pleiotropic role of vegf-a in regulating fetal pulmonary ... · pleiotropic role of vegf-a in...

Pleiotropic role of VEGF-A in regulating fetal pulmonary mesenchymalcell turnover

S. Majka,1,2 K. Fox,1,3,4 B. McGuire,5 J. Crossno, Jr.,1,3,4 P. McGuire,5 and A. Izzo6

1Department of Medicine, Cardiovascular Pulmonary Research Section, Divisions of 2Cardiologyand 3Pulmonary Sciences and Critical Care Medicine, and 4Denver Veterans Administration MedicalCenter, University of Colorado Health Sciences Center, Denver, Colorado; 5Department of Cell Biologyand Physiology, University of New Mexico School of Medicine, Albuquerque, New Mexico; and 6Departmentof Microbiology, Immunology & Pathology, Colorado State University, Fort Collins, Colorado

Submitted 19 April 2005; accepted in final form 9 January 2006

Majka, S., K. Fox, B. McGuire, J. Crossno Jr., P. McGuire,and A. Izzo. Pleiotropic role of VEGF-A in regulating fetalpulmonary mesenchymal cell turnover. Am J Physiol Lung CellMol Physiol 290: L1183–L1192, 2006. First published January 20,2006; doi:10.1152/ajplung.00175.2005.—Tight regulation ofVEGF-A production and signaling is important for the mainte-nance of lung development and homeostasis. VEGF null mice haveprovided little insight into the role of VEGF during the later stagesof lung morphogenesis. Therefore, we examined the in vitro effectsof autocrine and paracrine VEGF-A production and the inhibitionof VEGF-A signaling on a Flk-1-negative subset of fetal pulmo-nary mesenchymal cells (pMC). We hypothesized that VEGF-Areceptor signaling regulates turnover of fetal lung mesenchyme ina cell cycle-dependent manner. VEGF receptor blockade withSU-5416 caused cell spreading and decreased proliferation andbcl-2 localization. Nuclear expression of the cell cycle inhibitoryprotein, p21, was increased with SU-5416 treatment, and p27 wasabsent. Autocrine VEGF production by pMC resulted in prolifer-ation and p21/p27-dependent contact inhibition. In contrast, exog-enous VEGF-A increased cell progression through the cell cycle.Selective activation of Flt by placental growth factor demonstratedthe importance of this receptor/kinase in the VEGF-A responsive-ness of pMC. The expression and localization of the survival factorbcl-2 was dependent on VEGF. These results provide evidence thatVEGF-A plays a critical role in the regulation of fetal pulmonarymesenchymal proliferation, survival, and the subsequent develop-ment of normal lung architecture through bcl-2 and p21/p27-dependent cell cycle control.

proliferation; bronchopulmonary dysplasia; SUGEN 5416

BRONCHOPULMONARY DYSPLASIA (BPD) is a chronic disease af-fecting the lung parenchyma and associated vasculature fol-lowing treatment of premature newborns with mechanicalventilation and high oxygen (1, 30, 46). BPD is associated withdeath and long-term morbidity and has been described as aconsequence of arrested normal development due to inappro-priate signaling arising from oxygen-induced injury (5, 12, 16,26, 29, 38, 39, 46, 50). Normally the lung reflects the gesta-tional stage of the neonate at the time of delivery. In BPD theformation of both mature alveoli and pulmonary vasculature isimpaired (18, 32, 38, 39). Although the specific cellular andmolecular mechanisms underlying BPD are incompletely un-derstood, VEGF-A has been shown to be involved in thisdisease. Levels of VEGF-A are reduced in the tracheal fluid of

neonates exposed to hyperoxia, as well as in clinical autopsymaterial of infants with BPD (7, 15).

Normal lung morphogenesis is regulated by autocrine andparacrine signals exchanged between the developing epithe-lium and mesenchyme. VEGF-A is produced by developingepithelium and is a critical mediator of both branching mor-phogenesis and alveolarization. The differential localizationand expression of the VEGF-A isoforms during developmentcreate a patterning gradient that guides cell proliferation andtissue assembly (3, 14, 25, 42). Increased VEGF expression inthe developing epithelium before the pseudoglandular andcanalicular stages disrupts branching morphogenesis and re-sults in mesenchymal thinning (17, 53). During the terminalsaccular stage of development, the mesenchyme continues tothin, and the blood vessel endothelium aligns in close proxim-ity to the respiratory epithelium, forming a surface for efficientgas exchange (43, 52, 53). At this time the proportion ofmesenchyme to epithelium decreases due to fetal breathingmovements that cause mesenchymal cell-cycle arrest and apop-tosis (22, 43).

In animal models, decreased VEGF-A levels result in thick-ening of the mesenchyme, while inhibition of VEGF with ablockade of VEGFRII signaling reduces pulmonary arterialdensity and radial alveolar counts in infant and adult rats (28,37, 52). Both increased and decreased levels of VEGF intransgenic animals cause apoptosis of mesenchyme (34, 51).Defects in mesenchymal proliferation and apoptosis have beenshown to result in lung hypoplasia and respiratory failure (49).Developing pulmonary mesenchymal cells (pMC) express theVEGF receptors Flt-1 (VEGFRI) and Flk-1 (VEGFRII) in aheterogeneous pattern. Flk-1-expressing mesenchyme is asso-ciated with endothelial precursors and vasculogenesis in thedeveloping distal lung (2, 3, 9, 20, 22, 23, 44). Flk-1 levelsdecrease into adulthood, while flt-1 expression is maintained(23, 44). VEGF-A modulates both Flt-1 and Flk-1 receptorexpression to maintain tissue homeostasis (22, 23, 41, 48).

The VEGF-dependent mechanisms that regulate pMC turn-over and resultant tissue architecture are unknown. However,because VEGF-A is associated with mesenchymal thickening,we hypothesized that alterations in VEGF-A protein levels willmodulate fetal pMC turnover in a cell cycle-dependent manner.We used an Flk-1-negative fetal pMC subpopulation culturedusing oxygen tensions similar to the fetal environment (3% O2)

Address for reprint requests and other correspondence: S. Majka, SON 3928,Mail stop B-133, 4200 E. 9th Ave., Denver, Colorado 80262 (e-mail:[email protected]).

