Identification and partial characterization of
angiogenesis bioactivity in the lower
respiratory tract after acute lung injury.
C Henke, … , J McCarthy, P Bitterman
J Clin Invest. 1991;88(4):1386-1395. https://doi.org/10.1172/JCI115445.
Survival after acute lung injury (ALI) depends on prompt alveolar repair, a process
frequently subverted by the development of granulation tissue within the alveolar airspace. Immunohistochemical examination of the intraalveolar granulation tissue confirmed that capillaries as well as myofibroblasts were the principal cellular constituents. We therefore hypothesized that angiogenesis factors would be present on the air-lung interface after ALI. To evaluate this hypothesis, bronchoalveolar lavage fluid from patients with ALI (n = 25) and patient controls (n = 8) was examined for angiogenesis bioactivity by its ability of induce endothelial cell migration. While lavage fluid from controls had no bioactivity, lavage fluid from 72% of patients with ALI promoted endothelial cell migration. Heparin affinity, ion exchange, and gel filtration chromatography resolved the bioactivity into at least two moieties. One appeared identical to the well characterized endothelial cell growth factor, basic fibroblast growth factor. The other was a 150-kD non-heparin binding protein that mediated endothelial cell migration and attachment in vitro, and the growth of new vessels in vivo. These data are consistent with the hypothesis that the growth of capillaries into the alveolar airspace results from angiogenesis factors present on the alveolar surface of the lung after ALI.
Research Article
Find the latest version:
Identification
and Partial Characterization
of Angiogenesis Bioactivity
in
the Lower Respiratory Tract after
Acute Lung Injury
Craig Henke,* Vance Fiegel,t MarkPeterson,*Mark Wick,11 DavidKnighton,*JamesMcCarthy,$and PeterBitterman* *Department of Medicine, *Department of Surgery, Department ofLaboratoryMedicine and Pathology, UniversityofMinnesota,
Minneapolis, Minnesota,55455;andItDepartmentofPathology, Washington University,St. Louis, Missouri 63130
Abstract
Survival afteracute lung injury(ALI)depends on prompt
al-veolarrepair, a process frequentlysubverted by the
develop-ment ofgranulation tissuewithin the alveolar airspace.
Immu-nohistochemical examination of the intraalveolar granulation tissue confirmedthatcapillariesas well asmyofibroblastswere
theprincipal cellular constituents.Wetherefore hypothesized
thatangiogenesis factorswould be present on theair-lung
in-terface afterALI. Toevaluate this hypothesis, bronchoalveolar lavage fluidfrompatientswith ALI(a= 25)andpatient
con-trols(a =8)wasexamined forangiogenesis bioactivity by its
ability of induce endothelialcellmigration.While lavagefluid from controls hadnobioactivity,lavage fluid from 72% of
pa-tients with ALIpromotedendothelial cellmigration.Heparin affinity, ion exchange, andgel filtration chromatography
re-solvedthebioactivity intoat least twomoieties. One appeared identicaltothewellcharacterized endothelial cell growth
fac-tor, basic fibroblast growth factor. The other was a 150-kD non-heparinbindingproteinthatmediated endothelialcell
mi-grationandattachmentinvitro,andthegrowth ofnewvessels
invivo. These dataareconsistent withthehypothesisthatthe growth of capillaries into the alveolar airspace results from angiogenesisfactors presentonthealveolar surface ofthelung afterALI.(J. Clin. Invest. 1991. 88:1386-1395.) Key words:
lung repair *granulation tissue *endothelial cell
migration.
angiogenesis*angiogenesis factors
Introduction
Acute lung injury is an abrupt alteration ofthepulmonary
parenchyma leadingto functionalimpairmentof the gas
ex-changeapparatus.Thisdisorderoccursafter directpulmonary injury suchasaspiration of gastriccontents,oraftersepsisor
traumaoccurringin anotherpart of thebody. Despite
signifi-cantadvances incriticalcaretechnology,initsmostfulminant form,the adultrespiratorydistresssyndrome (ARDS),' acute
Addresscorrespondence andreprintrequeststo Dr.CraigA.Henke, Pulmonary Division, Department ofMedicine,UMHC Box276, Uni-versity of Minnesota, 420 Delaware Street Southeast, Minneapolis,
MN55455.
Receivedforpublication9October1990andinrevisedform24June
1991.
1. Abbreviations used in this paper: ABC, avidin-biotin-peroxidase
complex; ARDS, adult respiratory distress syndrome; bFGF, basic fi-broblastgrowth factor;DEAE-S,diethylaminoethyl sepharose.
lunginjuryexacts amortalityin excess of 50%. ARDS
consti-tutesamajor publichealth problem with anestimated150,000
patients treated annuallyin the United States (1).
Anatomically,acute lung injury is anexplosive transmural
process. Thereis widespreaddeath of thealveolar epithelium
andmicrocirculatory endotheliumas well assignificant
inter-stitial edema (2, 3). When alveolarrepair isto be effective,a
preciselycoordinatedsequenceof cellular migrationand repli-cation reconstitutes the air-lung interface, interstitium, and
blood-lung interface. The repair process, however,frequently fails. Underthese circumstances, an acute fibroproliferative
response ensues in which granulation tissue rapidly fillsthe
alveolarairspace, precluding effectivegasexchange (4). Whilemostattentionhas beenfocusedonthe role of the
interstitialmyofibroblast in the evolution ofacute intra-alveo-largranulation,newblood vessel growth isaprominent
compo-nentofthefibroproliferativeresponse(5).Thedevelopmentof
newblood vessels is controlled byadiscrete set of peptides, collectivelyreferredtoasangiogenesisfactors (6, 7). Fornew
blood vesselstogrowinto the alveolar airspace, microvascular endothelialcellsmustmigrate through the injured alveolarwall
and attachtotheintra-alveolar provisionalmatrix. We there-foredesignedthis studytoexamine thehypothesisthat angio-genesis factorscapable ofstimulating endothelial cell migra-tionand attachment would bepresent ontheair-lung interface after acute lung injury. The data indicate that angiogenesis bioactivitywaspresentin thelungs ofmostpatientsafteracute
lunginjury. Two peptides accounted for themajorityofthe
bioactivity.Onehadamolecularmassof 18kDandappeared identicaltotheestablished endothelial cellgrowth factor,basic fibroblastgrowth factor (bFGF).Theotherwas a150-kD mole-cule that stimulated directedmigrationandattachment of
en-dothelial cells invitro, angiogenesisinvivo,andappearedtobe
distinct frompreviously characterized
angiogenesis
factors.Methods
Study
population
Twogroups ofpatientsunderwentbronchoscopywithbronchoalveolar
lavage for the purpose ofexaminingtheair-lunginterface for the
pres-enceofangiogenesis factors.
1. Acutelunginjury (n=25).Patients withsevereacutelunginjury
being supportedwithmechanical ventilation whometthe standard
clinical criteria for ARDSwereincluded in this cohort(8,9).All pa-tients hadanappropriateantecedenthistory,severehypoxemiadespite
highconcentrationsofsupplementaloxygen, chestradiographshowing
diffuseinfiltrates,andnootherexplanationfor theirrespiratory
insuffi-ciency. Patientswerestudiedwithin 21 d of diseaseonset toobtain
samplesfrom the alveolarepithelialsurface.Asubset ofacutelung
injury patientswho died fromrespiratory insufficiency14to21dafter
diseaseonset wereevaluatedpost-mortembyexvivo wholelung
la-vagetopermitpreparativebiochemicalstudies. Mostpatients (n=17)
werestudied earlyintheir clinicalcourse(day 2-4)assoonas
bron-1386 Henke,Fiegel, Peterson, Wick, Knighton, McCarthy, andBitterman
J.Clin. Invest.
©TheAmericanSociety forClinical Investigation, Inc.
0021-9738/91/10/1386/10 $2.00
choalveolar lavage could be safely performed. The remainder of the patients (n=8) were studied later in their clinical course (day 10-21) in order to provide samples which temporally corresponded to the post-mortem whole lung lavage samples obtained for preparative biochemi-calexamination.
2. Patient controls (n=8). This group consisted of normal individ-uals andpatients undergoing bronchoscopy for evaluation of a local-ized process. Bronchoalveolar lavage was performed in an area distant from the lesion.
