Artificial Organs 14(5):355-360, Raven Press, Ltd., New York 0 1990 International Society for Artificial Organs
Endothelial Cell Seeding of Small Diameter Vascular Grafts N. L. James, K. Schindhelm, P. Slowiaczek, B. K. Milthorpe, N. P. B. Dudman, *G. Johnson, and *J. G. Steele Centre for Biomedical Engineering, University of New South Wales, Kensington, Australia, and the *Laboratory f o r Molecular Biology, CSIRO Division of Biotechnology, North Ryde, Australia
Abstract: This study examines, under flow conditions, the adhesion of endothelial cells to 3 mm diameter fibronectin (Fn)-coated expanded polytetrafluoroethylene (PTFE) vascular grafts. Cultured ovine carotid artery endothelial cells were labelled with 3sS-methionine. The grafts were seeded with endothelial cells (1.5 x 10"lml) by rolling for 1 h at 37°C and then either cultured to confluence for 48 h or flow tested immediately. Cell attachment to grafts (n = 5) was evaluated in an in vitro flow circuit, using flows of up to 330 mYmin. Ex vivo studies (n = 5 grafts) were conducted without anticoagulant using autologous cells in a sheep model. Grafts were inserted into an
externalized carotid-jugular shunt and exposed to blood flows of -150 mumin for 3 h. One hour seeded and 48 h cultured grafts demonstrated >95% cell retention following in vitro flow studies. Ex vivo studies of 48 h cultured grafts gave endothelial cell retention of 81% with no sign of thrombogenicity. Furthermore, a preliminary 24 h ex vivo study has shown >95% retention. This study demonstrates the fum attachment of seeded endothelial cells to Fn-coated PTFE grafts in the sheep model. Key Words: Vascular grafts-Endothelial cells-Polytetrafluoroethylene-Cell adhesion-Flow-Sheep.
Experimental and clinical studies indicate that endothelial cell seeding of small diameter vascular prostheses reduces thrombogenicity and enhances patency (1,2). Cells are harvested and seeded immediately onto a graft, covering a comparatively small proportion of the total graft surface. Despite antiplatelet therapy, initial platelet accumulation and thrombus formation on the graft are not decreased. However, the explants at 2 weeks to 1 month postimplantation have begun to show reduced levels of platelet deposition and thrombosis (3-5). Improved initial responses, which may in turn influence longer term patency, may be achieved using grafts implanted with a preformed endothelial cell monolayer (6).To date, the majority
of vascular graft studies have been performed in dogs. Since sheep and human blood clotting systems are reported to be similar (7), the sheep was used as an alternative model for vascular studies. The maintenance of a confluent endothelial layer on a graft may be influenced by physical factors such as shear stress and also biochemical, hematological, and immunological factors. The effect of shear stress alone may be evaluated in the short term in cell culture using flowing medium (in vitro flow testing). The combined effects of pulsatile flow and the host response to the graft may also be determined in the short term using an extracorporeal circuit, provided anticoagulants can be avoided (ex vivo flow testing). In this study, a technique for seeding Fn-coated expanded polytetrafluoroethylene (PTFE) grafts with endothelial cells is described. The seeded grafts are cultured in vitro to form confluent cell monolayers. The adherence of cells to the grafts is evaluated under in vitro flow conditions and in an ex vivo model without administration of platelet inhibitors.
Received February 1990; revised April 1990. Address correspondence and reprint requests to Dr. N . L. James, Centre for Biomedical Engineering, University of New South Wales, P.O. Box 1, Kensington NSW 2033, Australia. Presented in part at the VIIth World Congress of the International Society for Artificial Organs, Sapporo, Japan.
N . L. JAMES ET AL.
