Acellular Vascular Matrix:A Natural Endothelial Cell Substrate

Stephen G. Lalka, MD, Lisette M. Delker, BS, James M. Malone, MD,Raymond C. Duhamel, PhD, Melissa A. Kevorkian, BS, Beth A. Raper, BA,J. Craig Nixon, MS, Karen J. Etchberger, PhD, Michael C. Dalsing, MD,Dolores F. Cikrit, MD, Joseph L. Unthank, PhD, Malcolm B. Herring, MD,Indianapolis, Indiana, Phoenix, Arizona, and Tucson, Arizona

A preliminary assessment was made of the acellular vascular matrix graft as asubstrate for endothelial cell seeding, with respect to surface pretreatment (noneversus fibronectin and/or serum) and presence of exogenous growth factor. Arterieswere harvested from greyhounds and exposed to a sequential detergent extractionprocess to produce the acellular vascular matrix. Human umbilical vein endothelialcells were grown In tissue culture, harvested in first passage, then seeded at 105

cells/cm2 on sections of acellular vascular matrix and on gel-coated polystyrenepositive controls. After 18 hour incubation, endothelial cell-seeded acellular matriceswere fixed and processed for histologic and planimetric analysis; control wells werefixed and endothelial cells were counted by planimetry. Pretreatment of the acellularvascular matrix was found to have no effect on the percentage of endothelial cellcoverage of the matrix. There was significantly better endothelial cell coverage of theacellular matrix than on matched gel-treated polystyrene control wells. Withdrawal ofgrowth factor resulted in a significant reduction in endothelial cell coverage for allacellular vascular matrix groups. Growth factor withdrawal also significantly reducedattachment of endothelial cells on gel-treated polystyrene. Cell surface area wassignificantly smaller when growth factor was withdrawn from all groups except fromthe acellular vascular matrix without pretreatment. We conclude that: (1) the acellularvascular matrix is conducive to endothelial cell adherence and spreading evenwithout pretreatment; and (2) sudden withdrawal of exogenous growth factor mayimpair early coverage of substrates by endothelial cells due to an effect on the)radherence or spreading.

KEY WORDS: Acellular vascular matrix; endothelial seeding.

The efficiency of attachment, retention, andgrowth of seeded endothelial cells (EC) is primarilydependent on the composition of the substrate onto

From the Department of Surgery, Indiana UniversitySchool of Medicine, Indianapolis, Indiana; DepartmentofSurgery, Maricopa Medical Center, Phoenix, Arizona;Department of Pharmacology, University of Arizona,Tuscon, Arizona; Endoteclt Corporation, Indianapolis,Indiana,' and Department ofSurgery, St. Vincent Hospi­tal, Indianapolis, Indiana.

which they are seeded [1-4]. The extracellular ma­trix beneath the EC has an integral role in determin­ing EC polarity, orientation, morphology, and re-

Presented at the Third Symposium on Endothelial Seed­ing, October 20-21, 1988, Indianapolis, Indiana.Reprint requests: Stephen G. Lalka, MD, Indiana Uni­versity Medical Center, Wishard Memorial Hospital,1001 West uut. Street, Indianapolis, Indiana 46202.

/HUVEC~

(+)RDGF (-)RDGF

I I10S/cm2 lO S/ c ni2

A AGEL

If&. ~ TREATED ~ I!f!!JJ..~ W/fI (+)CONTROL W/fI Wi?

VOLUME 3No 2 - 1989

GROUPS HV

ACELLULAR VASCULAR MATRIX

GROUPS V-VIII

109

Fig. 1. Protocol for endothelial seeding of matchedpositive polystyrene controls.

sponse to growth factors [5-7]. The more closely asubstrate for EC seeding resembles the native sub­endothelial basement membrane (SEBM), the morelikely EC on it will display normal phenotype.

The acellular vascular matrix (AVM) graft, de­rived by sequential detergent extraction of nativevessels, is a non-immunogenic bioprosthesis com­posed of collagen and elastin with a luminal lining ofbasement membrane containing Type IV collagen,fibronectin (FN), and laminin [8-11]. Such a sub­strate should, by virtue of its physicochemicalstructure resembling native SEBM, promote attach­ment, retention and proliferation of seeded Ee.This similarity to SEBM may account for the factthat the AVM graft is the only reported non-EC­seeded prosthetic conduit that has become com­pletely lined in vivo with true endothelium withouttransmural ingrowth [8,12-14].

