Vascular Trauma Induces Rapid but Transient Mobilizationof VEGFR21AC1331 Endothelial Precursor Cells

Muhammad Gill, Sergio Dias, Koichi Hattori, Mary Lee Rivera, Daniel Hicklin, Larry Witte,Leonard Girardi, Roger Yurt, Harvey Himel, Shahin Rafii

Abstract—Bone marrow (BM)–derived circulating endothelial precursor cells (CEPs) are thought to play a role in postnatalangiogenesis. Emerging evidence suggests that angiogenic stress of vascular trauma may induce mobilization of CEPsto the peripheral circulation. In this regard, we studied the kinetics of CEP mobilization in two groups of patients whoexperienced acute vascular insult secondary to burns or coronary artery bypass grafting (CABG). In both burn andCABG patients, there was a consistent, rapid increase in the number of CEPs, determined by their surface expressionpattern of vascular endothelial growth factor receptor 2 (VEGFR2), vascular endothelial cadherin (VE-cadherin), andAC133. Within the first 6 to 12 hours after injury, the percentage of CEPs in the peripheral blood of burn or CABGpatients increased almost 50-fold, returning to basal levels within 48 to 72 hours. Mobilized cells also formedlate-outgrowth endothelial colonies (CFU-ECs) in culture, indicating that a small, but significant, number of circulatingendothelial cells were BM-derived CEPs. In parallel to the mobilization of CEPs, there was also a rapid elevation ofVEGF plasma levels. Maximum VEGF levels were detected within 6 to 12 hours of vascular trauma and decreased tobaseline levels after 48 to 72 hours. Acute elevation of VEGF in the mice plasma resulted in a similar kinetics ofmobilization of VEGFR21 cells. On the basis of these results, we propose that vascular trauma may induce release ofchemokines, such as VEGF, that promotes rapid mobilization of CEPs to the peripheral circulation. Strategies toimprove the mobilization and incorporation of CEPs may contribute to the acceleration of vascularization of the injuredvascular tissue.(Circ Res. 2001;88:167-174.)

Key Words: circulating endothelial precursor cellsn mobilization n vascular endothelial growth factorn vascular endothelial growth factor receptor 2n AC133

Vascular trauma, induced by burn or by mechanicaldisruption, such as during surgical procedures, leads to

a cascade of events that result in the chemoattraction ofinflammatory cells and other cell types to the site of injury.1

Production and release of proangiogenic factors such asVEGF and basic fibroblast growth factor 2 (FGF-2) maypotentiate the recruitment of inflammatory cells includingmonocytes,2 as well as other nonhematopoietic cell typessuch as circulating endothelial precursor cells (CEPs),3,4 tothe site of injured vascular tissue, accelerating vascularhealing.

Vascular healing is a complex process and may requirerapid endothelialization to prevent fatal complications such asthrombosis or bleeding. Two possible sources of endotheli-alization are (1) migration and co-option of preexistingvascular wall endothelial cells (ECs) or (2) recruitment ofCEPs from the stem-cell reservoirs such as bone marrow(BM). CEPs may reflect the phenotype of embryonic angio-blasts, which are migratory ECs with the capacity to circulate,

proliferate, and differentiate into mature ECs but have neitheracquired characteristic markers of mature ECs nor formedlumina.

We have shown that allogeneic sex-mismatched BM trans-plantation resulted in the transfer of CEPs to recipient dogs.5

In humans, evidence for CEPs originates from patientsimplanted with a left ventricular assist device (LVAD), wherethe surface of the titanium housing of LVADs is colonizedwith BM-derived circulating CD341 EC-like cells.6 However,because BM contains both mature ECs as well as BM-derivedCEPs, these studies did not conclusively demonstrate theexistence of a phenotypically and functionally distinct popu-lation of CEPs. Moreover, discrimination between CEPs andmature ECs or hematopoietic cells is complicated by the factthat subsets of hematopoietic cells express markers similar tothose of ECs.

In search of a BM-derived, CEP-specific marker, we havefound that AC133, a novel hematopoietic stem-cell marker, isalso expressed on subsets of CD341 cells.7 AC1331 CEPs

Original received October 3, 2000; resubmission received November 16, 2000; revised resubmission received December 21, 2000; accepted December21, 2000.

From the Division of Hematology and Oncology (M.G., S.D., K.H., M.L.R., S.R.), Cornell University Medical College; W Randolph Burn Center(R.Y., H.H.) and Division of Cardiothoracic Surgery (L.G.), New York Presbyterian Hospital; and Imclone Systems Incorporated (D.H., L.W.), NewYork, NY.

Correspondence to Shahin Rafii, MD, Division of Hematology/Oncology, Cornell University Medical College, 1300 York Ave, Room C-606, NewYork, NY 10021. E-mail srafii@mail.med.cornell.edu

© 2001 American Heart Association, Inc.