356 MATERIALS AND METHODS
Graft preparation PTFE vascular grafts (3 mm internal diameter, W. L. Gore and Associates, Flagstaff, AZ, United States) were used. To facilitate seeding and flow testing, grafts were connected to rigid Teflon tubes and reinforced externally with borosilicate glass tubing (see Fig. 1). The Teflon tubes (Quinton Instrument Co., Seattle, WA, United States) were inserted 3 mm into each end of a 26 mm long graft, leaving 20 mm of graft lumen exposed. Each of the overlap sections was sealed with polyolefin heat shrink tubing (4.8 mm diameter, RS Components Ltd., Rosebery, NSW, Australia), and a 30 mm glass rod (5.2 mm internal diameter, 7.1 mm internal diameter) was sealed over the graft using additional heat shrink tubing (9.4 mm diameter). An 8 mm section of silicone rubber tubing (2.6 mm internal diameter, Silastic, Dow Corning, Midlands, MI, United States) was fitted to each end of the mounted graft. Following packaging and sterilization by autoclave, a sterile injection site (Tuta Labs, Lane Cove, NSW, Australia) was inserted into the silicone rubber sections. The lumen of each graft was coated by incubation for 1 h at 37°C with 10 to 40 pg/d ovine Fn in phosphate buffered saline. The Fn had been purified from plasma (8). Endothelial cell preparation Endothelial cells were harvested from ovine carotid artery using collagenase treatment adapted from the procedures of Maciag et al. (9). Each end of the artery was mounted onto a plastic connector. The lumen was flushed with 10 ml phosphate buffered Hartmann's solution (Baxter Healthcare, Toongabbie, NSW, Australia) prior to injection of 10 ml 0.1% collagenase (from Clostridium histolyticum, Boehringer Mannheim, West Germany) in
FIG. 1. Polytetrafluoroethylene (PTFE) vascular graft preparation.
Artif Organs, Vol. 14, No.5 , 1990
phosphate buffered Hartmann's solution. The connector ends were sealed and the artery was incubated for 20 min at 37°C. The cells were washed out of the artery in 6 ml phosphate buffered Hartmann's solution and pelleted for 5 min at 200 g. The cell pellet was resuspended in 2 ml McCoy's 5A medium (M5A) (Commonwealth Serum Laboratories, Melbourne, VIC, Australia) containing 20% fetal calf serum, and placed into tissue culture wells (Becton Dickinson, Lincoln Park, NJ, United States) which had been precoated with 40 pglml Fn. The cells were verified as endothelial cells by staining for von Willebrand antigen (10,ll) and used in seeding and flow studies at passage 6-9. Prior to seeding, cells were radiolabeled metabolically by incubation for 20 h in Dulbecco's Modified Eagles medium containing 2.5 pCi/ml 35S-methionine (Amersham, North Ryde, NSW, Australia). Nonincorporated radiolabel was removed in 2 x 2 h incubations in M5A prior to graft seeding. Graft seeding Fn-coated grafts were inoculated with 1.5 x lo6 cells/ml M5A and individually taped to the inside of a tissue culture roller bottle. The bottles were rolled on a roller culture apparatus (Wheaton Instruments, Millville, NJ, United States) at 20 r/h for 1 h at 37°C. The outside surface of the mounted grafts (glass and teflon) was flushed with 70% ethanol, the cell inoculum was removed, and the grafts were either flow tested immediately or cultured for 48 h in 7 ml M5A media in a sterile 10 ml centrifuge tube.
In vitro flow testing Cell attachment was evaluated in an in vitro flow circuit (see Fig. 2). Following seeding with 35Slabelled cells, the graft was inserted into the flow circuit, which contained complete culture medium. Glass microfiber filter disks (Whatman GF/C, Maidstone, England) in filters distal to the grafts caught cells washed off during flow tests. The dual filter system facilitated a change of filter at each flow rate examined. Grafts were exposed to nonpulsatile flow with stepwise increases in flow rate ranging from 66 to 330 mYmin, corresponding to 3.3 to 16 dyn/cm2 wall shear stress, over a total test period of 50 min as detailed in Table 1. Estimates of wall shear stress were determined by assuming a steady laminar axisymmetric flow profile far from the entrance of a cylindrical pipe, which is described by a parabolic velocity distribution. The estimates based on these assumptions are conservative since entrance effects and theoretical transition to turbulent flow which occurs for the in vitro tests only (at 11 dyn/cm2)
ENDOTHELIAL CELL SEEDING OF GRAFTS
==--= I =7
FIG. 2. In vitro flow circuit.
would increase wall shear stress levels. For the estimates of in vitro shear stress, the viscosity of the dyn.s/cm2. medium was assumed to be 7.8 X An initial 2 min exposure to 33 mYmin flow was used as a wash to remove extraneous cell debris. Following flow testing, grafts were fixed, stained with a stain similar to Wright-Giemsa (Diff-Quik, Bacto Laboratories, Liverpool, NSW, Australia), and photomicrographs were taken. The cell cover on half of the graft was digested in 1 ml 1 M NaOH, neutralized in 1.5 ml 1 M HCI, and assessed for 35S activity. Radioisotope activity in cell digests and on filters was determined using a liquid scintillation counter (Packard Tricarb, Downers Grove, IL, United States) with a quench correction. Radioisotope counting enabled quantitation of the percentage of cells retained during in vitro perfusion. The total radioisotope count on a graft was taken as the sum of counts on the five flow filters and on the flow tested graft. The isotope counts retained on the filter from the initial 2 min wash procedure were not incorporated into the final calculations of cell retention.
