improved biocompatibility of small intestinal submucosa (sis) following conditioning by human...

Biomaterials 25 (2004) 515–525

ARTICLE IN PRESS

*Correspondin

3318.

E-mail addres

0142-9612/$ - see

doi:10.1016/S014

Improved biocompatibility of small intestinal submucosa (SIS)following conditioning by human endothelial cells

A.M. Woodsa, E.J. Rodenberga, M.C. Hilesb, F.M. Pavalkoa,*aDepartment of Cellular and Integrative Physiology, School of Medicine, Indiana University, 635 Barnhill Drive, Medical Science Building, Rm. 2069,

Indianapolis, IN 46202-5120, USAbCook Biotech, Inc., 3055 Kent Ave., West Lafayette, IN 47906, USA

Received 16 January 2003; accepted 1 July 2003

Abstract

Small intestinal submucosa (SIS) is a naturally occurring tissue matrix composed of extracellular matrix proteins and various

growth factors. SIS is derived from the porcine jejunum and functions as a remodeling scaffold for tissue repair. While SIS has

proven to be a useful biomaterial for implants in vivo, problems associated with endothelialization and thrombogenicity of SIS

implants may limit its vascular utility. The goal of this study was to determine if the biological properties of SIS could be improved

by growing human umbilical vein endothelial cells (HUVEC) on SIS and allowing these cells to deposit human basement membrane

proteins on the porcine substrate to create what we have called ‘‘conditioned’’ SIS (c-SIS). Using an approach in which HUVEC

were grown for 2 weeks on SIS and then removed via a technique that leaves behind an intact basement membrane, we hypothesized

that the surface properties of SIS might be improved. We found that when re-seeded on c-SIS, HUVEC exhibited enhanced

organization of cell junctions and had increased metabolic activity compared to cells on native SIS (n-SIS). Furthermore, HUVEC

grown on c-SIS released lower amounts of the pro-inflammatory prostaglandin PGI2 into the media compared to cells grown on n-

SIS. Additionally, we found that adhesion of resting or activated human platelets to c-SIS was significantly decreased compared to

n-SIS suggesting that, in addition to improved cell growth characteristics, conditioning SIS with human basement membrane

proteins might decrease its thrombogenic potential. In summary, conditioning of porcine SIS by human endothelial cells improves

key biological properties of the material that may improve its usefulness as remodeling scaffold for tissue repair. Identification of

critical modifications of SIS by human endothelial cells should help guide future efforts to develop more biocompatible vascular

grafts.

r 2003 Elsevier Ltd. All rights reserved.

Keywords: Cell adhesion; Cell culture; Endothelial cell; Platelet adhesion; Inflammation

1. Introduction

Recent developments in the field of tissue engineeringhave been driven by demand for replacement and repairtissues for human use. Small intestinal submucosa (SIS)is a resorbable, acellular bioscaffold composed ofextracellular matrix (ECM) proteins derived from thejejunum of pigs (reviewed in [1,2]). SIS has character-istics of an ideal tissue engineered biomaterial and canact as a bioscaffold for remodeling of many body tissuesincluding skin [3], body wall [3–5], musculoskeletalstructure [6], urinary bladder [7,8], and also supports

g author. Tel.: +1-317-274-3140; fax: +1-317-274-

s: [email protected] (F.M. Pavalko).

front matter r 2003 Elsevier Ltd. All rights reserved.

2-9612(03)00552-0

new blood vessel growth [9–11]. SIS induces site-specificremodeling of both organs and tissues depending on thesite of implantation [1]. Host cells are stimulated toproliferate and differentiate into site-specific connectivetissue structures, which have been shown to completelyreplace the SIS material within 90 days [6]. SIS’s abilityto induce tissue remodeling is not completely under-stood but has been strongly associated with angiogen-esis, cell migration and differentiation, and deposition ofECM [6]. SIS is a naturally occurring ECM rich in manycomponents known to support angiogenesis, includinggrowth factors [12] such as VEGF [13].Although SIS appears to have many characteristics of

an ideal biomaterial for construction of vascular grafts,potential biocompatibility problems such as lack ofrapid and efficient endothelialization, inflammation and

ARTICLE IN PRESSA.M. Woods et al. / Biomaterials 25 (2004) 515–525516

thrombogenicity of newly implanted grafts threatenlong-term patency. In this study, we characterizedadhesion, growth and prostacyclin production of humanumbilical vein endothelial cells (HUVEC), and exam-ined platelet adhesion on untreated ‘‘native’’ SIS (n-SIS)and on SIS that was ‘‘conditioned’’ (c-SIS) by HUVEC.The surface of SIS was conditioned by growing HUVECfor 14 days to allow deposition of human matrixproteins onto the surface of porcine SIS followed byremoval of the cells using a published procedure thatleaves intact the matrix proteins [14]. In this study, wedescribe improved biological properties of HUVEC onSIS that had been conditioned.

2. Materials and methods

2.1. Cell culture and SIS

HUVEC were obtained from Clonetics (East Ruther-ford, NJ) and grown in endothelial growth media EGM(Clonetics). EGM contains 2% fetal bovine serum, 12mg/ml bovine brain extract, 1mg/ml human epidermal growthfactor, and 1mg/ml hydrocortisone, getamicin, andamphotericin B. HUVEC were maintained at 5% CO2

at 37�C, and passages 1–4 were used for experiments.SIS was obtained from Cook Biotech, Inc. (West

Lafayette, IN) in a dehydrated form and was rehydratedin endothelial basal media (EBM; Clonetics) at 37� forat least 10min. Once hydrated, the SIS was fastened towells with an approximate area of 0.44 cm2 available forcell seeding.