Immunohistochemistry
Representative lung tissue was obtained at autopsy from patients who died ofacutelung injury (n= 12), as well as 10 control cases with no pulmonary abnormalities. The tissue was fixed in 10% neutral-buffered formalin for6 to 18 h and embedded in paraffin. 5-Msections of all
specimenswere cutand mounted on glass slides that had been coated
with a chrome-alum mixture.
Deparaffinization was accomplished with Americlear (American Scientific Products, Inc., McGaw Park, IL) and absolute alcohol, and endogenousperoxidasewasblocked by incubation in methanol-0.6% hydrogen peroxide for 30 min. The sections were rehydrated in graded alcohols, distilled water, and PBS (pH 7.4).
Primary antibodies to factor VIII-related antigen and vimentin wereapplied for 18 h at4VCin moisture chambers. Localization of
antibody binding was visualized with the avidin-biotin-peroxidase
complex (ABC) procedure, asdescribed previously (10), using
3,3'-diaminobenzidine hydrochloride(0.25 mg/ml with 0.003% hydrogen
peroxide)asthechromogen. Sectionswerethen immersedbrieflyin
0.125% osmium tetroxide(wt/vol),counterstained with Harris
hema-toxylin,andcoverslippedwithasynthetic mountingmedium.
Positive controls consisted of unrelated tissue-banked specimens known tocontain the determinants ofinterest; "internal controls" such
asbronchialepitheliumandlarge pulmonary blood vesselswerealso assessedtoverify methodologic integrity. Sectionsofallstudycases
werealso labeled with ABC after substitution of nonimmune mouse ascitesfluidforprimary antibodies;these servedasnegativecontrols.
Allimmunohistochemicalpreparationswerereviewedandinterpreted
byoneobserver.
Bronchoalveolar
lavage
Bronchoalveolarlavagewasperformedin ordertoobtain fluid from
theair-lung interfaceaspreviouslydescribed(1 1).Inacutelunginjury
patients receiving mechanical ventilation, bronchoscopy was
per-formedthroughanendotracheal tubeusinganappropriate adapter.All
patientsreceived 100% oxygenduringtheprocedure,and oxygen
satu-ration andelectrocardiogramwerecontinuouslymonitored. The pro-cedurewasterminated ifthe oxygen saturation fell below 90% formore
than 1min.Uptothree separate anatomic sitesweresampledin each
patientwithacutelunginjury.Inpatientcontrols,after local anesthesia
of the upperairway,the bronchoscopewas wedgedin aperipheral
bronchus andfive, 20 mlaliquotsof 0.9% salinewereinstilled and removed by gentle suction. Uptothree separate anatomiclocations
weresampled in eachpatient.Inasubset ofacutelunginjurypatients
whodied fromrespiratory insufficiency, wholelung lavagewas per-formed within 2 hofdeathexvivo. Wholelung lavagewasperformed
bycannulatingthe mainstem bronchus withplastic tubingfollowedby
theinstillation of sterile PBS(30cmH20pressure).Fluidwas
recov-eredbygravityflow.
Processing
ofbronchoalveolar
lavage
specimensRecovered bronchoalveolarlavage effluentwasseparatedfrom cells by
centrifugation (600gX15min).The cellfreelavageeffluentwaseither
aliquotedandstoredat-70°C beforeassay,orkeptat4°Cif the assay
was tobeperformed immediately.Proteinconcentrationof cell free lavageeffluentwasdeterminedusingacolorimetric assay(Pierce Chem-ical Co., Rockford, IL), and albumin concentration was quantified by ELISAaspreviouslydescribed(12).
Biological
assayof
lavagefluid
Thebioassaysusedwerechosento assess twoof the cellular responses
integraltothe processofangiogenesis:(a) endothelialcellmigration,
and(b)the formation ofcapillarynetworks in vivo(6, 7).Toidentify
the presence ofangiogenesis bioactivity, lavageeffluent from all pa-tientswasanalyzed usinganendothelialcellmigrationassay. Confir-mation ofangiogenesisbioactivitydetected in themigrationassaywas
performed usinganinvivoangiogenesisassay.
MIGRATION ASSAY
Endothelial cellmigrationwasassessed inamodifiedBoyden chemo-taxis chamber(Nucleopore, Pleasanton,CA),aspreviouslydescribed
(13, 14).Toensure the generalizabilityof thebioassay results, both
microvascular andlargevessel endothelial cellswereused inthe che-motaxis assay. Rabbit wound microvascular endothelial cellswere iso-lated from spongesimplantedin the backs of New Zealand white rab-bitsbymodifyingtheisolationprocedureusedtoharvestrabbit brain
capillaryendothelialcells,asdescribed(14).Bovinepulmonaryartery endothelial cellswereisolatedbygentlescrapingof the luminalsurface of thepulmonaryartery and cultured inM199+20% FCS in 25cm2
Primariaflasks(BectonDickinson andCo.,LincolnPark, NJ)
(370C;
5%C02,95%air). Pulmonaryarteryendothelialcells,whenconfluent, displayed typicalcobblestonemorphology using phasecontrast micros-copy. Cellidentitywasconfirmedbyanti-factor VIIIstainingandby
specific labeling, usingafluorescentacetylatedLDLprobeasdescribed
(15, 16).
Briefly,for eachstudy,26glof thetestfluidatvarious
concentra-tionswereplacedin the wells of the lower chamber of the chemotaxis apparatus. 45 Ml of the cell suspension (750,000 cells/ml in M199
+0.2%lactalbumin)wereplacedin the upperwells,whichwere sepa-rated from the lower chamberbyafibronectin-coated(1
Mg/ml,
30 min eachside),porous(8-Mm
diameterpores)polycarbonatefilter(Nucleo-pore).Thechemotaxis chamberswereincubated
(370C,
4h;95%air,5%C02)permittingcellmigrationto occur.Migrationwasquantified
as the number of cellsonthe lower side of the filter per 100X field.
Degranulated platelet lysate servedas a positive control and M199
+0.2% lactalbuminas anegativecontrol.
INVIVO ANGIOGENESIS ASSAY
Confirmation ofangiogenesis bioactivitydetected in themigration
as-saywasperformed by assessingtheabilityofatestsampletoinduce neovascularization in thenormallyavascularcorneaof the New Zea-land white rabbit. Toaccomplishthis,thetestfluidwasintroduced into thestromaof thecorneaand vascularingrowthfrom the limbus into thecorneawasassessedaspreviouslydescribed(17).Bronchoalveolar
lavageeffluentorpartiallypurifiedangiogenesis bioactivitywas
clari-fied bycentrifugation(10,000rpmX 5min),dialyzed againstbuffer
(7.5mMNaCl,5.0 mMNaPO4, pH 7.4),andconcentrated 10to 20-foldbyspeed evaporation(SpeedVac Concentrator;Savant Instru-ments,Farmingdale, NY).Theconcentratedsamplewasresuspended
in25Mlof sterile distilledH20,incorporatedintoanequalvolume of sterileHydron castingsolution,and allowedtosolidify.TheHydron
pelletcontainingthetestfluidwasintroduced intoapouchformed in
the cornealstroma.Hydroncontainingbuffer servedas a
negative
con-trol. Thecorneas wereexamined every otherdaytomonitor for infec-tion andcapillary growth.Photographsweretakenonday2andday7
torecord theresults,andsomeanimalswerekilledonday2and the remainderonday7toevaluate thecorneashistologicallyfor
inflamma-tion. Thecorneas were removed and fixedin PBS
containing
10%formalin. Paraffin sectionswerepreparedand stained with
hematoxy-linand eosin forhistologicexamination ofnewvessel
growth.
Biophysical
andbiochemical
characterizationof
theangiogenesis
bioactivity
Aftereachprocedurecharacterizingthe
angiogenesis
bioactivity,themigrationassaywasusedtoidentifythebioactivity.Insomecases,as
indicated, resultswereconfirmedusing the in vivoangiogenesisassay.
