Superior Late Patency of Small-Diameter DacronGrafts Seeded With Omental Microvascular Cells:An Experimental StudyMiralem Pasic, MD, PhD, Werner Muller-Clauser, PhD, Ludwig K. von Segesser, MD,Mario Lachat, MD, Tomislav Mihaljevic, MD, and Marko I. Turina, MDClinic for Cardiovascular Surgery, Department of Surgery, University Hospital Zurich, Zurich, Switzerland

The purpose of the study was to investigate the effect ofomental microvascular cell seeding on the patency ofsmall-diameter Dacron prostheses usable for coronaryartery bypass grafting. In a canine carotid artery model,each dog (n = 64) received one seeded and one similarnonseeded Dacron prosthesis (internal diameter = 4 or6 mm). Enzymatically harvested omental microvascularcells (omentum = 27.6 ± 5.9 g [± the standard deviation];range, 17 to 50 g) were seeded prior to implantation. Theseeding density was 1.91 ± 0.26 [± the standard error] x106 cells/em? of graft surface. Dipyridamole (75 mg/d)and acetylsalicylic acid (325 mg/d) were administeredorally for 4 weeks postoperatively. The prostheses wereexplanted between 2 and 52 weeks after placement. Theresults were assessed by angiography; light, scanning

Although the patency of synthetic aortocoronary graftsis poor, such a graft should be used in certain

situations when no suitable vein or arterial graft is avail­able, especially in the setting of reoperation. Lack of anendothelial surface is one of the most important variablescausing poor patency of small-diameter prosthetic vasculargrafts compared with autologous saphenous vein or arte­rial grafts Itl.

It has been supposed that when seeded onto small­diameter prosthetic vascular grafts immediately beforeimplantation, microvascular endothelial cells from omen­tal tissue can proliferate rapidly to cover the graft andimprove the patency rate. The seeded cells exhibit at leastsome normal endothelial cell functions, such as antagoniz­ing coagulation, preventing platelet deposition, and lysingfibrin that forms in the graft lumen. Therefore, this exper­imental study was designed to determine whether endo­thelial cell seeding with enzymatically derived microvas­cular cells from omental tissue may improve the patencyrate of small-diameter Dacron grafts usable for coronaryartery bypass grafting.

Presented at the Thirtieth Annual Meeting of The Society of ThoracicSurgeons, New Orleans, LA, Jan 31-Feb 2, 1994.

Address reprint requests to Dr Pasic, Clinic for Cardiovascular Surgery,Department of Surgery, University Hospital Zurich, Ramistrasso 100, CH­8091 Zurich, Switzerland.

© 1994 by The Society of Thoracic Surgeons

electron, and transmission electron microscopy; and mor­phometry. The seeded grafts developed a uniform lumi­nal monolayer of endothelial cells with minimal plateletor cellular deposition. These grafts also had a signifi­cantly higher overall patency rate and significantly largerthrombus-free surface areas than the nonseeded grafts.The overall actuarial patency rates at 1 week, 5, 12, 26, and52 weeks were 100%, 98%, 93%, 93%, and 93%, respectively,for seeded Dacron grafts and 100%, 91%, 61%, 54%, and18%, respectively, for nonseeded grafts. The patencyrates of Dacron grafts usable for coronary artery bypassgrafting are significantly improved by seeding withomental microvascular cells in a canine model.

(Ann Thorae Surg 1994;58:677-84)

Material and Methods

The study was performed using 64 mongrel dogs initiallyweighing 20 to 30 kg. All animals received humane care incompliance with the European Convention on AnimalCare and the "Guide for the Care and Use of LaboratoryAnimals" published by the National Institutes of Health(NIH publication 85-23, revised 1985). The experimentswere reviewed and approved by the Ethics Committee ofthe University Hospital Zurich.

A prophylactic cephalosporin antibiotic (cefalexinummonohydrate [Cefaseptin; Chassot & Cie AG, Koniz Bern,Switzerland], 1,200 mg/d) was given for 4 days, the firstdose preoperatively. An analgesic (buprenorphine chloride[Temgesic], 0.9 mg/d) was given for 4 days postopera­tively. All dogs received orally dipyridamole (75 mg/d) for1 day preoperatively and acetylsalicylic acid (325 mg/ d)for 4 days preoperatively. These medications were contin­ued for 4 weeks postoperatively, except when the graftswere explanted earlier.

Surgical ProcedureThe dogs were anesthetized with sodium thiopental(20 mg/kg intravenously), intubated, and ventilated with amixture of oxygen, nitrous oxide, and halothane. Hydra­tion was maintained by infusion of lactated Ringer's solu­tion at the rate of 10 mL . kg J. h 1 for the duration of theoperation.

