Endothelia;l cell a~esion to vasc~ar prostheticsurfaces D. Gourevitch,CarolynE. Jones, J. Crackerand M. Goldman Department of Surgery and Pathology, East 3i~mingham Hospital (University of ~irmiflgham], Birmingham, UK
Bordesley Green East,
Presented at Bio;nteracrions ‘87, Cambridge, UK in Ju/y 1987
Improving the long term patency of small diameter prosthetic grafts remains an important hut elusive objective in vascular surgery. In pursuit of a non-thrombogenic surface, we have cultured human-adult endothelial cells, examined their adhesive properties and their ability to colonize the inner surface of a [email protected]
graft. To examine cell adhesion, endothelial cells wera labelled with “‘In-oxine and inoculated onto prosthetic wells previously prepared with either cold insoluble globulin (CIG), 1% gelatin, alginate or left untreated as a control. At 100 min, mean percentage adhesion to GIG and gelatin precoated wells was 88.0 I 9.9% (2 SD) and 81.8 ri7 2.9% respectively, wherees alginate at 57.5 f 7.5% and the control well at 48.3 f 9.9% showed significantly less adhesion. Further experiments examined the adbesion of Indiumoxine labelled endothelial cells to [email protected]
graft (2 cm X 8 mm ID) placed in an in vitro arterial circuit. An immediate loss of approximately 20% of the cells occurred within the first 30 s. whereafter, a stable population were adherent to the graft material. By 102 min, 73.4 k 1.8% of cells remained attached when exposed to tissue culture medium, but only 64.1 + 6.9% after exposure to blood. Cultured human adult endothelial cells adhere most effectively to prosthetic surfaces precoated with CIG or gelatin, and remain attached following exposure to shear forces. Keywords: Prostheses, Dacron, endorhelial cek
graft, saphenous vein
The saphenous vein provides a compliant arterial conduit, with an endothelial cell surface, which actively prevents platelet adhesion and clot formation through the production of prostacyclin (PGlz) and other antithrombogenic agents’, ‘. By comparison, any vascular prosthetic graft provides a poor alternative for both coronary artery and peripheral vascular reconstruction3-5. All modern prosthetic grafts fail to encouarge complete endothelialization of their surface; they develop a thrombogenic layer of compacted fibrin and cannot rival autologous vein for long-term patency. Especially long lengths of vein are required for femorodistal bypass to a patent artery at the ankle. Despite the undoubted success of this operatio#-*, the saphenous vein may require to be supplemented with an arm vein or by combination with a prosthetic graft. Some surgeons are reluctant to embark on a long and tedious operation, requiring perhaps, 2 h or more to isolate and prepare the vein. Others may feel that the duration of the complete procedure, and anaesthetic, place the patient at undue risk. The provision of a metabolically active, fully endothelialized, ‘off the shelf’ vascular prosthesis, would provide an attractive alternative and encourage limb salvage. In our effort to develop an endothelialized prosthesis, we have harvested human adult endothelial cells from autologous vein remnants and omentum. These have been ___I-Correspondence to Dr D. Gourevitch, Dept. of Surgery, East Birmingham Hospital, Bordesley Green East, Birmingham ES 5ST. UK. 0 1988
Et Co (Publtshers)
subcultured to provide sufficient numbers of cells to study alternative methods of maximizing their adhesion to prosthetic surfaces. We examined various substrates to facilitate endothelial cell adhesion to lengths of Dacron graft and assessed cell adhesion during exposure to pulsatile blood flow in an ‘in vitro’ circulation.
Human adult endothelial cells were acquired from saphenous vein remnants (2.5-10.5 cm), removed at the time of varicose vein surgery, and from human omentum (6.5-l 3.5 g), excised from patients undergoing laparotomy for non-malignant conditions. All specimens were placed in tissue culture medium (X 10 cont. modified M 199; Flow Lab; diluted 1 : 10 with double-distilled water), with the addition of streptomycin (Evans) 53 ug/ml, and penicillin (Glaxo) 32 IU/ml. All specimens were treated within 6 h.
vein redbaits. Our methods for endothelial cell isolation from saphenous vein remnants has been detailed previously’. Briefly, vein remnants were cannulated and flushed through with tissue culture medium. Collagenase (Sigma) (0.1%) was instilled and the specimens incubated for 20 min at 37”C, in an atmosphere of air with 5% COz added. The released cells were flushed out, centrifuged, and then resuspended in a complete tissue culture medium
1988, LM 9 January
Cell adhesion to prosthetic surface: 0. Gourevitch et al.