The
biophysical properties
assessedincluded:(a)solubility in nonpolar solvents(1:1 ethyl acetate); (b) heat stability
(1000C,15 min); (c) acid stability (pH 3, 12 h); and(d)protease
sensitiv-ity (trypsin, 0.2 mg/ml, 30min, 370C;inactivated with a twofold molar excessof soybeantrypsin inhibitor).
Biochemical characterization of the angiogenesis bioactivity con-sisted ofsequential chromatography using heparin affinity, ion ex-change, and gelfiltration. The starting material for these studies was 1 liter of lavagefluidobtained from whole lung lavage of postmortem lungs(processed in . 2 h). Allprocedures were carried out using this material and confirmed in other routinely obtained lavage fluid as quantities permitted.
One class of endothelial cell growth factors, the fibroblast growth
factors, avidly binds heparin (18-20).Toisolate members of this class,
bronchoalveolar lavage fluid was applied to a heparin-sepharose
col-umninequilibrationbuffer (0.05M NaCl,0.01 M Tris, pH 7.4). The
column was developed stepwise with buffers of increasing ionic strength(0.5 M NaCl, 1 MNaCl, 1.5 MNaCl, 2 M NaCl). Fractions weredialyzed against PBS before the assay. Bioactivity binding to hepa-rin-sepharose was further analyzed by SDS-PAGE and Western blot analysis (see below).
Bioactivefractionsoflavagefluid failing to bind to
heparin-sephar-oseweredirectly appliedin equilibration buffer (0.125 M NaCl,0.01M
Tris, pH7.4)to ananion exchange resin (diethylamino ethyl sepharose;
[DEA-S]).The columnwasdevelopedinastepwisefashion with buff-ered NaClsolutions(0.01 MTris,pH 7.4) ofincreasingionic strength: 0.25 MNaCl,0.50MNaCl, 0.75 M NaCl, 1.0MNaCl. Fractions were
dialyzed againstPBSbefore the assay.
TheDEAE-Sfractionswith biologicalactivity were further purified
usingastrong anionexchange resin (MonoQcolumn;Pharmacia,Inc.
Piscataway, NJ) and a linear gradient as dictated by the results of the stepwise elution. Activefractionsweredialyzed andapplied in starting buffer(0.125MNaCI, 0.01 MTris,pH7.4).The columnwas devel-oped withalinear saltgradient(0.125-0.300 MNaCl).Fractionswere
directly examined forbiological activitywithoutfurther handling. Toobtain an estimateofmolecular mass, the MonoQ fractions withbiological activity were applied to a gel filtration column
(Super-ose12,10X300 mm,flowrate=0.5ml/min;Pharmacia).The column wasdeveloped withPBS, pH 7.4,containing10% acetonitrile,to mini-mizenonspecific interaction with the gel filtration matrix. Pilot studies indicated that 10%acetonitrile had no effectonthebioactivityin the chemotaxis assay. Molecularweightstandardsincluded: humanIgG (166kD), human serum albumin(66 kD),ovalbumin (44kD), soybean
trypsininhibitor(21 kD),andaprotinin (6.5 kD). Fractionswere
di-rectly examined forbiologicalactivity.Of note,instabilityofthe
bioac-tivityinhigherconcentrations oforganicbuffer(>30%acetonitrile)
precluded furtherpurification byreversephasechromatography.
ADHESION ASSAY
Directed endothelial cellmigrationcan occur as aresultofmovement
of endothelialcells towardapositivechemicalgradient (chemotaxis)or
asaresultofcontactwithproteinsthat promote cellular adhesion
(hap-totaxis)(14, 21, 22).Todetermine iftheangiogenesis bioactivity
identi-fied by the chemotaxis assaypromotedcellularattachment,adhesion of endothelial cellstosubstratum-boundproteinswasdetermined
us-ingastandardadhesion assayasdescribed(23). Test proteinsused in
theadhesionassaywerepreparedatvarious concentrations in Vollers
buffer(15mMNa2CO3and 34.8mMNaHCO3, pH 9.6),asfollows:(a)
positive controls, soluble typeIV collagen, fibronectin;(b) negative
control,bovineserumalbumin;and(c)testprotein, partially purified
lavageangiogenesisfactor(activefractions aftergel filtration).
WESTERN BLOT ANALYSIS
Toanalyze theheparin binding bioactivityfor the presence of the well established endothelial cellgrowthfactor,bFGF,Westernblotanalysis
wasperformed.Proteins wereelectrophoresed in 15%polyacrylamide
gelscontaining SDSaccordingtothemethod of Laemmli(24). The
samples were mixed (2:1,vol/vol) with nonreducingsamplebuffer
(20%glycerol, 2.3% sodiumdodecylsulfate,0.625MTrisbase,0.001%
bromphenol blue, pH 6.8)and loaded(upto50Mgtotalprotein)into
each lane. Proteinswereelectrophoresed(3 h, 220C) by applying 15
mAofconstant current(ECPS 3000/150;Pharmacia).For transfer of
proteins, gelswereoverlayedwitha0.2 Mm nitrocellulosemembrane,
submergedintransfer buffer(25mMTris-base, 192 mMglycine,20%
methanol, pH8.3),andtransferredelectrophoretically (45min, 24V)
usingatank-typeplateelectrodeblotter(IdeaScientificCo.,
Minneapo-lis, MN). Afterelectrophoretictransfer ofproteins,the nitrocellulose blotwasdried anddevelopedasdescribed for the immunoblot assay (seebelow).
IMMUNOBLOT ASSAY
Animmunoblot assaywasusedtoquantifybFGF in selected column fractionsduringthepurificationprocedures,aspreviouslydescribed
(25).Theprimaryantibodyused in the immunoblot assaywasarabbit anti-bovine bFGFantibody (R&D SystemsMinneapolis, MN;20Mg/
ml)and the secondaryantibodywas swineanti-rabbitIgG coupledto
horseradish peroxidase (Dako Corp.,SantaBarbara, CA; 1:250
dilu-tion inTTBS, 150 mMNaCl, 100 mMTris, 0.05% Tween20).To eliminatecross-reactivitybetween thesecondaryantibodyand human
IgGintheprocessed lavageeffluent,before incubation thesecondary
antibodywaspreabsorbed (I h, 220C)with TTBScontaining1mg/ml
nonfatdrymilk, I mg/mlBSA,and Img/mlhumanIgG.
Recombi-nantbFGF(R&D Systems)servedas apositivecontrol.
Statistical methods
Groupswerecomparedusingachi-squaretestforproportions.
Results
Analytical
resultsof
bronchoalveolarlavage
Bronchoalveolar lavage was performed in critically ill,
me-chanically ventilated patients and patient controls without
complications. For the lung injurypatients,results for range
and median cell counts(RBC = range 0 to 24,000 cells/Mi,
median3,200
cells/,ul;
WBC =range 10to1,680cells/,Ml,
me-dian 140 cells/Ml), differential cell counts
(MO
=12+8%,
N=88±9%) aswell aslavage fluid totalprotein (689±397Mg! ml), and albumin(321±264 Mg/ml)concentrationswere simi-lar to previous reports (26, 27). (Values are reported as
mean±standard deviation.) The overallmortality ofour
pa-tient populationwas 44%.Therefore, ourstudygrouphad a
degreeoflunginjurycomparabletothose describedbyother
investigators. Allvaluesforpatientcontrols fell within estab-lishednorms(28) (datanotshown).
Histologic
andimmunohistochemical
resultsAll autopsytissues from fatalcasesofacute
lung injury
showeddiffusedamagetoalveolarliningepithelium,with foci oftype
IIepithelialcellhyperplasia. Theinterstitiumwaswidened
ir-regularly byloosefibrovascular
tissue,
which contained capil-lary-sizedbloodvessels,surroundedby bland, fusiform,fibro-blast-like cells.Remnants of intra-alveolar fibrin membranes werealso observed.
Multifocally,
theinterstitial fibrovascular(granulation)
tissueprotruded
intoalveoli, leaving only
anarrowrimof
airspace
around it.RareextravasatedRBCswereapparentatthese
sites,
as werescattered chronicinflammatory
cells. Therewas noevidenceof bronchiolitis obliteransor
in-flammatory
changeinlarge
pulmonary
blood vessels. Controlspecimensweremicroscopicallyunremarkablein all instances.