2.2. Confocal microscopy

Samples of SIS for confocal microscopy were rinsedtwice in phosphate-buffered saline (PBS) and fixed in4% paraformaldehyde for 15min, rinsed extensively inPBS, permeabilized using 0.2% Triton X-100 andlabeled with appropriate antibodies. For visualizationby confocal microscopy, samples were mounted on glassslides with Fluoromount-Gt (Southern BiotechnologyAssociates Inc., Birmingham, AL). For evaluation ofcell spreading and cell–cell junction formation onindividual human matrix proteins, acid washed glassslides were coated with the following concentrations ofmatrix proteins for 16 h: 10 mg/ml human collagen (typeI) (Chemicon International, Temecula, CA); 10 mg/mlhuman fibronectin (Sigma Chemcial Co., St. Louis,MO); 2.5 mg/ml human laminin (Sigma Chemical Co.,St. Loius, MO); 0.5 mg/ml human vitronectin (ChemiconInternational, Temecula, CA). For experiments on glassslides coated with human matrix proteins, HUVECs atpassage 2 were seeded onto the glass slides at75,000 cells/mm2 and cultured for 72 h prior to fixationwith 4% paraformaldehyde. Visualization of cell–cell

adherens was made using mouse anti-b-catenin (ZymedLaboratories Inc., San Francisco, CA); tyrosine phos-phorylated proteins were visualized using a mouse anti-phosphotyrosine antibody (Transduction Laboratories,Lexington, KY). Actin filaments were visualized usingTexas-Red Phalloidin (Molecular Probes, Eugene,Oregon). Secondary antibodies used were FITC-con-jugated AffiniPure Donkey Anti-Rabbit IgG (H+L)and FITC-conjugated AffiniPure Goat Anti-Mouse IgG(H+L) (Jackson ImmunoResearch Laboratories, Inc.,West Grove, PA). Samples were observed and recordedusing Confocal Laser Microscopy.

2.3. Western blot analysis

Cells were harvested by direct lysis in sodium dodecylsulfate (SDS) gel sample buffer. Protein concentrationswere determined using the amido black method [15].Equal protein (30 mg) was loaded onto a 10% SDS-polyacrylamide gel for separation and transferred tonitrocellulose for immunoblot analysis. The sameprimary antibodies used in immunocytochemistry ex-periments as well as anti-mitochondrial glycosylase(ab6491) (Abcam Ltd., Cambridge, UK) along withthe appropriate horseradish peroxidase-labeled second-ary antibodies obtained from Jackson Immunoresearch(West Grove, PA) were used. Each experiment wasconducted at least in triplicate.For analysis of basement membrane protein deposi-

tion, SIS was minced into tiny sections and extracted inSDS gel sample buffer. These samples were then boiledand centrifuged at 14,000g for 10min. The supernatantwas collected and the remaining pieces of SIS werediscarded. Protein levels were equalized by amido backassays and loaded onto a 5% SDS-PAGE gel forseparation and transferred to nitrocellulose for immu-noblot analysis. Human fibronectin was detected usingrabbit anti-human fibronectin (Cederlane Laboratories,Ltd., Ont., Canada) with the appropriate horseradishperoxidase-labeled secondary antibodies (Jackson Im-munoresearch, Inc.).

2.4. Preparation of conditioned SIS

A procedure developed by Gospadarowicz [14] wasused to remove HUVEC from SIS with minimaldisruption to the basement membrane deposited by thecells. HUVEC were seeded at 100,000 cells/cm2 andgrown on SIS for 2 weeks and then rinsed several timesin PBS. 20mm solution of high-grade ammonia hydro-xide (NH4OH) (Mallinckrodt AR Select Cat No. 6665)at 40�C was placed in each well for approximately10min in situ. Wells were then rinsed several times withsterile water and vigorously mixed to remove all cellsfrom the SIS. Native SIS (n-SIS) was also subjected tothe same cell-stripping protocol used to prepare c-SIS.

ARTICLE IN PRESSA.M. Woods et al. / Biomaterials 25 (2004) 515–525 517

2.5. Verification of cell removal

We verified that all HUVEC were efficiently removedwhen preparing c-SIS visually using fluorescence micro-scopy and scanning electron microscopy (SEM), andbiochemically by immunoblot analysis. For fluorescencemicroscopy, samples were fixed in 4% paraformalde-hyde for 15min and then stained with Hoechst 33258nuclear staining dye obtained from Molecular Probes(Eugene. Oregon). Images of stained cell nuclei wereviewed and recorded on an RT Color Spotlights camera(Diagnostic Instruments, Sterling heights, MI) using anOptiphot-2 Nikon epifluorescent microscope. For SEM,samples were prepared by fixation in 4% gluteraldehyde,dehydrated through increasing ethanol concentrationsto 100% ethanol, sputter coated with gold and viewedusing a JOEL scanning electron microscope. Forimmunoblot analysis, the surface of SIS was scrapeddirectly in SDS sample buffer and an antibody specificfor the cellular enzyme anti-mitochondrial-8-oxogua-nine DNA-glycosylase (ab6491), Abcam Limited (Cam-bridge, UK) or for b-catenin along with the appropriatehorseradish peroxidase-labeled secondary antibodies(Jackson) was used to detect the presence of cells on SIS.