The costs of publication of this article were defrayed in part by the paymentof page charges. The article must therefore be hereby marked “advertisement”in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

Am J Physiol Lung Cell Mol Physiol 290: L1183–L1192, 2006.First published January 20, 2006; doi:10.1152/ajplung.00175.2005.

http://www.ajplung.org L1183

by 10.220.33.5 on April 3, 2017

http://ajplung.physiology.org/D

ownloaded from

http://ajplung.physiology.org/

as our model. VEGF-A regulated progression through the cellcycle. The differential expression of p21, p27, and bcl-2following VEGF-A treatment or inhibition suggested a role forthese cell cycle-related molecules as downstream effectors ofVEGF-A in the regulation of pMC proliferation, survival, andapoptosis. These results demonstrate a direct effect of VEGF-Aon the turnover of late terminal saccular/early alveolar stagepulmonary mesenchyme.

METHODS

Isolation and Treatment of Fetal Lung Mesenchymal Cells

All procedures were approved by and performed according to theAnimal Care and Use Committee guidelines at the University ofColorado Health Sciences Center (UCHSC). Columbia-Rambouilletsheep at 131 days of gestation were killed with a lethal dose ofpentobarbital sodium, and the fetal lungs were harvested. The tracheawas then cannulated and infused with saline. Isolated lungs weretransferred to preequilibrated tissue culture medium (�-MEM, 20%fetal calf serum, penicillin-streptomycin; Invitrogen Life Technolo-gies, Carlsbad, CA). Explant tissue from the distal lung was sectionedand placed with media on plastic culture dishes for 48 h under normalculture conditions. After 48 h the tissue was removed from the plates,and additional medium was added. Mesenchymal cells, which hadmigrated from the explant cultures, were visible at this time andrequired an additional 2 wk of culture before confluence was reached.Three independent cell isolations were performed.

Cell Characterization

pMC at passage 0 (p0, fresh from explant) were expanded andpassaged to p2, at which time they were characterized by immuno-cytochemistry for the detection of smooth muscle �-actin (SM�A),platelet-derived growth factor receptor-� (PDGFR-�; mesenchymalmarkers), the absence of desmin (pericyte marker), cytokeratin (epi-thelial marker), factor 8, and CD34 (endothelial markers) (n � 3).Additionally, semiquantitative PCR was performed to detectVEGF-A, flk-1, and flt-1 mRNA. Phenotype and growth characteris-tics were documented. Characterization was repeated at p4 or p5 toensure that no phenotypic switches occurred during the course ofexperiments.

Experimental Design

pMC from three independent cell preparations at p4–p7 were usedfor all experiments. All experiments were performed with two or threereplicates and repeated with a different isolation of cells to ensurereproducibility. Cells were routinely passaged when they reached 80%confluence.

For all experiments reported here, pMC having reached 50%confluence were cultured at Denver altitude under either fetal oxygenlevels (3% O2) or relative hyperoxia (21% O2) in the presence orabsence of VEGF-A (100 ng/ml), placental growth factor (PLGF, 10ng/ml), or SUGEN 5416 (SU-5416, 10 or 25 �M) and correspondingDMSO controls. The SU-5416 concentrations were chosen basedupon the maximal inhibition of VEGF-A receptors reported in previ-ous cell studies (19). At this dose other tyrosine kinase receptors retaintheir function (19). The cells were collected at days 0, 2, and 7 andanalyzed for changes in total cell number and cell cycle, apoptosis,and changes in gene/protein expression as described below. The cellswere fixed with 4% paraformaldehyde and stored for analyses.

Cell Turnover

After treatment, viable pMC were counted with a hemocytometerand trypan blue exclusion and subsequently stained using the Krishanmethod to monitor changes in the cell cycle. In brief, 1 � 106 cells

were suspended in Krishan stain containing propidium iodide andallowed to stand overnight at 4°C. Data were collected with aFacsCalibur with Cellquest software (Becton Dickenson, San Jose,CA) and analyzed using Summit software (Cytomation, Ft. Collins,CO). Immunohistochemistry (IHC) to detect Ki67 proliferation anti-gen was also performed. Apoptotic pMC were identified by IHC todetect cleaved caspase 3 (CC3).

Message and Protein Expression Analysis

For determination of mRNA levels of VEGF-A and VEGFRI andII, total RNA was extracted from fetal lung-derived pMC, using therecommended protocol for TRIzol (Invitrogen Life Technologies,Carlsbad, CA). Total RNA was measured spectrophotometrically, andintegrity was confirmed by agarose-formaldehyde gel electrophoresis.Reverse transcription was performed with 1 �g of DNase-treatedRNA, using oligo dT and the suggested protocol for the Superscript IIkit (Invitrogen Life Technologies). The resulting cDNA was added asa template for each subsequent PCR reaction. PCR reactions werecarried out using a Perkin Elmer thermocycler under the followingparameters: 94°C 5 min 30 cycles of 94°C 30 s, 60°C 30 s, 72°C 30 s,72°C 15 min. The primers used to generate cDNA were as follows:VEGF-A: 5�-TCACCGCCTCGGCTTGTCACA-3�, 5�-TGTAATGA-CGAAAGTCTGCAG-3� (100, 250, 380 bp); Flt-1: 5�-CTA TAGCAC CAA GAG CGA CGT G-3�, 5�-GGC GTT GAG CGG AATGTA G-3� (550 bp); Flk-1: 5�-GGA GTT TTT GGC ATC ACG GAAGT-3�, 5�-GGA AAC AGG TGA GGT AGG CAG AG-3� (600 bp);GAPDH: 5�-TCACCATCTTCCAGGAGC-3�, 5�-CTGCTTCAC-CACCTTCTTGA-3� (650 bp). The reactions were then electropho-resed on 1.2% agarose gels using ethidium bromide or SYBR greenfor resolution of DNA amplicons. All PCR amplicons were clonedand sequenced to validate the primer pairs. Densitometry was per-formed using NIH Image analysis, and the experimental ampliconswere standardized to the housekeeping gene GAPDH for each sample.PCR was performed using primers for the following genes: VEGFRI(Flt-1), VEGFRII (Flk-1), VEGF-A, and GAPDH. Data were calcu-lated as the means � SE (n � 2; 3 replicates in each) of the relativeintegrated density value, which was obtained as area under the curvefollowing Image Quant Analysis. Ovine fetal pulmonary artery endo-thelial cells were used as controls for the PCR primers.