Combining the techniques of EC-seeding withAVM extraction should result in a prosthetic con­duit . that more rapidly becomes confluent withseeded EC than could be achieved using eithertechnique separately. This paper reports a his­tologic evaluation of AVM and the initial seedingstudies, examining the effect of surface pretreat­ment with fibronectin [1] or serum [IS], and growthfactor withdrawal [16] on EC coverage of AVM.

MATERIALS AND METHODS

Preparation of AVM

Adult greyhounds (25-35 kg) of both sexes wereused as blood vessel donors. The animals werepremedicated with intravenous xylazine (I mg/kg),anesthetized with intravenous pentobarbital sodium(25 mg/kg) , intubated and placed on a volumeventilator (Harvard pump). Femoral, iliac, abdom­inal aortic, thoracic aortic, subclavian and carotid

a

b

Fig. 2. (a) Plexiglass seeding chamber. (b) AVMpositioned in seeding chamber.

arteries were mobilized through standard surgicalexposure and excised between clamps. The vesselswere immediately rinsed in cold heparinized salinethen cut longitudinally to yield flat specimens.Within 30 minutes of harvest, vessels began incu­bation at room temperature in sequential baths(average 35 ml solution per gram tissue) in rectan­gular covered plastic containers (500-750 ml) whilegently shaking on a rocking agitator.

The first extraction solution consisted of: a non­ionic, nondenaturing detergent, Triton X-too (1%);a serine protease inhibitor, phenyl methyl sulfonylfluoride (PMSF, 0.4 mM); a metalloenzyme prote­ase inhibitor, sodium ethylene diamine tetraacetate(EDTA, 5 mM); and a bacteriostatic agent, sodiumazide (0.02%). The Triton X-IOO solubilizes plasmamembranes and renders nuclear membranes perme­able but will not solubilize nuclear membraneswithout the mechanical agitation. Since Triton X­100 is non-denaturing, the enzymes released fromcells are still activated, requiring addition of PMSF

110 ACELLULAR VASCULAR MATRIX ANNALS OFVASCULAR SURGERY

Endothelial cell culture

-----.. SEM ....--_...+

PLANIMETRY

I

tLM

Fig. 4. Protocol for histologic evaluation of AVM.

EC in CM were cultured at 37°C in a 5% CO2 ,

100% humidity environment on 60 mm gelatin­treated tissue culture dishes fed twice weekly withpassage to T-75 flasks when confluent (approxi­mately one week). The EC used for seeding ofAVM in each of the eight experiments were derivedfrom different umbilical veins. However, all ECused were harvested from T-75 flasks in first pas­sage and only if viable, free of contamination andgreater than 80% confluent as determined by in­verted phase-contrast microscopy.

EC were harvested with 2 ml of 0.17% trypsinafter two washes with 12ml of PBS . The trypsin-ECmixture was centrifuged (610 rpm x 5 min) afteraddition of 4 ml of FBS to inactivate trypsin. Thepellet was resuspended in 8 ml of7:1 PBS-FBS, 0.2ml was removed for celI counting, and the remain­der re-centrifuged (as above). The final pelIet wasresuspended in 5 ml CM, but in four of the eightexperiments RDGF was deleted from the CM usedto resuspend the EC; these EC were subsequentlyused to seed the experimental and control plates inGroups V-VlIl (to test EC coverage after the with­drawal of growth factor).

EC were harvested from human umbilical vein byenzymatic digestion with 0.133% collagenase (0.26ml/cm vein) with gentle fingertip massage. The veinwas then .flushed with 50 ml of 4: I solution ofphosphate-buffered saline (PBS)-heat inactivatedfetal bovine serum (FBS) to inactivate the collage­nase. After centrifugation (3100 rpm at 4°C for 10min), the EC were resuspended in 10 ml of 9: 1PBS-FBS and re-centrifuged (l100 rpm, 4°C x 10min). The final pellet was resuspended in 5 ml ofcomplete medium (CM), composed of culture me­dium 199, 10% FBS, 1% sodium pyruvate, 1% non­essential amino acids, 1% L-glutamine, 0.1% in­sulin-selenium-transferrin serum supplement, 1%penicillin-streptomycin, and 10 JLlIml retinal-de­rived growth factor (RDGF).