Circulation Researchis available at http://www.circresaha.org

167

by guest on July 14, 2018http://circres.ahajournals.org/

Dow

nloaded from

express vascular endothelial growth factor receptor 2(VEGFR2) and are dependent on VEGF for survival and invitro expansion. AC1331VEGFR21 CEPs isolated from cordblood, BM, or fetal liver proliferate in vitro in an anchorage-independent manner and can be induced to undergo migrationor differentiate into mature adherent AC1332 mature ECs.7

Several studies have capitalized on the in vitro growthcapacity and differentiation profile of CEPs to distinguishbetween mature fallout vessel-derived endothelium and BM-derived CEPs. Using fluorescent in situ hybridization analysisto detect the Y chromosome in blood samples from BMtransplant recipients who received gender-mismatched stemcells, Lin et al8 could distinguish between vessel wall andBM-derived endothelial cells. They show that in vitro–derived ECs form adherent colonies during early phases ofculture (9 days after plating) and undergo 6-fold expansionand are derived predominantly from the recipient vessel wall,whereas ECs derived from late-outgrowth endothelial colo-nies (CFU-ECs, 30 days after culture) undergo 98-foldexpansion and mostly generate from transplanted donor BMcells. In this regard, the potential of forming late-outgrowthcolonies and expression of AC133 provide for surrogatemarkers to identify and quantify the numbers of humanBM-derived CEPs.

Because rapid endothelialization of denuded injured ves-sels is essential to avoid fatal complications, we hypothesizedthat elevation of plasma VEGF levels immediately aftervascular trauma may promote CEP mobilization into theperipheral circulation. In the present study, we demonstratethat in burn patients and those who undergo coronary arterybypass grafting (CABG), there is a rapid elevation of VEGFlevels followed by immediate mobilization, within 6 to 12hours of vascular trauma, of VEGFR21AC1331 cells into theperipheral circulation. Given that a significant number ofVEGFR21 cells have the capacity to form late-outgrowthcolonies and also express the stem-cell marker AC133, theycan be considered as BM-derived CEPs. These data suggestthat CEPs may act as first-aid angiogenic modules rapidlymobilized to enhance the endothelialization of denuded pro-thrombotic injured vasculature. Isolation and ex vivo expan-sion of these cells may facilitate the development of strategiesto augment and accelerate vascular healing in patients withprofound vascular insufficiencies.

Materials and MethodsCollection of Peripheral Blood Samples andIsolation of Mononuclear CellsPermission to perform these studies was approved by the investiga-tional review board at Cornell University Medical College. Sampleswere taken from patients who had.15% burn injury (n58).Peripheral blood (4 to 5 mL) was taken at 12, 24, 48, and 72 hoursand on the 7th and 14th days. Blood (4 to 5 mL) was also taken frompatients (n57) who underwent CABG. Samples were taken preop-eratively at 6, 12, 24, 48, and 72 hours and on the 4th day.

The blood samples from burn and CABG patients were collectedin heparinized tubes. Blood (5 mL) was diluted with HBSS (LifeTechnologies) and overlaid on top of a Ficoll preparation (AccuePrep Lymphocytes, Accurate Chemical & Scientific Corp). Thebuffy coat, containing the peripheral blood mononuclear cells(PBMNCs), was isolated and washed twice in HBSS at 1800gfor 10minutes. Finally, the PBMNCs were resuspended in serum-free

medium (X-vivo 20, BioWhittaker). These cells were studied for theexpression of endothelial-specific markers and formation of late-outgrowth endothelial colonies.

Flow Cytometry StudiesPBMNCs, cultured in serum-free medium X-vivo 20, were incubatedwith 1.5 mL FITC (green fluorescence) labeled high-affinity non-neutralizing monoclonal antibody (MoAb) to VEGFR2 (clone 6.64,Imclone Systems Inc) and 1.5mL phycoerythrin (PE; red fluores-cence) labeled anti-CD15 antibody (Sigma), for 20 minutes at 4°C.Nonviable cells were excluded by propidium iodide (PI; 30mmol/L,Becton Dickinson). The number of positive cells was compared withIgG isotype control (FITC, PE; Immunotech) and quantified using aCoulter Elite flow cytometer (Coulter). Other endothelial markersstudied were vascular endothelial cadherin (VE-cadherin) (BV9clone, Imclone Systems Inc) and AC133 (Miltenyi Biotech). Foranalysis in burn patients, samples taken from normal age andgender-matched subjects were used as control, and preoperativesamples were used as control in the CABG patients.

Reverse Transcriptase–Polymerase Chain Reaction(RT-PCR) Analysis of PBMNCsTotal RNA from PBMNCs of burn and CABG patients at differenttime points was obtained using Tri-Zol (Life Technologies), accord-ing to the manufacturer’s instructions. The first-strand cDNA wassynthesized using the T-Primed First Strand Kit (Amersham Phar-macia Biotech Inc), amplified by Taq DNA polymerase (AdvantagecDNA Polymerase Mix, Clontech) in a 25-mL reaction mixture,including 10 mmol/L dNTP (MBI Fermentas Inc) and 10mmol/L ofeach primer (Genelink). PCR was performed using a PCR thermalcycler (MWG Biotech). The PCR program used to amplifyVEGFR2, VE-cadherin, andb-actin had the following parameters: 5minutes initial denaturation at 94°C, annealing at 60°C for 1 minute,and 30 seconds of elongation at 72°C. This initial cycle was followedby 35 cycles of denaturation at 94°C for 45 seconds, annealing at60°C for 45 seconds, and 2 minutes of elongation at 72°C, followedby 7 minutes of extension at 72°C. Subsequently, PCR products werevisualized in 1.5% ethidium bromide–stained agarose gels. Humanumbilical vein endothelial cell (HUVEC) cDNA was used as apositive control. RNA from normal donors for the burn patients andpreoperative samples from the CABG patients were used as a controlfor baseline mRNA expression.

