- Email: [email protected]
Early Attachment of Platelets, Leukocytes, and Fibrinogen in Endothelial Cell Seeded Dacron Grafts
Early Attachment of Platelets, Leukocytes, and Fibrinogen in Endothelial Cell Seeded Dacron Grafts Norman Jensen, MD, Bengt Lindblad, MD, PhD, June Ljungberg, BS, Sigrid Leide, BS, and David Bergqvist, MD, PhD, Malmi5 and Uppsala, Sweden
Endothelial cell seeding has been advocated as a method for reducing the thrombogenicity of prosthetic grafts. Principally two different techniques for endothelial cell seeding can be used: immediate seeding of grafts followed by implantation or initial growth and establishment of an endothelial cell-covered surface before subsequent late implantation. This study was designed to determine whether the immediate seeding technique altered thrombogenicity directly after graft implantation. Carotid arteries from 19 sheep were replaced with Dacron interposition grafts; one side was seeded with endothelial cells and the other side was left unseeded. The dynamics of thrombus formation involving radiolabeled platelets, leukocytes, and fibrinogen were studied for 4 hours with flow reduced to 35 ml/min. No difference in platelet uptake (-6-fold increase compared to baseline values) was found between endothelial cell seeded and unseeded grafts. Likewise, there were no differences in leukocyte uptake (-4-fold increase) or fibrinogen uptake ( - 1 0 - to 15-fold increase) between the two groups. No differences were demonstrated with regard to patency or thrombus weight. In this experimental investigation we were unable to verify any change in the uptake of platelets, white blood cells, or fibrinogen between endothelial cell seeded and unseeded Dacron grafts during the first 4 hours after graft placement. Immediate seeding does not affect the initial thrombogenicity of grafts. (Ann Vasc Surg 1996;10:530-536.)
The concept of lining a prosthetic blood conduit with endothelium is attractive because such a lining may provide a better nonthrombogenic surface than the graft itself. Results of long-term experimental studies in animals have been favorable, ~ but results in h u m a n s have been less convincing. 2~ Two different techniques for endothelial cell seeding can be used: immediate seeding From the Departments of Experimental Research and Surget'y, University Hospital, MatmS, Sweden, and the Department of Surgery (D.B.), University Hospital, Uppsala, Sweden. Supported by grants from the Swedish Heart and Lung Foundation, the Swedkh Medical Research Council (No. 00759), the Swedish Medical Society, the Faculty of Medicine, Lund University, and Malmi~ University Hospital. Reprint requests: Norman Jensen, MD, Department of Surgery, Central Hospital, S-432 81 Varberg, Sweden.
530
of grafts followed by implantation or initial growth in vitro onto the graft to establish a surface that is completely covered with endothelial ceils before subsequent late implantation. From a clinical and practical standpoint, the immediate seeding technique appears to be the most attractive. However, many problems remain to be solved before such a technique can be widely accepted. Endothelial cell harvesting must be efficient and the technique for breaking up cell-cell junctions should not alter cell membrane function or chromosome structure. Several reports have verified such effects. 6-s Additionally, optimizing cell attachment onto the graft material is equally important 9 after flow is restored. ~° Thus far, no attention has been paid to the functional properties of freshly seeded cells. Whether these cells actually reduce the initial
Vol. 10, No. 6 1996
uptake of platelets and leukocytes or limit the coagulation process has not yet been determined. If these processes are not altered, there is the chance that growth factors and other stimuli will induce a reparative response of the same type that exists in unseeded grafts. T M We therefore decided to evaluate whether the immediate seeding technique of endothelial cells altered thrombogenicity immediately after graft implantation.
MATERIAL A N D M E T H O D S Nineteen sheep, weighing between 24 and 39 kg (median 33 kg), were used. The erythrocyte volume fraction ranged from 22% to 34% (median 29%). The sheep were taken to the laboratory 3 to 5 days before the experiment so they could become acclimated to their surroundings. The experiment was approved by the Animal Ethics Committee of Lund University, and the animals were cared for in accordance with guidelines set forth by the European convention for laboratory animal care.