Omentum (27.6 ::'::: 5.9 g [::,::: the standard deviation];

0003-4975/94/$7.00

678 PASIC ET ALENDOTHELIAL CELL SEEDING OF PROSTHETIC GRAFTS

Ann Thorac Surg1994;58:677-84

Table 1. Types of Implanted Dacron Grafts

Internal ExternalDiameter Length Ring

Type of Craft (rnm) (ern) Crimped Support

Sulgraft T33' 4 6 No NoSuIgraft T5.23" 4 6 No NoMeadox Microvel" 4 6 No Yes

(double velourknitted Dacron)

Meadow Microvel'' 6 6 Yes No(double velourknitted Dacron)

Vascutek VP 1200 K' 6 6 Yes Yes(knitted Dacron)

Vascutek Celsoft CS 6 6 Yes Yes1200 K (knittedDacron)

a This graft is manufactured by Sulzer Brothers, Ltd, Winterthur, Switzer-land. b This graft is manufactured by Meadox Medicals, Inc, Oakland,NJ. c- This graft is manufactured by Vascutek Ltd, Inchinnan, Renfrew-shire, UK.

range, 17 to 50 g) was removed through a supraumbilicallaparotomy, placed in a sterile 150-mL plate containing60 mL of HEPES-buffered saline solution (142 mmol/L ofNaCl,4 mmol/L of KCl, 15 mmol/L of HEPES, 4 mmol/Lof D-glucose, 2.5 mmol/L of CaClz, 1 mmol/L of MgClz),and transferred to the cell laboratory for further process­ing. The abdominal incision was closed, a midline neckincision was made, and both common carotid arteries weredissected free. A 6-cm segment of each artery was replacedwith a 6-cm segment of a Dacron graft.

According to the protocol of each experiment, six differ­ent types of grafts with an internal diameter of either 4 mm(n = 36 dogs) or 6 mm (n = 28 dogs) were implanted(Table 1). The paired carotid artery model was usedthroughout the study; the control nonseeded graft wasimplanted into the left carotid artery and the seeded graft,into the right carotid artery in each dog. The control graftwas always implanted first. The oblique end-to-end anas­tomoses were performed with a continuous 7-0 polypro­pylene suture (Prolene; Ethicon, Somerville, NJ); underoptical magnification.

The seeding procedure was performed just before im­plantation. According to the protocols, some of the graftswere seeded with autologous endothelial cells but notpreclotted (n = 22), and some were seeded during preclot­ting with either autologous plasma (n = 17) or autologousblood (n = 25) prior to implantation. The preclottingprocedure for the control nonseeded graft was identical tothe procedure for the seeded graft but without the additionof the endothelial cell suspension. Care was taken to keepthe seeded prostheses misted during the operation. Theneck wound was closed in layers, transfemoral arteriogra­phy was performed using the Seldinger method, and theanimals were allowed to recover.

Simultaneous seeding with autologous endothelial cellsand preclotting of a graft with autologous plasma wasperformed with a 2-mL suspension of the endothelial cells,which was added to 15 mL of autologous plasma and

centrifuged onto the prostheses using a rotation device(Sulzer Brothers, Ltd, Winterthur, Switzerland). Theseeded grafts were rotated at 45 g for 20 minutes prior toimplantation.

Preclotting with autologous blood was performed ac­cording to the modified four-step procedure of Yates andcolleagues [2]. One end of the graft was clamped, andabout 15 mL of arterial blood was injected into the lumenof the graft so that the blood was forced through the graftinterstices. After five passes of blood, 1 mL of the cellsuspension was injected into the graft lumen, followed bytwo more passes of whole blood. The entire procedure wasthen repeated through the other end of the graft. After thepreclotting procedure, a Fogarty balloon catheter wasinserted into the graft to remove any blood clot from thegraft lumen. The control grafts were preclotted using thesame methods but without the addition of the endothelialcell suspension.

The animals were not heparinized either preoperativelyor during the operation. Starch-free gloves were usedduring the operative procedure to avoid the cytotoxic effectof glove powder [3].

Microvascular Endothelial Cell HarvestingEndothelial cells were harvested from omental tissue ac­cording to the method of Schmidt and associates [4] withsome modifications. Cell harvesting and seeding of theprostheses were performed by a technician in the celllaboratory. The omentum was rinsed with HEPES­buffered saline solution, trimmed, minced with two scal­pels into very small pieces, and aspirated into a 10-mLpipette. The tissue was then transferred into a 50-mLErlenmeyer flask containing 1,500 U/mL of collagenasetype II (Sigma Chemical Co, St. Louis, MO). The ratio was1 gr of omental tissue to 2 mL of collagenase. The suspen­sion of omental tissue and collagenase was incubated for40 minutes in a 37°C water bath at 100 shaking motions perminute. The digested tissue was homogenized by repeti­tive pipetting, transferred into a 15-mL tube, and centri­fuged twice at 100 g for 5 minutes. The supernatantcontained mainly adipocytes and collagenase solution.