Immunohistochemically,
inARDSspecimens,small capil-lary-sizedblood vessels labeled by anti-factorVIII-relatedanti-gen
(Fig.
1)werepresentinintraalveolarfibrovascularprotru-sions. Inaddition, vimentinwasintenselyexpressed in
44A..I
.7:..
X,. ts4: i:s
A,
'Ki,k.
lw w
_A_~111
.. ._ .,
AL,&... ..
_Fv _ Fly
&~~~~~~~~~~~~~~~~~~~~~~~~~~~~~4
K *~~~~~
Figure 1.Immunoreactivity for factor VIII-relatedantigenwithin the endothelial cells of intra-alveolar capillaries inautopsylung from a fatal caseof ARDS. The capillaries (arrow) are supported by a loose fibrovascular matrix which protrudes from the interstitium into the airspace.
Shownisabudofgranulationtissuereplacingalveolarairspace(magnification, 250) (anti-factorVIII-relatedantigen/ABC Immunostain).
blast-like interstitial cells, vascular endothelium, and supporting stromal tissue of the lung, and it highlightedthe
contoursof smallcapillary-sizedblood vessels in intra-alveolar
granulation tissue. Specimens of normal control lung failedto
exhibit thepresenceof intra-alveolar fibrovascularprotrusions
orinterstitialgranulationtissue.Allexternaland internal
con-trolsstainedappropriately.
Evaluation
oflavage effluent for
angiogenesisbioactivity
Invitromigration. Endothelial cellmigrationtowardan
angio-genic stimulus isoneof the principalcellularprocesses
asso-ciated with new capillary growth in vivo.When lavagefluid frompatients withacutelunginjurywastestedfor itsabilityto
stimulate endothelialcell migration, samples from 18 of25
patientswithacutelunginjury (72%)stimulated themigration of pulmonaryarteryendothelial cells(Fig. 2 A).Itshould be notedthatnoroutinely assessed clinicalparameters,including disease severity, clinical course, or outcome, were found to
distinguishthosepatients' sampleswithangiogenesis bioactiv-ity fromthosewithout demonstrable bioactivity. In contrast, noneof the samples frompatientcontrolsstimulated endothe-lial cellmigration(P<0.01). Becausethelikelytargetcells in vivoaremicrovascularendothelialcells,asubsetofthepositive specimens stimulatingpulmonary artery endothelialcell mi-grationwasassessed forbioactivityusing capillaryendothelial cells. Ofnote,consistent with the resultsusing largevessel endo-thelialcells,allof the lavagesamplesfrompatientswithacute
lung injury (n = 9), and none of the samples from patient controls(n =4), stimulated microvascularendothelialcell mi-gration (P<0.01; Fig. 2 B). Checkerboard analysisrevealed thatendothelial cellmigrationinduced by lavage fluidwasdue
both to directedcell migration aswell asrandommigration
(Fig.
3).
300-la
-'a
"-
200-x
0
-j 100
.5
U 0
-A. Large Vessel
C
0
S
0 0 0
P~
4
au
.2
L N
0
0
V-C,
To
300-
200-
10
0-AcuteLung Nomals
Injuy (n-B)
(n-25)
B. Capillary
*
0
0
AcuteLung Normal.
Injuy (n-4)
(n-9)
Figure 2. Endothelial cell migration in response to bronchoalveolar lavage fluid. Undiluted bronchoalveolar lavage fluid obtained from
patientswith acute lunginjury or normal individuals was evaluated
for itsabilitytoinduceendothelial cellmigrationin amodified
Boy-denchamber apparatus. Shown is(A) large vessel and (B) capillary endothelial cellmigration quantifiedasthe numberof cellsmigrating
per IOOXfield.
Angiogenesis Bioactivity IsPresentinthe LungafterLungInjury 1389
0
.f.*
c
-J
.2
.'a: X
co 0
.ojo
Dilution of Lavage Fluid In Upper Chamber
M199 1/100 1/10 1/3
M1990\
0 0 201/100 32 30 25 40
1/10 153 153 110 150
1/3 278 248 213 173
.1
\
eachcombination oflavagefluid dilutior agonal representdirected cell migration.
Figure 3.Checkerboard
analysis of endothelial cellmigration in re-sponse to bronchoal-veolarlavage fluid from patients with acute lung injury.The upper and lower chamberof the assay apparatus was filledwith varying
dilu-tionsofravagefluid
from patientswith acute
lunginjury as indicated. Shownis the number of endothelial cells migrat-ing per IOOX field for as.The values below the
di-Anestimate of thequantityofthemigration bioactivityina
givensample wasprovidedbyevaluating endothelialcell mi-grationin response todefined dilutions of bronchoalveolar la-vagefluid. Undilutedlavagefluidstimulated maximum
endo-thelial cellmigration. Onaverage,the half-maximumincrease
inendothelialcellmigrationoccurredinresponse to a1:3 dilu-tion of lavage fluid. Ofnote, noneof thepatient controlfluids
demonstratedbioactivityatany of the tested dilutions. In
addi-tion,thepresenceof inhibitors ofendothelial cell migrationin
patient control sampleswasformally excludedwhen
quantita-tiverecoveryofbioactivitywasobservedusingtwoapproaches:
(a) addition of thepositivecontrol(platelet lysate) in parallel
assays,and(b) equalmixexperimentswith selectedacutelung
injurypatient samples (datanotshown). Likewise, in selected
acutelunginjury samplesinwhich endothelial cell migration
was notobserved, thelavage fluidwasstudiedusinganequal
mixexperimentin which thepositivecontrolwasaddedtothe lavagesample.Usingthis approach,significantinhibitionwas notobserved.
Invivoneovascularization.Confirmation of the angiogene-sisbioactivitydetectedinthemigrationassaywasobtainedby theability ofselectedbioactive samplestoinduce corneal
neo-vascularization. When concentrated lavage fluid from the
pa-tientsdemonstratingthegreatest amountof in vitro
chemotac-ticbioactivity(n=2)wasimplanted inthe avascular cornea of
rabbits, bothsamplesinduced capillarygrowth.Histologically,
inflammationcharacterized by accumulation of macrophages
into theinflammatory sitewasassociatedwith the new
micro-vessels,leavingopen thepossibilitythat thelavagefluid did not
induce angiogenesisdirectly, but rather by inducing an
inflam-matory response. Significantly, no growth ofnew vessels was
induced by lavage fluid from patientcontrols (n=2).
Biophysical properties ofthe angiogenesis bioactivity. Bio-physical characterization ofthe angiogenesis bioactivity
re-vealedpropertiesexpectedofapeptide(s). The bioactivity was
destroyed by boiling(
1000C,
15min),acidicconditions (pH 3),and protease exposure(trypsin 0.2mg/ml, 30
min,
370C).
Soy-beantrypsin inhibitorwas used to neutralize trypsin activity.
Whensoybean trypsin inhibitor was mixed equally with the
positive control, platelet releasate, chemotactic activity was preserved. In addition, the bioactivity was not extracted into
nonpolar solvents(TableI).These results served as the starting
pointforfurtherpurificationandcharacterization.
Table I. Biophysical Characterization of the Angiogenesis Bioactivity
Number of endothelial cells
Treatment migrating/lOOX field
No treatment 150
Aqueousphase 125
Lipid phase 0
Trypsindigestion 0
Acid (pH 3) 0
Heat 0
Biochemical characterization of theangiogenesis
bioactiv-ity. A characteristic of several extracellular matrix proteins
(e.g., fibronectin, laminin,andthrombospondin)that promote
cellmigration and adhesion and a class of endothelial cell
pep-tide growth factors (fibroblast growth factor family) is their
ability tobind to heparin. When bronchoalveolar lavage
ef-fluentwasappliedtoheparin-sepharose,30% of the angiogene-sisbioactivity boundatlow ionicstrengthandremainedbound
up to 1.5 MNaCl,asdetermined in thelargevessel endothelial
cellmigration assay. The bound bioactivity eluted from
hepa-rin at 2 M NaCl.
Among thecharacterized heparin bindingendothelial cell
growth factors, only bFGF binds avidly at 1.5 M NaCl.