Ml 99, containing 30% human serum, L-glutamine, endothelial cell growth factor, streptomycin and penicillin. The cells were subcultured onto plastic Petri dishes. Approximately 25 min is required for this harvesting procedure. Human omentum. The omental samples were minced and placed in a 50 ml flask containing 10 ml of cation-free buffer (CFB), pH 7.4, with collagenase (Cooper’s) 4 mg/ml, and bovine serum albumin (BSA) (Sigma), 4 mg/ml. The specimens were incubated at 37°C with constant agitation, for 25 min. The flask was then centrifuged at X 200 g for 7 min. The resultant pellet was washed with CFB, containing 0.1% BSA and centrifuged for 7 min at X 200 g. This washing and centrifugation was repeated once. The resultant pellet was suspended in 45% Percoll (Sigma) and centrifuged at X 20 000 g for 20 min at 4°C. Capillary endothelial cells formed a creamy layer at the top of the density gradient. These were then washed in CFB/O, 1% BSA and centrifuged at X 200 g, for 5 min. This process was repeated twice before the cells were inoculated onto tissue culture dishes with medium M 199 and 20% foetal calf serum, and incubated at 37°C in an atmosphere of air with 5% COz added. The isolation time for omental endothelial cells was approx. 85 min.
Endothelial cell identification”*
Positive identification for endothelial cells was made through phase-contrast microscopy and immunohistochemical staining for Factor VIII Related-antigen and with Ulex Europeaus-lectin (Figure I).
oxine labelling of endothelial
Endothelial cell cultures were washed twice with PBS prior to incubation with 0.02% trypsin at 37°C for 3 min. The released endothelial cells were resuspended in 20 ml of complete medium M 199 and centrifuged at X 350 g for 10 min. The endothelial cell pellet was resuspended in 1 ml of medium M 199 and a cell count performed with a modified Neubauer haemocytometer, in conjunction with the trypan blue exclusion test. 10 uCi of “‘lndium oxine (Amersham) was added to the suspension, and allowed to incubate for 5 min at room temp. The specimen was then diluted as required. Labelling efficiency was determined for each sample.
Preparation of cold insoluble globulin (CIG) CIG was produced from human plasma removed from patients during plasmaphoresis. The plasma was decanted into 50 ml tubes and centrifuged at 3500 rpm for 20 min at 1 “C. The precipitate was dissolved in 50 ml of 0.85% saline and the process of centrifugation and resuspension repeated. The final precipitate was dissolved in 0.85% saline (40 ml), and 2 ml aliquots were stored at - 20°C. After thawing, the CIG was pipetted onto surfaces in combination with tissue culture medium M 199.
Endothelial cell adhesion to prosthetic wells Wells, previously prepared with one of 3 substrates i.e. (i) CIG, (ii) 1% gelatin (Sigma) and (iii) alginate or left untreated as a control, were inoculated with 50 ~1 of “‘lndium oxine labelled endothelial cells. At 20 min intervals, up to 100 min. successive wells were irrigated with PBS and the radioactivity in the supernatant measured. Percentage cell adhesion was subsequently determined. Elusion free “‘lndium oxine was
1988, Vol9 January
Figure 1 Positive endothelial call identification VI 11 R-antigen, and (b) with UEA- 1.
staining for (a) Factor
estimated by centrifugation of the supernatant and measuring the radioactivity in the resultant, cell-free medium.
Endothelial cell adhesion to Dacron tubes Closed lengths (2 cm) of 6 mm (ID) Dacron graft were prepared prior to injecting the substrates CIG or 1% gelatin. These substrates were removed immediately before injecting 0.5 ml of “‘lndium oxine labelled endothelial cells. The grafts were then rotated for 30 min. to produce an even distribution of cells before incubation at 37°C in an atmosphere of air with 5% COP added, for 30 min.
