ARTICLE IN PRESS

0142-9612/$ - se

doi:10.1016/j.bi

�Correspondfax: +335 56 90

E-mail addr

(P. Fernandez).

Biomaterials 26 (2005) 5042–5047

www.elsevier.com/locate/biomaterials

Technical note

Quality control assessment of ePTFE precoating procedure for invitro endothelial cell seeding

Philippe Fernandeza,�, Audrey Deguettea, Laurent Pothuauda, Genevieve Belleanneeb,Pierre Costec, Laurence Bordenavea

aINSERM-U577, Universite Victor Segalen Bordeaux 2, 146, Rue Leo Saignat, 33320 Bordeaux, FrancebService d’Anatomie et Cytologie Pathologiques, Hopital du Haut-Leveque, CHU, Bordeaux, France

cINSERM-U441 et Service de Cardiologie et Maladies Vasculaires, Soins Intensifs et Cardiologie Interventionnelle,

Hopital du Haut-Leveque, CHU, Bordeaux, France

Received 2 September 2004; accepted 5 January 2005

Abstract

Over the past two decades, in vitro autologous endothelial cell (EC) coverage of expanded polytetrafluoroethylene (ePTFE) graft

has been developed and clinically applied with success in infrainguinal bypasses. Before endothelialization, the luminal surface of the

graft has to be coated with a currently used fibrinolytically inhibited fibrin glue. The aim of this work is to validate the precoating of

the ePTFE (4mm ID) ringed graft with fibrin. Twenty cm-long grafts were precoated with fibrin glue (2 operators) then fixed for

microscopy investigations. Grafts were sliced into 3 regions. Thickness analysis was evaluated by image processing. Three grafts

have been tested for endothelialization and observed at days 3, 8. Cell-free coated ePTFE were imaged using high-frequency

ultrasound modality. Whatever the examined segment an overall homogeneous covering protein is shown. Fibrin thickness after

image processing is 8.570.25 and 4.170.4 mm for two operators (Po.001). We have evaluated reproducibility and inter- and intra-

variability of the operator, assessed quality controls and quality assurance all along the prosthesis and finally endothelialization and

subsequent behaviour under shear stress conditions.

r 2005 Elsevier Ltd. All rights reserved.

Keywords: Vascular grafts; Polytetrafluoroethylene; Fibrin; Endothelialisation; Autologous cells; Shear

1. Introduction

Bypass surgery is a common treatment for cardio-vascular disease, including coronary artery and periph-eral vascular disease which is the leading cause ofmortality in both the USA and Europe. Autologousvessels are the preferred replacement grafts for diseasedsegments smaller than 5mm in diameter but manypatients do not have veins suitable for vascularreconstruction. In these patients, the development of aviable, compliant-elasticity similar to arterial, small-

e front matter r 2005 Elsevier Ltd. All rights reserved.

omaterials.2005.01.011

ing author. Tel.: +335 57 57 14 83;

05 17.

ess: philippe.fernandez@u-bordeaux2.fr

diameter vascular bypass graft has been an intense areaof focus.

Over the past two decades, the group of Zilla [1–3] hasdeveloped and clinically applied an in vitro endothelia-lization procedure whereby infrainguinal expandedpolytetrafluoroethylene (ePTFE) prostheses are conflu-ently lined with cultured autologous endothelial cellsbefore implantation. In the context of coronary bypassoperations, emerging strategies involving recipient auto-logous cells seeded on ePTFE vascular grafts [4] are nowreported to overcome the limited availability or insuffi-cient quality of autologous bypass material. Becauseseeded endothelial cells are not usually retained onsynthetic substrates, the luminal surface of the graft hasto be coated before cell seeding to promote endo-thelial cell adhesion and spreading : the fibrinolytically

ARTICLE IN PRESSP. Fernandez et al. / Biomaterials 26 (2005) 5042–5047 5043

inhibited fibrin glue is currently used in clinical trials[3,5]. However, if the coating procedure appearsrather standardized for 6 or 7mm ID, it seems fromour point of view much more difficult to achieve insmaller calibres as well as to perform in a reproduciblemanner. Thus, with the focus of creating small-dia-meter blood-vessel substitutes, the aims of the presentpaper are to validate the precoating of the ePTFE (4mmID) ringed vascular graft with the fibrin matrix,by (i) evaluating the coating process : feasibility,reproducibility and inter- and intra-variability of theoperator, (ii) assessing quality controls and qualityassurance all along the prosthesis in order to verify that coating is homogeneous throughout the entirelength of the graft, finally, (iii) by an in vitro endo-thelialization cell seeding and a control of the main-tainence of endothelial cell monolayer under shear stressconditions.