There-fore,theheparinbinding bioactivity eluting at 2 MNaClwas
subjectedtoWestern blotanalysis usingananti-bFGF primary
antibody.An 1 8-kD bandcomigratingwith a bFGFstandard
(recombinant bFGF) was identified (Fig. 4). In addition, a
lower molecularweight (- 12to 13kD)bFGF
immunoreac-tiveband was present. NobFGFimmunoreactivitywas
demon-stratedintheheparin-sepharoseunboundfraction,norin any
of the fractions eluting at lower ionic strengths (data not
shown).
Themajority (70%)oftheangiogenesis bioactivityfailed to bindheparin,andwasevaluated for itsabilitytobindto
ion-ex-changeresins. 50%of thebioactivityboundtothe weak anion
Heparin-S Eluate
18kD
-12kD- By?
>.s.rlg
bFGF Standard
.v.*
'.T, Qis ''
.Ma,;A'
*.3s A.,
Ok'
an 18 kD
IF
W:'
Beg.;
;Ws; ::
,,,4...
Figure 4. Western blot
analysisof the heparin
bindingbioactivity.
Bioactive material elutedfrom heparin sepharose (2 MNaCI)
wassubjectedto
SDS-PAGE(15%)and
trans-ferredtonitrocellulose membranes.Shown are transferredproteins
rec-ognizedbyanti-bFGF polyclonalantibody.A
recombinant bFGF standardwasincluded
as astandard.
exchange resin (diethylamino ethyl sepharose; 0.05 M NaCl, 0.01 M Tris, pH 7.4), eluting at 0.25 M NaCl.
Chromato-graphic analysis ofthe bioactivity using a strong anion ex-changer (Mono Q column; Pharmacia) and a linear salt
gra-dientrevealed asingle peakof bioactivity eluting from the
col-umn at - 260mMNaCQ(Fig.5).
Gel filtration chromatography of the Mono Q peakwas
carriedout to further purifythebioactivity and to obtain an
estimateofits molecular weight. The endothelial cell migration bioactivity eluted just after the human IgGstandard, corre-sponding to an approximate molecular weight of 150 kD
(Fig. 6).
The biochemical characteristics (heparin affinity, ion
ex-changeproperties) of the angiogenesis bioactivitywere
exam-ined in relationto thetimepoint in theclinicalcourse that a
patientwasstudied. Thebiochemicalproperties ofthe angio-genesis bioactivitypresentin bronchoalveolar lavage fluid
ob-tainedfrom acute lung injury patients early in their clinical
course(days 2-4)wereidenticaltothose present in the lavage
fluid obtained frompatients later in their clinicalcourse (days 10-21).
Biologicalproperties ofthe 150-kD bioactivity.The molecu-larweight estimate of 150kDindicatedthat theangiogenesis bioactivitywaslarger than the majority ofcellular-derived
pep-tide growth factors. However, it fell within therangeofa
vari-etyof extracellular matrix protein and degradation products
that promote both endothelial cell migration and adhesion.
Therefore,the 150-kD bioactivitywasevaluatedfor its ability
topromoteendothelial cell attachment. When aplastic
sub-stratumwascoated withincreasing concentrations of the
150-kDpeptide, therewas adose-dependent increase in endothelial cell adhesion(Fig. 7). TypeIVcollagen and fibronectin, two
extracellular matrixproteinsthatpromote endothelial cell ad-hesion inadose-dependentfashion,wereusedaspositive
con-trols.Bovineserumalbumin,.whichdoesnotpromote endothe-lial cell attachment,wasusedas anegativecontrol.In mediat-ing endothelial cell attachment the150-kDpeptidewas at least
aspotent as type IVcollagen andoneto twologsmorepotent
thanintact fibronectin.
Basedontheseresults, itbecame apparentthat the endothe-lial cell migrationobserved inresponsetocrude lavagefluid
- 0.3
-0.2 '.
I
-0.1 z
z
0
ax
0
-0.3 s
002 oom
0
C--0.1I
-o
I I I I I I I
0 5 10 15 20 25 30 Retention Time (min)
Figure 5. Anion exchange chromatographyof theheparin sepharose
unboundbioactivity.Thebioactivity failingtobindheparinwas
ap-pliedtoaMonoQ column,whichwasdevelopedwithalinearNaCl gradient.Shownisendothelialcellmigrationinresponsetoundiluted aliquotsfromeach column fraction.
300 275 250
0. 225
I 200
x 175
° 150
d 125
=100
0
50 25 0
1+ 1
0 5 10 15 20 25 30 35 40 4550
Fraction No.
-0.6 -0.5 -0.4 -0.3 -0.2 0.1
Figure 6. Gel filtrationchromatographyof thebioactivity.A500-Al
aliquotfrom the MonoQ peakofbioactivitywasappliedto a
Super-ose 12column anddevelopedinPBS, pH 7.4, 10% acetonitrileat0.5 ml/min. Shown isendothelialcellmigrationin responsetoundiluted
aliquots from each column fraction. Theelutionprofileoftwoofthe molecularweightstandards is indicated:IgG (166 kD)and human serum albumin(66 kD).
mighthave beendependenton thefibronectin usedtocoatthe polycarbonate filters. Therefore, the migrationassay was
re-peatedusingthe 150-kDbioactivity andfilters rinsed in PBS
but notcoated with fibronectin. Significantly,
concentration-dependent endothelial cellmigration occurredin responseto
the 150-kD bioactivity, indicating that the observed effects
were notdependenton thepresenceof exogenous fibronectin.
100 90 80 70 60 50 40 30 20 10 0
0.01 0.1 1 10 100
Coating Concentration
(jig/ml)
Figure7. Endothelial cell attachment inresponsetothe 150-kD
bioactivity.Tissueculture wellswerecoated with solubletestproteins
orcontrols andwereexaminedfor theirabilitytomediate the
attach-mentofradiolabeled endothelial cells. Shown is the endothelialcell adhesion inresponsetoincreasingconcentrations of the 150-kD
pep-tide(o); positivecontrolproteins,typeIVcollagen(o) and fibronectin (A);orthenegativecontrolprotein, BSA (i).Datais expressedas
percentage adhesion=[DPM attached/DPM added]x 100.
75
-I
la
.2
50-l.
x
0 0
25-C 25
-=
To plausibly be involved in neovascularization of the
air-space after acute lunginjury, the 150-kDbioactivity should
promote new capillary growth in vivo. When , 10 ngof the
150-kD bioactivity was incorporated into Hydron and
im-plantedintheavascular cornea ofrabbits, new capillary growth was observed. Vessels were observed to be growing from the
limbusto theHydron pellet containing the 150-kD
angiogene-sisbioactivityimplanted withinthe corneal stroma. Notably,
histological examination ofthe corneas on day 2 and 7 did not
show thepresence ofaninflammatoryinfiltrate.Thus,partial purification ofthe angiogenesis bioactivitypresentinthecrude
lavage fluid removed proinflammatory molecules but
pre-served theangiogenesis activity. (Fig. 8, Aand B).
Discussion
Replacement ofthe alveolarairspace with granulation tissue is the centralmorphologicfeaturecharacterizingthe lungsof
pa-tientsdying afteracutelunginjury (4). Angiogenesis,aprocess
consistingofendothelial cellmigration andreplication in
re-sponse to aspecificsetofexogenoussignals, isanintegral
fea-tureof thisgranulation tissue formation. This study
demon-strates the presence of growth regulatory signals recovered
from the air-lung interface that have theabilitytomodulate endothelial cellfunction.Analysisof theangiogenesis
bioacti-vityreveals the presenceoftwodiscrete moieties: (a)an 18-kD
heparin binding factor thatappears to be identicaltobFGF; and(b)a 150-kDfactor thatpromotesendothelial cell migra-tion and adhesion inadose-dependentfashion invitro, and induces theformationof microvessels in vivo.
Wound repair:
aparadigm
for studying
thegranulation
tissue responseTheprimary tissuehealingresponse towoundingisthe forma-tion ofgranulation tissue followed by epithelialization (29).