The cell pellet was resuspended in lO-mL of phosphate­buffered saline solution, filtered through a polyester meshwith a pore size of 250 f.Lm (Bolting Cloth Weaving, Zurich,Switzerland) to remove nondigested large tissue frag­ments, and then washed two times with HEPES-bufferedsaline solution. Finally, the collected cells either wereresuspended in autologous plasma or autologous blood forsubsequent simultaneous preclotting and seeding or wereaspirated into a syringe for graft seeding.

Cell IdentificationAn approximately 0.5-mL aliquot of the cells was removedfor the cell count, cell identification, and tissue culture [5].The cell suspension was diluted 1:20 with culture mediumM (Amined, Basel, Switzerland), containing 20% fetal calfserum (Biological Industries, D.N. Maale Hagalil, Israel),60 f.Lg/mL of heparin sodium (Sigma Chemical Co), 2.5mmol/L of L-glutamine, 50 U/mL of penicillin, 50 f.Lg/mLof streptomycin sulfate, 50 f.Lg/ mL of endothelial-cell

Ann Thorac Surg1994;58:677-84

PASIC ET ALENDOTHELIAL CELL SEEDING OF PROSTHETIC GRAFTS

679

growth supplement (Collaborative Research Inc, Bedford,MA), and 10 ng/mL of epidermal growth factor (Collabo­rative Research Inc), with a pH of 7.4. The suspension ofmicrovascular cells was plated onto plastic tissue culturewells precoated with 5 /Lg/mL of human fibronectin (Col­laborative Research, Inc). The primary cultures were heldat 37°C in a humidified 0.7% carbon dioxide environment,with a change of culture medium after 24 hours and laterafter every third day [5].

Cell viability was assessed during cell culture and wasdefined as the number of attached endothelial cellscounted after 18 hours in the cell culture. Cell number wascalculated from fluorometric deoxyribonucleic acid mea­surements [6].

Endothelial cells in cultures or on explanted graft seg­ments were identified by their typical cobblestone mor­phology at confluence, by indirect immunofluorescencestaining for factor VIII-related antigen present on theirsurface, and by biochemical testing for the specific rapidreceptor-mediated uptake of acetylated low-density li­poprotein labeled with a fluorescent probe (Dil-Ac-LDL,200 mg/mL; Biomedical Technologies Inc, MA). They werethen examined with phase-contrast microscopy and epi­fluorescence [7].

Smooth muscle cells were identified by marking withalpha-smooth muscle actin l'-globulin asm-1 (Progen;Readysystem Inc, Zurzach, Switzerland) and visualizingusing a fluorescent-labeled secondary antibody (fluoresce­in-conjugated anti-mouse IgG goat y-globulin, diluted 1:20with phosphate-buffered saline solution [8].

ExplanationThe prostheses were explanted between 2 and 52 weeksafter operation: at 2 weeks (n = 5 dogs), 5 weeks (n = 31dogs), 12 weeks (n = 10 dogs), 26 weeks (n = 12 dogs), and52 weeks (n = 6 dogs). Graft patency was determined byarteriography immediately after implantation followed byangiography at 1 week,S, 12, 26, and 52 weeks. Aftertransfemoral angiography, the grafts were dissected freeand then excised with the adjacent arterial segments. Afterexcision, the graft with adjacent parts of the artery wasperfused in salt solution, opened longitudinally, pinnedflat on a Petri dish, photographed for planimetric studies ofthe thrombus-free surface area, and then stained withEvans blue to differentiate between endothelium (white)and nonendothelialized intima (blue). Extent of endotheli­alization was determined by photomorphometry of theluminal surface using a computerized image analyzer(videoplanimeter; Carl Zeiss Inc, Switzerland). Finally, theprostheses were divided into three parts (proximal, central,distal), and each segment was cut into four pieces for light,scanning electron, and transmission electron microscopicobservations as well as for testing for specific uptake ofDil-Ac-LDL.