Endothelial cell harvest

and EDTA inhibitors at pH 8 (which inhibits theacid proteases, such as lysosomal enzymes). Thevessels were placed in fresh solution hourly for atotal of four hours for the initial detergent extrac­tion; the frequent exchanges wash away intracellularproteases released as cells are lysed by the detergent.

Vessel segments were then incubated in a dena­turing detergent, sodium dodecyl sulfate (SDS,10%). After 12-18 hours, specimens were immersedin baths of 1% SDS for 96 hours with six exchanges.The vessels were then washed in sequential baths ofdistilled water (36 hours) followed by 70% ethanol(36 hours), with frequent exchanges to remove alItraces of detergent. The AVM segments were fi­nalIy stored in 70% ethanol in sterile polypropylenecontainers at 4°C. Prior to EC seeding, the AVMsegments were rehydrated for 90 minutes in hepa­rinized normal saline (l J.L'ml) with two exchanges.

During the processing of one set of vessels,segments of AVM were removed after each se­quence of the extraction protocol and processed asbelow for evaluation by light and scanning electronmicroscopy.

III IV

AVM

(.'ROGFSEEDING

IllV'C< CHAMBERAVM+FN

(-lRDGF

SEM(CONTROL)

V VI VII VIII

Fig. 3. Protocol for EC seeding of AVM.

VOLUME 3No 2-1989 ACELLULAR VASCULAR MATRIX III

-,. ".:

, '. - .. ," -," ". . ":'.. ,--

" .",1.,

. ~ ,.( -

. ~ "...: ",-::", .

- - ... ,....-: .... ' .-- ........ -- -::---"1

Fig. 5. (a-f) Histologic evaluation of sequential detergent extraction process for AVM (lM stained withtoluidine blue; lumen at top). Native greyhound artery (a) (lM x 80), (b) (SEM x 2800; E, endothelial cells). AfterTriton x-100 and 10% SOS, note vacuolated appearance (c) (lM x 80), (d) (SEM x 2800). After distilled waterwashes; note native collagen-elastin lamina maintained (e) (lM x 80), (f) (SEM x 2800).

112 ACELLULAR VASCULAR MATRIX ANNALS OFVASCULAR SURGERY

...-1.I~!

~ :~~ .. - . ,

;g~-~-.: ~_._.~.~_. : .; ,~:..~.:-.J

t

..- --[ .. ----~- .. ..:.-t"'---=-~~ .... _- -ti 6L --. -- .... ". - -L._~, .,'=- __:' '" __-_.:._ -~.-'--'-- .

Fig. 5. (g-j) After 70% ethanol washes; note shrinkage of matrix (g) (LM x 80), (h) (SEM x 2800). After salinerehydration, AVM resembles native vessel matrix with intact subendothelial basement membrane (arrow) (i)(LM x 80), (j) (SEM x 2800).

Cell counts and dilutions

To determine the number of viable cells har­vested from the T-75 flask, the 0.2 ml aliquotobtained above was combined with 0.2 ml trypanblue (staining non-viable EC) and pipetted into ahemocytometer. The final EC concentration of the0.5 ml and 0.87 ml aliquots used to seed eachexperimental and control well, respectively, wasachieved by dilution in CM, with or without thedeletion of RDGF.

Control plates

For each EC culture used in each of the eightexperimental groups, two gel-treated wells of a24-well polystyrene tissue culture plate were seededwith 0.87 ml of EC at 105/cm2 concentration(matched positive control for EC growth) (Fig. I).

The plates were incubated for 18 hours at 37°C in a5% CO2, 100% humidity environment. The wellswere stained with crystal violet and counted underinverted phase-contrast microscope by planimetricanalysis [17]..

EC seeding of AVM

Plexiglass seeding chambers consisted of threesets of three wells, each 12 mm in diameter (Fig.2a). The flat AVM specimens were cut into 16 mmx 50 mm segments, laid on similar-sized flat poly­tetrafluoroethylene (PTFE) segments, and the wellplate was tightly fastened to the base plate withscrews (Fig. 2b). (The PTFE cushion provided aseal to prevent leakage of EC culture from the edgesof the well or through branches in the AVM seg­ment.)