Whenitoccursafterasurface wound,theformation of
granu-lation tissue isanadaptiveresponse,bringingabout closureof the wounded integument. However, when this highly
con-servedresponse totissueinjuryoccursafterdisruptionof the alveolarwall, the alveolarairspacefillswithgranulationtissue, obliteratingthegasexchangeapparatus. Inthisanatomic loca-tion, granulationtissue representsa maladaptiveresponseto
tissueinjury,preventingeffectivegasexchange.Despiteits
ad-verseeffectonlungfunction,theanatomicsimilarityof
intra-alveolargranulationtonormal woundhealingsuggests thatthe extensive body of experimental information pertaining to
woundhealingmayprovideinsight intothe processesleading
tointra-alveolar fibrosisfollowing lunginjury.
Inthe lungasintheskin,granulationtissue developsina
dynamic inflammatory milieu.
Anatomically,
the injuredspaceisrapidlyfilled withthecontentsof
degranulating
plate-lets as well asplasmaproteins (30, 31).Thesediverse elements
develop into aprovisional solid phase matrix. Migrationof mesenchymal cells and microvascular endothelial cells into the
wound space ensues, followed by formation ofcapillary
net-works anddeposition of connective tissue elements (30-35). The anatomic precision ofgranulation tissue development likely depends on an orderly sequence of soluble and solid phase molecularsignalscoordinated bothspatiallyand
tempo-rally. Thismayexplain whyno apparentrelationshipbetween diseaseseverityorclinicalcourseand thepresenceof
angiogen-esis activity was noted.Inaccord with the woundhealing
para-digm,this lack ofrelationship likely reflectsthebiasintroduced
by samplingahighlycoordinatedresponseat asingle pointin
time. Longitudinal studiesin whichsamplesare sequentially obtained throughout the clinical course oflung injury will
likely berequired to showarelationship between angiogenic activityand theultimateoutcome.Nevertheless,this studyhas
identified two signalsthat mayhave a role in the integrated response thatleadsto new blood vessel formation.
The process of angiogenesis,which is highlycoordinated, is
likely to be under the control ofmultiple molecular signals
(36). Newvessel formation involves themigrationof
endothe-lialcells toward achemotacticstimulus. The migrating
endo-thelial cells maythen become aligned withoneanotherand
undergoreplicationtoformasolidsprout(36-39). Basedon
this information and the results of this study it is useful to
consider these signals intwogroups:first, signalsthatfunction
asclassical peptide growth factors (e.g., bFGF), and second, signals thatmayfunction in solid phaseasmediators ofchemo-taxis andadhesion (e.g., the 150-kD peptide).
Peptide growthfactors capable ofstimulating endothelial cellgrowth, migration and/or angiogenesis. A defined set of peptide growth factorsarecapable ofstimulatingendothelial cellgrowthormigrationand/orangiogenesis. In general they derive from both recruitedinflammatory cellsaswellas
paren-chymal. cells and include platelet-derived endothelial cell growth factor (40-42), andmacrophage-derived transforming
growthfactoralpha(TGFa) (43, 44). The relative importance
of thesetwoangiogenesis factorstothenewvesselgrowth ob-served afteracutelunginjuryremains uncertain.
Oneof thetwomoieties identifiedatthe alveolar surface in thisstudy, bFGF, has been reportedtostimulate the replication ofendothelial cells in vitro andnewmicrovesselgrowthin vivo
(45,46).However, recentinvestigation has failedtoshow
signif-icantchemoattractant or in vivo angiogenic activity using
re-combinant bFGF (47).Theresultsofourinvestigation indicate thatlavagefluidelutedfrom heparin-sepharose with2 M NaCl promotes large vessel endothelial cell migration and contains
bFGF as identified by Western analysis. However, because
bFGFwas notpurified, itsrole in endothelial cell migration
remains inferential. Data linking bFGF with the process of
woundhealing include itsability to cause rapid
neovasculariza-tion in thecorneaandskin,aswell asfibroplasia inthedermis
after wounding(45, 48-50). These observationssupport the
notionthatbFGFisinvolved in neovascularizationand wound
repair.
Theorigin ofbiologically activebFGF inwounded tissue is
incompletelyunderstood.Accordingto onehypothesis, bFGF produced by endothelial cells is released and sequestered in the
basement membrane bound toheparansulfate (andpossibly otherglycosaminoglycans)(51-54).Intheinflammatory
envi-ronmentofwoundedtissue,bFGF storedwithinthe basement
membraneisreleasedenzymatically(e.g.,plasmin or
heparin-ase) where itcanact on targetmicrovascular cellstopromote
angiogenesis (53).While the origin ofthe bFGF recovered from
thealveolarepithelialsurface afteracute lunginjury remainsto be determined, this study provides clear evidence that this knownendothelialcell growth factor ispresentat asitewhere newbloodvessels areforming.
Moieties mediating endothelial cell migration and
adhe-sion. Asignificant portion oftheidentifiedangiogenesis
bioac-tivityrecoveredfrom thelungs ofpatients afteracutelung in-jury isa 150-kDmoiety thatservesas a potentendothelial cell
viS
^ : ..
a A: 85 4 tso Xe so e; Bus:
iwa::^
>:<.i:::::::,:..:.
...^:
::::s.:::i::::i.
Or.;::..
. :: - .; <>:
r' I+>: '; &g5 ; ,',
e" add' ;<>'<>'?1 ::' ax,. sir - _. , Or,Ji, *s&t,<5&. a> <ws
;-lSS He
|'I,.;.:.:..<<.i.,2Si='^r#.;Z!<...He:-.:::.W
sP;-A
!*_,sm^
i""" "'v""-...^i':
'' . ... *1 S Ad hi. :i.P.512g-U^U'...s, ,sgL..H
.. i .Or .,. :, '.'.'A. ., ., i-tiSS t;,,so?-: :: ..A.. ,.:.
E.,,?.2...,....its0
|AndeeSse;
insI:RfsaY:ii..:tj;#.,,._...
;_ !Xx i.-' 9,_. !' :,Figure 8. Corneal neovascularization. The 150-kDpeptide was incorporated into sterile Hydron and implanted in the avascular cornea of a rabbit. At day7the eyes were examined. (A) In situ photograph showing new capillary growth (arrow). (B) Histology of cornea displaying new vessel growth (arrow) and lack of inflammatory cell infiltrate, H and E, magnification, 200.
the LungafterLungInjury 1393
ka
migration factorand adhesogen. In addition, it promotes in vivo angiogenesis without causing inflammation. Its biochemi-caland biological propertiessuggestthat it belongs to the
fam-ily of extracellular matrixelementsthat promote endothelial cell attachment and migration. Several studies support the con-cept that components ofthe extracellular matrix and basement
membrane areimportant modulators of those endothelialcell
functions necessary for wound healing and new vessel growth
and includeType IVcollagen, thrombospondin, fibronectin,
and/orfragments of fibronectin (22, 55-63).
Granulation
of
thealveolar
airspaceafter
acutelung
injury: anhypothesis
Morphological studies of lung tissue after acute lung injury have documented thedevelopmentandprogression
ofgranula-tion tissueasit obliterates the alveolarairspace.The anatomic
similarity of thisprocess to woundhealingisstriking.
Disrup-tionof the endothelial andepithelialcelllayersand their
base-mentmembranesresultsindirect communicationbetween the
microvascular bed, interstitium, and alveolar airspace. As a
consequenceof thisdisruption, anintra-alveolarexudate rich
in plasma proteins and cells, including platelets and
macro-phages, accumulates in the alveolarairspace(64).As the pro-cessevolves, theintra-alveolar exudate becomesorganized by themigrationofmesenchymalcellsand endothelial cells from theinterstitiumthroughrentsinthedisrupted epithelialbarrier
into theairspace,resultingintheformation ofgranulation tis-suewithin the airspace.