Primersb-actin: b-actin primer sequence (513 bp): sense: TCA TGT TTG

AGA CCT TCA A; antisense: GTC TTT GCG GAT GTC CAC GVEGFR2: VEGFR2 primer sequence (790 bp): sense: CTG GCA

TGG TCT TCT GTG AAG CA; antisense: AAT ACC AGT GGATGT GAT GCG G

VE-cadherin: VE-cadherin primer sequence (596 bp): sense: ACGGGA TGA CCA AGT ACA GC; antisense: ACA CAC TTT GGGCTG GTA GG

Late-Outgrowth Endothelial Colony AssayTo evaluate the potential of mobilized PBMNCs as CEPs, we used aprotocol developed by Lin et al,8 with minor modifications. Freshlyisolated PBMNCs, obtained 12 hours after trauma, were cultured inendothelial growth medium (M199, Gibco BRL) supplemented with20% FBS (HyClone), endothelial cell growth factor (30mg/mL),FGF (5 ng/mL, human recombinant basic FGF [Sigma]), heparin (5U/mL), penicillin (100 U/mL), streptomycin (100mg/mL), andfungizone (0.25mg/mL). These cells were placed on 12-well platesand coated with 0.2% gelatin. The plates were incubated at 37°C ina humidified environment with 5% CO2. This process resulted inattachment of cells consisting mostly of monocytes or matureendothelial colonies on the well plates. Nonadherent cells weretransferred after 4 to 5 days to other wells coated with 0.2% gelatinand grown in endothelial growth medium. Each week, for up to 2weeks, the nonadherent cells were transferred to fresh plates toassess for the formation of late-outgrowth colonies.

168 Circulation Research February 2, 2001

by guest on July 14, 2018http://circres.ahajournals.org/

Dow

nloaded from

Endothelial colonies were identified by metabolic uptake ofDiI-acetylated-LDL (DiI-Ac-LDL) by incubating the cells with 1mg/mL of human DiI-Ac-LDL (PerImmune Inc) at 37°C for 5 hoursand visualizing by fluorescence microscopy. For von Willebrandfactor (vWF) staining, cells were rinsed with HBSS and fixedimmediately with 4% paraformaldehyde for 10 minutes. A primarymouse antibody against human vWF (Dako) was used at 1:250dilution. Biotinylated rat anti-mouse IgG (Vector Laboratories Inc)was used at 1:200 dilution. Streptavidin linked to FITC (JacksonImmunoResearch Laboratories Inc) was used at a dilution of 1:200.Cells were visualized by fluorescence microscopy. HUVECs wereused as positive control for DiI-Ac-LDL uptake and vWF staining.Adherent colonies incorporating DiI-Ac-LDL and coexpressingvWF, representing early outgrowths, were quantified every week.DiI-Ac-LDL 1vWF1 colonies derived from passaged nonadherentcells, which formed between 6 to 8 weeks, were considered aslate-outgrowth colonies (CFU-ECs). Samples from the healthydonors were used as control in the burn patients and preoperativesamples were used as control in the CABG patients.

VEGF Plasma DeterminationQuantitative determination of VEGF in the plasma of patients orAdVEGF165-injected mice was determined by a specific human ormurine VEGF ELISA using the Quantikine VEGF-ELISA kit (R&DSystems Inc), according to the manufacturer’s instructions. Periph-eral blood plasma samples were obtained at different time pointsfrom the burn and CABG patients. Plasma samples from healthydonors were used to determine the baseline VEGF levels for the burnpatients, whereas in CABG patients, this was determined by analyz-ing the preoperative plasma levels of the patients. Plasma of miceinjected with AdNull was used as control. Both assays have asensitivity limit of 7 pg/mL. Samples were analyzed in triplicate.

In Vivo Adenoviral Delivery of VEGFImmunocompetent mice (on BALB/C background), aged 8 weeks(.20 g) and sex-matched, were purchased from the Jackson Labo-ratory (Bar Harber, Maine) and maintained in germ-free conditions.Mice (n56) received AdVEGF165 (1.53108 pfu) or AdNull (1.53108

pfu) in a volume of 100mL by single intravenous administration onday 0. AdVEGF165 is an Ad5-derived, E1a-, E3-deficient (E1a–E3–

E41) vector with an expression cassette in the E1a region containingthe mouse AdVEGF165 cDNA and driven by the cytomegalovirusmajor immediate/early promoter/enhancer. The control vector, Ad-Null, is similar in design, except that it contains no transgene in theexpression cassette.