IHC and Western Blot Analyses

Differences in protein expression were analyzed by IHC andWestern blot analysis. IHC was performed on pMC cultured onchamber slides. Subconfluent pMC were fixed in 4% paraformalde-hyde, rinsed with PBS, and incubated with primary antibody, eitherusing the recommended protocol for the M.O.M. kit (Vector) or at [5�g/ml] in phosphate-buffered saline Tween for 16 h at 4°C. After awash in PBS, secondary antibody was added (1:500) for 30 min at RT.Secondary antibodies were fluorescent, conjugated to Alexa dyes 488and 594 (Molecular Probes), or biotinylated for use with the ABC/DAB system (Vector) of signal amplification and detection. Followingthe IHC procedures, we rinsed slides with PBS to remove unboundantibodies, coverslipped them with DAPI-Vectashield (Vector Labs),and viewed them under an Olympus iX71 scope equipped withfluorescence. To quantify the intensity of bcl-2 IHC in the SU-5416-treated groups, we imported digital images into Adobe Photoshop andperformed histogram analyses in the red channel, counting 331,000pixels. Western blot analysis was used to detect semiquantitativechanges in protein levels for the various experimental conditionsdescribed. Briefly, cell lysates were standardized for protein content(BCA assay; Bio-Rad), and 20 �g of protein were loaded into eachwell. The membranes were stained with Ponceau S as a secondarymeans of ensuring equal sample loading. They were then blocked with5% normal goat serum and inoculated with primary followed bysecondary-biotinylated antibody incubated in Tris-buffered saline

L1184 VEGF REGULATES FETAL PULMONARY MESENCHYMAL TURNOVER

AJP-Lung Cell Mol Physiol • VOL 290 • JUNE 2006 • www.ajplung.org

by 10.220.33.5 on April 3, 2017


ownloaded from


Tween, and then the ABC/ECL detection was performed (Vector,Amersham). Film was scanned on the STORM Imager and densitom-etry performed using Image Quant. Primary antibodies for proceduresdescribed included the following: SM�A (SM�A1A4, DAKO);desmin (DE-R11, DAKO); von Willebrand factor (vWF, F8-86;DAKO); PDGFR-� (clone 28, BD Transduction); BCL, Bax (N-19,P19; Santa Cruz); p21, VEGF-A, Flt-1 (sc397, sc152, sc316; SantaCruz); p27 (BD 554069); CC3 (Asp175) (no. 9661, Antibody CellSignaling Technology); and Ki67 (catalog no. RM-9106, Lab VisionNeoMarkers). Controls for each antibody/protein analysis included amatched isotype control and secondary antibody only as well as wholelung lysate as a tissue positive for the specific antigens.

Statistical Analysis

All experiments followed a randomized block design with theuse of cells from a least three different animals. Assays werecompleted from at least two independent experiments consisting oftwo to six replicates in each. Data are expressed as means � SE,and significance between groups was determined by one-wayANOVA or Student’s t-test using a statistical software packageJMP5 (SAS Institute, Cary, NC). Statistical significance was set atP � 0.05, � � 0.05.

RESULTS

Isolation and Characterization of Pulmonary Mesenchyme

To examine the direct effects of VEGF-A on pMC, weisolated a Flk-1neg distal lung cell population representative ofthe late canalicular/early terminal saccular phase of lung de-velopment. This is the stage at which supplemental O2 therapyis typically administered in neonates (1, 13, 36). We identifiedpMC in vivo using IHC to detect SM�A (Fig. 1, A and B).They also express VEGF-A message (20, 21) and protein,which can be localized to the cell surface (Fig. 1, C and D).IHC was performed on isolated cell populations to confirm theidentity of pMC as mesenchyme (45, 47). VEGF-A was local-ized to the pMC cell surface in vitro but was not detectable byELISA in the conditioned medium (Fig. 2A). Our data showedpMC expressed mRNA for the 3 VEGF-A isoforms and lowlevels of Flt-1 protein (Fig. 2, C–E). Low levels of Flt-1mRNA expression reflected the protein expression (Fig. 2, B,D, and E). Flk-1 (VEGFRII) message was not detected follow-ing isolation or in response to VEGF-A treatment. CulturedpMC expressed uniform basal levels of SM�A, PDGFR-�,

Fig. 1. Identification and characterization of pulmonary mesen-chyme in situ. Immunohistochemistry (IHC) was performed onsections of paraffin-embedded fetal day 131 lung tissue to detectsmooth muscle �-actin (SM�A; A, B) or VEGF-A (C, D) using3,3�-diaminobenzidine (DAB) detection. A, B: SM�A was local-ized to the lung mesenchyme, the smooth muscle layers aroundblood vessels and airways (AW) including the alveolar SMC. C,D: VEGF-A localized to lung mesenchyme, vascular endothelium,and epithelium. Black arrows, mesenchyme; red arrows, endothe-lium. Magnification: A and C, �40; B and D, �200.

L1185VEGF REGULATES FETAL PULMONARY MESENCHYMAL TURNOVER


by 10.220.33.5 on April 3, 2017


ownloaded from


lacked markers of differentiated vasculature, desmin and vWF,as well as cytokeratin (Fig. 2, E–H). These results confirm thatthe fetal cell isolates were pMC. They also show that ourmodel cell population did not contain endothelial precursorcells and that the subpopulation under study was indeed Flk-1negative.

VEGF-Dependent pMC Proliferation

VEGF receptor blockade decreased pMC proliferation. Wetreated pMC with SU-5416, VEGF-A receptor (flt-1 and flk-1)blockade, under 3% O2 concentration. SU-5416 treatmentsignificantly decreased the total cell number compared withuntreated controls at days 2 and 7 (Fig. 3A). Ki67 expression,which identifies cells active in the cell cycle, was quantitatedand showed significantly lower numbers of actively proliferat-ing cells following SU-5416 treatment compared with un-treated control cells at days 2 and 7 (Fig. 3B). There was alsoa reduction in proliferation between the two SU-5416-treated

groups at days 2 and 7. Cell cycle analysis was performed todetermine whether the SU-5416 response was dependent ongrowth arrest. VEGF receptor inhibition by 25 �M SU-5416treatment resulted in a 13% increase of cells arrested in G1(P � 0.05) compared with untreated controls, with a corre-sponding 65% decrease in G2 number (P � 0.013) by day 2.Proliferation was decreased significantly in response toSU-5416.