Ex vivo flow studies Ex vivo studies were conducted without anticoagulant in a sheep model using an externalized shunt between the carotid artery and the jugular vein. Grafts (n = 5 ) were seeded with autologous cells, connected in series using silicone rubber tubing, and inserted into the shunt. A 5 mm noncannulated flow probe (Zepeda Instruments, Seattle, WA, United States) was attached to a section of glutaraldehyde-fixed carotid artery. This was inserted into the shunt at the beginning and end of the study for measurement of blood flow. Tests were conducted on 1 h seeded grafts that were cultured for 48 h prior to testing. The grafts were exposed to blood flows of approximately 150 mYmin at a mean wall shear stress of 25 dyn/cm2 for 3 h. Blood viscosity was measured in a viscometer (Brookfield Engineering Laboratories, Stoughton, MA, United States) at the end of each ex vivo flow test. A preliminary ex vivo study was also conducted on a series of three grafts over 24 h. To obtain a quantitative assessment of cell losses from the grafts exposed to ex vivo flow, these grafts were compared to a similar set of control grafts that had not been exposed to blood flow. Each of the flow and control grafts was fixed, stained, and examined prior to bisection for digestion and radioisotopic counting as described for the in vitro studies.
RESULTS Cell seeding After 1 h seeding, cells on the graft were evenly distributed at a subconfluent concentration. There was a range in cell attachment from rounded, focally attached cells to those which were more flattened and spread over the surface. Following a further 48 h culture, the grafts were covered in a complete, even monolayer of endothelial cells.
In vitro and ex vivo flow studies The levels of cell retention from in vitro and ex vivo flow studies are given in Table 2. Following TABLE 2. Results of in vitro and ex vivo flow tests
TABLE 1. I n vitroflow test scheme Time (rnin) 2 10 10 10 10 10
Flow rate (mumin)
Shear stress (dyn/crn2)
33 66 132
1.6 3.3 6.5
9.8 13 16
Culture time (after 1 h seed) (ht
Flow test time
Cell retention % t SD 96.3 2 1.8 98.7 2 0.2 81“
In vitro In vitro Ex vivo
50 rnin 50 min 3h
5 5 flow 4 control
3 flow 2 control
No signifcant difference from control: Mann-Whitney u-test.
Artif Organs, Vol. 14. N o . 5 , 1990
N . L . JAMES ET AL.
1 h seeding, grafts that were immediately flow tested in vitro yielded cell retentions of 96.3 1.8% (n = 5). The cells on these grafts were subconfluent and did not appear to be spread fully (see Fig. 3). In vitro flow testing of seeded grafts which had been cultured for 48 h gave cell retentions of 98.7 0.2% (n = 5). These grafts were covered in confluent cell monolayers (see Fig. 4). Ex vivo flow studies on seeded grafts which had been cultured subsequently for 48 h showed high levels of cell retention on exposure to pulsatile blood flow of 150 mYmin at a mean wall shear stress of approximately 25 dynlcm2. Comparison of the mean isotopic count from the flow grafts (n = 5) with control grafts (n = 4) gave 81% cell retention over 3 h ex vivo flow. A Mann-Whitney u-test indicated no significant difference between flow tested and control groups (p > 0.1). All of the grafts were lined with confluent cell monolayers; Figs 5 and 6 show the cell cover on flow tested and control grafts, respectively. A low number of small holes, the area of a few cells each, were seen on three flow grafts but not on control grafts. There was no evidence of thrombogenicity . The preliminary ex vivo test conducted over 24 h gave a cell retention of 97% for flow tested grafts (n = 3) compared with control grafts (n = 2). Mann-Whitney comparison
FIG. 3. Endothelial cells on fibronectin (Fn)-coated polytetrafluoroethylene (PTFE) graft after 1 h seeding and imrnediate in vitro flow testing. Diff-Quik.