2. Materials and methods

Thin-walled ePTFE ringed grafts with 4mm innerdiameter and standard 30 mm internodal distance werekindly provided by W.L. Gore & Associates, Inc,Flagstaff, Ariz.

Nineteen ePTFE grafts 20 cm in length were pre-coated with the clinically approved and fibrinolyticallyinhibited fibrin glue (Tissucols, kindly provided byBaxter AG, Maurepas, France) as previously described[1,3,4].

For scanning electron microscopy (SEM) wholesamples from four experiments were cut off and thespecimens fixed, then dehydrated, critical point driedmetallized and investigated with a Hitachi S2500.

Twelve coated prostheses (6 coated by operator A and6 by operator B) were immersed in 10% neutralformalin for 30min, then placed in .9% NaCl forimaging modality (see below). Thereafter thesegrafts were sliced and two segments of proximal, midand distal regions were processed. Then, sampleswere embedded in paraffin, stained with hematoxylin–eosin–safran (HES) for light microscopy.

Thickness analysis processings were performed using‘‘Image J’’ software (image processing and analysis inJava) [http://rsb.info.nih.gov/ij/]. The ROI was manu-ally drawn around fibrin coating and converted intobinary image. Then, the skeleton of the ROI wasobtained. The thickness distribution of the ROI wasevaluated as the distribution of distance values encodedalong the skeleton. After normalization of the data, thisdistribution related the probability function of thick-ness, that is the probabilty to find a particular thicknessalong the overall ROI.

Fibrin coatings of cell-free ePTFE prostheses wereimaged using high-frequency ultrasound modality. A

mechanical transducer (UltracrossTM coronary catheter,Boston Scientific Corporation, Fremont, CA) wasselected because of image quality. The imaging catheter(0.9mm) was carefully centered into the ePTFE tube inorder to avoid image distortion. Studies were imaged ona ClearviewTM console (Cardiovascular Imaging Sys-tem, Boston Scientific, Fremont, CA) and two sequen-tial pullbacks were recorded on S-VHS videotape foroff-line analysis by a senior radiologist.

Three grafts have been tested for endothelialization.Primary human saphenous vein endothelial cells

(HSVEC) were harvested as described by Golledgeet al. [6] grown in a complete culture medium composedof M199 medium with glutamax (Invitrogen Corp,Cergy Pontoise, France) supplemented with 20% humanserum (Eurobio, les Ulis, France), heparin 25,000 IU/mL (Sanofi Synthelabo, Paris, France), gentamycin50mg/mL (Biomedia, Boussens, France) , 10 ng/mLbFGF (Sigma Aldrich, St Quentin Fallavier, France).After confluency was reached, HSVEC were seeded oninner graft surface at 15� 104 cells/cm2. Then, ECseeded graft is placed in a rotative device (Endostrabil-isator, Biegler, Mauerbach, Austria) for 5 h in 5% CO2

incubator to obtain homogeneous coverage.Endothelialized prostheses were submitted at day 8 to

static and flow conditions during 2 h in a haemodynamicbench which delivered a laminar pulsatile flow with ashear stress of 15 dyn/cm2, and a pressure of 120/60mmHg. Prostheses samples were observed in epi-fluorescence (Live/Dead Viability/Cytotoxicity Kit,Molecular Probes, Leiden, The Netherlands).

Immunohistochemical stainings were performed onparaffin-embedded material after 3 and 8 days ofculture. For cell-free grafts, the antibody used wasanti-fibrinogen (DakoCytomation, Trappes, France).For cell seeded grafts, the antibodies used wereantiCD31 (DakoCytomation, Trappes, France) andantiCD34 (Immunotech, Marseille, France).