2.6. Cell adhesion assays

Cell adhesion to SIS was measured using metaboli-cally labeled HUVEC. Cells were grown in 150mm2

tissue culture dishes and labeled for 16 h in 25 mCi of 35S-methionine, rinsed in PBS and then removed from thedish using trypsin. 35S-methionine labeled HUVEC wereresuspended in growth media and equal numbers of cells(100,000/cm2) were added to n-SIS or c-SIS and allowedto adhere for varying lengths of time as indicated. Afterthe adhesion period, SIS was washed with PBS toremove unbound cells and the amount of radioactivityremaining on the SIS was measured using a liquidscintillation counter.

2.7. Metabolic activity assessment

To estimate the relative metabolic rates of cells grownon n-SIS and c-SIS, cells were seeded at 100,000 or200,000 cells/cm2 and cultured for 1 h to 2 weeks.Metabolic activity was assessed using CellTiter 96Aqueos Once Solution Cell Proliferation Assay obtainedfrom Promega (Madison, WI). The assay is a colori-metric method, which measures the amount of NADPHproduced by dehydrogenase enzymes in metabolicallyactive cells by determining the amount of coloredreaction product (Formazan) that can be producedfrom MTS tetrazolium compound (Owen’s reagent) inmedia samples. 200 ml of a solution containing 20 ml ofCellTiter 96 Aqueous Once Solution Reagent was addedto 100 ml of culture media and incubated at 37�C for 3 h

according to the manufacturer’s instructions. After 3 hOD490 was recorded on an ELISA plate reader.

2.8. Prostacyclin measurement

Prostacyclin release was measured by incubating SISwells in 200 ml of EGM media for 1 h. The media wascollected and centrifuged at 14,000g for 2min to pelletany particulates, and the supernatant was retained forprostacyclin measurement. 6-keto Prostaglandin F1a

EIA kit was obtained from Cayman Chemical (AnnArbor, MI) and used to determine the amount of PGI2present in each sample.

2.9. Platelet adhesion assays

Preparation of platelets: Whole blood was drawn byvenipuncture from healthy aspirin-free human donorsand collected in ACD (83mm sodium citrate, 111mm

glucose, 71.4mm citric acid, pH 4.5; 1:7, v/v). Platelet-rich plasma was isolated by centrifugation at 250g for20min, and prostaglandin E1 (0.1mm) or prostaglandinI2 (50 ng/ml) was added. The platelet-rich plasma wascentrifuged for 15min at 850g; and the platelets wereresuspended in Hepes-buffered saline ((Buffer A) 10mm

Hepes, pH 7.4, 138mm NaCl, 12mm NaHCO3, 10mm

KCl, 5.5mm glucose, 0.35%BSA, 2 units/ml heparinand 1 unit/ml apyrase). Resuspended platelets werewashed three times in Buffer A without BSA orinhibitors and the concentration was adjusted to (0.5–1.0)� 109 platelets/ml. The platelets were then labeledwith 51Cr (0.5mCi) for 60min. They were washed twotimes in Buffer A without inhibitors. For plateletactivation, to 6–8ml of resuspended platelets, 2 unitsof thrombin were added. 51Cr-labeled platelets wereallowed to incubate for 1 h at room temperature, andthen platelets were aspirated off, followed by asubsequent 5ml PBS wash. Adherent platelets werelysed using 200 ml of 2% SDS twice. The lysates werecollected, and a g-counter was used to determine theradioactivity in each sample. All means calculated havesix replicates.

2.10. Statistical analysis

Statistical analysis was made by ANOVA and po0:05was considered significant.

3. Results

3.1. Characterization of HUVEC growth on native SIS

A variety of cell types have been grown successfullyon SIS including: mouse Swiss 3T3 fibroblasts, primaryhuman urinary bladder stromal cells, primary canine


prostate carcinoma cells, rat pulmonary endothelialcells, human epidermal keratinocytes, and humanmicrovascular endothelial cells [16–18], but not HU-VEC. To verify that HUVEC would also grow on SIS,cells were seeded at a density of 100,000 cells/cm2 andallowed to grow for various lengths of time. Fig. 1 showsconfocal microscope images illustrating the appearanceof HUVEC on SIS and glass at various times afterseeding using an antibody against the cell–cell junctionprotein b-catenin, followed by FITC-labeled secondaryantibody and with Texas-Red-phalloidin to visualizeactin. HUVEC grown on glass slides were confluent by 2days after seeding and exhibited a typical cobblestoneappearance characteristic of endothelial cells in cultureand similar to the morphology of endothelial cells in vivo[19]. However, HUVEC grown on SIS at the same initialseeding density of 100,000 cells/cm2 for 2 days did notform a confluent monolayer with well-developed cell–cell junctions. We found that 5 days after seeding,HUVEC formed a confluent monolayer when grown onSIS

Fig. 1. HUVEC grown on SIS form a confluent monolayer more

slowly than when grown on glass slides. HUVEC were seeded onto

glass slides or untreated SIS at a concentration of 100,000 cells/mm2

and analyzed by immunofluorescence confocal microscopy to deter-

mine the length of time needed for cells to form a confluent monolayer.