Flow Cytometry Studies on Mouse PBMNCsCells were costained with 1.5mL of FITC-labeled high-affinityMoAb to VEGFR2 (clone DC101, Imclone Systems Inc) and 1mLof PE (red fluorescence) labeled anti-CD11b antibody (Pharmingen)for 20 minutes at 4°C. The number of positive cells was comparedwith IgG isotype control (FITC, PE; Pharmingen). PI (30mmol/L)was used to exclude the nonviable cells. The cells were analyzed byusing the Coulter Elite flow cytometer.

Statistical AnalysisA standard Student’st test was used to determine statisticaldifferences.

ResultsVEGF Plasma Levels Increase AfterVascular TraumaVEGF plasma levels from both burn and CABG patientsincreased immediately after vascular trauma, which peakedbetween 6 to 12 hours after injury, and then decreasedgradually over a 2-day period (Figures 1A and 1B). VEGFlevels were lower after 3 to 4 days of injury. These resultssuggest that vascular trauma induces a rapid but transientincrease in plasma VEGF levels after trauma.

Vascular Trauma Induces Mobilization ofPBMNCs Expressing mRNA for VEGFR2and VE-CadherinWe investigated whether the rapid increase in VEGF levelscorrelated with an increase in the number of circulatingVEGFR21 PBMNCs. Peripheral blood samples from burnand CABG patients demonstrated VEGFR2 and VE-cadherinexpression at the mRNA level, as determined by RT-PCR(Figures 2A and 2B). VEGFR2 and VE-cadherin mRNAwere present in the PBMNCs between 6 and 12 hours andthen gradually decreased over time. Control samples for burnand CABG patients had virtually undetectable mRNA levelsfor any of these markers.

Characterization of Mobilized CEPsThe RT-PCR analysis of PBMNCs from burn and CABGpatients suggested that after vascular trauma, a population ofcells expressing VEGFR2 and VE-cadherin may be rapidlymobilized into the peripheral circulation. These results weresubsequently confirmed by analysis of surface expressionusing flow cytometry (FACS).

Given that monocytes express Fc receptors and may give riseto false-positive results, the PBMNCs were also costained withCD15 to exclude mobilized cells of myeloid-monocytic lineage.

Figure 1. VEGF levels are rapidly increased in burn and CABGpatients. Plasma VEGF levels in burn (A) and CABG (B) patientswere determined by ELISA. There was a rapid elevation (6 to 12hours) of VEGF levels after trauma, followed by a gradualdecrease over time (n55 for burn, n54 for CABG). Resultsshown are mean6SD for plasma VEGF levels, in both burn andCABG patient samples.

Gill et al CEP Mobilization in Vascular Trauma Patients 169

by guest on July 14, 2018http://circres.ahajournals.org/

Dow

nloaded from

Confirming the RT-PCR results, FACS analysis of PBMNCsobtained from burn and CABG patients showed a rapid increasein circulating VEGFR21 cells after 6 to 12 hours (Figures 3 and4). These cells were negative for the myeloid marker CD15 (seeFigures 3A and 3B for a typical FACS profile). Samples fromburn patients showed an average of 20 65564077VEGFR21CD152 cells (5.161.04% of the total PBMNCs)mobilized 12 hours after burn injury (Figure 4A) and decreasingafter 72 hours. On the other hand, healthy donors had0.10%60.01 (5356127 cells) circulating VEGFR21CD152

cells (P,0.05). Similarly, in CABG patients, although thepreoperative samples showed 9156107 (0.0260.02% ofthe total PBMNCs) VEGFR21CD152 cells, there was anaverage of 34 843613 515 (8.463.2% of the totalPBMNCs) VEGFR21CD152 cells mobilized 6 hours aftersurgery (Figure 4B). In these patients, at 12 and 24 hoursafter surgery, there was a sustained elevation ofVEGFR21CD152 cells with an average of 17 80065823(4.561.45% of the total PBMNCs) and 11 46764097 (2.861%of the total PBMNCs) cells (Figure 4B), respectively. Thenumber of VEGFR21CD152 cells gradually decreased over a 3-to 4-day period, in both groups of patients. The last peripheralblood sample from burn patients was taken 14 days after injuryand had 6646257 (0.2060.03% of the total PBMNCs)VEGFR21CD152 cells (Figure 4A). Similarly, the last samplefrom CABG patients was taken on the fourth day after surgeryand showed 13296408 (0.360.06% of the total PBMNCs)VEGFR21CD152 cells (Figure 4B). These results demonstratethe presence of a small, but distinct, population ofVEGFR21CD152 cells within the PBMNC population of vas-cular trauma patients, which rapidly gets mobilized within 6 to

Figure 2. Expression of VEGFR2 and VE-cadherin mRNA inmobilized PBMNCs by RT-PCR. Representative sample from aburn (A) and CABG (B) patient, showing VEGFR2 (KDR) andVE-cadherin mRNA expression as determined by RT-PCR analy-sis of the PBMNCs.

Figure 3. Quantification of the mobilized VEGFR21 andVE-cadherin1 cells by flow cytometry. Representative two-colorflow cytometry analysis of mobilized PBMNCs, quantifying thenumber of VEGFR21 and VE-cadherin1 cells 12 hours after burninjury (A) and 6 hours after CABG (B), demonstrating the numberof VEGFR21 (5.3% in burn, 11.1% in CABG) and VE-cadherin1

(4.7% in burn, 8.7% in CABG) cells.