Microscopic ExaminationAfter macroscopic examination, the grafts underwent mi­croscopic studies with light microscopy after staining withhematoxylin and eosin and elastin. For scanning electronmicroscopy, the grafts were processed by standard method

[9]. Specimens were prepared with 1% osmium tetroxidesolution, dehydrated in a graded series of ethanol andamyl acetate, and then dried to the critical point withcarbon dioxide and sputter-coated with 20 /Lm of goldpalladium. Examinations were performed with a PhilipsSEM 505 scanning electron microscope.

For transmission electron microscopy, the portions of thegrafts were prefixed with 2% glutaraldehyde and 0.8%paraformaldehyde in sodium cacodylate buffer (0.1 mol/L;pH 7.2) followed by scratching of the cell layers from theDacron using a razor blade. The scratched cell layer waspostfixed with 1% osmium tetroxide solution containing1.5% K4Fe(CN)6' After standard embedding into Epon,ultra-thin sections (50 /Lm) were contrasted either withuranyl acetate and lead citrate or with tannic acid fordetection of elastic material. The specimens were examinedwith a Philips TEM 420 transmission electron microscope.

Statistical AnalysisThe data are presented as the mean::'::: the standard error ofthe mean, unless otherwise stated, with 95% confidencelimits in parentheses.

Life table for graft patency rate (event-free rate) wasestimated by the method of Cutler and Ederer [10]. Event­free proportions at 1 week,S, 12, 26, and 52 weeks postoper­atively are reported with 95% confidence limits. Differencesbetween the seeded and nonseeded control groups weretested by a nonparametric log-rank test. Differences inmean values between the groups were assessed by use ofStudent's unpaired t test. A p value of less than 0.05 wasconsidered significant. Statistical analyses were performedwith BMDP statistical software (BMDP, Los Angeles, CA).

Results

During the surgical implantation procedures, all graftswere successfully seeded. All dogs survived the studyperiod without wound infection or other complications.

Cell Harvesting and IdentificationA mean of 2.58 ::'::: 0.4 x 106 cells / g of omental tissue (95%confidence limits, 1.71 to 3.44) was harvested from 27.6 ::':::5.9 g (::'::: the standard deviation) of omental tissue. Endo­thelial seeding density was 1.91 ::'::: 0.26 x 106 cells/ern" ofgraft surface (l.40 to 2.44) and ranged from 0.10 x 106

cells /cm" of graft surface. The mean percentage of viablecells after 18 hours in culture was 41%.

The short-term cultures of the suspension of harvestedcells revealed not only endothelial cells but also other typesof cells. Endothelial cells were identified by their typicalcobblestone contact-inhibited morphology at confluenceunder phase-contrast microscopy, indirect immunofluores­cence staining with factor VIII-related antigen, and specificperinuclear uptake of Dil-Ac-LDL. After graft explanta­tion, the seeded grafts exhibited a monolayer of perinu­clear labeled cells on the luminal surface, whereas thenonseeded grafts usually demonstrated diffuse nonspecificstaining with only small areas positively stained.

680 PASIC ET ALENDOTHELIAL CELL SEEDING OF PROSTHETIC GRAFTS

Ann Thorac Surg1994;58:677-84

PATENCY (%)

(I)(U)

I

40

20

60

100 LIaI! (..)l7OT)·····'i«'-

80 (07 r·....

··············I(~~~ · 1(11)................................ I.......................... (I)

........

oL- ---l

12 26 52

WEEKS

I 95% C.L. ... SEEDED ....... NON-SEEDED IFig 1. Actuarial patency rates of all seeded Dacron grafts and all non­seeded Dacron grafts. All experimentalanimals received dipyridamole(75 mgjd) and acetylsalicylic acid (325 mg/d) orally for 4 weeksafterimplantation. Numbers of grafts at risk are in parentheses. (CL. =

confidence limits.)

Fig 2. Macroscopic appearance of luminal surfaceof nonseeded graft(top) and seeded graft (bottom) explantedafter 5 weeks. The endothe­lial cell-seeded graft lumen was almost thrombus free at 5 weeks. Notepannus ingrowth of endothelium across anastomoses of nonsecdedgraft; white, shiny areas are spreading from both anastomotic ends,but central region is covered with red thrombus.

Fig 3. Thrombus-free surface of all patent seeded and nonseededgrafts. (* = p < 0.05 between groups.)

was not obvious in the seeded grafts. The nonseeded graftsthat remained patent after this period showed a differentdegree of spontaneous endothelialization.

Thrombus-Free SurfaceSeeded grafts demonstrated a significantly larger throm­bus-free surface than control nonseeded grafts at everyinterval after operation (p = 0.0002 to 0.02) except at 1 yearafter implantation, but few nonseeded grafts were patentat this point (Figs 3, 4). Thrombus-free surface increasedprogressively with the duration of graft implantation forboth seeded and nonseeded grafts.