For statistical analysis, ten Random fields from sixprostheses were observed to measure fibrin glue thick-ness for each operator. The comparison of thicknessdistributions was performed by using a paired t-testapplied to the corresponding repartition functions.

3. Results

The average time necessary for coating procedure was2575min. Concerning light microscopy of cell-freecoated grafts, a total of 72 sections were examined.Histologic cross sections of the different coated graftsbefore cell seeding showed that whatever the examinedsegment, an overall qualitative homogeneous proteinmatrix was observed as shown by SEM. The averagethickness of the fibrin glue layer from proximal to distalregions obtained after image processing is 8.570.25 and

ARTICLE IN PRESS

Fig. 1. Histologic cross sections of fibrin glue coated ePTFE from proximal (A), mid (B) to distal end (C) showing the inner (I) surface being

completely covered by the covering matrix protein as well as the structure of ePTFE graft, clearly visible underneath (HES).

Table 1

Values of thicknesses (mm) for both operators and for the different

segments of the coated grafts

Operator A Operator B P

Proximal 5.970.3 3.670.5 .01

Mid 8.670.3 4.170.4 .0003

Distal 11.470.3 4.870.5 .0008

P Significant NS

Concerning operator A, significant differences appear between

proximal and mid segments (Po.007) and between proximal and

distal segments (Po.006). Concerning operator B, no significant

difference is pointed out.

P. Fernandez et al. / Biomaterials 26 (2005) 5042–50475044

4.170.4 mm for operators A and B, respectively(Po.0001). Fig. 1 shows qualitative results of crosssections of two grafts coated by operators A and B,representative of most observed sections in two sets of6-coated grafts. Table 1 summarizes values of thick-nesses after calculations have been performed on thedifferent segments of the grafts depending on theoperator showing that there is a significant differencealong the prosthesis for operator A and none foroperator B but also a significant difference for each ofthe 3 regions comparing the two operators (Table 1). Insome specimens from different sites, bigger thicknessesof matrix could be observed and were identified as fibrindeposits after immunohistochemical staining. By meansof intragraft ultrasound, we assessed luminal dimensionsand putative irregularities. Luminal irregularities werechecked throughout the entire pullback and measure-ment of the thickness was done when visual examinationrevealed 4100 mm of echo-reflective material inside theePTFE tubing. Coated grafts (Fig. 2A) were comparedwith uncoated specimens (Fig. 2B) to measure thehomogeneity of the inner fibrin layer. Significant fibrindeposits were found in 15% of the imaged grafts with athickness reaching sometimes 800 mm (Fig. 2C and D).Histologic cross sections of freshly lined grafts showed

the inner surface being completely covered by a singlelayer of flattened EC strongly positive for CD31 andCD34 (Fig. 3). After 2 h of a physiological shear stressapplication, endothelial cell monolayer integrity wasmaintained (Fig. 4).

4. Discussion

Different strategies have been developed over theyears to improve the performance of vascular grafts,

ARTICLE IN PRESS

Fig. 2. Intragraft ultrasound imaging of coated (A), uncoated grafts (B), without (A) or with dense echoes (C, D) corresponding to fibrin deposits of

thickness reaching sometimes 800mm (D) (c: catheter, w: wall prosthesis).

P. Fernandez et al. / Biomaterials 26 (2005) 5042–5047 5045

from tissue engineering [7] to the design of novelvascular grafts [8–12] including binding of endothelialcells [13,14] that have antithrombogenic properties ontothe luminal surface of the synthetic grafts. However, asfar as ePTFE is concerned, one prerequisitive for celllining is surface pretreatment to promote cell adhesionand spreading. It has been shown previously that aprecoating substrate should provide sufficient attach-ment area [15]. In this respect fibrin glue provides anunderlying protein matrix that fulfills two majoradvantages: it retains the endothelium and it is approvedfor clinical use.