VOLUME 3No 2 - 1989 ACELLULAR VASCULAR MATRIX 113

~ .'lI1j......~-~

1- ......

..

I.ll..•

B

'.- ..­r,·! -

ri,;..I' _~1

a a

b bFig. 6. Cross-section of AVM, (a) x 400 and (b) x10,000 demonstrates preservation of native hlstoar­chltecture by extraction process (B, arrow, basementmembrane; M, collagen-elastin matrix).

Fig. 7. EC-seeded AVM (Group III: AVM-FN+FBS+RDGF). (a) LM x 160. Note endothelial cell monolayer(arrow) on subendothelial basement membrane (B)and collagen-elastin matrix (M). (b) SEM x 1200.Note uncovered AVM substrate (M), endothelial cell(E) and normal morphology of Isolated endothelialcell with nucleus visible (N).

Eight different Ee-seeding conditions, each at105/cm2 seeding density were tested in triplicate; 0.5ml of EC culture was added to each well of theseeding chamber (Fig. 3):

Group I.

Group II.

AVM, without pretreatment of theluminal surface by fibronectin (FN)or serum (FBS), seeded by EC inthe presence of RDGF (10 j..tIlml)(AVM-FN-FBS+RDGF);AVM, pretreated by one hour incu­bation with 0.2 mllwell fibronectin(17.7 j..tg/cm2) , seeded with EC inpresence of RDGF (AVM+ FN­FBS+RDGF);

Group III.

Group IV.

Group V.

Group VI.Group VII.Group VIII.

AVM pretreated by one hour incu­bation with 0.5 ml/well of 10% FBS(44.2 j..tI/cm2

) , seeded with EC in thepresence. of RDGF (AVM-FN +FBS+RDGF);AVM pretreated with the combina­tion of FN (0.2mIlwell) and FBS(0.5 mllwell), then seeded with ECin the presence of RDGF (AVM+FN + FBS+RDGF);AVM, without pretreatment, seededby EC in CM with RDGF deleted(AVM-FN-FBS-RDGF);AVM+ FN-FBS-RDGF;AVM-FN + FBS-RDGF;AVM+FN+FBS-RDGF.

114 ACELLULAR VASCULAR MATRIX

TABLE I.-Effect of substrate"

ANNALS OFVASCULAR SURGERY

+RDGF-RDGF

AVM54.4 ± 18.637.8 ± 11.1

AVM + FN57.6 ± 14.033.5 ± 11.7

AVM + FBS56.8 ± 18.726.1 ± 18.2

AVM + FN + FBS61.8 ± 18.130.6 ± 11.4

pt

NSSNSS

'Percent EC coverage. rneans e standard deviationtOne-way analysis of variance

The seeding chambers were loosely covered withpolystyrene lids and incubated for 18 hours at 37°Cin a 5% CO2 , 100% humidity environment. Thewells were decanted, rinsed with 0.5 ml PBS, thenfilled with 0.5 ml of 2.5% glutaraldehyde (diluted insodium cacodylate buffer) and fixed overnight.

Histologic evaluation

After fixation, the AVM discs from each wellwere cut into five tissue blocks (Fig. 4). One non­seeded, non-pretreated control block went for lightmicroscopy (LM) and the other for scanning elec­tron microscopy (SEM). A center strip of AVM wasembedded in plastic and stained with toluidine bluefor LM. The remaining blocks of the seeded AVMwere prepared for SEM by dehydration in ethanol,critical-point drying and gold-palladium coating.Electron micrographs (800X) from six predeter­mined sites from each AVM well were assessed byplanimetry [17]. The sampling sites on the SEMspecimens were designated before planimetric anal­ysis to ensure unbiased planimetric counting. Thecell surface areas of those EC with identifiableborders on the micrographs were determined by theQuantigraph™ morphometric digitizing microcom­puter system (Novus Instruments, Inc.), assuminghomogeneity of the cell population with respect tocell height.

Statistical analysis

The fraction of AVM covered by EC (percent ECcoverage) was determined by planimetry (from sixsampling sites per well x three wells per experi­ment) and the mean and standard deviation werecomputed (n= 18) for each experimental group. The

percent EC coverage and percent EC attachment ofthe gel-treated control wells were also determinedby planimetry (six predetermined sampling sites perwell x two wells per experiment) and the mean andstandard deviation were computed (n= 12) for eachcontrol group. The mean and standard deviation ofthe cell surface area were computed (the number ofcells that could be morphometrically analyzed var­ied among the groups, n=26 to 73). Differencesbetween groups were assessed by one-way analysisof variance (ANaVA 1) and unpaired T-test.