Surgical Procedure Anesthesia was induced with thiopental sodium (Abbott Laboratories, North Chicago, Ill.) and maintained with a slow, continuous infusion. Animals were intubated and respirations maintained with a ventilator (model 900, Siemens Elema, Stockholm, Sweden) using 50% oxygen and 50% nitrous oxide at a rate of 7.8 L/min for a 30 kg animal. A continuous infusion of 1000 ml of Ringer's solution was a d n ~ i s t e r e d . A midline neck incision was made, and the left jugular vein and the carotid arteries were dissected free. Blood was collected for platelet and leukocyte retrieval. The left jugular vein was resected for endothelial cell harvest. Segments of the carotid arteries measuring 4 cm were excised and replaced with externally supported knitted double-velour grafts (Microvel supported 100, Meadox Medicals, Inc., Oakland, N.J.) measuring 6 cm in length and 6 m m in diameter. Anastomoses were oblique and fashioned using running polypropylene sutures. On one side, chosen at random, a preclotted graft was inserted, and on the other side a preclotted and endothelial cell seeded graft was placed. Ten milliliters of heparinized saline solution (Pharmacia, Uppsala, Sweden), 10 IU/ml, was injected into each occluded artery; no other antithrombotic medication was administered. Blood flow was reduced to 35 ml/min and checked regularly with a transit time
Endothelial cell seeding of Dacron grafts
531
flowmeter (Transonic TC 101, Transonic Systems Inc., Ithaca, N.Y.).
Endothefial Cell Harvest and Seeding After careful dissection with ligation of all tributaries, approximately 18 cm of the left jugular vein was excised, and small cannulas were introduced and tied at each end. The vein was flushed with phosphate-buffered saline (PBS; Nordvac Medica, Stockholm, Sweden) and then filled with a solution of neutral protease (Dispase II, Boerhinger-Mannheim, Mannheim, Germany), 5 mg/ml in PBS, and incubated for 60 minutes in 37 ° C. The veins were then emptied, filled with PBS, and slightly massaged between fingers to loosen any residual cells. Both cell suspensions were centrifuged at 200 x g for 7 minutes and the pellet was suspended in 1 ml of medium 199 (M-199, Gibco Laboratories, Grand Island, N.Y.), an aliquot was counted, and viability was assessed with the trypan blue exclusion test. Ceils from this experiment were not subcultured, but ceils harvested in an identical manner have been cultured and stained for factor VIII-related antigen and found to be positive. The grafts were preclotted with repeated injections of blood and the blood was allowed to coagulate. The graft was cleared with an embolectomy catheter, and the process was then repeated until the graft was totally leak proof. One of the two preclotted grafts was occluded at one end with a clamp, filled with the cell suspension, topped up with M-199, and occluded at the other end. It was left for 45 minutes at 37 ° C and was rotated 90 degrees every 10 minutes. After the adherence period, the fluid was emptied and the graft inserted.
Separation and Labeling of Leukocytes A total of 50 ml of blood was drawn without anticoagulant. Platelets were removed and the blood was defibfinated by swirling it in a 100 ml flask containing 30 to 40 glass beads (2 to 3 inin in diameter) for i0 minutes. Defibrinated blood was mixed with an equal amount of PBS, and 6 ml of the mixed blood was layered over 3 ml of Lymphoprep (Nyegaard & Co., Oslo, Norway). The tubes were centrifuged for 20 minutes at 600 × g. The pelleted fraction contained erythrocytes and granulocytes. After removing the separation fluid down to 2 m m above the erythrocyte mass, the cells in the pellet were resuspended in 2 rnl of PBS, transferred to
532
Annals of Vascular Surgery
Jensen et al.