Cells were labeled with a mouse monoclonal antibody against b-catenin followed by an anti-mouse Ig antibody conjugated to FITC to

visualize cell–cell junctions. Cells were also labeled with phalloidin

conjugated Texas-Red to visualize filamentous actin.

3.2. HUVEC condition SIS

To determine if the surface of SIS could be modifiedin such a way as to render it more favorable forendothelial cell growth, HUVEC were seeded on SISand grown for 14 days to allow sufficient time forthe cells to deposit human matrix proteins ontothe surface of the porcine SIS. Thus, HUVEC werecultured for approximately 9 days after reachingconfluency. HUVEC were then removed from theSIS using NH4OH as has been described previously[20] to prepare extracellular matrices from endothelialcells according to a procedure originally describedby Gospodarowicz [14]. We referred to this processas ‘‘conditioning’’ of the SIS. For all experiments,native SIS (n-SIS) was treated with the same cellremoval protocol. Fig. 2A shows microscopy imagesof untreated SIS, SIS with cells grown for 14 daysand left intact, and SIS with cells removed (conditioned)that were stained with Hoechst dye to label cell nucleito verify that all cells were removed from the SIS.

Fig. 2. HUVEC are efficiently removed from SIS following ‘‘con-

ditioning.’’ Native SIS without cells, SIS with HUVEC grown for 14

days, and SIS with HUVEC grown for 14 days and then removed

(conditioned) were stained with Hoechst dye to label cell nuclei and

viewed in the fluorescence microscope (A) or analyzed by immuno-

blotting using antibodies against either of two cellular proteins, b-catenin or mitochondrial glycosylase (B). Both visual (microscopy) and

biochemical (immunoblot) analysis demonstrate that no cells are

detected on the SIS following the conditioning treatment.


Confirmation that all cells were removed was alsomade by immunoblotting for the cellular proteinsb-catenin and mitochondrial-DNA-8-oxanine-glycosy-lase (Fig. 2B).To demonstrate that HUVEC deposit proteins

that remain associated with SIS, an antibody specificfor human fibronectin (hFN) was identified that didnot cross-react with the porcine fibronectin. Westernblot analysis revealed that after allowing HUVEC tobe grown and cultured on n-SIS for 2 weeks and thenremoved, hFN was detected on the surface in theabsence of cells. Fig. 3 shows a western blot to illustratethe presence of hFN on the surface of SIS followingthe conditioning treatment. An antibody specific forhuman laminin that did not cross-react with porcinelaminin was also used. In this case, no evidence ofsecretion of laminin onto conditioned SIS was found(not shown).To verify that the cell removal treatment did not cause

obvious structural damage to the SIS, SEM was used tovisualize the SIS surface (Fig. 4). SEM of native SISsubjected to the cell removal treatment and of condi-tioned SIS demonstrated that the SIS surface was notstructurally damaged by the cell removal treatment incomparison to native SIS that did not undergo the cellremoval treatment.

Fig. 3. HUVEC deposit human fibronectin onto the surface of

conditioned SIS. Immunoblot analysis of native SIS without cells,

SIS with HUVEC grown for 14 days, and SIS with HUVEC grown

for 14 days and then removed (conditioned) using an antibody that

is specific for human fibronectin and has little or no cross reactivity

with porcine fibronectin. Human fibronectin deposited by HUVEC

is retained on the conditioned SIS following removal of the

cells.

Fig. 4. Scanning electron microcopy of the surface of SIS. SEM of untreated

conditioned SIS indicted there was no obvious structural damage to the SIS

3.3. HUVEC attachment is not different on c-SIS

compared to n-SIS

We next evaluated whether there was a difference ininitial cell adhesion during the first 1–48 h after seedingHUVEC onto n-SIS and c-SIS (Fig. 5). HUVECgrowing on tissue culture plastic were labeled with35S-methionine for 18 h during log phase growth toallow incorporation of the radioactive label into newlysynthesized proteins. The cells were washed free ofunincorporated radioactivity in the media and thelabeled cells were collected by brief trypsinization.Equal numbers of cells were seeded onto n-SIS orc-SIS and allowed to adhere. At each time pointexamined, unbound cells were washed away and theadherent cells were fixed to the SIS using 4%paraformaldehyde and bound radioactivity was mea-sured using a g-counter. Fig. 5 shows there were nostatistically significant differences in the number of cellspresent on n-SIS vs. c-SIS at any of the time pointsexamined between 1 and 48 h.

SIS, native SIS (subjected to the NH4OH cell removal treatment) and

as a result of the cell removal treatment.

Fig. 5. Cell attachment assays on native and conditioned SIS.

HUVEC were labeled with 35S-methonine overnight and then seeded

onto native and conditioned SIS to determine if there was a difference

in the relative number of cells that were able to adhere to the two

surfaces. There was no statistically significant difference in the relative

numbers of HUVEC adhered to native or conditioned SIS at times

ranging from 1 to 48 hr after seeding.