Artiforguns, Vol. 14, No. 5 , 1990
FIG. 4. Confluent endothelial cell monolayer on fibronectin (Fn)-coated polytetrafluoroethylene (PTFE) graft which was seeded for 1 h, cultured further for 48 h, and flow tested in vitro. Diff-Quik.
of the flow and control grafts indicated no difference between the two groups (p = 0.8). The ex vivo shunt remained patent over the length of the study and blood flow was maintained at 140 ml/min.
DISCUSSION In this study, a technique for establishing a fully confluent lining of endothelial cells on small diameter vascular grafts was developed. Seeded grafts
FIG. 5. Fibronectin (Fn)-coated polytetrafluoroethylene (PTFE) graft with confluent endothelial cell lining after exposure to 3 h ex vivo blood flow. Diff-Quik.
ENDOTHELIAL CELL SEEDING OF GRAFTS
FIG. 6. Endothelial cell monolayer on fibronectin (Fn)-coated polytetrafluoroethylene (PTFE) graft not exposed to blood flow, used as a control in 3 h ex vivo study. Diff-Quik.
retained confluent cell monolayers with exposure to flow in both in vitro flow studies and an ex vivo shunt without antiplatelet therapy. Alternative seeding techniques that aim at obtaining even cell coating (12-15) involve rotation of the graft by 90 to 180" at fixed intervals for up to 2 h. Subsequent culturing may be conducted for up to 48 h. The procedure developed in this study involves continuous rotation of the grafts for 1 h, ensuring even cell coating and eliminating the need for intermittent attention to the grafts. Further culture for 48 h was conducted to obtain confluent cell layers. Grafts that were seeded for 1 h and immediately flow tested and those tested following a further 48 h culture gave >95% cell retention in in vitro flow studies at physiological levels of wall shear stress. Ludgren et al. (16) report 87% retention for comparable flow studies on cultured canine endothelial cells grown on PTFE grafts overnight, and Sentissi et al. (17) report minimal loss of cultured bovine endothelial cells grown on PTFE grafts over a 2week period. Recent in vitro flow studies conducted with human umbilical vein cells on PTFE grafts following 2 h seeding gave 92% cell retention following exposure to similar shear stress conditions (6). The optimal benefit from endothelialization of
grafts prior to implantation should occur when the graft is covered with a confluent cell layer, since exposed regions of Fn-coated graft may elicit immediate platelet and thrombus accumulation on contact with blood (14). The limited harvest numbers from present intraoperative harvesting and seeding techniques severely restricts the proportion of the total graft surface that is coated with cells. In addition, grafts which are inoculated and immediately implanted lose a large proportion of the deposited cells within 24 h of implantation (18). Furthermore, grafts are coated to enhance cell attachment with clot, Fn, gelatin or collagen-materials that are highly thrombogenic. These factors could contribute to platelet deposition and thrombogenicity on seeded grafts in the initial stages following implantation. For these reasons, ex vivo studies were conducted on confluent cell monolayers. In the ex vivo studies, grafts retained confluent cell layers and maintained patency without antiplatelet therapy when exposed to pulsatile blood flow for 3- and 24-h periods. Schneider et al. (6) report a similar high level of cell retention in ex vivo studies using confluent human umbilical vein cells seeded for 2 h onto PTFE grafts in a baboon ex vivo shunt over a 1 h flow period. The sheep model, which is a relatively stringent test of graft performance, was successfully applied in these studies. Ortenwall et al. (19) seeded PTFE grafts in sheep and dogs. These authors obtained the lowest graft patency rates in the sheep model, indicating that this model is more conservative than the dog, and suggesting that it may be more difficult to maintain patent grafts in sheep than in dogs. In summary, a dynamic seeding method which gives even, confluent endothelial cell lining of Fncoated PTFE grafts has been developed. High levels of cell retention were obtained on exposure of endothelialized grafts to in vitro flow. Grafts seeded with autologous cells and exposed to ex vivo blood flow in a sheep model showed good cell retention and excellent patency without antiplatelet therapy in short-term studies. Thus the loss of endothelial cells from grafts in vivo may be influenced by factors other than the flow dynamics of the blood. These results suggest that examination of the performance of endothelial cell seeded grafts in the sheep model over a longer period would be appropriate. Acknowledgment:The authors thank the Department of Industry, Technology, and Commerce, Commonwealth of Australia, and Telectronics Pty. Ltd. for their support of this project and Dr. C. D. Bertram for his advice.