RESULTS

The sequential histologic evaluation of the AVM,beginning with the native greyhound artery,through the detergent extraction process, and fi­nally after EC seeding, is shown in Figs. 5-7. Afterthe detergent extraction steps (Triton X-IOO andSDS), AVM has a vacuolated appearance from lossof cells from the matrix (Fig. 5c). After the distilledwater wash, the maintenance of the native collagenand elastin histoarchitecture is apparent (Fig. 5e).With 70% ethanol treatment, the dehydrated AVMshrinks significantly (Fig. 5g,h). After rehydrationin saline, the AVM resembles the native vesselmatrix with what appears to be an intact subendo­thelial basement membrane (Fig. 5ij). In cross­section on SEM, one can clearly see how theextraction process preserves the normal histoarchi­tecture of the native vessel (Fig. 6). Fig. 7a,b showsone of the seeded AVM (Group III) by LM andSEM.

When considering the effect of pretreatment ofthe substrate before EC seeding, no difference wasfound between non-coated AVM and AVM pretreat-

TABLE II.-Gel-treated polystyrene controt'

+RDGFCONTROLpt

-RDGFCONTROLpt

AVM54.4 ± 18.638.2 ± 9.30.00937.8 ± 11.129.9 ± 7.00.038

AVM + FN57.6 ± 14.016.2 ± 3.7§

33.5±11.717.6 ± 7.8< 0.001

AVM + FBS56.8 ± 18.727.1 ± 10.5< 0.00126.1 ± 18.212.5 ± 6.20.019

AVM + FN + FBS61.8 ± 18.124.5 ± 10.2§

30.6 ± 11.417.6 ± 5.70.001

'Percent EC coverage. means e standard deviation"r-test§Statisticalcomparison not valid (see text)

VOLUME 3No 2 - 1989 ACELLULAR VASCULAR MATRIX 115

TABLE III.-Effect of RDGF withdrawal*

+RDGF-RDGFpt

AVM54.4 ± 18.637.8 ± 11.1

0.002

AVM + FN57.6 ± 14.033.5±11.7<0.001

AVM + FBS56.8 ± 18.726.1 ± 18.2<0.001

AVM + FN + FBS61.8 ± 18.130.6 ± 11.4<0.001

'Percent EC coverage. means e standard deviationtT-test

DISCUSSION

The ideal vascular prosthetic conduit would beone with a complete luminal lining of true, function­ing and therefore highly thromboresistant EC [18].Currently, EC seeding is the best method for pro­ducing such a graft [19,20]. However, EC seedingremains an inefficient process due to inadequaciesof the prosthetic substrate. There are poor initialadherence of EC to the substrate and additional EClosses from the substrate under the shear stresses of

ed with FN, FBS or a combination of FN+ FBS(Table I). When AVM was compared to matchedgel-treated controls, there was a significantly betterEC coverage of AVM than gel-treated polystyrenefor all groups that could be statistically evaluated(Table II) (AVM+FN+RDGF and AVM+FN+FBS+RDGF control wells inadvertently seeded at5.75 x 104EC/cm2 instead of I x lOsEC/cm2

; there­fore, comparison to the experimental group is notvalid). When growth factor was suddenly with­drawn from EC just prior to seeding, there was asignificant fall in EC coverage on AVM for allgroups by T-test (AVM: P=0.002; AVM+FN:P<O.OOI; AVM+FBS: P<O.OOI; AVM+FN+FBS:P<O.OOI; Table III). When all gel-treated controlwells, seeded with EC from which RDGF had beensuddenly withdrawn, were compared to thoseseeded in the presence of RDGF, there was asignificantly worse attachment after growth factorwithdrawal (P=0.002 by T-test; Table IV). Cellsurface area varied widely between groups (TableV). However, when comparing substrate groupswith and without growth factor, EC were found tobe significantly smaller in all groups after RDGFwithdrawal, except on non-coated AVM.