Exogenous VEGF-A increases pMC proliferation throughFlt-1. Overexpression in the lung of VEGF-A in transgenicmice using an epithelial surfactant protein C promoter disruptsbranching morphogenesis and mesenchymal thinning (52, 53).The effects of exogenous VEGF-A treatment on pMC culturedunder fetal O2 conditions (3%) were therefore examined. Therewas no significant increase in total cell number when pMCwere treated with VEGF (Fig. 3C). No apoptosis was detectedusing cleaved caspase IHC. Ki67 expression was correlatedwith the total cell numbers (Fig. 3D). On day 7 Ki67 expres-

Fig. 2. Isolation and characterization of Flk-1neg pulmonary mesen-chyme. IHC using fluorescent secondary antibodies or DAB detectionwas performed to confirm the isolated population of cells as pulmo-nary mesenchymal cells (pMC). A: VEGF-A localized to fetal day131 lung mesenchyme in vitro (�200). B: low levels of Flt-1 werelocalized to the pMC by IHC. C–E: Semiquantitative RT-PCRanalysis was performed to detect temporal changes in VEGF-A andFlt-1 gene expression in the presence (VF) or absence (UT) ofexogenous VEGF-A. C, E: pMC express mRNA for 3 VEGF iso-forms (188, 164, 120). Levels of expression were not affected byVEGF-A treatment. D, E: temporal expression of Flt-1 messagefollowing VEGF treatment (P � 0.1). Data were calculated asmeans � SE from 2 independent experiments with 3 replicates ineach. F, G: IHC analysis confirmed the expression of SM�A andplatelet-derived growth factor receptor (PDGFR)-�. H: von Wille-brand factor (vWF) was not detected. IDV, integrated density values.



by 10.220.33.5 on April 3, 2017


ownloaded from


sion was decreased in untreated control groups compared withthe VEGF-A-treated cells, which were actively proliferating(Fig. 3D). These data were further supported by cell cycleanalysis, which detected differences in stage of cell cyclebetween groups

We found that greater numbers of VEGF-treated pMC wereactively proliferating compared with controls. Forty-eighthours following VEGF treatment, cells appeared active in allthree phases of growth/synthesis, as evidenced by the 7.6%decrease of cells in G1 (P � 0.0021) and 44% and 35%increase in cells in the S (P � 0.0046) and the G2M (P �0.0046) phases, respectively, compared with controls. On day7, the untreated control cells were at rest in G1/G2M, while theVEGF-treated cells were progressing from DNA synthesis, S,to G2M. This progression is evident from the 55% increase incells in S phase and the 39% decrease of cells in G2M phase(P � 0.020) in the VEGF-treated group. The growth arrest of

the untreated control cells correlated with decreased Ki67expression at day 7, likely as a result of confluence. In responseto VEGF treatment, no significant changes in VEGF isoforms(Fig. 2, C and E), Flt-1 (Fig. 2, D and E), or Flk-1 messagewere identified by RT-PCR.

To define a pathway for pMC responsiveness to paracrineVEGF-A, we evaluated the potential role of Flt-1. We repeatedthe experiments previously outlined substituting PLGF forVEGF-A, because this factor selectively binds Flt-1 (4, 11).The results were comparable to VEGF-A treatment for criteriaanalyzed (Fig. 3E). There was no significant change in totalcell number until day 7. Cell cycle analyses further illustratedthat on day 7 the PLGF pMC progression from G1 (P � 0.060)into S (P � 0.04) was 5–6% higher than controls.

VEGF Regulates the Nuclear Expression of the Cell CycleModulator p21 by pMC

The cellular expression and distribution of the cell cycleregulatory proteins p21/p27 in control and SU-5416 (25 �M)treated pMC are shown in Fig. 4. When the untreated controlcells reached confluence, p27 expression was localized to thenuclei, whereas p21 was detected in the nucleus and cytoplasm(Fig. 4, K and N). In the SU-5416-treated cells, p27 was notdetected (Fig. 4G), but p21 was present in the cytoplasm andnuclei of cells (Fig. 4, J and M). VEGF-A-treated cells exhib-ited the same expression pattern as untreated control cells (Fig.4, I, L, O). These results suggest that a subset of cells survivesSU-5416 treatment by expressing nuclear p21 and downregu-lating p27 expression, possibly using this as a potential mech-anism of survival.

Alterations in the VEGF Levels Decrease Bcl-2 Expression

Analysis of protein expression and localization of bcl-2(survival-G1 arrest) was performed to examine a possiblequantitative change in the protein levels. Bcl-2 staining wasmore intense in the untreated pMC (Fig. 5, B and F) comparedwith the VEGF-A (Fig. 5D)- and SU-5416 (Fig. 5H)-treatedgroups. The SU-5416 treatment decreased bcl-2 intensity 50%.Protein levels of bcl-2 were decreased at 7 days followingVEGF-A treatment compared with untreated controls (Fig.6A). PLGF treatment caused a similar decrease in bcl-2 proteinlevels at day 7 (Fig. 6B), similar to the decrease in bcl-2observed following VEGF treatment.

DISCUSSION

The VEGF-dependent mechanisms that influence pMC turn-over and resultant lung architecture are unclear; therefore, weexamined the effects of VEGF-A signaling on the fetal lateterminal saccular/early alveolar stage of lung developmentusing an Flk-1-negative subset of pMC in vitro. Our datademonstrate that pMC isolated from distal lung during theterminal saccular stage responded to changes in VEGF-A viaVEGFRI (Flt-1) with alterations in proliferation and cell cycleprogression compared with untreated cells (Table 1). Thesefindings support the hypothesis that changes in the level ofVEGF-A during the late stages of lung development maycontribute to structural changes of the mesenchyme that un-derlie persistent lung diseases like BPD.