Figure 4. Time course of the total number of mobilizedVEGFR21CD152 cells in the PBMNCs of burn and CABGpatients. The total number of mobilized VEGFR21CD152 cellsover time in burn (n58) (A) and CABG (n57) (B) patients wasdetermined by two-color flow cytometry in 1 mL of peripheralblood. Nonviable cells were excluded by PI staining.

170 Circulation Research February 2, 2001

by guest on July 14, 2018http://circres.ahajournals.org/

Dow

nloaded from

12 hours after injury and then decreases gradually over a periodof days.

Phenotypic Analysis of the VEGFR21CD152 CellsThe peripheral blood samples from burn and CABG patientswere further analyzed for the presence of other hematopoieticand endothelial markers by flow cytometry. The 12-hour timepoint was chosen, because it consistently showed, in bothgroups of patients, the highest number of mobilized CEPs.Twelve hours after injury, in both burn and CABG patients,there was a maximum increase in cells expressing VE-cadherin and AC133 (Table), which accompanied the maxi-mum increase in VEGFR2 expression. There was a meanincrease of 15 98764555 (5.161.04% of the total PBMNCs)and 15 54463390 (4.561.45% of the total PBMNCs) VE-cadherin1 cells at 12 hours, in burn and CABG patients,respectively (Table). Similarly, the increase in AC1331 cellsin burn and CABG patients was 876962019 (2.260.5% ofthe total PBMNCs) and 834262109 (2.160.51% of the totalPBMNCs), respectively (Table). Furthermore, as shown forVEGFR2, expression of these markers in the PBMNC pop-ulation gradually decreased over a 2- to 3-day period (data notshown). These results suggest that the cell population mobi-lized after vascular trauma consists of VEGFR21 cells ex-pressing VE-cadherin and AC133.

BM-Derived CEPs Are Present in a SignificantNumber of Mobilized PBMNCsEmerging data suggest that late-outgrowth endothelial cellsdifferentiating from the nonadherent population of platedPBMNCs represent a population of BM-derived anchorage-independent BM-derived CEPs characterized by high prolif-erative potential.

To assess the number of mobilized PBMNCs with thecapacity of generating late-outgrowth colonies, PBMNCs mobi-

lized from burn and CABG patients 12 hours after trauma weregrown in vitro in conditions defined to support the formation oflate-outgrowth colonies.8 Typical endothelial colonies, withcobblestone morphology, appeared after 7 to 10 days of culture(Figures 5A and 5B) and showed both the metabolic incorpora-tion of DiI-Ac-LDL as well as vWF staining (Figure 5C). Thesewere the early-outgrowth endothelial colonies, which are pre-sumably derived form mature fallout circulating ECs. Fewendothelial colonies formed between 10 and 14 days of culture(Figures 5A and 5B) and were positive for both DiI-Ac-LDL andvWF (Figure 5C). These were also scored as the early out-growths. On day 15, the nonadherent population was passed fora third time. The colonies formed between 6 and 8 weeks(Figures 5A and 5B) from this passaged nonadherent populationwere referred to as the late-outgrowth endothelial colonies(CFU-ECs). These late-outgrowth colonies were quantifiedbased on their capacity to metabolically incorporate DiI-Ac-LDL and stain positively with vWF (Figure 5C). PBMNCsderived from control samples were able to form the earlyoutgrowths but not the late outgrowths. We and others7,9 haveshown that purified populations of AC133 have the capacity toform late-outgrowth colonies with a plating efficiency of 5%. Onthe basis of these results, we estimated that close to 12% of themobilized AC133 cells were capable of forming late CFU-ECs.Collectively, these data show that a small, but significant,number of the mobilized circulating endothelial cells from burnand CABG patients, 12 hours after trauma, were BM-derivedCEPs.

Elevation of VEGF Plasma Levels in MiceResulted in Increased Number of RapidlyMobilized VEGFR21CD11b2 PBMNCsIntravenous administration of Ad vectors expressing solubleVEGF165 (AdVEGF165) (1.53108 pfu) induced a rapid (24hour) mobilization of VEGFR21CD11b2 (14.760.57%)

Quantification of the Number of Circulating Endothelial Cells in Burn and CABG Patients

