in-travenous administration of autologous bone marrow after cerebralischemia in adult rats. Brain Res 2004; 1007: 1.

9. Pavlichenko N, Sokolova I, Vijde S, et al. Mesenchymal stem cells trans-plantation could be beneficial for treatment of experimental ischemicstroke in rats. Brain Res 2008; 1233: 203.

10. Pratt PF, Medhora M, Harder DR. Mechanisms regulating cerebralblood flow as therapeutic targets. Curr Opin Investig Drugs 2004; 5: 952.

11. Tanaka R, Miyasaka Y, Yada K, et al. Basic fibroblast growth factorincreases regional cerebral blood flow and reduces infarct size afterexperimental ischemia in a rat model. Stroke 1995; 26: 2154.

12. Kaya D, Gursoy-Ozdemir Y, Yemisci M, et al. VEGF protects brainagainst focal ischemia without increasing blood– brain permeabilitywhen administered intracerebroventricularly. J Cereb Blood FlowMetab 2005; 25: 1111.

13. ChenX,LiY,WangL,etal. Ischemicratbrainextracts inducehumanmarrowstromal cell growth factor production. Neuropathology 2002; 22: 275.

14. Oswald J, Boxberger S, Jorgensen B, et al. Mesenchymal stem cells canbe differentiated into endothelial cells in vitro. Stem Cells 2004; 22: 377.

15. Friedman R, Betancur M, Boissel L, et al. Umbilical cord mesenchymalstem cells: Adjuvants for human cell transplantation. Biol Blood Mar-row Transplant 2007; 13: 1477.

16. Kinnaird T, Stabile E, Burnett MS, et al. Local delivery of marrow-derived stromal cells augments collateral perfusion through paracrinemechanisms. Circulation 2004; 109: 1543.

17. Silva GV, Litovsky S, Assad JA, et al. Mesenchymal stem cells differen-tiate into an endothelial phenotype, enhance vascular density, and im-prove heart function in a canine chronic ischemia model. Circulation2005; 111: 150.

18. Tang J, Xie Q, Pan G, et al. Mesenchymal stem cells participate inangiogenesis and improve heart function in rat model of myocardialischemia with reperfusion. Eur J Cardiothorac Surg 2006; 30: 353.

19. Lu LL, Liu YJ, Yang SG, et al. Isolation and characterization of humanumbilical cord mesenchymal stem cells with hematopoiesis-supportivefunction and other potentials. Haematologica 2006; 91: 1017.

20. Baksh D, Yao R, Tuan RS. Comparison of proliferative and multilin-eage differentiation potential of human mesenchymal stem cells de-rived from umbilical cord and bone marrow. Stem Cells 2007; 25: 1384.

21. Kim JW, Kim SY, Park SY, et al. Mesenchymal progenitor cells in thehuman umbilical cord. Ann Hematol 2004; 83: 733.

22. Wang HS, Hung SC, Peng ST, et al. Mesenchymal stem cells in theWharton’s jelly of the human umbilical cord. Stem Cells 2004; 22: 1330.

23. Weiss ML, Medicetty S, Bledsoe AR, et al. Human umbilical cord ma-trix stem cells: Preliminary characterization and effect of transplanta-tion in a rodent model of Parkinson’s disease. Stem Cells 2006; 24: 781.

24. Jomura S, Uy M, Mitchell K, et al. Potential treatment of cerebral globalischemia with Oct-4� umbilical cord matrix cells. Stem Cells 2007; 25: 98.

25. Wu KH, Zhou B, Mo XM, et al. Therapeutic potential of human umbilicalcord-derived stem cells in ischemic diseases. Transplant Proc 2007; 39: 1620.

26. Wu KH, Zhou B, Yu CT, et al. Therapeutic potential of human umbil-ical cord derived stem cells in a rat myocardial infarction model. AnnThorac Surg 2007; 83: 1491.

27. Paxinos G, Watson C. The rat brain in stereotaxic coordinates, ed. 5.San Diego, Elsevier Academic Press 2005.

28. Voyta JC, Via DP, Butterfield CE, et al. Identification and isolation ofendothelial cells based on their increased uptake of acetylated-low den-sity lipoprotein. J Cell Biol 1984; 99: 2034.

29. Ilancheran S, Michalska A, Peh G, et al. Stem cells derived from humanfetal membranes display multilineage differentiation potential. Biol Re-prod 2007; 77: 577.

30. Romanov YA, Svintsitskaya VA, Smirnov VN. Searching for alternativesources of postnatal human mesenchymal stem cells: Candidate MSC-like cells from umbilical cord. Stem Cells 2003; 21: 105.