TABLE IV.-Effect of RDGF withdrawal*

59.6 ± 13.4

flow [1,2,21]. Limited EC attachment and retentionmake the lag time between seeding and completeluminal EC coverage of a graft inordinately depen­dent upon the inherent proliferative ability of theremaining EC and growth-promoting factors in theperiendothelial milieu.

Various components of subendothelial extracel­lular matrix, such as basement membrane collagensType IV and V, fibronectin, and laminin, are knownto enhance EC attachment, spreading, growth, dif­ferentiation and migration [3,22,23]. Exogenouscoating of vascular prostheses with collagen, fibro­nectin, and other components of SEBM can dra­matically improve EC adherence and retention[1,2,24]. It therefore seems apparent that just as theideal vascular conduit should have an endothelialmonolayer, so too those EC should ideally besubtended by a substrate that mimics SEBM.

Malone, and associates [8] in 1983 demonstratedthat canine carotid arteries can be completely de­nuded of cellular components by a series of deter­gent extractions that yield purified extracellularmatrix in which the collagen and elastin laminae ofthe vascular wall are maintained in their naturalhistoarchitecture. Immunohistochemical and bio­chemical analyses [9] of AVM revealed that cellmembranes, nuclei, and most cytoplasmic compo­nents are removed by detergent extraction. Aminoacid analysis documented that collagen and elastinare the primary components of AVM. Severalextracellular matrix proteins are detectable by im­munohistology in addition to the interstitial collagenand elastin, including fibronectin, laminin, andType IV (basement membrane) collagen; these lat­ter three noninterstitial proteins are present on theluminal surface of the inner elastic membrane.Electron microscopy confirms that the subendo­thelial basement membrane layer is intact [9],

We postulated that the close resemblance of th.eAVM luminal lining to native SEBM WOUld. make .Ita natural substrate for Ee-seeding. By histologicevaluation with LM and SEM, we confirmed thepreservation of the native vessel hist~architecture

by the sequential detergent extraction process(Figs. 5,6). We also demonstrated that EC adhereto, spread and assume normal morphology on AVM(Fig. 7). . .

Pretreatment of vascular prosthetic matenals(PTFE, Dacron, Hytrel) with FN or collagen, mim-

pt

0.002-RDGF

37.1 ± 11.9+RDGF

Percentage of ECattachment ongel-treatedpolystyrenecontrols

'Means:!: standard deviationtT-test

116 ACELLULAR VASCULAR MATRIX

TABLE V.-Effect of RDGF withdrawal"

ANNALS OFVASCULAR SURGERY

+RDGF-RDGFpt

AVM

435 = 132410 = 141NSS

AVM + FN

600 = 223368 = 138< 0.001

AVM + FBS

544=218341 = 143< 0.001

AVM + FN + FBS

501 = 213367 = 1160.001

'Cell surface area (JLlTl2), rneans e standard deviationtT·test

icking the native SEBM surface, has been shown tosignificantly improve EC attachment, retention andgrowth [1,2,23,24]. We therefore tested AVM withno-pretreatment (since it should mimic intactSEBM with FN, laminin and Type IV collagen andshould not require pre-coating) as well as with FNor serum pre-coating. (Serum contains FN, hor­mones and other growth factors that should en­hance EC adherence and spreading on a substrateby mimicking the native EC environment) [3,15,25].We hypothesized that AVM should be as conduciveto EC seeding with no pretreatment as with surfacepretreatment.

In this study, non-coated AVM performed aswell, in terms of coverage by seeded EC at 18hours, as AVM coated with FN, serum or both(Table I). That contrasts with the performance ofthe other synthetic vascular substrates used inEC-seeding studies wherein precoating improvedEC adherence and retention [l,2,23,24]. EC cover­age of AVM was even better than on gel-treatedtissue culture plastic used as positive-controls forthe EC culture (Table 11). Gel-treated polystyrene iswell known to be highly conducive to EC attach­ment and proliferation and is the substrate of choicefor tissue culture.

There is indirect evidence to suggest that AVMcontains a protease-sensitive, Triton-resistant ECmitogen, basic fibroblast growth factor [26-28]. IfAVM indeed contains an active growth factor, theabsence of an exogenous mitogen should not havean effect on EC attachment and spreading. Wetested this hypothesis by seeding EC· with andwithout RDGF.