To our knowledge, the present study is the first onethat reports the reliability of fibrin coating particularlyin very small-diameter vascular grafts. In addition, weprovide by image processing and analysis the quantita-tive fibrin glue thickness determination for the entiregraft length and for 2 operators. Actually, the reliabilityof the technique, originally developed by Zilla et al. [1]for 6mm expanded PTFE grafts, seems reflected by aconstant thickness of the fibrin glue layer achieved. Inbrief, Zilla et al. [16] found that the thickness of thefibrin glue layer was an average of 8.076.1 mm beforeendothelial cell lining compared with 10.079.5 mm ofthe preimplantation specimens that have been main-tained in culture for approximately 12 days. Deutschet al. [17] obtained thickness values of 7 and 8 mm beforeclinical implantation. When considering the samecoating procedure for 4mm ID, the same group [16]showed the following results: a varying thickness(18.4712.9 mm) immediately after cell lining and14.679.5 mm in the preimplantation specimens. How-ever, all these reported results rely on a mean value that

does not reflect putative variations from proximal todistal ends. In our study, the average thicknesses gainedfor 4mm ID by both operators are thinner than thoseobtained for identical ID but comparable to thoseobtained for 6mm ID. When considering inter-operatorreproducibility, significant differences have been pointedout between operator A and B; these differences remainnegligible when compared with ID. For a given operatorand a given segment, reproducibility assessed bycoefficient variations provided in Table 1 is satisfactory.The inter-segment variations that reflect the quality ofcoating procedure all along the graft are more repro-ducible for operator B than for operator A. The formeris a trained senior physician used to performing such aprocedure while operator A is a pre-doc researcher.

Anyway, the achieved substratum, whatever theconcerned operator, is efficient because it allows seededECs to adhere and spread as expected. Otherwise, theendothelial cell monolayer integrity is maintained underphysiological shear stress conditions. It is unlikely thatin clinical trials, such a difference (8.5 vs. 4.1 mm) hasdetrimental consequences.

In some cases, an accumulation of deposits could beobserved with an original application of ultrasoundimaging, corroborated by subsequent histologic exam-ination. These protrusions were randomly distributedalong the prostheses that corresponded to fibrin asshown by immunostaining. Their origin is a matter ofdiscussion and could not be elucidated: fibrin glueinflation or polymerization induced by fixative reagent?Such deposits could not be observed on the threeendothelialized grafts. Do EC metabolize the excessfibrin deposits?

ARTICLE IN PRESS

Fig. 3. Histologic cross sections showing the inner surface of coated grafts being completely covered by a single confluent layer of flattened cells

typical for an endothelium (A–F), a few hours after cell seeding (A, B) at days 3 (C, D) and 8 (E, F). The cells resting on glue were strongly positive

for CD 31 at day 0 (B) and CD 34 (D and F at days 3 and 8, respectively). Between the endothelium and the graft a continuous layer of fibrin glue was

demonstrable (arrow) (original magnifications : � 10 for A, C, E, respectively; � 20 for B, D, F, respectively) (A, C, E: HES; B, D, F: APAAP

technique).

Fig. 4. Epifluorescence results from endothelialized grafts under static (A) or flow (B) conditions (original magnifications : � 40).

P. Fernandez et al. / Biomaterials 26 (2005) 5042–50475046

5. Conclusion

In summary, we consider that the objectives aimed bythis work have been achieved, (i) feasability of fibringlue coating of 4mm ID vascular grafts with a very low

operator dependant variability of coating thickness (ii)ability to assess quality control all along the procedurewith different techniques, but would it be reasonable topropose a systematic ultrasound imaging quality controlprior to each implantation taking into account sterile

ARTICLE IN PRESSP. Fernandez et al. / Biomaterials 26 (2005) 5042–5047 5047

requirements to perform the imaging as well as theresolution limits of the technique ? (iii) the endothelialcell monolayer maintain under arterial shear stressconditions. As far as such a clinical application willbe performed in routine coronary bypass operations,we recommend to entrust the coating procedure to awell-trained technician’s care.

Acknowledgements

The authors thank GORE (Flagstaff, USA) forgrafts gifts but also, BAXTER (Maurepas, France) forTissucols kits gifts.

References

[1] Zilla P, Fasol R, Preiss P, Kadletz M, Deutsch M, Schima H,

Tsangaris S, Groscurth P. Use of fibrin glue as a substrate for in

vitro endothelialization of PTFE vascular grafts. Surgery

1989;105(4):515–22.