Artif Organs, Vol. 14, No. 5 , 1990
N . L. JAMES ET A L .
1. Belden TA, Schmidt SP, Falkow LJ, Sharp WV. Endothelial cell seeding of small diameter vascular grafts. Trans A m Soc Artiflntern Organs 1982;28:173-7. 2. Hening MB, Compton RS, Le Grand DR, Gardner AL, Madison DL, Glover JL. Endothelial seeding of polytetrafluoroethylene popliteal bypasses. A preliminary report. J Vasc Surg 1987;6:114-8. 3. Stanley JC, Burke WE, Ford JW, et al. Enhanced patency of small-diameter, externally supported Dacron iliofemoral grafts seeded with endothelial cells. Surgery 1982;92:9941005. 4. Allen BT, Long JA, Clark RE, Sicard GA, Hopkins KT, Welch MJ. Influence of endothelial cell seeding on platelet deposition and patency in small-diameter Dacron arterial grafts. J Vasc Surg 1984;1:224-33. 5. Hirko MK, Schmidt SP, Hunter TJ, Evancho M, Sharp WV, Donovan DL. Endothelid cell seeding improves 4 mm PTFE vascular graft performance in antiplatelet medicated dogs. Artery 1987;14:137-53. 6. Schneider PA, Hanson SR, Price TM, Harker LA. Preformed confluent endothelial cell monolayers prevent early platelet deposition on vascular prostheses in baboon. J Vasc Surg 1988;8:229-35. 7. Tillman P, Carson SN, Talken L. Platelet function and coagulation parameters in sheep during experimental vascular surgery. Lab Anim Sci 1981;31:263-7. 8. Hannan GN, Redmond JW, McAuslan BR. Similarity of the carbohydrate moieties of fibronectins derived from blood plasma and synthesised by cultured endothelial cells. Biochim Biophys Acfa 1984;801:396402. 9. Maciag T, Hoover GA, Stemerman MB, Weinstein R. Serial propagation of human endothelial cells in vitro. J Cell Biof 1981;9 1 :420-6.
Art$ Organs, Vol. 14, No. 5 , 1990
10. Kumar S, West DC, Ager A. Heterogeneity in endothelial
cells from large vessels and microvessels. Differentiation 1987;36:57-70. 11. Ryan US, Clements E, Habliston D, Ryan JW. Isolation and culture of pulmonary artery endothelial cells. Tissue Cell 1987;10:535-54. 12. Kesler KA, Herring MB, Arnold MP, et al. Enhanced strength of endothelid attachment on polyester elastomer and polytetrafluoroethylene graft surfaces with fibronectin substrate. J Vasc Surg 1986;3:58-64. 13. Anderson JS, Price TM, Hanson SR, Harker LA. In vitro endothelialization of small-caliber vascular grafts. Surgery 1987;101:577-86. 14. Shindo S , Takagi A, Whittemore AD. Improved patency of collagen-impregnated grafts after in vitro autogenous endothelid cell seeding. J Vasc Surg 1987;6:325-32. 15. Schneider PA, Hanson SR, Price TM, Harker LA. Durability of confluent endothelial cell monolayers on small-caliber vascular prostheses in vitro. Surgery 1988;103:45642. 16. Ludgren CH, Hening MB, Arnold MP, Glover JL, Bendick PJ. Fluid shear disruption of cultured endothelium: the effect of cell species, fibronectin cross-linking and supporting polymer. Trans A m Soc Artiflntern Organs 1986;32:334-8. 17. Sentissi JM, Ramberg K, O'Donnell TF, Connolly RJ, Callow AD. The effect of flow on vascular endothelial cells grown in tissue culture on polytetrafluoroethylene grafts. Surgery 1986;99:33743. 18. Ramalanjaona G, Kempczinski RF, Rosenman JE, Douville EC, Silberstein EB. The effect of fibronectin coating on endothelial cell kinetics in polytetrafluoroethylene grafts. J Vasc Surg 1986;3:264-72. 19. Ortenwall P, Bylock A, Kjelstrom T, Risberg B. Seeding of PTFE carotid interposition grafts in sheep and dogs: species-dependent results. Surgery 1988;103:19s205.