Microscopic ExaminationsAfter implantation, the seeded grafts demonstrated clus­ters of cells adhering to the Dacron fibers and covered witha formation of red thrombus on the prosthetic lumen,preventing washout and subsequent loss of the seeded

522612

WEEKS

5

[Ii. SEEDED GRAFTS 0 NON-SEEDED GRAFTS

2

THROMBUS-FREE SURFACE-----------'----'---

60

o

40

80

20

100

Macroscopic AppearanceAt explantation, all types of grafts revealed some degree oftissue incorporation on the external surfaces whether thegrafts had been seeded or not. The anastomotic regionsappeared to have more incorporation than the central partsof all grafts. The luminal surfaces of the seeded grafts werewhite, smooth, and shiny. At the anastomoses, the luminallining was contiguous with the native vessel wall, therebycreating a smooth transition. Some small areas of redthrombus were also found, depending on the duration ofimplantation (Fig 2). The seeded prostheses explanted 2weeks after operation demonstrated large areas of redthrombus, which was significantly reduced at 5 weekspostoperatively, and could only occasionally be found inthe later follow-up.

Pannus ingrowth of endothelium could be identifiedacross both anastomoses only in the nonseeded grafts (seeFig 2). Ingrowth of arterial endothelium was seen approx­imately 0.5 ern from each anastomosis 5 weeks afterimplantation. The difference between anastomotic regionswith pannus ingrowth and endothelialized central parts

PatencyThe seeded grafts exhibited significantly better overallshort-term and long-term patency rates than the controlnonseeded grafts regardless of type and internal diameterof the grafts implanted, duration of implantation, preco­agulation, or seeding procedure. The difference betweenthe seeded and nonseeded grafts was more significant forthe long-term than the short-term results in the two groups(Fig 1). The overall actuarial patency rates of the seededgrafts were 100%, 98% 000% to 93%), 93% (100% to 83%),93%, and 93% at 1 week,S, 12, 26, and 52 weeks, respec­tively. The overall actuarial patency rates of the controlgrafts were 100%, 91% 000% to 81%), 61% (81% to 42%),54% (76% to 31%), and 18% (48% to 0%) at 1 week,S, 12, 26,and 52 weeks, respectively.

Ann Thorac SurgI994;58:677~ 84

PASIC ET ALENDOTHELIAL CELL SEEDING OF PROSTHETIC GRAFTS

681

522612

WEEKS52

o

THROMBUS-FREE SURFACE100,-----

20

40

60

80

Fig 4. Thrombus-free surface of central part of all patent seeded andnonsecdcd grafts. (* = p < 0.05 between groups.)

Fig 6. Light photomicrograph of cross section of seeded graft explanied5 weeks after implantation. The highly organized multilayer is under­neath the endothelial cell layer. (Hematoxylin and eosin; original mag­nification. x20.)

cells (Fig 5). Five weeks postoperatively, the seeded graftsexhibited smooth, cellular luminal surfaces adherent to thegraft by healthy connective tissue (Fig 6). This could alsobe found in the nonseeded grafts, but to a lesser extent. Theluminal side of the nonseeded grafts was covered mainlywith a coagulum of fibrin and entrapped red blood cells(Fig 7). The microscopic examinations of the anastomoticareas of the seeded grafts revealed smooth graft-vesselinterfaces.

Scanning electron microscopy of the seeded graftsshowed complete coverage by a surface monolayer ofconfluent endothelial cells with minimal platelet or cellulardeposition (Fig 8). The cells were arranged in a cobblestonepattern with tight junctions between adjacent cells (Fig 9).Occasional platelets and leukocytes were seen attached tothe regions of incomplete endothelialization. Transmissionelectron microscopy revealed that the luminal monolayercoverage of the seeded grafts was formed by ovally shapedendothelial cells with marginal flaps, slightly indented

nuclei, numerous well-defined organelles, thick bundles ofactin filaments, and typical Weibel-Palade bodies (Fig 10).

Comment

Our study clearly demonstrated that one-stage endothelialcell seeding with autologous omental microvascular cellsresulted in excellent long-term patency of small-diametergrafts in dogs. Despite the encouraging results of endothe­lial cell seeding in animal studies [11], the results inhumans have been disappointing. Although some trials[12] have demonstrated that endothelium could developon seeded grafts, most human trials were not able to showendothelialization of the grafts [13]. One of the mainreasons for failed endothelialization was a very low seed­ing density [13, 14], with a consequent rapid loss of theinitially attached endothelial cells [15]. A substantial lengthof donor vein is necessary to seed a graft, and large

Fig 5. Light photomicrograph of cross section of seeded Dacron graftcxplanied after 2 hours in circulation. The luminal surface is coveredwith red thrombus. prel'mting washout of undcrluing' microvascularcells by blood flOw. (Hematoxylin and eosin; original magnification.x20.)