other tubes, and an additional 5 ml of PBS was added prior to centrifugation for 7 minutes at 600 x g. The pelleted cells were resuspended three times in 1.8 ml of distilled water followed by 200 ~,1 of PBS in order to lyse the red blood cells. The tubes were filled up to 10 ml with PBS and centrifuged for 7 minutes at 600 x g. The pellets, which now contained granulocytes, were resuspended in 0.9% sodium chloride and combined into one 12 ml tube; 2 x 25 fzl was removed from the suspension to obtain white blood cell and platelet counts. The tube was then centrifuged for 7 minutes at 600 x g, and the pellet was added to 4 ml of technetium-99m-HMPAO, 400 MBq (HM-PAO-Ceretec, Amersham International, Amersham, U.K.). The leukocytes were incubated for 20 minutes, and in order to stop the labeling process, 6 ml of cell-free plasma was added. After centrifugation (600 x g for 5 minutes), the supernate was removed and the leukocyte preparation was resuspended in 5 to 10 ml of cell-free plasma. Before intravenous injection, the radioactivity in the cells and in the supernate was measured, and the labeling efficiency was calculated.
Separation and Labeling of Platelets The labeling procedure was modified from that described by Thakur et al. ~5 Whole blood (51 ml) was drawn and mixed with 9 ml of acid citrate dextrose solution. The mixture was centrifuged at 200 × g for 15 minutes. The supernate (plateletrich plasma [PRP] ) was decanted and centrifuged at 1000 x g for 10 minutes, and the supernate (platelet-poor plasma [PPP]) was removed and saved. The platelet buttons were resuspended in 2 ml of modified Tyrode's solution and combined into one tube and centrifuged once again at 1000 × g for i0 minutes. The platelet button was then resuspended in 4 ml of modified Tyrode's solution, samples were taken to obtain platelet and white blood cell counts, and the mixture was incubated in 12.5 MBq indium-111--oxine (DRN 4908 indium oxinate, Mallinckrodt Diagnostica BV, Petten, The Netherlands) for 15 minutes. To remove any indium not bound to the platelets, 2 ml of PPP was added to the labeled platelet suspension followed by centrifugation at 1000 × g and removal of the supemate. The platelet button was resuspended in PPP and before reinfusion, the radioactivity in the cells and the supernate was measured for calculation of the labeling efficiency.
Separation and Labeling of Fibrinogen Homologous fibrinogen was prepared prior to the experiments. From one sheep approximately 1000 ml of plasma was obtained. Fibrinogen was purified according to the method of Blomb~ck and Blomb~ck, 16 and labeling was performed using iodine-125 according to the method described by McFarlane~7; aliquots containing 0.2 mg of fibrinogen in each were frozen for later use. Iodine125 has a half-life of 60 days, which makes this prefabrication possible.
Radioactivity Measurements and Data Processing Four g a m m a radiation sodium iodide detectors (NaI IT1], model 0.91 × M.510/.75 BLP-X, Bicron Corp. Inc., Newbury, Ohio) placed in excavations in two lead collimators, two detectors in each colimator, were used simultaneously. The detector was connected to a linear amplifier and the signals were analyzed using a multichannel analyzer (model 918 A, 476-8, EG & G Ortec Nuclear Systems, Oak Ridge, Tenn.), which was convected to a computor (Vectra 45940 A, Hewlett-Packard, Sunnyvale, Calif.). Since measurements were carried out simultaneously with four detectors, a multichannel buffer and a multiplexer/router were used to handle the signals from the detectors before they were analyzed by means of the multichannel analyzer. The count rates of the radionuclides were corrected to allow for the physical decay, the influence of scattered radiation, and the limited energy resolution of the detectors. After correction of the count rates, the changes in these rates were representative of the changes in the a m o u n t of platelets, leukocytes, and fibrinogen. During the first measurement period (60 seconds × 17) all detectors were placed over the intact arteries. Measurements were carried out to detect any radionuclide contamination. Thereafter the radiolabeled leukocytes, platelets, and fibrinogen were injected and activity in the native artery was measured. These activity measurements were used as a baseline. After the grafts were inserted, blood flow was reduced to 35 ml/min by a distal clamp. One detector was placed proximally over the artery and one over the graft. Measurements were then carried out at 5-minute periods for 4 hours. The arteries as well as the grafts were wrapped in plastic film, and physiologic saline solution was added underneath to keep them moist.