[2] Zilla P, Deutsch M, Meinhart J, Puschmann R, Eberl T, Minar E,

Dudczak R, Lugmaier H, Schmidt P, Noszian I, et al. Clinical in

vitro endothelialization of femoropopliteal bypass grafts: an

actuarial follow-up over three years. J Vasc Surg 1994;19(3):

540–8.

[3] Meinhart JG, Deutsch M, Fischlein T, Howanietz N, Froschl A,

Zilla P. Clinical autologous in vitro endothelialization of 153

infrainguinal ePTFE grafts. Ann Thorac Surg 2001;71(5 Suppl):

S327–31.

[4] Laube HR, Duwe J, Rutsch W, Konertz W. Clinical experience

with autologous endothelial cell-seeded polytetrafluoroethylene

coronary artery bypass grafts. J Thorac Cardiovasc Surg 2000;

120(1):134–41.

[5] Lamm P, Juchem G, Milz S, Schuffenhauer M, Reichart B.

Autologous endothelialized vein allograft: a solution in the search

for small-caliber grafts in coronary artery bypass graft operations.

Circulation 2001;104(12 Suppl 1):I108–14.

[6] Golledge J, Tumer RJ, Harley SL, Springall DR, Powell JT.

Development of an in vitro model to study the response of

saphenous vein endothelium to pulsatile arterial flow and

circumferential deformation. Eur J Vasc Endovasc Surg

1997;13(6):605–12.

[7] Ratcliffe A. Tissue engineering of vascular grafts. Matrix Biol

2000;19(4):353–7.

[8] Salacinski HJ, Tiwari A, Hamilton G, Seifalian AM. Cellular

engineering of vascular bypass grafts: role of chemical coatings

for enhancing endothelial cell attachment. Med Biol Eng Comput

2001;39(6):609–18.

[9] Lin YS, Wang SS, Chung TW, Wang YH, Chiou SH, Hsu JJ,

Chou NK, Hsieh KH, Chu SH. Growth of endothelial cells on

different concentrations of Gly–Arg–Gly–Asp photochemically

grafted in polyethylene glycol modified polyurethane. Artif

Organs 2001;25(8):617–21.

[10] Birchall IE, Field PL, Ketharanathan V. Adherence of human

saphenous vein endothelial cell monolayers to tissue-engineered

biomatrix vascular conduits. J Biomed Mater Res 2001;56(3):

437–43.

[11] Krijgsman B, Seifalian AM, Salacinski HJ, Tai NR, Punshon G,

Fuller BJ, Hamilton G. An assessment of covalent grafting of

RGD peptides to the surface of a compliant poly(carbonate-

urea)urethane vascular conduit versus conventional biological

coatings: its role in enhancing cellular retention. Tissue Eng 2002;

8(4):673–80.

[12] Tiwari A, Salacinski HJ, Punshon G, Hamilton G, Seifalian AM.

Development of a hybrid cardiovascular graft using a tissue

engineering approach. FASEB J 2002;16(8):791–6.

[13] Bordenave L, Remy-Zolghadri M, Fernandez P, Bareille R, Midy

D. Clinical performance of vascular grafts lined with endothelial

cells. Endothelium 1999;6(4):267–75.

[14] Seifalian AM, Tiwari A, Hamilton G, Salacinski HJ. Improving

the clinical patency of prosthetic vascular and coronary bypass

grafts: the role of seeding and tissue engineering. Artif Organs

2002;26(4):307–20.

[15] Kaehler J, Zilla P, Fasol R, Deutsch M, Kadletz M. Precoating

substrate and surface configuration determine adherence and

spreading of seeded endothelial cells on polytetrafluoroethylene

grafts. J Vasc Surg 1989;9(4):535–41.

[16] Zilla P, Preiss P, Groscurth P, Rosemeier F, Deutsch M, Odell J,

Heidinger C, Fasol R, von Oppell U. In vitro-lined endothelium:

initial integrity and ultrastructural events. Surgery

1994;116(3):524–34.

[17] Deutsch M, Meinhart J, Vesely M, Fischlein T, Groscurth P, von

Oppell U, Zilla P. In vitro endothelialization of expanded

polytetrafluoroethylene grafts: a clinical case report after 41

months of implantation. J Vasc Surg 1997;25(4):757–63.