EC coverage of AVM and percent EC attachmenton gel-treated polystyrene were both adversely af­fected by withdrawal of retinal-derived growth fac­tor (Tables III,IV). This effect was irrespective ofAVM surface pretreatment. Based on observationsof human umbilical vein EC in culture in our lab,EC remain in lag phase at 18 hours of incubation.Since EC should not have been replicating, onewould not expect RDGF withdrawal to have anyeffect in this experiment. Moreover, when Gospo­darowicz and associates [29] grew bovine vascularEC in the presence of fibroblast growth factor for 12days before withdrawal, the EC continued to pro­liferate and were healthy seven days later. How­ever, Limanni and coworkers [16] demonstrated

that growth factor withdrawal may lead to physio­logic and morphologic alterations in EC. Our results(Tables III,IV) were consistent with the latter re­port. In addition, we found that EC size was signif­icantly reduced in the absence of RDGF for allAVM groups except for non-coated AVM (TableV). This further suggests that growth factor with­drawal inhibits EC attachment and spreading byaltering cell physiology and morphology; the mech­anism by which this occurs remains to be e1uci- .dated. .

CONCLUSIONS

We found that AVM is a natural EC substratethat is conducive to adherence and spreading er nceven without surface pretreatment by fibronectin,serum, or both. Sudden withdrawal of exogenousgrowth factor may impair early coverage of sub­strates by EC; this may be due to physiologic andmorphologic alterations of EC that affect adherenceor cell spreading. Based on this study, AVM war­rants further trials as an EC-seeding substrate.Further delineation of the components of the AVMluminal lining are necessary. Finally the effects ofgrowth factor withdrawal on seeded EC need morein-depth investigation.

ACKNOWLEDGMENTS

The authors gratefully acknowledge the assis­tance of John B. Sharefkin, MD, Department ofSurgery, Uniformed Services University of theHealth Sciences, Bethesda, Maryland and AndrewP. Evan, PhD, Department of Anatomy, IndianaUniversity School of Medicine, Indianapolis, Indi­ana.

The sequential detergent-extraction process forproducing the acellular vascular matrix graft has apatent pending.

This work was supported by a Grant-in-Aid fromthe American Heart Association, Indiana Affiliate,Inc.

VOLUME 3No 2 - 1989 ACELLULAR VASCULAR MATRIX 117

REFERENCES

I. KESLER KA, HERRING MB, ARNOLD MP, GLOVERJL, PARK HM, HELr-.mS MN, BENDICK PJ. Enhancedstrength of endothelial attachment on polyester elastomerand polytetrafluoroethylene graft surfaces with fibronectinsubstrate. J Vase Surg 1986;3:58-64.

2. RAMALANlAONA G, KEMPCZINSKI RF, ROSENMANJE, DOUVILLE EC, SILBERSTEIN EB. The effect offibronectin coating on endothelial cell kinetics in poly tetra­fluoroethylene grafts. J Vasc Surg 1986;3:264-272.

3. KLEINMAN HK, KLEBE RJ, MARTIN GR. Role ofcollagenous matrices in the adhesion and growth of cells. JCell BioI 1981;88:473-485.

4. GORDON PB, LEVITf MA, JENKINS CSP, HATCHERVB. The effect of the extracellular matrix on the detachmentof human endothelial cells. J Cell PhysioI1984;12/:467-475.

5. GOSPODAROWICZ D, ILL C. Extracellular matrix andcontrol of proliferation of vascular endothelial cells. J ClinInvest 1980;65:1351-1364.

6. MADRI lA. Endothelial cell-matrix interactions in he­mostasis. Prog Hemo Thromb 1982;1-24.

7. PRATf 8M, HARRIS AS, MAORROW JS, MADRI lA.Mechanisms of cytoskeletal regulation. Am J Pat/101 1984;1/7:349-354.

8. MALONE JM, BRENDEL K, DUHAMEL RC, REINERTRL. Detergent-extracted small diameter vascular prosthe­ses. J Vase Surg 1984;/:181-191.

9. DUHAMEL RC, BRENDEL K, REINERT RL, MALONEJM. Properties of acellular vascular matrix. Proceedings,Society for Biomaterials, 10th Annual Meeting. April 27­May I, 1984.

10. MALONE JM, REINERT RL, BRENDEL K, DUHAMELRC. Thrornbogenicity of small diameter vascular grafts.Proceedings, Society of Biornaterials, 10th Annual Meeting,April 27-May I, 1984.