Vol. 10, No. 6 1996
Endothelial cellseeding of Dacrongrafts
Statistics
A lead a p r o n w a s u s e d to shield t h e s h e e p ' s b o d y to reduce r a d i a t i o n scatter. After the m e a s u r e m e n t period, the grafts w e r e excised, o p e n e d lengthwise, a n d t h e t h r o m b u s free surface w a s assessed. Excess liquid b l o o d w a s a b s o r b e d f r o m t h e a d h e r e n t t h r o m b u s u s i n g blotting paper, a n d the film of t h r o m b u s w a s r e m o v e d w i t h a scalpel a n d w e i g h e d . The a n i m a l s w e r e t h e n killed.
The values p r e s e n t e d in Figs. 1 to 3 are b a s e d o n the corrected c o u n t rates a n d c o r r e s p o n d e d to c o u n t r a t e s in t h e n a t i v e arteries as m e a s u r e d before i n s e r t i o n of t h e grafts. The results are p r e s e n t e d as m e d i a n s a n d interquartfle ranges. For analysis t h e M a n n - W h i t n e y U test a n d the W i l c o x o n test for paired o b s e r v a t i o n s w e r e used; p < 0.05 w a s c o n s i d e r e d significant.
10 ] Endothelialcell seeded grafts ] Non-seeded grafts .= 7,5-
tTw[1 T
,~
5-
2:3
E
~ 2,5-
I
I
I
I
I
!
I un
I
t
xJD
I
I
1
I~
Time (min.) Fig. 1. Comparison of platelet uptake of indium during a 4-hour period in endothelial cell seeded and unseeded Dacron grafts.
10 [] Endothelialcell seeded grafts 7,5-
Non-seeded grafts
[]
c.)
[
O
5-
ml
8
I
Cq
533
I
I
I
I
I
rch
Time (rain.) Fig. 2. Comparison of leukocyte uptake of technetium during a 4-hour period in endothelial cell seeded and unseeded Dacron grafts.
534
Annals of Vascular Surgery-
Jensen et aL
30
' ' " ""="
25~>+ -~
[]
Endothelial cell seeded grafts
[]
Non-seeded grafts
20-
T
TI T
T T
t
a
:
r,
I
|
z
I
I
I
_
Time (min.)
Fig. 3. Comparison of fibrinogen uptake of iodine during a 4-hour period in endothelial cell seeded and unseeded Dacron grafts.
RESULTS A median of 4.8 x l0 s endothelial cells were harvested (interquartile range 2 to 6.4 x i05). The cell harvest allowed a seeding density of 4.2 x 104 cells/cm2 of graft surface (interquartile range 1.7 to 5.6 x 104 cells/cm2). A median of 54 x 106 leukocytes were used for labeling (interquartile range 40 to 61 x 106) of which 62% were polymorphonuclear cells (interquartile range 55% to 70%) that had a median value for contamination of platelets of 19 x 106 (interquartile range 1.4 to 27 x 106). Platelet contamination varied among experiments and was greater t h a n 10% in eight experiments and less t h a n 10% in six. A median of 640 MBq 99m-technetium was used and 76% of the activity was found to be located intracellularly (interquartile range 63% to 89%). Figs. 1 to 3 show the attachments of the various thrombus components as activity increased compared to baseline values. No differences could be demonstrated in the attachment of platelets, leukocytes, and fibrinogen. With regard to platelets (Fig. 1), there was a great deal of interindividual variability in platelet attachment, but there was less variation within each experiment. During the 4 hours of measurement a slight increase in platelet attachment was noted, which was found to be approximately 6-fold greater t h a n the baseline value by the end of the measurement period. As for leukocytes (Fig. 2), there was only a slight increase during the measurement period. The m a x i m u m activity was approximately 3.5
times that of the baseline value. Both platelets and leukocytes showed an immediate attachment. Fibrinogen uptake (Fig. 3) was initially low and showed a continuous increase with a maxim u m activity approximately 10 to 15 times that of baseline measurements. One of the 19 endothelial cell seeded grafts occluded after 115 minutes compared to three unseeded grafts (68, 178, and 194 minutes) (not significant). The median thrombus weight was 311 mg (interquartile range 218 to 408 mg) in the endothelial cell seeded group and 299 mg (interquartile range 210 to 395 mg) (not significant) in the unseeded group.