PatientSamples

VEGFR2(KDR) VE-Cadherin AC133 CD15

AC133/KDRRatio, %

Burn Patients 1 14 760 (3.6)* 13 918 (3.5) 6314 (1.6) 1271 (0.3) 44.4

2 19 044 (4.6) 12 006 (3.0) 7164 (1.8) 1233 (0.3) 39.1

3 18 104 (4.5) 16 450 (4.1) 7642 (1.9) 1608 (0.4) 42.2

4 24 962 (6.2) 23 208 (5.8) 10 782 (2.7) 2508 (0.6) 43.5

5 26 412 (6.6) 21 182 (5.3) 11 652 (2.9) 3572 (0.9) 43.9

6 23 648 (5.9) 15 620 (3.9) 10 080 (2.5) 2892 (0.7) 42.3

7 17 236 (4.3) 16 268 (4.1) 6836 (1.7) 1198 (0.3) 39.5

8 21 076 (5.2) 9240 (4.6) 9682 (2.4) 3301 (0.9) 46.1

CABG Patients 1 20 382 (5.1) 10 402 (2.6) 9160 (2.3) 1968 (0.4) 45.1

2 9484 (2.4) 14 228 (3.6) 5962 (1.5) 2480 (0.6) 62.5

3 18 076 (4.5) 13 058 (3.2) 7296 (1.8) 2096 (0.5) 40.0

4 12 452 (3.1) 18 048 (4.5) 8120 (2.1) 1280 (0.3) 67.7

5 15 118 (3.8) 16 400 (4.1) 5988 (1.5) 1574 (0.3) 39.4

6 23 916 (6) 15 886 (4.0) 10 600 (2.6) 2486 (0.6) 43.3

7 25 174 (6.3) 20 788 (5.2) 11 268 (2.8) 2918 (0.7) 44.4

*Peripheral blood obtained from burn and CABG patients at 12 hours after trauma. Values shown represent theabsolute number of cells in 1 mL of peripheral blood. Numbers in parentheses represent the correspondingpercentages in the total PBMNCs in 1 mL of peripheral blood.

Gill et al CEP Mobilization in Vascular Trauma Patients 171

by guest on July 14, 2018http://circres.ahajournals.org/

Dow

nloaded from

PBMNCs (Figures 6A and 6B). CD11b/CD18 (Mac-1) isprimarily expressed on the myeloid-monocytic cells. InAdVEGF165-inoculated mice, there was a maximum increasein VEGF plasma levels at 24 hours (Figure 7), whichcorrelated with the number of rapidly mobilizedVEGFR21CD11b2 cells. VEGF levels gradually decreasedover the next 2 weeks (Figure 7). AdNull vector (no trans-gene), which was used as a control, did not induce significantmobilization of VEGFR21CD11b2 (1.960.28%) cells (Fig-ures 6A and 6B). Circulating VEGFR21CD11b2 cells werestill abundant (2.160.28%) in the AdVEGF165-injected miceafter 7 days compared with control (AdNull-treated) mice(0.660.14%) (Figure 6B). Doses above 1.53108 pfu weretoxic and resulted in the demise of the treated mice (data notshown). There were no signs of toxicity in mice treated withAdNull vectors at doses of up to 109 pfu. These data suggestthat elevation of VEGF plasma levels induce a rapid mobili-zation of VEGFR21CD11b2 cells in mice within a similartime frame to that seen for the mobilization ofVEGFR21CD152 cells in burn and CABG patients. PlasmaVEGF levels remained low in AdNull-injected mice (Figure7). These results suggest that, similar to burn or CABGpatients, an acute but transient increase in mouse plasmaVEGF levels promotes a rapid mobilization ofVEGFR21CD11b2 cells into the peripheral circulation.

DiscussionPostnatal neovascularization is the process of new bloodvessel formation from preexisting endothelial cells.10 How-ever, the origin of endothelial cells incorporated into thenewly formed blood vessels has been the subject of intensivescrutiny. Emerging data have suggested that a subpopulationof BM-derived CEPs may contribute to new blood vesselformation.5,11–13 It has been shown that CEPs have thecapacity of being recruited into ischemic tissues or growing

tumors.4,13–15 In the present study, we have extended theseobservations by demonstrating that vascular trauma induces avery rapid but transient mobilization of a significant numberof BM-derived VEGFR21AC1331 cells.

Given the rapid entry of the CEPs into the peripheralcirculation, it has been suggested that the chemocytokinesreleased as a result of the vascular injury may inducemobilization of CEPs. Among the known chemocytokines,VEGF has been shown to be effective in mobilizing CEPsinto the peripheral circulation.3 Indeed, patients undergoingCABG or those who have suffered extensive burns, haveelevated VEGF plasma levels, peaking 6 to 12 hours aftervascular insult and gradually decreasing to baseline after 3 to4 days. Elevated VEGF levels have also been measured inburn patients in a previous study.16 However, it remainedunclear whether this increase in circulating VEGF levels hadany physiological significance. In this report, we demonstratethat a rise in circulating VEGF levels is accompanied by anincrease in PBMNCs comprised of a significant number ofBM-derived CEPs. There was a 50-fold increase, comparedwith preoperative level, in the number of circulating cells 6 to12 hours after surgery in CABG patients, and a similar resultwas seen in those patients suffering from extensive burns.Similarly, we demonstrate that acute elevation of plasmaVEGF in vivo in mice leads to increased circulation ofVEGFR21 cells. Mice inoculated with adenoviral vectorsencoding the VEGF165 transgene showed a rapid but transientincrease in circulating VEGFR21CD11b2 PBMNCs. Thesedata suggest that, similar to the murine VEGF-inducedmobilization of CEPs, acute elevation of VEGF levels invascular trauma patients may be the primary factor inducingmobilization of CEPs. However, because vascular traumaalso promotes the release of numerous known or as-yetunrecognized chemocytokines, other factors in addition toVEGF may also contribute to the mobilization of CEPs.