Fig 7. Light photomicrograph of cross section of nonseeded Dacrongraft cxplanted 5 weeks after implantation. The luminal surface is cov­ered with a coagulum of fibrin with entrapped erythrocytes and littleor no cellular organization. (Hematoxylin and eosin; original magnifi­cation. x20.)

682 PASIC ET ALENDOTHELIAL CELL SEEDING OF PROSTHETIC GRAFTS

Ann Thorac Surg1994;58:677-84

Fig 8. Scanning electron photomicrograph of seeded graft 12 weeksafter implantation. There is a confluent and athrombogenic monolayerof endothelial cells (Original magnification, X 170.)

quantities of inoculated cells are necessary to overcomeearly cellular loss caused by shearing forces of the blood.

Two different possibilities for improving seeding densityare available. The first is a multistage technique thatcomprises endothelial cell harvesting from a smaller lengthof vein followed by cell growth in culture through severalpassages and then immediate seeding on either syntheticgrafts [16] or in vitro lining where the cells grow toconfluence before implantation [8]. The second possibilityis to use alternative sources with abundant endothelialcells, such as perinephric tissue, subcutaneous fat, oromentum [4, 17].

Theoretically, both techniques have several disadvan­tages. In vitro endothelialization of prosthetic grafts is notpossible in patients who require urgent operation; re­peated passages of endothelial cells have a detrimentaleffect on endothelial cell growth, function, and morphol-

Fig 9. Scanning electron photomicrograph of seeded graft 5 weeks af­ter implantation. Note tight junctions between cells. (Original magni­fication, X 1,310.)

Fig 10. Transmission electron photomicrograph of endothelial cellswith typical Weibel-Palade bodies from luminal surface of seeded Da­cron graft. (Original magnification, X22,000.)

ogy [18]; and endothelial cell culture and in vitro endothe­lialization are technically cumbersome and not successfulin every patient [19]. With the high number of harvestedcells from omentum that we achieved, we were able toaccomplish a high seeding density of 1.91 ± 0.26 X 106

cells/crrr' of graft surface. However, the isolates containednot only endothelial cells but a variety of cell types,including smooth muscle cells, mesothelial cells, fibro­blasts, and pericites [20]. Despite approximately 95% en­dothelialization at 4 weeks, an inner capsule continued toaccumulate beneath the endothelial monolayer [4, 17] andrepresented the potential for late occlusion during long­term follow-up [21]. To date, there has been only limitedexperience with long-term patency after seeding withendothelial cells derived from omental tissue. Our experi­mental study examined this point and showed no risk oflate occlusion of small-diameter Dacron grafts that hadbeen seeded with enzymatically derived omental micro­vascular cells.

There are several important aspects of our study. Itdemonstrated the efficacy and safety of endothelial cellseeding with cells derived from the omentum and anexcellent patency rate. It was possible to reach a very highseeding density and perform successful seeding in everyanimal with one operation only. Moreover, the studyshowed the safety for prospective clinical trials.

Some limitations of the study warrant comment. Al­though the study demonstrated promising results in thecanine model, the clinical value of seeding is not yetknown. Animal data are not transferable directly to theclinical setting because of species differences in the healingof synthetic grafts; spontaneous endothelialization of non­seeded small-diameter grafts is usually seen in animals butnever in humans. Therefore this experimental study re­quires validation by clinical trial because there has beenonly limited experience with endothelial cell seeding in

Ann Thorac Surg1994;58:677-84

PASIC ET ALENDOTHELIAL CELL SEEDING OF PROSTHETIC GRAFTS

683

humans. Although some clinical studies [19, 22] suggesteda possible advantage for endothelial cell seeding, espe­cially in nonsmokers [23], preliminary results are difficultto interpret because of lack of standardization of harvest­ing techniques, different sources of cells, variations in sizeof the cell inoculum, small numbers of patients, and lack ofan accurate method of surveillance of endothelial cell­seeded grafts [11].

At present, the endothelial cell-seeding procedure is notfeasible at every institution. Although already simplified, itis still a demanding procedure that requires a specializedcell laboratory and technicians. Therefore, for a largerclinical application, the method should be further modifiedfor easier clinical use without the need of a cell laboratoryand a specialized staff. An automated seeding device thatallows the processing of tissue in a closed, sterile systemcould be a possible solution.