3.4. Growing HUVEC on c-SIS improves cell–cell

junction organization

We next sought to determine if there were anydifferences in the morphological characteristics ofHUVEC grown on n-SIS vs. c-SIS. Fig. 6 illustratesthe dramatic improvement in the organization of cell–cell junctional complexes (adherens junctions) when cellsare grown on c-SIS compared to cells grown on n-SIS at48 h after seeding. Cells were labeled for b-catenin andfor phosphotyrosine containing proteins and imagedusing confocal microscopy.Cells grown on n-SIS appeared to be visible in

multiple different focal planes indicating the surface ofn-SIS was relatively rough. In contrast, cells grown onc-SIS were visible across the surface of the SIS primarilywithin a single focal plane suggesting that conditioningof the SIS might modify the surface of the SIS bydeposition of a layer of matrix proteins that creates asmooth surface that allows cells to spread moreuniformly. This result suggests that SIS conditionedwith human basement membrane proteins provides animproved substrate for the establishment of endothelial

Fig. 6. Formation of cell–cell adherens junctions is improved in

HUVEC grown on conditioned SIS compared to cells grown on native

SIS. HUVEC were seeded onto native or conditioned SIS at a

concentration of 100,000 cells/cm2 and analyzed by immunofluores-

cence confocal microscopy 48 h after seeding. Cells were labeled with a

mouse monoclonal antibody against b-catenin or with mouse

monoclonal antibody against phosphotyrosine containing proteins

followed by an anti-mouse Ig antibody conjugated to FITC to

visualize cell–cell junctions. Cells grown on conditioned SIS form well

organized cell–cell cell junctions as indicated by the appearance of b-catenin staining at the membrane whereas cells grown on native SIS

have patchy, discontinuous membrane staining. Phosphotyrosine

staining, which can be an indicator of active cell–cell and cell–matrix

signal transduction activity, is also more abundant in cells grown on

conditioned SIS compared to native SIS.

monolayers with well-organized cell–cell junctions. Wealso tested whether coating glass slides with each of fourdifferent purified human matrix proteins could providean improved substrate for establishment of well-organized cell–cell junctions. Consistent with ourobservation that conditioned SIS contained humanfibronectin, but not human laminin, cell–cell junctionformation (Fig. 7) and phosphotyrosine staining (Fig. 8)at cell–cell junctions was improved when HUVEC wereplated on to human fibronectin and was similar to thatseen on c-SIS. In contrast, cell–cell junction formationand phosphotyrosine staining was weaker when HU-VEC were plated on to human laminin. Interestingly,both cell–cell junction formation and phosphotyrosinestaining at cell junctions were also more similar to thatseen on c-SIS when HUVEC were plated on to eitherhuman type I collagen or human vitronectin, suggestingthat these two proteins may be present in c-SIS. Wewere, however, unable to verify this due to the lack ofappropriate human-specific antibodies against collagenor vitronectin that did not also exhibit cross reactivitywith the porcine protein.

3.5. Metabolic activity of HUVEC is higher on c-SIS

To evaluate the metabolic activity of HUVEC grownon n-SIS and c-SIS, metabolic activity assays, whichmeasured the production and release of NADPH intothe culture media, were conducted. HUVEC wereseeded onto n-SIS and c-SIS (100,000 cells/cm2 and200,000 cells/cm2) and allowed to grow over a period of1–240 h. At both seeding densities, HUVEC grown on n-SIS exhibited a significant decrease in NADPH produc-tion after 24 and 96 h in culture, despite the fact equalnumbers of cells adhered to n-SIS and c-SIS (see Fig. 5),when compared to their initial (1 h) levels (Fig. 9). Insharp contrast, HUVEC grown on c-SIS exhibited asignificant increase in NADPH production after 24 h inculture. NADPH production by HUVEC remainedsignificantly higher in cells grown on c-SIS comparedto n-SIS through the first 96 h. By 168 h culture,NADPH levels were not different between groups; after240 h in culture NADPH production had droppedsignificantly below initial levels for both substrates andwere not significantly different from each other.

3.6. PGI2 production is reduced when HUVEC are grown

on c-SIS

To evaluate the inflammatory response of HUVECgrown on n-SIS and c-SIS, production and release of theinflammatory prostaglandin, prostacyclin (PGI2), wasmeasured using an enzyme-linked immunosorbant assay(EIA). HUVEC were seeded at an initial density of100,000 cells/cm2 and grown on n-SIS or c-SIS for 24 or48 h. After 24 or 48 h, fresh media was incubated with

ARTICLE IN PRESS

Fig. 7. Formation of cell–cell adherens junctions is improved and similar to c-SIS when HUVEC are grown on purified human fibronectin, human

type I collagen, or human vitronectin, but not when grown on human laminin. HUVEC were seeded onto glass slides coated with either human type I

collagen, human fibronectin, human laminin or human vitronectin at a concentration of 75,000 cells/mm2 and analyzed by immunofluorescence

confocal microscopy 72 h after seeding. Cells were labeled with a mouse monoclonal antibody against b-catenin followed by an anti-mouse Ig

antibody conjugated to FITC to visualize cell–cell junctions. Cells grown on collagen, fibronectin or vitronectin formed well organized cell–cell

junctions as indicated by the appearance of b-catenin staining at the membrane whereas cells grown on human laminin had weak membrane staining.

A.M. Woods et al. / Biomaterials 25 (2004) 515–525 521

the wells for 1 h and the media was collected.Significantly higher levels of PGI2 were releasedinto the media by cells grown on n-SIS compared tocells grown on c-SIS (Fig. 10). This difference (approxi-mately 3-fold more PGI2 (pg/ml) than HUVEC onc-SIS), increased to approximately 4-fold at the 48 htime point.