During lung development, the pulmonary mesenchyme het-erogeneously expresses VEGF receptors and can be divided

Fig. 3. VEGF-A regulates pMC cell proliferation. pMC were cultured in 3%oxygen and collected at days (d) 0, 2, and 7 following SU-5416, VEGF-A, orplacental growth factor (PLGF) treatment. A, B: VEGF-A receptor blockadedecreased pMC proliferation. SU-5416 treatment resulted in a significantdecrease in total cell numbers compared with untreated control (UT) at d2(P � 0.0001) and 7 (P � 0.0001). B: this decrease in cell number correlatedwith decreased Ki67 expression at d2 (P � 0.03) and 7 (P � 0.005). Asignificant decrease in Ki67 was also detected between SU-5416-treatedgroups over time (P � 0.0008). The addition of VEGF-A increased pMCproliferation. C: total cell number was not significantly different betweenuntreated control and VEGF-A-treated groups (d7; P � 0.0532). D: however,Ki67 expression was significantly increased at d7 following VEGF-A treat-ment (P � 0.015). E: PLGF treatment increased cell numbers by d7 (P �0.037). Data are presented as means � SE (n � 4). *P � 0.05; **P � 0.01.



by 10.220.33.5 on April 3, 2017


ownloaded from


into Flk-1-positive and Flk-1-negative populations (20, 23, 44,45). Cells expressing Flk-1 are vascular precursors in the distallung, and the expression of this receptor decreases in adulthood(23, 41, 48). In contrast, Flt-1 is expressed at higher levels than

Flk-1 over the entire course of lung development (22, 41, 48).The expression of Flt-1 is maintained into adulthood, in con-trast to Flk-1 (41, 48). pMC were isolated from distal fetalovine lung explants corresponding to late terminal saccular

Fig. 4. pMC expression and localization of cell cycle regulatory proteins. A–F: phase contrast and nuclear staining of pMC on d7. Immunohistochemical analyses of cellularexpression and distribution of the cell cycle regulatory proteins p21/p27. G, J, M: SU-5416 reduced p21 expression, and p27 was absent. G, H, I: p27 localized to the nucleusof pMC in untreated control and VEGF-A-treated cells. J–O: p21 was present in both the nucleus and cytoplasm. Magnification �100. M–O: enlarged areas.



by 10.220.33.5 on April 3, 2017


ownloaded from


Fig. 5. VEGF-A affects pMC Bcl-2 expression. Photomi-crograph (A, C, E, G) images depict the morphology ofpMC in the untreated control and VEGF-A- and SU-5416-treated samples. B, D, F, H: Bcl-2 expression was shownby IHC. Bcl-2 reactivity was most intense in the untreatedcontrol cells (B, F) and visible to a lesser extent in theVEGF-A (D)- and SU-5416 (H)-treated samples. Magni-fication �100.

Fig. 6. VEGF-A and VEGF-A receptor inhibitiondecreased Bcl-2 expression in pMC. Western blotanalysis was performed to quantitate changes inBcl-2 expression. Representative blots of bcl-2expression are shown above corresponding treat-ments. pMC were cultured in the presence orabsence of VEGF-A (A) or PLGF (B) treatment.Bars show the mean IDV � SE from Western blotanalyses (n � 4–6). pMC responded to the addi-tion of VEGF-A and PLGF with decreased bcl-2on d7 (P � 0.05, P � 0.03). *P � 0.05.



by 10.220.33.5 on April 3, 2017


ownloaded from


stage, the stage at which supplemental oxygen therapy isadministered to premature human neonates (1, 13, 36). TheFlk-1neg pMC migrated from the distal lung tissue and formedadherent colonies in culture (21). This in vitro system lacks thestructural complexity of lung tissue, including cell-cell inter-actions as well as paracrine secreted factors; however, itssimplicity allowed us to address the effects of VEGF on thespecific population of Flt-1-positive pMC. Another consider-ation with the cell culture system is the inconsistency of VEGFreceptor expression over length of time in culture, possibly dueto a perceived injury sustained by the cells during isolation andtheir resulting compensatory survival mechanisms. To controlfor this, the pMC were isolated on the basis of their propertiesas migratory mesenchyme, which characteristically involves anautocrine Flt-1 and VEGF-A pathway (6). Importantly, we didnot observe any changes in VEGF-A, Flt-1, or Flk-1 mRNAexpression over time (Fig. 2, B and D). We are thereforeconfident that these cells represent a valid model system foraddressing our hypothesis.

The effects of exogenous VEGF-A on pMC under fetaloxygen tension (3% O2) were studied. Changes in VEGF-Aexpression both increase and decrease lung mesenchyme thick-ness in transgenic mice, although the mechanisms remainunknown (10, 34, 49, 51, 53). In vivo, proper temporal andspatial distribution of VEGF-A is important for the mesen-chyme and adjacent interactions with epithelium or endothe-lium, required for normal lung development (21, 42). In BPD,VEGF-A and Flt-1 levels are decreased (7, 15, 38). BecauseFlt-1 and VEGF-A are important in pMC survival, this de-crease may lead to increased apoptosis and improper tissuemorphogenesis (6). Therefore, we initially studied the effectsof inhibition of VEGF-A receptor signaling on pMC usingSU-5416, which inhibits the VEGF-A receptor tyrosine kinasesFlt-1 and Flk-1. VEGF-A receptor inhibition caused a decreasein cell proliferation and density. Cell cycle analysis and lowlevels of Ki67 expression showed that greater numbers ofSU-5416-treated pMC arrested in G1 after 48 h compared withcontrols.

Cell cycle arrest in response to VEGF receptor blockade wasfurther supported by increased nuclear p21 expression withoutp27 expression in SU-5416-treated cells. p21 expression ischaracteristic of apoptosis resistance and is important for cellsurvival through cell cycle arrest, during which the cellsmaintain the ability to proliferate and repopulate (24, 27, 31,33). The expression of p21 independent of p27 has been shownto protect mesenchymal cells from programmed cell death in

low-density cell culture (47). Loss of p21 proceeds the onset ofapoptosis (37). The expression of p21 and cell cycle arrest, inaddition to the large surface area of the cells, was indicative ofrepair or survival through increased contact with the extracel-lular matrix (9). Downregulation of the G1 checkpoint regula-tors p21 and p27, decreases in bcl-2, and subsequent increasedapoptosis in response to SU-5416 have previously been dem-onstrated in tumor cells (54). The pMC expressing p21 werepresumably arrested in G1 to promote survival. It is possiblethat VEGF inhibition selected cells specifically resistant to thisparticular stress and caused p21-dependent growth arrest. Thegrowth-arrested cells may then remain viable and subsequentlyproliferate.