‘In vitro’ arterial circuit A circuit was constructed to provide a pulsatile flow of 140 ml/min, at 140 mmHg systolic pressure, with a temperature of 37°C (Figure 2). The Dacron grafts (Bard/USCI) 2 cm X 6 mm (ID) were tightly fitted into short lengths of perspex, and these were then inserted between cut sections of the siliconized rubber tubing. Arterial pressure was generated from a Sarns pump, and adjusted with a gate-clip positioned on the afferent tubing proximal to the reservoir. The circuit was initially primed with tissue culture medium, and 4 consecutive experiments were performed; 7 further experiments were conducted with the circuit filled with heparinized human blood. Continuous, sequential, 5 s counts of y-emission from the grafts were recorded using a well-crystal at 30 s, 1 and 5 min intervals, and 60 min thereafter. Counts from the tubing, in addition to the background, were recorded during the experiments.
Cell adhesion to prosthetic surface: 0. Gourevitch et al.
CIG 1% Gelatin
80 E 70 ‘I ;
40 Penpex tubing
of he in vitro arterial circuit
Well adhesion results of 4 consecutive experiments comparing endothelial cell adhesion to wells are shown in Figure 3. Increased endothelial cell adhesion to CIG and gelatin coated wells was apparent by 20 min compared with the alginate and untreated controls, and the separation between the 2 groups was statistically significant (p < 0.001). This difference persisted until conclusion of the 100 min, when endothelial cell adhesion to CIG and gelatin precoated wells was 86.0 + 9.9% (+ SD) and 8 1.6 f 2.9% respectively, which contrasted with adhesion to alginate at 57.5 rt 7.5% (p < 0.001) and the control at 48.3 t- 9.9% (p < 0.001). No statistically significant differences of endothelial cell adhesion were achieved between CIG and gelatin, or alginate and the untreated control wells. The
Figure 3 Percentage ” ‘Indium precoated Dacron wells.
Circuit expe~ments Maximal adhesion was considered to have occurred after incubating endothelial cells with gelatin coated Dacron tube graft for 100 min. The grafts were then inserted into the ‘in vitro’ arterial circuit. The results of 7 consecutive experiments are shown in Figure 4. This illustrates the percentage of “‘lndium oxine labelled endothelial cell adhesion over 162 min. An early loss of 20.5 f 1.9% (+ SD) of cells occurred within the first 30 s, but endothelial cell adhesion persisted with 73.4 t 1.8% of the cells remaining attached to the Dacron at 102 min, and with no further loss of cells by the conclusion of the experiment. A similar pattern of endothelial cell adhesion was demonstrated on exposure of the cell-coated Dacron grafts to blood, in the arterial circuit (Figure 5). Despite an initial loss of 17.6 t- 9.1% of cells by 30 s, a stable population of adherent cells was achieved by 5 min. By 102 min, at the conclusion of the experiment, 64.1 + 6.9% of endothelial cells remained adherent.
DISCUSSID~ Surgeons remain dependent on prosthetic vascular grafts as the optimal bypass material, autologous vein, is unavailable in approximately 20% of their patients. The thrombogenic nature of the artificial surface, and its failure to develop an plays an important role in endothelial cell coveringI subsequent early and late occlusion. ln vivo animal experiments have demonstrated that endothelialization can be encouraged by autologous, endothelial cell seeding of the graft, at the time of implantation’4, 15. It remains unknown whether this can be repeated in humans. Jarrell eta/.16 have
Time (min) Figure 4 Percentage ” 'lndium labelled endothelial cell adhesion to precoated Dacron (140 mmtfg; 140 ml/min] graft exposed to tissue culture medium4 in the in vitro arterial circoit.