II. DUHAMEL RC, MALONE 1M, REINERT RL, BREN­DEL K. The effect of glutaraldehyde crosslinking on endo­thelialization of acellular vascular matrix grafts. Proceed­ings, Society for Biomaterials, 11th Annual Meeting, April25-28, 1985.

12. BERGER K, SAUVAGE LR, RAO AM, WOOD S1. Heal­ing of arterial prostheses in man: its completeness. Alln Surg1972;/75: 118-127.

13. SAUVAGE LR, BERGER KE, WOOD sr, YATES SG,SMITH JC, MANSFIELD PB. Interspecies healing of po­rous arterial prostheses. Arch Surg 1974;109:698-705.

14. SAUVAGE LR, BERGER K, BEILIN LB, SMITH rc,WOOD Sl, MANSFIELD PB. Presence of endothelium inan axillary-femoral graft of knitted Dacron with an externalvelour surface. Alln Surg 1975;/82:749-753.

15. GAJDUSEK CM, SCHWARTZ SM. Ability of endothelial

cells to condition culture medium. Cell Pltysiol 1982;/10;35-42.

16. LIMANNI A, FLEMING T, MOLINA R, HUFNAGEL H,CUNNINGHAM RE, CRUESS DF, SHAREFKIN JB.Expression of genes for platelet-derived growth factor inadult human venous endothelium. J Vasc Surg 1988;7:10-20.

17. HERRING M, EVANS D, BAUGHMAN S, GLOVER 1.The quantitation of cultured cellular surface coverage: ap­plications for transparent and opaque surfaces. J BiomedMater Res 1984;/8:567-575.

18. CALLOW AD. Presidential address: The microcosm of thearterial wall-A plea for research. J Vasc Surg 1987;5:1-18.

19. HERRING M, GARDNER A, GLOVER J. A single-stagedtechnique for seeding vascular grafts with autogenous endo­thelium. Surgery 1978;84:498-504.

20. CALLOW AD. Endothelial cell seeding: problems and ex­pectations. J Vase Surg 1987;6:318-319.

21. ROSENMAN lE, KEMPCZINSKI RF, PEARCE WH,SILBERSTEIN EB. Kinetics of endothelial cell seeding. JVase Surg 1985;2:778-784.

22. BAKER KS, WILLIAMS SK, JARRELL BE, KOOLPEEA, LEVINE E. Endothelialization of human collagensurfaces with human adult endothelial cells. Am J Surg1985;150: 197-200.

23. WILLIAMS SK, JARRELL BE, FRIEND L, RADOMSKIJS, CARABASI RA, KOOLPE E, MUELLER SN,THORNTON SC, MARRINUCCI T, LEVINE E. Adulthuman endothelial cell compatibility with prosthetic graftmaterial. J Surg ReJ 1985;38:618-629.

24. FOXALL TZ, AUGER KR, CALLOW AD, LIBBY P.Adult human endothelial cell coverage of small-caliber Da­cron and polytetrafluoroethylene vascular prostheses in vi­tro. Surg Res 1986;41:158-172.

25. HAYASHI I, SATO GH. Replacement of serum by hor­mones permits growth of cells in a defined medium. Nature1976;259:132-134.

26. KAN M, KATO M, YAMANE I. Fibroblast products ad­sorb to the culture substrate and stimulate growth of humanumbilical vein endothelial cells, J Cell Physiol 1985;/24:125-130.

27. VLODAVSKY I, SULUVAN R, FRIDMAN R, SASSE r.FOLKMAN 1, KALGSBRUN M. Heparin binding endothe­lial cell growth factor produced by endothelial cells andsequestered by the subendothelial extracellular matrix. JCell BioI 1986;103(5) Part 2:98a (Abst),

28. FOLKMAN 1, KLAGSBRUN M, SASSE 1, WADZINSKIM, INGBER D, VLODAVSKY I. A heparin-binding angio­genic protein-basic fibroblast growth factor is stored withinbasement membrane. Am J Pathol 1988;/30:393-399.

29. GOSPODAROWICZ D, MORAN J, BRAUN D, BIRD­WELL C. Clonal growth of bovine vascular endothelialcells: fibroblast growth factor as a survival agent. Proe NatlAcad Sci USA 1976;73:4120-4124.

•••