Based on thewoundhealing paradigm, both cellular and
matrix derived proteinproducts arecapable of providing
sig-nalsfor neovascularization ofthe alveolarairspace afteracute
lunginjury. One of these signals identified in thisstudy is the
endothelial cell growthfactor bFGF. Possiblecellular sources
for bFGFinclude the alveolar macrophage and microvascular endothelial cells.Inaddition, it is possible that bFGF boundto
basement membranesis released by molecules generated by the explosiveinflammatory responseassociated with earlyphases
ofacutelunginjury.Alargeportion ofthe angiogenesis
bioacti-vity withinthe alveolarairspace afteracute lunginjuryis an
-1 50-kDpeptide. Althoughtheprecisechemicalstructureof
the 150-kDmoleculeangiogenesis factor remainsto be
estab-lished, its ability to promote cell attachment in addition to
migrationandangiogenesis in vivo,suggests thatitbelongs to
thefamily of adhesogenicextracellularmatrix elements. Can-didate molecules include fragments ofcollagen (type IV) or
fibronectin. Alternatively,itmaybe relatedtothecell-derived protein thrombospondin. Collectively these molecules share with the150-kDangiogenesis factortheabilityto promote
en-dothelial cell adhesion. Purificationtohomogeneitywill be
re-quiredforadefinitiveanswer toitsidentity.Independent ofthe precise identity of this biologically-active moiety,this
investi-gationsupportstheconclusion that neovascularizationof the airspace, stimulatedby atleasttwodistinct peptide moieties,
maybeoneofthe central processesinthe pathogenesis
ofintra-alveolargranulation afteracutelung injury. Acknowledgments
Theauthors thank BarbaraPennerand MaryChelberg for their excel-lenttechnical assistance in the chemotaxis assay and adhesion assay,
respectively. In addition, we thank Sylvia Danielson and Funmi
Thompson for their excellent secretarial assistance.
This work was supported by grant F32 HL-08051-02 (to Dr. Henke) and grant ROI HL-39833 (to Dr. Bitterman) from the Na-tional Institutes of Health.
References
1.Rinaldo, J. E., and R. M. Rogers. 1982. Adultrespiratorydistress syn-drome.N.Engl.J.Med. 306:900-909.
2. Bachofen, M., and E. R.Weibel. 1977. Alterations of the gasexchange
apparatus inadult respiratory insufficiency associated with septicemia. Am. Rev. Resp. Dis. 116:589-615.
3.Tomashefski,J. F., P.Davies, C.Boggis, R. Greene, W. M. Zapol, and L. M.Reid.1983. The pulmonary vascularlesions ofthe adult respiratory distress
syndrome.Am. J.Pathol.112:112-126.
4.Fukuda, V., M. Ishizaki, Y. Masuda, G. Kimura, 0. Kawanami, and Y.
Masugi. 1987. The role of intraalveolarfibrosisin the process of pulmonary structural remodeling in patients withdiffealveolardamage. Am. J. Pathol. 126:171-182.
5.Hasleton, P. S. 1983. Adult respiratory distress syndrome-a review.
Histo-pathology(O04) 7:307-332.
6.Folkman, J.,and M. Klagsbrun. 1987.Angiogenicfactors. Science(Wash. DC). 235:442-447.
7.Zetter, B. R. 1988.Angiogenesis:stateof the art Chest. 93:1595-1665. 8.Petty, T. L. 1982.Adultrespiratorydistresssyndrome:
definitionandhistor-ical perspective.Clin. Chest Med. 3:3-8.
9. Pepe, P. E., R. T. Potkin, D. H. Reus, L D. Hudson, and C. J. Carrico. 1982.Clinical predictorsofthe adult respiratorydistresssyndrome.Am.J.Surg. 144:124-129.
10. Hsu, S. M.,L. Raine, and H. Fanger. 1981.Useof
avidin-biotin-peroxi-dasecomplex(ABC)inimmunoperoxidase techniques:acomparisonbetween
ABC and unlabeled antibody (PAP) procedures. J. Histochem. Cytochem. 29:577-580.
11.Saltini,C., A. J. Hance, V. J. Ferrans, F.Bassett,P. Bitterman, and R. G. Crystal. 1984. Accurate quantification of cells recovered by bronchoalveolar
la-vage. Am. Rev.Respir. Dis. 130:650-658.
12. Davis, W. B., S. L Rennard, P. B. Bitterman, and R. G. Crystal. 1983. Pulmonary oxygen toxicity: demonstration of early, reversible biologic changes in the human alveolar structures induced by hyperoxia. N. Engl. J. Med. 309:879-883.
13. Postlethwaite, A. E., R. Snyderman, and A. H. Kang- 1976. The
chemo-tacticattractionofhumanfibroblaststo alymphocyte-derivedfactor. J. Exp.
Med.144:1188-1203.
14.Banda, M. J., D. R. Knighton, T. K. Hunt, and Z. Werb. 1982. Isolation of anonmitogenic angiogenesis factor from wound fluid. Proc. Nadl. Acad. Sci.
USA. 79:7773-7777.
15.Stein, O., andY. Stein. 1980.
Bovineaorticendothelialcellsdisplaymacro-phage-likeproperties towards acetylatedl2SI-abelledlow densitylipoprotein.
Biochim. Biophys. Acta. 620:631-635.
16. Voyta, L. C., P. A. Netland, D. P. Via, and B. R. Zetter. 1984. Specific labelling of endothelial cells using fluorescent acetylated-low density lipoprotein.
J. Cell Biol. 99:81a. (Abstr.)
17.Gimbrone, M. A., R. S. Cotran, S. B. Leapman, and J.Folkman.1974. Tumorgrowth and neovascularization: anexperimentalmodel using the rabbit cornea. J.Nail. Cancer Inst. 52:413-419.
18. Shing, Y., J.Follaman,R.Sullivan, C. Butterfield, J. Murray, and M. Klagsbrun. 1984.Heparin affinity:
purificationofatumor-derivedcapillaryendo-thelialcellgrowthfactor.Science(Wash. DC). 223:1296-1299.
19.Gospodarowicz,D., J.Cheng, G.Lui,A.Baird,and P.B6hlent. 1984.
Isolation of brain fibroblastgrowthfactorbyheparin-sepharoseaffinity
chroma-tography:identifywithpituitary fibroblast growth factor. Proc. Natl. Acad. Sci. USA. 81:6963-6967.
20.Maciag,T., T. Mehlman, and R. Friesel. 1984. Heparin binds endothelial cellgrowthfactor,theprincipal endothelialcellmitogeninbovine brain.Science
(Wash. DC). 225:932-934.
21. Bowersox, J. C., and N. Sorgente. 1982.Chemotaxisofaortic endothelial cells inresponsetofibronectin.Cancer Res.42:25-47.
22.Herbst,T.J., J. B.McCarthy, E. D. Tsilibary, and L. T. Furcht. 1988.
Differentialeffects oflaminin,intact type IV collagen, andspecific domains of type IV collagen onendothelial cell adhesion andmigration. J. CellBiol. 106:1365-1373.
23. McCarthy, J. B., S. T. Hagen, and L. T. Furcht. 1986. Humanfibronectin
containsdistinctadhesion- andmotility-promotingdomains formetastatic
mela-noma cells. J.Cell Biol.102:179-188.
24.Laemmli,U.K.1970.Cleavage ofstructuralproteinsduringtheassembly of the headof bacteriophage T4.Nature(Lond.).227:680-685.
25.Wadzinski,M.G., J. Folkman, J.Sasse,K.Devey, D.Ingber, and M. Klagsbrun.1986.Heparin-bindingangiogenesisfactors:
detectionbyimmunolog-ical methods.Clin.Physiol.Biochem.5:200-209.
26.Holter,J.F.,J. E.Weiland,E. R.Pacht,J. E.Gadek,and W. B.Davis. 1986. Proteinpermeabilityin the adultrespiratorydistress syndrome.Lossofsize
selectivityof the alveolarepithelium.J.Clin. Invest. 78:1513-1522.
27.Weiland,J.E.,W. B.Davis,J. F.Holter,J. R.Mohammed,P.M.
Dorin-ski,and J. R. Gadek. 1986.Lung neutrophilsin the adultrespiratorydistress
syndrome:clinical andpathophysiologicalsignificance.Am. Rev.Respir.Dis.
133:216-225.
28.Cherniak,R.M.(Coordinator).1990. TheBALCooperativeGroup
Steer-ingCommittee.Bronchoalveolarlavageconstituentsinhealthy individuals,
idio-pathic pulmonary fibrosis,and selectedcomparisongroups.Am.Rev. Respir.
Dis. 141(Suppl):S169-S202.