Previous studies using transgenic mice demonstrated thatVEGF overexpression in the fetal lung disrupts lung morpho-genesis. This disruption is characterized by mesenchymal thin-ning, decreased myofibroblast differentiation, acinar matura-tion, and sacculation (53). These pathological changes may beattributed to an inappropriate organization of the mesenchymalcell compartment (53). VEGF-A regulates the expression ofFlk-1 and Flt-1 and subsequent cell proliferation or survival,resulting in tissue homeostasis (48). We therefore examinedthe effects of exogenous VEGF-A on pMC. VEGF-A treatmentdid not affect the total cell number or apoptosis. However, cellcycle analysis and Ki67 expression showed that greater num-bers of VEGF-A-treated pMC were actively proliferating com-pared with confluent contact-inhibited control cells on day 7.The similarity in the proliferative rate of the groups until day7 suggests that the effects of exogenous VEGF-A were notdetectable until day 7 when confluence was reached in theuntreated controls. The continued proliferation of VEGF-treated cells beyond the cell density of control pMC suggests aderegulation of contact inhibition.

We evaluated changes in contact inhibition as a function ofcell cycle by examining qualitative changes in protein levels ofbcl-2 associated with cell survival (22). Both SU-5416 andVEGF-A decreased levels of bcl-2 protein and altered itslocalization in pMC. Bcl-2 not only delays the onset of apop-tosis but functions to arrest cells in G1 phase of the cell cycleand slow the transition between G1 and S phase (8, 40).Increased levels of bcl-2 protein decrease apoptosis in miceexposed to hyperoxia (5, 22, 39). Therefore, VEGF-A ismitogenic in this instance and regulates pMC progressionthrough the cell cycle, likely through decreases in bcl-2. Bcl-2is also involved in patterning during development. Its expres-sion is pronounced at sites of ectoderm/mesenchymal interac-tion (35). VEGF-A may therefore regulate cell-cell interactionsinfluencing the two-dimensional organization of pMC in cul-ture. The proliferative responses of pMC to VEGF-A weremimicked by PLGF treatment, which selectively activates theFlt-1 receptor (4, 25). SU-5416 treatment as VEGF receptorblockade may also inhibit to a lesser extent PDGF receptors(PDGFR-� and -�). PDGFR-� is also involved in mesenchy-mal differentiation and proliferation (6, 45) and is expressed bythe pMC. Future studies will require more specific inhibitors ofFlt-1 signaling, inhibition of alternate VEGF receptors, andVEGF-A inhibition.

In conclusion, we demonstrated that Flk-1neg pMC re-sponded to VEGF-A inhibition or addition with significantchanges in cell proliferation and cell cycle progression.VEGF-A therefore plays a multifunctional role at the levels of

Table 1. Summary of results

Inhibition of VEGFReceptors SU-5416

Autocrine VEGF-A(Untreated) Paracrine VEGF-A

Growth Arrest Normal Growth Additional Growthflt-1 flt-1 flt-1 2 cell-cycle progression cell-cycle progression 1 cell-cycle progression2 bcl-2 protein cell survival (Bcl-2)/

apoptosis (Bax)2 Bcl-2 protein

nuclear p21 p21/p27-dependentcontact inhibition

2p21 expression

absent p27 expression 2p27 expressionp21-dependent growth

arrest/selective cellsurvival

increased cell turnoverand alteredmorphology



by 10.220.33.5 on April 3, 2017


ownloaded from


both regulation of cell cycle progression through bcl-2 expres-sion as well as the regulation of survival (Table 1). By studyingthe effects of VEGF on the properties of the pulmonarymesenchyme, we are beginning to understand developmentalprocesses that contribute to BPD, a disease that involveslong-term mesenchymal, vascular, and alveolar irregularities,including alveolar hyperplasia, vascular wall thickening, andfibrosis. Because functional mesenchymal remodeling is nec-essary for normal alveogenesis and vasculogenesis, alterationsin the regulation of mesenchyme must likely play a key role inthe underlying architectural defects of this tissue. Furtherstudies are necessary to evaluate the mechanisms by whichchanges in VEGF-A influence pulmonary mesenchymal archi-tecture at later stages of development and growth within thecontext of the lung parenchyma.

ACKNOWLEDGMENTS

We thank Drs. Edward C. Dempsey, Vijaya Karoor, James West, andKatherine Young for input and critical review of this manuscript. We thankDrs. Steven Abman, Theresa Grover, Thomas Parker, and Christine Hunt-Peacock for providing ovine lung tissue; Dr. Jay Westcott, Jon Geske (ELISATechnologies), and Patsy Ruegg (IHCtech Histopathology Services Aurora,CO) for VEGF-A ELISAs and input regarding histochemistry and apoptosisstudies; Karen Helm and Michael Ashton for expertise performing the cellsorting experiments; the University of Colorado Cancer Center Flow Cytom-etry Core (supported by National Institutes of Health Grant 5 P30 CA-46934-15); and Drs. Norbert Voelkel, Ivor Douglas, and Neil Markham for technicalassistance.

GRANTS

This work was funded in part by American Heart Association GrantSDG-0335052N [principal investigator (PI) S. Majka], the Children’s HospitalResearch Institute of Denver (pilot PI: S. Majka), National Heart, Lung, andBlood Institute Grants HL-68702 and HL-57149 (PI: S. Abman), and theUCHSC Department of Pediatrics, Pulmonary Section.

REFERENCES

1. Abman S and Groothius J. Pathophysiology and treatment of BPD.Respir Med 41: 277–315, 1994.

2. Akeson AL, Wetzel B, Thompson FY, Brooks SK, Paradis H, Gen-dron RL, and Greenberg JM. Embryonic vasculogenesis by endothelialprecursor cells derived from lung mesenchyme. Dev Dyn 217: 11–23,2000.

3. Akeson A, Greenberg J, Cameron J, Thompson F, Brooks S, WigintonD, and Whitsett J. Temporal and spatial regulation of VEGF a controlsvascular patterning in embryonic lung. Dev Biol 264: 443–455, 2003.