The study was supported by the Kommision zur F6rderung derWissenschaftlichen Forschung, Bern, Switzerland, project nos.1576,1724.1, and 2178.1, and by Sulzer Medical Technology, Ltd,Winterthur, Switzerland.

We are indebted to Peter Groscurth, MD, and [org Auer, MD Vet,from the University of Zurich for their active support and contri­bution to the study.

References

1. Jones DN, Rutherford RB, Ikezawa T, Nishikimi N, IshibashiH, Whitehill TA. Factors affecting the patency of small-caliberprostheses: observations in suitable canine model. J Vase Surg1991;14:441-51.

2. Yates SG, Barros D'Sa AAB, Berger K, et al. The preclotting ofporous arterial prostheses. Ann Surg 1978;188:611-22.

3. Sharefkin JB, Fairchild KD, Albus RA, Cruess DF, Rich NM.The cytotoxic effect of surgical glove powder particles on adulthuman vascular endothelial cell cultures: implications forclinical uses of tissue culture techniques. J Surg Res 1986;41:463-72.

4. Schmidt SP, Monajjem N, Evancho MM, Pippert TR, SharpWV. Microvascular endothelial cell seeding of small-diameterDacron vascular grafts. J Invest Surg 1988;1:35-44.

5. Muller-Clauser W, Bay U, Lehmann KH, Turina M. Animproved procedure for enzymatic harvesting of highly puri­fied canine venous endothelial cells for experimental smalldiameter vascular prostheses. Ann Vase Surg 1989;3:134-9.

6. Anderson JS, Price TM, Hanson SR, Harker LA. In vitroendothelialization of small-caliber vascular grafts. Surgery1987;101:577-86.

7. Vojta JC, Via DP, Butterfield CE, Zetter BR. Identification and

DISCUSSION

DR JAY IVES LABOURENE (Ann Arbor, MI): I thought that wasa terrific and very stimulating paper.

I wonder if you looked at any synthetic function of the endo­thelial cell layer with respect to prostacyclin or other factors toshow that the cells are normally functioning endothelial cells.

DR PASIC: Grafts, whether seeded or nonseeded, produce pros­tacyclin in small amounts. Seeded grafts have a significantlyhigher production of prostacyclin than nonseeded grafts, but it islower than that of the native arteries. We noted that the antiplate-

isolation of endothelial cells based on their increased uptake ofacetylated-Iow density lipoprotein. J Cell BioI 1984;99:2034-40.

8. Zilla P, Fasol R, Dudeck U, et al. In situ cannulation, microgridfollow-up and low-density plating provide first passage en­dothelial cell mass cultures for in vitro lining. J Vase Surg1990;12:180-9.

9. Schrotter D, Spiess E, Paweletz N, Benker R. A procedure forrupture free preparation of confluently grown monolayer cellsfor scanning electron microscopy. J Electr Microsc Techn1984;1:219-25.

10. Cutler SJ, Ederer F. Minimum utilization of the life tablemethod in analyzing survival. J Chronic Dis 1958;8:699-712.

11. Mosquera DA, Goldman M. Endothelial cell seeding. Br J Surg1991;78:656-60.

12. Herring M, Baughman S, Glover J. Endothelium develops onseeded human arterial prosthesis: a brief clinical note. J VaseSurg 1985;2:727-30.

13. Zilla P, Fasol R, Deutsch M, et al. Endothelial cell seeding ofpolytetrafluoroethylene vascular grafts in humans: a prelimi­nary report. J Vase Surg 1987;6:535-41.

14. Ortenwall P, Wadenvik H, Risberg B. Reduced platelet depo­sition on seeded versus unseeded segments of expandedpolytetrafluoroethylene grafts: clinical observations after a6-month follow up. J Vase Surg 1989;10:374-80.

15. Rosenman JE, Kempczinski RF, Pearce WH, Silberstein EB.Kinetics of endothelial cell seeding. J Vasc Surg 1985;2:778-84.

16. Graham LM, Brothers TE, Vincent CK, Burkel WE, Stanley JCThe role of an endothelial cell lining in limiting distal anasto­motic intimal hyperplasia of 4-mm-LD. Dacron grafts in acanine model. J Biomed Mater Res 1991;25:525-33.

17. Pearce WH, Rutherford RB, Whitehill TA, et al. Successfulendothelial seeding with omentally derived microvascularendothelial cells. J Vase Surg 1987;5:203-10.

18. Watkins MT, Sharefkin JB, Zajtchuk R, et al. Adult humansaphenous vein endothelial cells: assessment of their repro­ductive capacity for use in endothelial seeding of vascularprostheses. J Surg Res 1985;36:588-96.