31. Sarugaser R, Lickorish D, Baksh D, et al. Human umbilical cordperivascular (HUCPV) cells: A source of mesenchymal progenitors.Stem Cells 2005; 23: 220.

32. Can A, Karahuseyinoglu S. Concise review: Human umbilical cordstroma with regard to the source of fetus-derived stem cells. Stem Cells2007; 25: 2886.

358 Transplantation • Volume 87, Number 3, February 15, 2009

33. Dominici M, Le Blanc K, Mueller I, et al. Minimal criteria for definingmultipotent mesenchymal stromal cells. The International Society forCellular Therapy position statement. Cytotherapy 2006; 8: 315.

34. Lund RD, Wang S, Lu B, et al. Cells isolated from umbilical cord tissuerescue photoreceptors and visual functions in a rodent model of retinaldisease. Stem Cells 2007; 25: 602.

35. Ennis J, Gotherstrom C, Le Blanc K, et al. In vitro immunologic propertiesof human umbilical cord perivascular cells. Cytotherapy 2008; 10: 174.

36. Grinnemo KH, Mansson A, Dellgren G, et al. Xenoreactivityand engraftment of human mesenchymal stem cells transplantedinto infarcted rat myocardium. J Thorac Cardiovasc Surg 2004;127: 1293.

37. Matyszak MK. Inflammation in the CNS: Balance between immuno-logical privilege and immune responses. Prog Neurobiol 1998; 56: 19.

38. Pachter JS, de Vries HE, Fabry Z. The blood-brain barrier and its role inimmune privilege in the central nervous system. J Neuropathol ExpNeurol 2003; 62: 593.

39. Chen J, Sanberg PR, Li Y, et al. Intravenous administration of humanumbilical cord blood reduces behavioral deficits after stroke in rats.Stroke 2001; 32: 2682.

40. Krupinski J, Kaluza J, Kumar P, et al. Role of angiogenesis in patientswith cerebral ischemic stroke. Stroke 1994; 25: 1794.

41. Chen J, Li Y, Katakowski M, et al. Intravenous bone marrow stromalcell therapy reduces apoptosis and promotes endogenous cell prolifer-ation after stroke in female rat. J Neurosci Res 2003; 73: 778.

42. Chen J, Zhang ZG, Li Y, et al. Intravenous administration of humanbone marrow stromal cells induces angiogenesis in the ischemicboundary zone after stroke in rats. Circ Res 2003; 92: 692.

43. Zacharek A, Chen J, Cui X, et al. Angiopoietin1/Tie2 and VEGF/Flk1 induced by MSC treatment amplifies angiogenesis and vascularstabilization after stroke. J Cereb Blood Flow Metab 2007; 27: 1684.

44. Risau W. Mechanisms of angiogenesis. Nature 1997; 386: 671.45. Karahuseyinoglu S, Cinar O, Kilic E, et al. Biology of stem cells in

human umbilical cord stroma: In situ and in vitro surveys. Stem Cells2007; 25: 319.

46. da Silva Meirelles L, Caplan AI, Nardi NB. In search of the in vivoidentity of mesenchymal stem cells. Stem Cells 2008; 26: 2287.

47. Eliceiri BP, Cheresh DA. Adhesion events in angiogenesis. Curr OpinCell Biol 2001; 13: 563.

48. Hung SC, Pochampally RR, Hsu SC, et al. Short-term exposure ofmultipotent stromal cells to low oxygen increases their expressionof CX3CR1 and CXCR4 and their engraftment in vivo. PLoS ONE2007; 2: e416.

49. Potier E, Ferreira E, Andriamanalijaona R, et al. Hypoxia affects mes-enchymal stromal cell osteogenic differentiation and angiogenic factorexpression. Bone 2007; 40: 1078.

50. Hu X, Yu SP, Fraser JL, et al. Transplantation of hypoxia-precondi-tioned mesenchymal stem cells improves infarcted heart function viaenhanced survival of implanted cells and angiogenesis. J Thorac Car-diovasc Surg 2008; 135: 799.

Customer Service Contact Information

All correspondence concerning business matters, including subscription information, orders, or changes of address, should be directed to:

Lippincott Wilkins & Williams 16522 Hunters Green Parkway Hagerstown, MD 21740-2116 Tel: 800-638-3030 (North America); +44 (0) 20-7981-0525 (Europe); 1-301-223-2300 (RoW) Fax: 1-301-223-2400 Email: customerservice@Wolterskluwer.com

© 2009 Lippincott Williams & Wilkins 359Liao et al.