4. Autiero M, Waltenberger J, Communi D, Kranz A, Moons L, Lam-brechts D, Kroll J, Plaisance S, De Mol M, Bono F, Kliche S, FellbrichG, Ballmer-Hofer K, Maglione D, Mayr-Beyrle U, Dewerchin M,Dombrowski S, Stanimirovic D, Van Hummelen P, Dehio C, HicklinDJ, Persico G, Herbert JM, Communi D, Shibuya M, Collen D,Conway EM, and Carmeliet P. Role of PLGF in the intra- and intermo-lecular cross talk between the VEGF receptors Flt-1 and Flk1. Nat Med 9:936–943, 2003.

5. Barazzone C, Horowitz S, Donati Y, Rodriguez I, and Piguet P.Oxygen toxicity in mouse lung: pathways to cell death. Am J Respir CellMol Biol 19: 573–581, 1998.

6. Bates RC, Goldsmith JD, Bachelder RE, Brown C, Shibuya M,Oettgen P, and Mercurio AM. Flt-1 dependent survival characterizes theepithelial mesenchymal transition of colonic organiods. Curr Biol 13:1721–1727, 2003.

7. Bhatt AJ, Pryhuber GS, Huyck H, Watkins RH, Metlay LA, andManiscalco WM. Disrupted pulmonary vasculature and decreased VEGF,Flt-1 and Tie-2 in human infants dying with BPD. Am J Respir Crit CareMed 164: 1971–1980, 2001.

8. Bonnefoy-Berard N, Aouacheria A, Verschelde C, Quemeneur L,Marcais A, and Marvel J. Control of proliferation by bcl-2 familymembers. Biochim Biophys Acta 1644: 159–168, 2003.

9. Borges E, Jan Y, and Ruslahti E. PDGF receptor beta and VEGFR-2bind to the beta-3 integrin through its extracellular domain. J Biol Chem275: 39867–39873, 2000.

10. Burridge K and Wennerberg K. Rho and Rac take center stage. Cell116: 167–179, 2004.

11. Cantley L. The PI3 kinase pathway. Science 296: 1655–1657, 2002.12. Carlo WA, Stark AR, Wright LL, Tyson JE, Papile LA, Shankaran S,

Donovan EF, Oh W, Bauer CR, Saha S, Poole WK, and Stoll B.Minimal ventilation to prevent bronchopulmonary dysplasia in extremelylow birth weight infants. J Pediatr 141: 370–375, 2002.

13. Coalson J, Winter V, Siler-Khodr T, and Yoder B. Neonatal chroniclung disease in extremely immature baboons. Am J Respir Crit Care Med160: 1333–1346, 1999.

14. Compernolle V, Brusselmans K, Acker T, Hoet P, Tjwa M, Beck H,Plaisance S, Dor Y, Keshet E, Lupu F, Nemery B, Dewerchin M, VanVeldhoven P, Plate K, Moons L, Collen D, and Carmeliet P. Loss ofHIF2a and inhibition of VEGF impair fetal lung maturation whereastreatment with VEGF prevents fatal respiratory distress in premature mice.Nat Med 8: 702–710, 2002.

15. D’Angio C and Maniscalco W. The role of vascular growth factors inhyperoxia induced injury to the developing lung. Front Biosci 7: d1609–d1623, 2002.

16. Demayo F, Minoo P, Plopper CG, Schuger L, Shannon J, and TordayJS. Mesenchymal-epithelial interactions in lung development and repair:are modeling and remodeling the same process? Am J Physiol Lung CellMol Physiol 283: L510–L517, 2002.

17. Dziadek M, Darling P, Bakker M, Overall M, Zhang RZ, Pan TC,Tillet E, Timpl R, and Chu ML. Deposition of collagen VI in ECMduring mouse embryogenesis correlates with the expression of the a3(VI)subunit gene. Exp Cell Res 226: 302–315, 1996.

18. Fong GH, Rossant J, Gertsenstein M, and Breitman ML. Role of theflt receptor tyrosine kinase in regulating the assembly of vascular endo-thelium. Nature 376: 66–70, 1995.

19. Fong TA, Shawver LK, Sun L, Tang C, App H, Powell TJ, Kim YH,Schreck R, Wang X, Risau W, Ullrich A, Hirth KP, and McMahon G.SU5416 is a potent selective inhibitor of the VEGF receptor (Flk/KDR)that inhibits tyrosine kinase catalysis, tumor vascularization and growth ofmultiple tumor types. Cancer Res 59: 99–106, 1999.

20. Gebb S and Shannon J. Tissue interactions mediate early events inpulmonary vasculogenesis. Dev Dyn 217: 159–169, 2000.

21. Greenberg JM, Thompson FY, Brooks SK, Shannon JM, McCor-mick-Shannon K, Cameron JE, Mallory BP, and Akeson AL. Mesen-chymal expression of VEGF D & A defines a vascular patterning indeveloping lung. Dev Dyn 224: 144–153, 2002.

22. Haddad J, Choudhary K, and Land S. The ex vivo differential expres-sion of apoptosis signaling cofactors in the developing perinatal lung.Biochem Biophys Res Commun 281: 311–316, 2001.

23. Hara A, Chapin CJ, Ertsey R, and Kitterman JA. Changes in fetal lungdistension alter expression of VEGF and its isoforms in developing ratlung. Pediatr Res 58: 30–37, 2005.

24. Hipfner D and Cohen S. Connecting proliferation and apoptosis indevelopment and disease. Nature 5: 805–816, 2004.

25. Hirashima M, Ogawa M, Nishikawa S, Matsumura K, Kawasaki K,Shibuya M, and Nishikawa S. A chemically defined culture ofVEGFR2 cells derived from embryonic stem cells reveals the role ofVEGFR1 in tuning the threshold for VEGF in developing EC. Blood 101:2261–2267, 2003.

26. Ho L. BPD dysplasia and CLD of infancy. Ann Acad Med Singapore 31:119–130, 2002.

27. Ilyin GP, Glaise D, Gilot D, Baffet G, and Guguen-Guillouzo C.Regulation and role of p21 and p27 cyclin dependent kinase inhibitorsduring hepatocyte differentiation. Am J Physiol Gastrointest Liver Physiol285: G115–G127, 2003.