3.7. Platelet adhesion is reduced on c-SIS

To determine whether conditioning of SIS mightaffect the thrombogenic potential of SIS, plateletadhesion assays were performed. Using freshly isolatedhuman platelets, we found that conditioning the SISresulted in a significant decrease in platelet adhesionof approximately 40% when compared to n-SIS(Fig. 11A). This value of a 40% decrease was typicalof four independent experiments using platelets from

multiple donors in which the decrease in plateletadhesion on c-SIS ranged from 30–43%. In eachindependent experiment this decrease was statisticallysignificant (po0:05 by ANOVA). Interestingly, plateletadhesion was also decreased (by approximately 21–26%) on SIS with HUVEC still present, compared to n-SIS, but this difference was not statistically significant inany of four independent experiments (data not shown).The platelet adhesion assay was repeated using

platelets that were activated by the addition of thrombinto more accurately mimic the conditions likely found inthe vicinity of an implanted graft (Fig. 11B). Whencompared to n-SIS, c-SIS significantly reduced adhesionof activated platelets by approximately 30%. As withunactivated platelets, in each of four independentexperiments, adhesion of thrombin-activated plateletsto SIS with HUVEC still present was consistently lowerby 15–20%, but this difference was not statistically

ARTICLE IN PRESS

Fig. 8. Phosphotyrosine staining at cell–cell adherens junctions is improved and similar to c-SIS when HUVEC are grown on purified human

fibronectin, human type I collagen, or human vitronectin, but not when grown on human laminin. HUVEC were seeded onto glass slides coated with

either human type I collagen, human fibronectin, human laminin or human vitronectin at a concentration of 75,000 cells/mm2 and analyzed by

immunofluorescence confocal microscopy 72 h after seeding. Cells were labeled with a mouse monoclonal antibody against phosphotyrosine followed

by an anti-mouse Ig antibody conjugated to FITC to visualize cell–cell junctions. Phosphotyrosine staining, which can be an indicator of active cell–

cell and cell–matrix signal transduction activity, was more abundant in cells grown on collagen, fibronectin or vitronectin whereas cells grown on

human laminin had weak phosphotyrosine staining.

A.M. Woods et al. / Biomaterials 25 (2004) 515–525522

significant. We also coated n-SIS with commerciallyavailable human fibronectin to determine if coating withthis single ECM protein that we know is present in c-SIScould mimic the decrease in platelet adhesion seen on c-SIS. Although coating n-SIS with human fibronectinconsistently reduced platelet adhesion by approximately6–15% in each of three independent experimentscompared to uncoated n-SIS, this difference failed toconsistently achieve statistical significance (not shown).

4. Discussion

In this study we investigated the biological surfaceproperties of SIS that was ‘‘conditioned’’ by the

deposition of human matrix proteins by humanendothelial cells in comparison to the ‘‘native’’ SIS.The interface between host tissues and biomaterials usedfor tissue grafts and implants plays a critical role indetermining the success of grafts in vivo. Futureapplications of SIS as a tissue matrix may include itsuse as a vascular graft material. However, potentialproblems associated with the biological surface proper-ties of SIS, such as inefficient endothelialization andthrombus formation, could restrict its usefulness as avascular matrix [1,2]. We show here that, not only canSIS support adhesion and growth of HUVEC, but alsothat these cells are capable of modifying the surface ofSIS, possibly by the deposition of ECM proteins. Whenthe cells were removed from the SIS, the modified or

ARTICLE IN PRESS

Fig. 9. NADPH production is greater at 24 and 96 h after plating in

HUVEC grown on conditioned SIS compared to native SIS. Metabolic

activity assays that measure the production of NADPH by cells were

carried out with native and conditioned SIS onto which 100,000 or

200,000 cells/cm2 were seeded for 1–240h. At 24 and 96 h after seeding,

NADPH production by HUVEC grown on conditioned SIS was

significantly higher than by cells grown on native SIS. At the 168 h

time point there was no difference between the groups. Note that error

bars have been omitted to simplify the appearance of the graph.

Conditioned vs. native at 24 and 96 h, po0:005:

Fig. 10. Prostacyclin (PGI2) release from HUVEC is lower when cells

are grown on conditioned SIS compared to cells grown on native SIS.

HUVEC were seeded at 100,000 cells/cm2 on native and conditioned

SIS and grown for 24 or 48 h. After 24 or 48 h the media was replaced

and HUVEC were cultured for an additional 1 h and PGI2 in the media

was measured. HUVEC grown on conditioned SIS released signifi-

cantly less prostacyclin (po0:0001) compared to cells grown on native

SIS at both 24 and 48 h time points.

Fig. 11. Adhesion of human platelets is lower on conditioned SIS

compared to native SIS. Adhesion of resting (A) or thrombin activated

(B) human platelets labeled with 51Cr was measured on native SIS and

conditioned SIS. Platelet adhesion was significantly reduced by 40.4%

when resting platelets were used, and by 29.6% when thrombin-

activated activated platelets were used. �po0:05 vs native SIS.