Figure 5. Early and late endothelial out-growth colony formation by the circulat-ing mononuclear cells from burn (A) andCABG (B) patients. PBMNCs (105 cells)obtained from the peripheral blood werecultured on gelatin-coated plates in thepresence of endothelial growth medium.Adherent proliferating colonies thatformed in the first 2 weeks and late colo-nies that formed in the passaged cellsduring 6 to 8 weeks of initial culture werequantified for the metabolic incorporationof DiI-Ac-LDL (red) as well as vWF(green) staining (C). DiI-Ac-LDL1vWF1-adherent colonies that formed from pas-saged nonadherent cells, between 6 to 8weeks of initial culture, were scored aslate-outgrowth colonies (CFU-ECs),whereas those that formed from initialplating, within the first 3 weeks, wereconsidered as early outgrowth colonies.

172 Circulation Research February 2, 2001

by guest on July 14, 2018http://circres.ahajournals.org/

Dow

nloaded from

Vascular trauma may also result in the nonspecific intro-duction of mature vascular wall–derived ECs into the circu-lation. We have used the expression pattern of AC133 and thepotential of late-outgrowth CFU-EC colony formation toestimate the number of BM-derived CEPs. We demonstratethat, on average, close to 40% of the mobilized VEGFR21

cells express AC133. In addition, we estimated that close to12% of mobilized AC1331 cells form late-outgrowth CFU-ECs. Collectively, these data suggest that a small, butsignificant, number of the mobilized PBMNCs are BM-derived CEPs. It is intriguing that only the mobilizedPBMNCs from earlier phases after the trauma were AC1331

or formed late-outgrowth colonies, suggesting that the rapidincrease in circulating BM-derived CEPs is the result of achemokinetic process, rather than differentiation or matura-tion of a population of stem cells. Furthermore, the rapidincrease in PBMNCs with endothelial precursor potentialsuggests that the number of CEPs in the adult BM may begreater than expected. Taken together, these results suggestthat the rapid mobilization of CEPs, in response to a rise incirculating VEGF levels after trauma, may contribute to therevascularization of the injured tissues.

Strategies to enhance mobilization and incorporation of CEPsinto impaired vascular beds may have important therapeuticimplications. However, it remains to be demonstrated to whatextent these mobilized CEPs contribute to the compromisedvascular tissue and whether acceleration of their mobilizationand homing may promote the vascular healing process. Never-theless, as shown in the present study, therapies aimed atblocking mobilization of CEPs and other BM-derived cells intothe peripheral circulation and subsequently into neovascularizedtissues may have undesirable clinical complications. Finally,VEGF-induced mobilization of VEGFR21AC1331 cells mayfacilitate the isolation of an enriched population of CEPs fromthe peripheral circulation, which may ultimately be used for genetherapy or ex vivo expansion.

AcknowledgmentsS. Rafii is supported by National Heart, Lung, and Blood Institute(NHLBI) Grants R01 HL-58707, R01 HL-61849, P01-HL 66592(project 2), an American Heart Association Grant-in-Aid, the Dor-othy Rodbell Foundation for Sarcoma Research, and the RichFoundation.

References1. Nissen NN, Polverini PJ, Koch AE, Volin MV, Gamelli RL, Di Pietro LA.

Vascular endothelial growth factor mediates angiogenic activity during theproliferative phase of wound healing.Am J Pathol. 1998;152:1445–1452.

2. Barleon B, Sozzani S, Zhou D, Weich HA, Mantovani A, Marme D.Migration of human monocytes in response to vascular endothelialgrowth factor (VEGF) is mediated via the VEGF receptor flt-1.Blood.1996;87:3336–3343.

3. Asahara T, Takahashi T, Masuda H, Kalka C, Chen D, Iwaguro H, InaiY, Silver M, Isner JM. VEGF contributes to postnatal neovascularizationby mobilizing bone marrow-derived endothelial progenitor cells.EMBOJ. 1999;18:3964–3972.

4. Takahashi T, Kalka C, Masuda H, Chen D, Silver M, Kearney M, MagnerM, Isner JM, Asahara T. Ischemia- and cytokine-induced mobilization ofbone marrow-derived endothelial progenitor cells for neovascularization.Nat Med. 1999;5:434–438.

5. Shi Q, Rafii S, Wu MH, Wijelath ES, Yu C, Ishida A, Fujita Y, KothariS, Mohle R, Sauvage LR, Moore MA, Storb RF, Hammond WP.Evidence for circulating bone marrow-derived endothelial cells.Blood.1998;92:362–367.

6. Rafii S, Oz MC, Seldomridge JA, Ferris B, Asch AS, Nachman RL,Shapiro F, Rose EA, Levin HR. Characterization of hematopoietic cellsarising on the textured surface of left ventricular assist devices.AnnThorac Surg. 1995;60:1627–1632.