suggested that the enzymatic harvesting and seeding of endothelial cells can be performed during a vascular reconstruction. This would require the primary isolation of a large quantity of endothelial cells, and their effective bonding to the graft surface. Enzymatic cell harvesting from adult veins provides a pure culture of endothelial cells, yet it is estimated that only 15% of the available cells are liberated17. Hence, only relatively small numbers of cells are made available for immediate seeding unless they are subcultured in vitro. Applying such techniques, we achieved a 4-fold increase in cell numbers, which would be sufficient quantity (> IO5 cells/cm*) for high density seeding. McCall et a/.” have previously established that a minimum of 2.5 X IO5 cells/cm’ are necessary to produce a confluent growth of porcine aortic endothelium, and it is unlikely, therefore, that a lesser density of cells would fare better in an atherosclerotic patient. cannot
occur during Omentum
1988. Vol9 January
Cell ~hesjon to prosthetic surface: D. G~revjich
adherence to its surface when circulation.
exposed to the arterial
Time (min) Figure 5 Percentage ” ‘lndium labelled endothelial cell adhesion to precoated Dacron (140 mmHg: 140 mi/minj graft exposed to blood in the in vitro arterial circuit7.
cell~‘“~*~, and offers increased cell numbers. Unfo~unately, this may be at the expense of a pure endothelial cell isolation. Other ceil types, such as mesothelial cells and fibroblasts, L may be included in the harvest”. With respect to endothelial cell adhesion, Jarrell et a1.16 reported that only lo-40% of these omentally derived endothelial cells adhered to prosthetic surfaces, and that approximately, only 60% of these cells remained adherent by 60 min. Our studies showed that at least 100 min were required for maximal cell adhesion to occur. When this period is summed with the ceil harvesting times from vein and omental specimens, we calculate that between 125 and 185 min are required to allow maximal cell adhesion to the graft surface. We feel this would produce an unacceptable increase in the operating time for a vascular reconstruction. The application of fibronectin to prosthetic surfaces, has been advocated as a means of improving cell bonding2’* 22. Yet, in spite of using fibronectin in our experiments, 20% of cells were lost from the graft surface, immediately it was exposed to shear forces. Ramalanjaona et at.23 noted a similar loss of endothelial cells from fibronectin coated PTFE grafts interposed in the dog carotid artery, as did Sentissi et a1.24, who recorded the loss of bovine aortic endothelium in an in vitro circuit. Fibronectin, though, is known to be thrombogenic25 and this loss of cells would expose an even greater area of unseeded, precoated graft. Here, gelatin may be a better substrate, offering equal ability in terms of cell bonding without increasing platelet deposition on their surfaces. Finally, several weeks must elapse before a complete endothelial cell cover is formed: a period during which a significant number of grafts fail. An alternative is to provide a fully endothelialized graft at the time of surgery. By subculturing endothelial cells in vitro, a sufficient density of cells can be provided for confluent growth, and the covering of a graft surface prior to implantatioP. The development of a fully lined graft would offer advantages in terms of early reduction in thromb~enici~ and a greater chance of maintaining optimal endothelial cell
1988, Vol9 January
12 13 14