A.M. Woods et al. / Biomaterials 25 (2004) 515–525 523

‘‘conditioned’’ surface provided an improved biologicalsurface. The surface of conditioned SIS has severaldistinct biological advantages over the native SIS whenendothelial cells were seeded onto the SIS including: (1)more rapid formation of endothelial monolayers, (2)improved organization of the adherens junctions thatform between endothelial cells, (3) increased metabolicactivity of endothelial cells, (4) and decreased release ofinflammatory prostacyclin (PGI2) from endothelial cells.Together, these findings indicate that conditioning ofSIS, possibly by deposition of human matrix proteins,by HUVEC renders the surface of porcine SIS morebiologically compatible with human cells and probablythe human vasculature. Results of this study should helpguide future efforts to utilize human ECM proteins,most notably human fibronectin, in surface coatings forvascular grafts to improve graft biocompatibility.In order to improve the chances for long-term patency

of vascular grafts it is desirable for the graft surface tobe able to support rapid and efficient endothelializationof the graft surface and platelet adhesion and thrombusformation in vascular grafts must be minimized [2,21].Therefore, we also examined the effect of ‘‘condition-ing’’ SIS on the tendency of human platelets to adhere tothe surface of the material. We found that either restingor activated platelets adhered less to conditioned SISthan to native SIS providing strong evidence thatmodification of the surface of SIS with human basementmembrane proteins might result in a less thrombogenicsurface.The results described in these studies using primary

cultured human endothelial cells cannot be directlyequated to SIS function as a vascular graft. However,these studies do provide valuable insights into thepotential improvements that can be made to the surfaceof SIS to improve its biological properties and allow usto make predictions about the requirements for design-ing an ideal graft that will be successful in vivo. Forexample, a key requirement for proper endothelialbarrier function is the formation of well-organizedadherens junctions between cells. Our results indicatethat HUVEC grown on c-SIS more rapidly formorganized cell–cell junctions than do cells grown on n-SIS. This could have important implications with respectto the ability of SIS vascular grafts to form an efficientpermeability barrier after implantation. Similarly, evi-dence of increased metabolic activity during the first 96hafter seeding of HUVEC on c-SIS compared to n-SIScould be extrapolated to predict that endothelial cellsin vivo might exhibit improved anabolic properties ifallowed to be grown on SIS that was conditioned withhuman matrix proteins. Results of this study shouldguide future efforts to develop more biocompatiblevascular grafts. For example, our results suggest thatcells grown on type I collagen, fibronectin or humanvitronectin exhibit well organized cell–cell junctions and


high levels of tyrosine kinase activity as evidenced byphosphotyrosine staining at cell–cell junctions, similar tocells grown on c-SIS. In contrast, cells grown on lamininexhibit poorly organized cell–cell junctions and have lowlevels of phosphotyrosine staining. Together, these newdata suggest that coating vascular grafts with eitherhuman type I collagen, human fibronectin, or humanvitronectin, or perhaps a combination of these proteins,but not human laminin, might provide a substrate thatpromotes cell–cell junction formation and activation oftyrosine kinase activity similar to that seen on c-SIS.Previous studies in which human epidermal cells or

fibroblasts were plated onto SIS demonstrated severalbasement membrane proteins including fibronectin, col-lagen (type IV and VII), and laminin were produced bythe cells [22]. However, large scale conditioning of SISusing human endothelial cells would be problematic bothin terms of cost and time. Therefore, it would bepreferable to identify key matrix components secreted byendothelial cells that improve biological surface propertiesof SIS. Identification of the human matrix proteinspresent in conditioned SIS will facilitate this process. Inthis study we have identified human fibronectin as acomponent of the conditioned SIS surface. Preliminarystudies using human fibronectin adsorbed to the surface ofnative SIS have indicated that the presence of humanfibronectin alone does not recapitulate the properties ofconditioned SIS (data not shown). We did not investigatethe possible deposition of human growth factors onto SISby HUVEC in this study. However, it is possible thatfuture efforts can identify a combination of matrixproteins and growth factors produced by human en-dothelial cells that can be adsorbed directly to the surfaceof SIS to improve its biological properties and obviate theneed to undergo the expensive and time consumingprocess of growing endothelial cells on the SIS.

5. Conclusions

We have shown that the surface properties of SIS areimproved by culturing human endothelial cells on theSIS surface and then removing the cells in a process werefer to as ‘‘conditioning’’. Because of the important roleplayed by tissue-implant interfaces in grafts in vivo,

creation of a biocompatible surface is critical to thesuccess of implanted materials. This study shows thatSIS supports adhesion and growth of human umbilicalvein endothelial cells and that these cells are capable ofmodifying the surface of SIS, possibly by the depositionof extracellular matrix proteins. Advantages of condi-tioned SIS over the native SIS include more rapidformation of endothelial monolayers, improved organi-zation of the adherens junctions that form betweenendothelial cells, increased metabolic activity of en-dothelial cells, and decreased release of inflammatory

prostacyclin (PGI2) from endothelial cells. Althoughhuman fibronectin alone cannot recapitulate all of thebeneficial effects of conditioned by HUVECs, fibronec-tin is deposited by HUVEC onto SIS and may be animportant factor in creating a more biocompatible SISsurface. These findings will guide future efforts to utilizeendothelial cells and human ECM proteins, mostnotably human fibronectin, in surface coatings forvascular grafts to improve graft biocompatibility.

Acknowledgements

The authors thank Dr. Jonathan Jones for helpfuladvice with the cell removal technique. We thank Dr.Suzanne Norvell for critical reading of the manuscriptand valuable advice with experiments, and Rita Gerardfor assistance with experiments. Thanks to Dr. SaraSawyer for help and advice on the growth of endothelialcells on SIS. This work was support by NIH grantsAR45218 and AR45831 and by a grant from the Indiana21st Century Fund to FMP.

References

[1] Hodde J. Naturally occurring scaffolds for soft tissue repair and

regeneration. Tissue Eng 2002;8(2):295–308.