7. Peichev M, Naiyer AJ, Pereira D, Zhu Z, Lane WJ, Williams M, Oz MC,Hicklin DJ, Witte L, Moore MA, Rafii S. Expression of VEGFR-2 and

Figure 6. Acute VEGF elevation promotes rapid but transientmobilization of VEGFR21 cells. A, Two-color flow cytometryshowing the number of VEGFR21CD11b2 cells mobilized 24hours after AdNull or AdVEGF165 intravenous injections, demon-strating that 1.7% and 14.7% of VEGFR21CD11b2 cells weremobilized, respectively. B, Time course of the mobilizedVEGFR21CD11b2 cells in the AdNull- and AdVEGF165-injectedmice (n53). The number of VEGFR21CD11b2 cells was deter-mined by flow cytometry. Nonviable cells were excluded by PIstaining.

Figure 7. Rapid elevation of VEGF in mice by injection ofAdVEGF165. Plasma VEGF levels in the mice injected withAdVEGF165 and AdNull vectors were determined by ELISA.There was a rapid elevation (24 hours) of VEGF levels afterAdVEGF165 injection, followed by a gradual decrease over time(n53). Results are the mean6SD for plasma VEGF levels.

Gill et al CEP Mobilization in Vascular Trauma Patients 173

by guest on July 14, 2018http://circres.ahajournals.org/

Dow

nloaded from

AC133 by circulating human CD341 cells identifies a population offunctional endothelial precursors.Blood. 2000;95:952–958.

8. Lin Y, Weisdorf DJ, Solovey A, Hebbel RP. Origins of circulating endo-thelial cells and endothelial outgrowth from.J Clin Invest. 2000;105:71–77.

9. Gehling UM, Ergun S, Schumacher U, Wagener C, Pantel K, Otte M,Schuch G, Schafhausen P, Mende T, Kilic N, Kluge K, Schafer B,Hossfeld DK, Fiedler W. In vitro differentiation of endothelial cells fromAC133-positive.Blood. 2000;95:3106–3112.

10. Folkman J, Shing Y. Angiogenesis.J Biol Chem. 1992;267:10931–10934.11. Shi Q, Wu MH, Onuki Y, Ghali R, Hunter GC, Johansen KH, Sauvage

LR. Endothelium on the flow surface of human aortic Dacron vasculargrafts.J Vasc Surg. 1997;25:736–742.

12. Asahara T, Murohara T, Sullivan A, Silver M, van der Zee R, Li T,Witzenbichler B, Schatteman G, Isner JM. Isolation of putative progenitorendothelial cells for angiogenesis.Science. 1997;275:964–967.

13. Asahara T, Masuda H, Takahashi T, Kalka C, Pastore C, Silver M,Kearne M, Magner M, Isner JM. Bone marrow origin of endothelialprogenitor cells responsible for postnatal vasculogenesis in physiologicaland pathological neovascularization.Circ Res. 1999;85:221–228.

14. Vale PR, Losordo DW, Milliken CE, Maysky M, Esakof DD, Symes JF,Isner JM. Left ventricular electromechanical mapping to assess efficacyof phVEGF165 gene transfer for therapeutic angiogenesis in chronic myo-cardial ischemia.Circulation. 2000;102:965–974.

15. Kalka C, Masuda H, Takahashi T, Gordon R, Tepper O, Gravereaux E,Pieczek A, Iwaguro H, Hayashi SI, Isner JM, Asahara T. Vascularendothelial growth factor165 gene transfer augments circulating endo-thelial progenitor cells in human subjects.Circ Res. 2000;86:1198–1202.

16. Grad S, Ertel W, Keel M, Infanger M, Vonderschmitt DJ, Maly FE.Strongly enhanced serum levels of vascular endothelial growth factor.Clin Chem Lab Med. 1998;36:379–383.

174 Circulation Research February 2, 2001

by guest on July 14, 2018http://circres.ahajournals.org/

Dow

nloaded from

Girardi, Roger Yurt, Harvey Himel and Shahin RafiiMuhammad Gill, Sergio Dias, Koichi Hattori, Mary Lee Rivera, Daniel Hicklin, Larry Witte, Leonard

Precursor Cells Endothelial+AC133+Vascular Trauma Induces Rapid but Transient Mobilization of VEGFR2

Print ISSN: 0009-7330. Online ISSN: 1524-4571 Copyright © 2001 American Heart Association, Inc. All rights reserved.is published by the American Heart Association, 7272 Greenville Avenue, Dallas, TX 75231Circulation Research

doi: 10.1161/01.RES.88.2.1672001;88:167-174Circ Res. 

http://circres.ahajournals.org/content/88/2/167World Wide Web at:

The online version of this article, along with updated information and services, is located on the

  http://circres.ahajournals.org//subscriptions/

is online at: Circulation Research Information about subscribing to Subscriptions: 

http://www.lww.com/reprints Information about reprints can be found online at: Reprints:

  document. Permissions and Rights Question and Answer available in the

Request Permissions in the middle column of the Web page under Services. Further information about this process isOffice. Once the online version of the published article for which permission is being requested is located, click

can be obtained via RightsLink, a service of the Copyright Clearance Center, not the EditorialCirculation Research Requests for permissions to reproduce figures, tables, or portions of articles originally published inPermissions:

by guest on July 14, 2018http://circres.ahajournals.org/

Dow

nloaded from