DISCUSSION In animal experiments, endothelial cell seeding shows a beneficial effect after 3 to 5 weeks/As but a longer healing time may diminish the differences. 19'2° The possible effects, beneficial or negative, of seeded endothelial cells during the first hours or days have received very little attention. When seeding endarterectomized baboon aortas with a very high density of endothelial cells (6 X l0 s cells/cm2), Schneider et al. 2' noted a decrease in platelet deposition during the first hour after seeding and also demonstrated the spreading out of endothelial cells at this time. This shows that seeded cells can regain their functional properties very soon after seeding. In comparison, our seeding densities were lower. Another difference was the seeding base, preclot-
Vol. 10, No. 6 1996
ted Dacron, which differed from endarterectomized artery. Enzymatic methods for harvesting endothelial cells could induce alterations in endothelial cell function. 6 Enzymatic harvesting from different species has been shown to be fairly efficient but in cell harvesting from h u m a n veins this efficiency is suboptimal. We have achieved the best harvesting efficiency with h u m a n saphenous vein using Dispase II. Various collagenases have been less effective in our hands. Another potential drawback of collagenase-harvested endothelial cells is the possibility of functional alterations, which have been reported. 22 It may be desirable to evaluate the grafts from a morphologic point of view, but after 4 hours under reduced flow, the Dacron surface is covered by a thin film of thrombus. The seeded endothelial cells are hidden underneath. It is extremely difficult to evaluate the surface by means of electron microscopy, given these conditions. We chose preclotted Dacron grafts because of their higher cell attachment rate, about 50%, which is approximately five times higher than what we achieved with expanded polytetrafluoroethylene (PTFE) grafts. 9 After 2 hours of flow, about 50% of seeded cells remain on the graft, but most of the cell loss takes place during the first 20 minutes) ° We also know that the same seeding technology produces measurable improved endothelialization after 3 weeks. ~9 The time allowed for endothelial cells to adhere and spread out was limited in this experimental series to reflect the clinical situation. There are several possible explanations for the lack of measurable effects. The most likely one is that the endothelial cells are in a proliferative state and display fewer nonthrombogenetic properties. T M This raises the question of whether immediate seeding can be of benefit in the clinical situation. Will a nonfunctioning endothelium on a graft lining induce the same reparative process as on a nonseeded graft? No benefit is derived from a nonfunctioning endothelium unless the reparative process is altered. Supraconfluent density seeding 2~ might shorten the time during which cell function is interrupted. Endothelial cells have a distinct contact inhibition and as a result functional variations m a y occur. The incidence of contamination during platelet harvest was low, but during leukocyte harvest the proportion of mononuclear cells was high and platelet contamination was high. This caused some uncertainty regarding the uptake of leukocytes onto grafts. Although there is a possibility
Endothelial cell seeding of Dacron grafts
535
that determinations of leukocyte uptake were inaccurate, the uptake curves were similar w h e n only experiments with low platelet contamination were analyzed. Platelet attachment is used as a quantifiable marker of thrombus formation. Leukocytes have not been used, partly because of methodologic difficulties. Their behavior in relation to platelets, synthetic grafts, and endothelium does warrant some attention. Leukocyte-depleted dogs have a higher rate of endothelial cell retention than control dogs, which suggests that leukocytes cause increased endothelial sloughing. 