Grabowski, E.F.,Jaffe. E.A. and Weksler, B.B.. Prostacyclin production by cultured endotheiial cell monolayers exposed to step increase in shear stress, J. tab. C/in. bled. 1985, 105. 36-43 Moncada. S.. Herman, A.G.. Higgs, E.A. and Vane, J.R., Differential formation of prostacyclin (PGx or PGIZ) by layers of the arterial wall; an explanation for the antithrombogenic properties of vascular endothelium. Thromb. Res. 1972, 11, 323-344 Bergan. J.J., Veith, F.J. and Berhard, V.M. et a/., Randomisation of autogenous vein and polytetrafluoroethylene grafts in femoral distal reconstruction, Surgery 1982, 92, 921-929 Cranley, J.J. and Hafner, CD., Newer prosthetic material compared with autogenous saphenous vein for occlusive vascular disease of the lower extremity. Surgery 198 1, 89, 2-7 Yeager. R.A., Hobson, R.W.. Jamil, Z., Lynch,T.G., Lee, B.D. and Jain, K., Differential patency and lower extremity ischaemia,Surge~ 1982, 91,99-103 Leather. R.P.. Powers, S.R. and Karmody, A.M., A reappraisal of the ‘in situ’ saphenous vein arterial bypass; its use in limb salvage, Surgery 1979,86.453-6 1 DeWeese, J.A. and Rob, C.G., Autogenous vein grafts ten years later, Surgery 1977,82. 775-84 Simms. M.H. and Slaney, G.. In situ vein grafts to the distal peroneal artery, Br. J. Surg. 1984, 71, 308 Gourevitch. D.. Jones, Carolyn E., Cracker, J. and Goldman, M., An in vitro technique to cover Dacron arterial grafts with human adult endotheliai ceils, Surg&a/ Res. Comm. 1987 (in Press) Cracker, J. and Smith, PJ., lmmunohist~hemi~l localisation of Factor VIII-related antigen in Hodgkins disease, J. C/in. Pafh. 1984, 37.37-44 Walker, R.A., The binding of peroxidase-labelled lectins to human breast epithelium. 1. Normal, hyperplastic and lactating breast, J. Path. 1964, 142, 279-291 Wilkinson, A.R., Hawker, R.J. and Hawker, Linda M., “‘lndium labelled canine platelets, Thromb Res. 1978, 13, 175-l 82 Sauvage. L.R., Berger, K.E. and Wood, S.J. et a/., interspecies healing of porous arterial prostheses, Arch. Surg. 1974, 109, 698-705 Stanley, J.C., Burkel, W.E., Ford, J.W. et a/., Enhanced patency of small-diameter, externally supported Dacron iliofemoral grafts seedad with endothefial cells, Surgery 1982, 92, 994-l 005 Sharefkin, J.B., Latker. C. and Smith, M. et a/., Early normalisation of platelet survival by endothelial seeding of Dacron arterial prostheses in dogs, Surgery 1982,92.385-393 Jarrell, B.E., Williams, S.K. and Stokes, G. eta/., Use of freshly isolated capillary endothelial cells for the immediate establishment of a monolayer on a vascular graft at surgery, Surgery 1986, 100. 392-39s Rosenman. J.E.. Kempczinski, RF., Pearce. W.H. and Silberstein, E.B., Kinetics of endothelial cell seeding,J. Vast. Surg 1985,2,778-784 McCall, E., Povey, J. and Dumonde, D.C., The culture of vascular endothelial cells to confluence on micro~rous membranes, Thromb. Res. 1981.243,417-431 Kern, P.A., Knedler. A. and Eckel. R.H., Isolation and culture of microvascular endothelium from human adipose tissue, ./. C/in. invest. 1983.7, 1822-1829 Pearce, W.H., Rutherford, R.B. and Whitehall, T.A. et a/., Successful endothelial seeding with omentally derived microvascular endothelial cells, J. Vase. Surg. 1987, 5, 203-206 Seeger, J.M. and Klingman, N.. Improved endothelialcell seeding with cultured cells and fibronectin-coated grafts, J. Surg. Res. 1985, 38, 641-647 Pearlstein, E. and Hoffstein, ST., Fibr~~in-m~iat~ cellular adhesion to vascular su~ndothelial matrices, Exper. Cell Res. 1981, 134,161-170 Ramalanjaona, G.. Kempczinski, RF. and Rosenman, J.E. ef a/.. The effect of fibronectin coating on endothelial cell kinetics in polytetrsfluoroethylene grafts, J. Vast. Surg. 1986, 3, 264-272 Sentissi. J.M., Ramberg, K. and O’Donnell.T.F.etal.,Theeffectof flow on vascular endothelial cells grown in tissue culture on polytetrafluoroethylene grafts, Surgery 1986, 99, 337-343 Kempczinski, R.F.. Ramalanjaona, G.R., Douville, C. and Silberstein, ES., Thrombogenicity of a fibronectin-coated experimental polytetrafluoroethyiene graft, Surgery 1987, 101.439-444 Foxall. T.L., Auger, K.R., Callow, A.D. and Libby, P., Adult human endotheiiaf cell cover of small-calibre Dacron and ~~etr~luor5 172 ethylene vascular prostheses in vifm, J. Surg. Res. 1986.41,158-