28. Jakkula M, Le Cras TD, Gebb S, Hirth KP, Tuder RM, Voelkel NF,and Abman SH. Inhibition of angiogenesis decreases alveolization in thedeveloping rat lung. Am J Physiol Lung Cell Mol Physiol 279: L600–L607, 2000.

29. Jobe A and Bancalari E. Bronchopulmonary dysplasia. Am J Respir CritCare Med 163: 1723–1729, 2001.

30. Jobe A. The new BPD: an arrest of lung development. Pediatr Res 46:641–643, 1999.

31. Karnezis A, Dorokhov M, Grompe M, and Zhu L. Loss of p27enhances the transplantation efficiency of hepatocytes transferred intodiseased livers. J Clin Invest 108: 383–390, 2001.



by 10.220.33.5 on April 3, 2017


ownloaded from


32. Khaliq A, Li XF, Shams M, Sisi P, Acevedo CA, Whittle MJ, WeichH, and Ahmed A. Localization of PLGF in human term placenta. GrowthFactors 13: 243–250, 1996.

33. Kippin T, Martens D, and van der Kooy D. P21 loss compromises therelative quiescence of forebrain stem cell proliferation leading to exhaus-tion of their proliferation capacity. Genes Dev 19: 756–767, 2005.

34. Kotch L, Iyer N, Laughner E, and Semenza G. Defective vasculariza-tion of HIF-1a null embryos is not associated with VEGF deficiency butwith mesenchymal cell death. Dev Biol 209: 254–267, 1999.

35. Krajewski S, Hugger A, Krajewska M, Reed J, and Mai J. Develop-mental expression patterns of bcl-2, bcl-x, bax, and bak in teeth. CellDeath Differ 5: 408–415, 1998.

36. Langston C, Kida K, Reed M, and Thurlbeck WM. Human lung growthin late gestation and in the neonate. Am Rev Respir Dis 129: 607–613,1984.

37. Le Cras TD, Markham NE, Tuder RM, Voelkel NF, and Abman SH.Treatment of newborn rats with a VEGF receptor inhibitor causes pulmo-nary hypertension and abnormal lung structure. Am J Physiol Lung CellMol Physiol 283: L555–L562, 2002.

38. Maniscalco W, Watkins R, Pryhuber G, Bhatt A, Shea C, and HuyckH. Angiogenic factors and alveolar vasculature: development and alter-ations by injury in very premature baboons. Am J Physiol Lung Cell MolPhysiol 282: L811–L823, 2002.

39. Mantell L and Lee P. Signal transduction pathways in hyperoxia inducedlung cell death. Mol Genet Metab 71: 359–370, 2000.

40. Marone M, Bonanno G, Rutella S, Leone G, Scambia G, and PierelliL. Survival and cell cycle control in early hematopoiesis: role of bcl-2 andthe cyclin dependent kinase inhibitors p27 and p21. Leuk Lymphoma 43:51–57, 2002.

41. Mata-Greenwood E, Meyrick B, Steinhorn RH, Fineman JR, andBlack SM. Expression of VEGF and its receptors is altered in lambs withincreased pulmonary blood flow and pulmonary hypertension. Am JPhysiol Lung Cell Mol Physiol 285: L222–L231, 2003.

42. Ruhrberg C, Gerhardt H, Golding M, Watson R, Ioannidou S, Fuji-sawa H, Betsholtz C, and Shima DT. Spatially restricted patterning cuesprovided by heparin binding VEGF-A controls blood vessel branchingmorphogenesis. Genes Dev 16: 2684–2698, 2002.

43. Sanchez-Esteban J, Wang Y, Cicchiello L, and Rubin L. Cyclicmechanical stretch inhibits cell proliferation and induces apoptosis in fetalrat lung fibroblasts. Am J Physiol Lung Cell Mol Physiol 282: L448–L456,2002.

44. Schachtner S, Wang Y, and Baldwin S. Qualitative and quantitativeanalysis of embryonic pulmonary vessel formation. Am J Respir Cell MolBiol 22: 157–165, 2000.

45. Sousa P, Tanswell A, and Post M. Different roles for PDGFR-� and -�receptors in embryonic lung development. Am J Respir Cell Mol Biol 15:551–562, 1996.

46. Stark A. High frequency oscillatory ventilation to prevent BPD–Are wethere yet? N Engl J Med 347: 682–684, 2002.

47. Van den Bos C, Silverstetter S, Murphy M, and Connolly T. p21rescues human mesenchymal stem cells from apoptosis induced by lowdensity culture. Cell Tissue Res 293: 463–470, 1998.

48. Wang D, Donner D, and Warren R. Homeostatic modulation of KDRand Flt expression and VEGF mRNAs by VEGF-A. J Biol Chem 275:15905–15911, 2000.

49. Weiguo S, Jiang Y, Lu M, and Morrisey E. Wnt7b regulates mesen-chymal proliferation and vascular development in the lung. Development129: 4831–4842, 2002.

50. Weinberger B, Laskin DL, Heck DE, and Laskin JD. Oxygen toxicityin premature infants. Toxicol Appl Pharmacol 181: 60–67, 2002.

51. Ylikorkala A, Rossi DJ, Korsisaari N, Luukko K, Alitalo K, Henke-meyer M, and Makela TP. Vascular abnormalities and deregulation ofVEGF in lkb1 deficient mice. Science 293: 1323–1326, 2001.

52. Zeng X, Gray M, Stahlman M, and Whitsett J. TGFb1 perturbsvascular development and inhibits epithelial differentiation in fetal lung invivo. Dev Dyn 221: 289–301, 2001.

53. Zeng X, Wert SE, Federici R, Peters KG, and Whitsett JA. VEGFenhances pulmonary vasculogenesis and disrupts lung morphogenesis invivo. Dev Dyn 211: 215–227, 1998.

54. Zhong X, Li X, Wang G, Zhu Y, Hu G, Zhao J, Neace C, Ding H, ReedE, and Li QQ. Mechanisms underlying the synergistic effect of SU5416and cisplatin on cytoxicity in human ovarian tumor cells. Int J Oncol 25:445–451, 2004.



by 10.220.33.5 on April 3, 2017


ownloaded from


pleiotropic role of vegf-a in regulating fetal pulmonary ... · pleiotropic role of vegf-a in...

Documents