23 In our study there was no increased attachment of leukocytes to the seeded grafts. Whether the platelet uptake in this experimental model corresponds to conditions in humans is unknown. Without a doubt there are differences in the behavior of platelets between species and sheep platelets do react differently to epinephrine in comparison to h u m a n platelets; adhesion is lower and thromboxane production is almost absent in sheep. On the other hand, this difference is more pronounced in other species such as rabbits and goats where platelet reactivity is lower than it is in humans and sheep. Platelet reactivity is higher in monkeys than in humans. Thus there is no ideal experimental animal. 24 In this series a substantial increase in fibrinogen uptake was noted. This increased uptake is greater than what we have found w h e n blood flow is reduced to mimic a poor runoff state in PTFE grafts. 25 In sheep such a large increase in the uptake of fibrinogen can be the result of reduced fibrinolytic activity. It can also depend on differences in the type of graft material. The reason for the difference between platelet and leukocyte uptake on the one hand and fibrinogen uptake on the other is unclear.
CONCLUSION The multichannel analyzer setup, which uses several radionuclides in parallel, has led to the development of the present-day experimental setring for graft thrombogenicity studies. A conventional gamma camera measures the count rate from one and occasionally two radionuclides, whereas the multichannel analyzer can analyze the gamma count rate from up to three radionuclides simultaneously with good accuracy for prolonged periods of time. The gamma camera uses one detector but the multichannel analyzer uses four. Because each detector obtains a complete energy spectrum and each detector covers one
536
Jensen et al.
region of interest, two grafts and native vessels can be examined simultaneously. Our experimental investigation, however, was unable to demonstrate any reduction in uptake of platelets, leukocytes, or fibrinogen in grafts immediately seeded with endothelial ceils compared to unseeded control grafts during the first 4 hours after graft placement. Immediate seeding does not initially produce any reduction in graft thrombogenicity. REFERENCES 1. Stanley JC, Burkel WE, Graham LM, et aL Endothelial cell seeding of synthetic vasollar prosthesis. Acta Chir Scand 1985;529( Suppl): 17-28. 2. Jensen N, Lindblad B, Bergqvist D. Endothelial cell seeded Dacron aortobifurcated grafts: Platelet deposition and longterm follow-up. J Cardiovasc Surg 1994;35:425-429. 3. Ortenwall E Wadenvik H, Risberg B. Reduced platelet deposition on seeded versus unseeded segments of expanded polytetrafluoroethylene grafts: Clinical observations after a 6-month follow-up. J Vasc Surg 1989;10:374-380. 4. Zilla E Deutsch M, Meinhart J, et al. Clinical in vitro endothelialization of femoropopliteal bypass grafts: An actuarial follow-up over three years. J Vasc Surg 1994;19:540-548. 5. Magometschnigg H, Kadletz M, Vodrazka M, et al. Changes following in vitro endothelial cell lining of ePTFE prostheses: Late morphologic evaluation of six failed grafts. Eur J Vasc Surg 1994;8:502-507. 6. Ford JW, Burkel WE. Isolation of adult canine venous endothelium for tissue culture. In Vitro Cell Dev Biol Anim 1981;17:44-50. 7. Engfeldt E A m e r E Ostman J. Nature of the inhibitory effect of collagenase on phosphodlesterase activity. J Lipid Res 1985;26:977-981. 8. Cesarone CE Fugasa E, Gallo G, et al. Collagenase perfusion of rat liver induces DNA damage and repair in hepatocytes. Mutation Res 1984;141:113-116. 9. Jensen N, Lindblad B, Bergqvist D, In vitro attachment of endothelial cells to different graft materials, Eur Surg Res 1996;28:49-54. 10. Jensen N, Lindblad B, Leide S, et ak Loss of seeded endothelial cells in vivo. A study of Dacron grafts under different flow conditions. Eur J Vasc Surg 1994;8:690-693. 11. DiCorletto PE. Cultured endothelial cells produce multiple growth factors for connective tissue cells. Exp Cell Res 1984;153:167-172.