29.Leibovich,S.J.,and D. M.Wiseman. 1988.Macrophages,woundrepair
andangiogenesis. Prog.Clin. Biol.Res.266:131-146.
30.Hunt,T.K. 1976. WoundHealing:DisordersofRepair. Chirurgecom, Inc.,SouthPlainfield,NJ. 99pp.
31.Knighton,D.R.,K. F.Ciresi,V. D.Fiegel,L.L.Austin,and E. L. Butler.
1986.Classificationandtreatmentofchronicnonhealingwounds. Ann.Surg.
294:322-328.
32.Leibovich,S.J.,and R. Ross. 1975. The role ofthemacrophagein wound
repair.Astudywithhydrocortisoneandanti-macrophageserum.Am. J.Pathol.
78:71-100.
33.Knighton, D.R.,and V. D.Fiegel. 1989.Macrophage-derivedgrowth
factors in woundhealing: regulationofgrowthfactorproduction bythe oxygen
environment.Am.Rev.Respir.Dis.140:1108-1111.
34.Hunt,T. K.,D.R.Knighton,K.K.Thakral,W. H. GoodsonIII,and W.S. Andrews. 1984. Studiesoninflammationandwoundhealing: angiogenesis
andcollagen synthesisstimulatedinvivobyresident andactivatedwound
macro-phages. Surgery.96:48-54.
35.Ross,R.1968.Thefibroblastandwoundrepair.Biol.Rev.43:51-96.
36.Schoefl,G. I. 1963. Studiesoninflammation.Growingcapillaries:their
structureand permeability. Virch. Arch. Pathol. Anat. Physiol. Kin. Med.
337:97-141.
37.Ausprunk,D.H.,and J. Folkman. 1977.Migrationandproliferationof
endothelial cells inpreformedandnewlyformedblood vesselsduringtumor
angiogenesis.Miscrovasc. Res. 14:53-65.
38.Scholley,M.M.,G. P.Ferguson,H. R.Seibel,J. L.Montour,and J. D.
Wilson.1984.Mechanismsofneovascularization.Vascularsproutingcan occur
withoutproliferationofendothelialcells.Lab.Invest.51:624-634.
39.Presta,M.,and D. B.Rifkin.1988. Newaspects ofblood vesselgrowth:
tumorandtissue derivedangiogenesisfactors.Haemostasis. 18:6-17.
40.Miyazono,K.,T.Okabe,A.Urabe,F.Takaku,andC.H.Heldin.1987.
Purificationand propertiesofanendothelialcellgrowth factor from human
platelets.J.Biol.Chem.262:4098-4103.
41.Miyazono,K.,and C. H.Heldin. 1989.High-yield purificationof
platelet-derivedendothelial cellgrowthfactor:structuralcharacterizationand
establish-mentofaspecificantiserum.Biochemistry.28:1704-1710.
42.Ishikawa, F.,K.Miyazono,U. Hellman,H.Drexler,C.Wernstedt,K.
Hagiwara,KUsuki,F.Takaku,W.Risau,and C. H.Heldin.1989.Identification
ofangiogenicactivityand thecloningandexpressionofplatelet-derived
endothe-lialcellgrowthfactor.Nature(Lond.).338:557-562.
43.Schreiber,A. B.,M. E.Winkler,and R.Derynck. 1986.Transforming
growthfactor-a:a morepotentangiogenicmediatorthanepidermal growth
fac-tor.Science(Wash.DC).232:1250-1253.
44. Rappolee,D. A.,D.Mark,M. J.Banda,and Z. Werb. 1988. Wound
macrophagesexpress TGF-a and othergrowthfactorsin vivo:analysis bymRNA
phenotyping.Science(Wash.DC).241:705-712.
45.Shing,Y.,J.Folkman,C.Haudenschild,D.Lund,R.Crum,and M.
Klagsbrun. 1985.Angiogenesisis stimulatedbyatumor-derivedendothelial cell
growthfactor. J. Cell.Biochem.29:275-287.
46.Joseph-Silverstein, J.,and D. B. Rifkin. 1987.Endothelial andgrowth
factor and the vessel wall. Semin.Thromb.Haemostasis. 13:504-513. 47.Knighton, D. R., G. D.Philips, and V. D.Fiegel. 1990. Woundhealing
angiogenesis: indirect stimulationbybasic fibroblastgrowthfactor. J. Trauma. 30:S134-S144.
48.Dividson, J. M., M. Klagsbrun, K. E. Hill, A. Buckley, R. Sullivan, P. S. Brewer, and S. C. Woodward.1975. Accelerated woundrepair,cellproliferation,
andcollagen accumulationareproducedbyacartilage-derivedgrowth factor. J. CellBiol. 100:1219-1227.
49.Hayek, A., F. L. Culler, G. M.Beattie,A.D.Lopez, P. Cuevas, andA.
Baird.1987. An in vivo model for study of the angiogenic effects of basic fibro-blastgrowth factor. Biochem.Biophys.Res.Commun. 147:876-880.
50. Fredji-Reygrobellet, D., J. Plouet, T. Delagre, C.Baudouin, F. Bourret, and P. Lapalus. 1987. Effects of aFGF and bFGF on woundhealinginrabbit corneas.Curr.Eye Res. 6:1205-1209.
51.Vlodavsky, I., J. Folkman, R. Sullivan, R. Fridman, R. Ishai-Michaeli, J. Sasse, and M. Klagsbrun. 1987.Endothelial cell-derived basic fibroblast growth factor.synthesisanddeposition into subendothelial extracellular matrix. Proc. Nati. Acad. Sci. USA. 84:2292-2296.
52. Folkman, J., M. Klagsbrun, J. Sasse, M. Wadzinski, D. Ingber, and I. Vlodavsky. 1988. Aheparin-binding angiogenicprotein-basic fibroblast growth factor-is stored within basement membrane. Am. J. Pathol.130:393-400.
53.Klagsbrun, M., and I. Vlodavsky. 1988.Biosynthesisand storage of basic fibroblastgrowth factor (BFGF) by endothelial cells: implication for the mecha-nismof angiogenesis. Prog.Clin. Biol. Res. 116:55-62.
54.Sommer, A., and D. B. Rifkin. 1989. Interaction of heparin with human basic fibroblast growth factor. protection of the angiogenic protein from proteo-lytic degradation by a glycoasminoglycan. J. Cell. Physiol. 138:215-220.
55.Saksela,O., and D. B. Rifkin. 1990. Release of basic fibroblast growth factor-heparan sulfate complexes from endothelial cells by plasminogen activa-tor-mediated proteolytic activity. J. Cell Biol. 110:767-775.
56.Mosher, D. F. 1990. Physiology of thrombospondin. Annu. Rev. Med. 41:85-97.
57. Mosher, D. F., M. J. Doyle, and E. Jaffe.1982. Synthesis and secretion of
thrombospondinby cultured human endothelial cells. J. Cell Biol. 93:343-348. 58. Lawler, J., R. Weinstein, and R. 0. Hynes. 1988. Cell attachment to thrombospondin: the role of arg-gly-asp, calcium, andintegrinreceptors. J. Cell Biol. 107:2351-2361.
59.Murphy-Ullrich,J.E., and M.Hodk. 1989.Thrombospondinmodulates
focaladhesions in endothelial cells. J.CellBiol. 109:1309-1319.
60.Raugi,G. J., J. E.Olerud, and A. M. Gown. 1987. Thrombospondin in
earlyhuman woundtissue. J. Invest.Dermatol. 89:551-554.
61.Schwarz, U., D. Schuppan, I. Oberbaumer, R. W. Glanville, R. Deutz-mann, R.Timpl, and K. Kiihn. 1986. Structure of mouse type IV collagen. Eur. J.Biochem. 157:49-56.
62.Ruoslahti,E. 1988. Fibronectin and its receptors. Ann. Rev. Biochem.
57:375-413.
63.Clark, R.A.F., P.DellaPelle,E.Manseau,J.M.Lanigan,H. F. Dvorak, and R. B. Colvin. Blood vessel fibronectin increases inconjunctionwith endothe-Hal cellproliferationandcapillary ingrowthduringwoundhealing.J. Invest. Dermat.79:269-276.