19. Magometschnigg H, Kadletz M, Vodrazka M, et al. Prospec­tive clinical study with in vitro endothelial cell lining ofexpanded polytetrafluoroethylene grafts in crural repeat re­construction. J Vase Surg 1992;15:527-35.

20. Sterpetti AV, Hunter WJ, Schultz RD, et al. Seeding withendothelial cells derived from the microvessels of the omen­tum and from the jugular vein: a comparative study. J VaseSurg 1988;7:677-84.

21. Herring M, Baughman S, Glover J, et al. Endothelial seeding ofDacron and polytetrafluoroethylene grafts: the cellular eventsof healing. Surgery 1984;96:745-54.

22. Kadletz M, Magometschnigg H, Minar E, et al. Implantation ofin vitro endothelialized polytetrafluoroethylene grafts in hu­man beings. A preliminary report. J Thorac Cardiovasc Surg1992;104:736-42.

23. Herring M, Gardner A, Glover J. Seeding human arterialprostheses with mechanically derived endothelium. J VascSurg 1984;1:279-89.

let agents that we administered for the first month after implan­tation impaired prostacyclin synthesis, which was significantlylower in the seeded grafts from dogs given antiplatelet drugs thanin the seeded grafts of dogs not given this therapy.

DR HIEP C NGUYEN (New York, NY): I enjoyed your paper. Inthe endothelial cells that you isolated, what percentage is contam­inated with the smooth muscle cell? We tried to do the same thing,except we isolated endothelial cells from the umbilical cord, and asignificant amount of our isolates were contaminated with smooth

684 PASIC ET ALENDOTHELIAL CELL SEEDING OF PROSTHETIC GRAFTS

Ann Thorac Surg1994;58:677-84

muscle cells. And, as you know, once contamination occurs, thesmooth muscle cells will take over in the culture, and they behavedifferently from the endothelial cells.

DR PASIC: Although the seeding suspension is a mixture ofdifferent nonendothelial types of cells, we supposed that themicrovascular endothelial cells from the omental tissue seededonto small-diameter prosthetic vascular grafts immediately beforeimplantation could proliferate rapidly to cover the graft andimprove the patency rate by exhibiting at least some normalendothelial-cell functions, such as antagonizing coagulation, pre­venting platelet deposition, and lysing fibrin formed in the graftlumen.

A large number of microvascular endothelial cells can beharvested from the omentum. We identified these cells by theircobblestone morphology, by the uptake of diacetylated low­density lipoprotein, and by positive immunofluorescent stainingfor von Willebrand's factor (factor VIII) antigen. However, theseendothelial cells are mixed in a different ratio with other cells,including mesothelial cells, fibroblasts, smooth muscle cells, peri­cytes, and blood cells. Contamination with nonendothelial cellsdepends on the harvesting method. A purer endothelial-cell yieldcan be achieved by lowering the separation density during har­vesting, but this leads to lowering the number of all cells available

for seeding. We know the number of endothelial cells in oursuspension, but we do not know the percentage of endothelialcells in the suspension.

It is difficult to distinguish the isolated endothelial cells fromother cells, especially from mesothelial cells, because both types ofcells are present in omental tissue. In a tissue culture, they havesimilar growth patterns under light microscopy. However, it wasshown that in a culture of mixed cells derived from omentum,endothelial cells were rapidly displaced by mesothelial cells,resulting in a pure culture of mesothelial cells. Similarly, in atissue culture of endothelial cells and fibroblasts, endothelial cellswere suppressed by fibroblasts that proliferated and formed aconfluent layer. Like endothelial cells, mesothelial cells produceprostacyclin and possess fibrinolytic activity. These cells producein vitro large amounts of tissue-type plasminogen activator to­gether with types 1 and 2 plasminogen activator inhibitor. More­over, when seeded onto vascular prostheses, these cells acquire aconfluent monolayer lining with no adherent platelets or amor­phous material within 1 month after operation.

The results of our studies have demonstrated an excellent latepatency rate for grafts seeded with this suspension of mixed cells.Therefore, we do not tend to achieve a suspension of purelyendothelial cells.

Notice From The Thoracic Surgery Foundationfor Research and Education

Applications for grants from The Thoracic Surgery Founda­tion for Research and Education, including the Nina StarrBraunwald Fund, are now available. Guidelines and applica­tion materials may be obtained by writing to The Thoracic

Surgery Foundation for Research and Education, 401 NMichigan Ave, Chicago, IL 60611-4267. You may also placeyour request by phone (312/644-6610) or fax (312/527-6635).Deadline for receipt of grant applications is October 1, 1994.