Annals of Vascular Surgery
12. Barrett TB, Gajdusek CM, Schwartz SM, et al. Expression of the sis gene by endothelial cells in culture and in vivo. Proc Natl Acad Sci USA 1984;81:6772-6774. 13. Collins T, Ginsburg D, Boss JM, et al. Cultured human endothelial cells express platelet derived growth factor B chain: cDNA cloning and structural analysis. Nature 1985; 316:748-750. 14. Clowes AW, Gown AM, Hanson SR, et aL Mechanisms of arterial graft failure: 1. Role of cellular proliferation in early healing of PTFE prostheses. Am J Pathol 1985;118:43-54. 15. Thakur ML, Walsh L, Malech HL, et al. Indium-ll 1--labeled human platelets: Improved method, efficacy and evaluation. J Nucl Med 1981;22:381-385. 16. Blomb/ick B, Blomb~ick M. Purification of human and bovine fibrinogen. Arkiv Kemi 1956;10:415-443. 17. McFarlane AS. Efficient trace labeling of proteins with iodine. Nature 1958;182:53. 18. Jensen N, Brunkwall J, F~ilt K, et al. Prostacyclins are produced from endothelial cell-seeded grafts: An experimental study in sheep. Eur J Vasc Surg 1992;6:499-504. 19. Jensen N, Brunkwall J, F~ilt K, et al. Recovery of endothelial cells and prostanoid production in endothelial cell seeded grafts. Eur J Vasc Endovasc Surg (in press). 20. Burkel WE, Ford JW, Vinter DW, et al. Fate of knitted Dacron velour vascular grafts seeded with enzymatically derived autologous canine endothelium. Trans Am Soc Artif Intern Organs 1982;28:178-182. 21. Schneider PA, Hanson SR, Price TM, et al. Confluent durable endothellalization of endarterectomized baboon aorta by early attachment of cultured endothelial cells. J Vasc Surg t990; 11:365-372. 22. Sharefkii1 JB, Fairchild KD, Albus RA, et al. The cytotoxic effect of surgical glove powder particles on adult human vascular endothelial cell cultures: Implications for clinical uses of tissue culture techniques. J Surg Res 1986;41:463472. 23. Emerick S, Herring M, Amold M, et al. Leukocyte depletion enhances cultured endothelial retention on vascular prostheses. J Vasc Surg 1987;5:342-347. 24. Didisheim P. Comparative hematology in the human, call sheep, and goat: Relevance to implantable blood pump evaluation. ASAIO J 1985;8:123-127. 25. Lundell A, Bergqvist D, Lindblad B. The uptake of platelets, fibrinogen and leukocytes in eFrFE vascular grafts in relation to blood f l o w - A n experimental study in the sheep. Eur J Vasc Surg 1993;7:698-703.
Endothelial cell seeding has been advocated as a method for reducing the thrombogenicity of prosthetic grafts. Principally two different techniques for endothelial cell seeding can be used: immediate seeding of grafts followed by implantation or initial growth and establishment of an endothelial cell-covered surface before subsequent late implantation. This study was designed to determine whether the immediate seeding technique altered thrombogenicity directly after graft implantation. Carotid arteries from 19 sheep were replaced with Dacron interposition grafts; one side was seeded with endothelial cells and the other side was left unseeded. The dynamics of thrombus formation involving radiolabeled platelets, leukocytes, and fibrinogen were studied for 4 hours with flow reduced to 35 ml/min. No difference in platelet uptake (-6-fold increase compared to baseline values) was found between endothelial cell seeded and unseeded grafts. Likewise, there were no differences in leukocyte uptake (-4-fold increase) or fibrinogen uptake ( - 1 0 - to 15-fold increase) between the two groups. No differences were demonstrated with regard to patency or thrombus weight. In this experimental investigation we were unable to verify any change in the uptake of platelets, white blood cells, or fibrinogen between endothelial cell seeded and unseeded Dacron grafts during the first 4 hours after graft placement. Immediate seeding does not affect the initial thrombogenicity of grafts. (Ann Vasc Surg 1996;10:530-536.)