[2] Lantz GC, Badylak SF, Hiles MC, Coffey AC, Geddes LA,

Kokini K, Sandusky GE, Morff RJ. Small intestinal submucosa

as a vascular graft: a review. J Invest Surg 1993;6(3):297–310.

[3] Prevel CD, Eppley BL, Summerlin DJ, Sidner R, Jackson JR,

McCarty M, Badylak SF. Small intestinal submucosa: utilization

as a wound dressing in full-thickness rodent wounds. Ann Plast

Surg 1995;35(4):381–8.

[4] Clarke KM, Lantz GC, Salisbury SK, Badylak SF, Hiles MC,

Voytik SL. Intestine submucosa and polypropylene mesh for

abdominal wall repair in dogs. J Surg Res 1996;60(1):107–14.

[5] Prevel CD, Eppley BL, Summerlin DJ, Jackson JR, McCarty M,

Badylak SF. Small intestinal submucosa: utilization for repair of

rodent abdominal wall defects. Ann Plast Surg 1995;35(4):374–80.

[6] Badylak SF, Tullius R, Kokini K, Shelbourne KD, Klootwyk T,

Voytik SL, Kraine MR, Simmons C. The use of xenogeneic small

intestinal submucosa as a biomaterial for Achilles tendon repair in

a dog model. J Biomed Mater Res 1995;29(8):977–85.

[7] Kropp BP, Eppley BL, Prevel CD, Rippy MK, Harruff RC,

Badylak SF, Adams MC, Rink RC, Keating MA. Experimental

assessment of small intestinal submucosa as a bladder wall

substitute. Urology 1995;46(3):396–400.

[8] Kropp BP, Rippy MK, Badylak SF, Adams MC, Keating MA,

Rink RC, Thor KB. Regenerative urinary bladder augmentation

using small intestinal submucosa: urodynamic and histopatholo-

gic assessment in long-term canine bladder augmentations. J Urol

1996;155(6):2098–104.

[9] Sandusky Jr GE, Badylak SF, Morff RJ, Johnson WD, Lantz G.

Histologic findings after in vivo placement of small intestine

submucosal vascular grafts and saphenous vein grafts in the

carotid artery in dogs. Am J Pathol 1992;140(2):317–24.

[10] Sandusky GE, Lantz GC, Badylak SF. Healing comparison of

small intestine submucosa and ePTFE grafts in the canine carotid

artery. J Surg Res 1995;58(4):415–20.


[11] Knapp PM, Lingeman JE, Siegel YI, Badylak SF, Demeter RJ.

Biocompatibility of small-intestinal submucosa in urinary tract as

augmentation cystoplasty graft and injectable suspension.

J Endourol 1994;8(2):125–30.

[12] Voytik-Harbin SL, Brightman AO, Kraine MR, Waisner B,

Badylak SF. Identification of extractable growth factors from

small intestinal submucosa. J Cell Biochem 1997;67(4):478–91.

[13] Hodde JP, Record RD, Liang HA, Badylak SF. Vascular

endothelial growth factor in porcine-derived extracellular matrix.

Endothelium 2001;8(1):11–24.

[14] Gospodarowicz D. Preparation of extracellular matrices produced

by cultured bovine corneal endothelial cells and PF-HR-9 endo-

dermal cells: their use in cell culture. In: Barnes DW, Sirbasku DA,

Stao GH, editors. Methods for preparation of media. Supplements

and Substrata. New York: Alan R Liss; 1984. p. 275–93.

[15] Sheffield JB, Graff D, Li HP. A solid-phase method for the

quantitation of protein in the presence of sodium dodecyl sulfate

and other interfering substances. Anal Biochem 1987;166(1):49–54.

[16] Badylak SF, Record R, Lindberg K, Hodde J, Park K. Small

intestinal submucosa: a substrate for in vitro cell growth.

J Biomater Sci Polym Ed 1998;9(8):863–78.

[17] Hodde JP, Record RD, Tullius RS, Badylak SF. Retention

of endothelial cell adherence to porcine-derived extracellular

matrix after disinfection and sterilization. Tissue Eng 2002;8(2):

225–34.

[18] Badylak S, Liang A, Record R, Tullius R, Hodde J. Endothelial

cell adherence to small intestinal submucosa: an acellular

bioscaffold. Biomaterials 1999;20(23–24):2257–63.

[19] Fung YC. Biomechanics: motion, flow, stress, and growth. New

York: Springer-Verlag New York Inc.; 1990. p. 196–225.

[20] Gonzales M, Weksler B, Tsuruta D, Goldman RD, Yoon KJ,

Hopkinson SB, Flitney FW, Jones JC. Structure and function of a

vimentin-associated matrix adhesion in endothelial cells. Mol Biol

Cell 2001;12(1):85–100.

[21] Prevel CD, Eppley BL, McCarty M, Jackson JR, Voytik SL, Hiles

MC, Badylak SF. Experimental evaluation of small intestinal

submucosa as a microvascular graft material. Microsurgery

1994;15(8):586–591; Discussion 592–3.

[22] Lindberg K, Badylak SF. Porcine small intestinal submucosa

(SIS): a bioscaffold supporting in vitro primary human epidermal

cell differentiation and synthesis of basement membrane proteins.

Burns 2001;27(3):254–66.

improved biocompatibility of small intestinal submucosa (sis) following conditioning by human...

Documents