The concept of lining a prosthetic blood conduit with endothelium is attractive because such a lining may provide a better nonthrombogenic surface than the graft itself. Results of long-term experimental studies in animals have been favorable, ~ but results in h u m a n s have been less convincing. 2~ Two different techniques for endothelial cell seeding can be used: immediate seeding From the Departments of Experimental Research and Surget'y, University Hospital, MatmS, Sweden, and the Department of Surgery (D.B.), University Hospital, Uppsala, Sweden. Supported by grants from the Swedish Heart and Lung Foundation, the Swedkh Medical Research Council (No. 00759), the Swedish Medical Society, the Faculty of Medicine, Lund University, and Malmi~ University Hospital. Reprint requests: Norman Jensen, MD, Department of Surgery, Central Hospital, S-432 81 Varberg, Sweden.
530
of grafts followed by implantation or initial growth in vitro onto the graft to establish a surface that is completely covered with endothelial ceils before subsequent late implantation. From a clinical and practical standpoint, the immediate seeding technique appears to be the most attractive. However, many problems remain to be solved before such a technique can be widely accepted. Endothelial cell harvesting must be efficient and the technique for breaking up cell-cell junctions should not alter cell membrane function or chromosome structure. Several reports have verified such effects. 6-s Additionally, optimizing cell attachment onto the graft material is equally important 9 after flow is restored. ~° Thus far, no attention has been paid to the functional properties of freshly seeded cells. Whether these cells actually reduce the initial
Vol. 10, No. 6 1996
uptake of platelets and leukocytes or limit the coagulation process has not yet been determined. If these processes are not altered, there is the chance that growth factors and other stimuli will induce a reparative response of the same type that exists in unseeded grafts. T M We therefore decided to evaluate whether the immediate seeding technique of endothelial cells altered thrombogenicity immediately after graft implantation.
MATERIAL A N D M E T H O D S Nineteen sheep, weighing between 24 and 39 kg (median 33 kg), were used. The erythrocyte volume fraction ranged from 22% to 34% (median 29%). The sheep were taken to the laboratory 3 to 5 days before the experiment so they could become acclimated to their surroundings. The experiment was approved by the Animal Ethics Committee of Lund University, and the animals were cared for in accordance with guidelines set forth by the European convention for laboratory animal care.
Surgical Procedure Anesthesia was induced with thiopental sodium (Abbott Laboratories, North Chicago, Ill.) and maintained with a slow, continuous infusion. Animals were intubated and respirations maintained with a ventilator (model 900, Siemens Elema, Stockholm, Sweden) using 50% oxygen and 50% nitrous oxide at a rate of 7.8 L/min for a 30 kg animal. A continuous infusion of 1000 ml of Ringer's solution was a d n ~ i s t e r e d . A midline neck incision was made, and the left jugular vein and the carotid arteries were dissected free. Blood was collected for platelet and leukocyte retrieval. The left jugular vein was resected for endothelial cell harvest. Segments of the carotid arteries measuring 4 cm were excised and replaced with externally supported knitted double-velour grafts (Microvel supported 100, Meadox Medicals, Inc., Oakland, N.J.) measuring 6 cm in length and 6 m m in diameter. Anastomoses were oblique and fashioned using running polypropylene sutures. On one side, chosen at random, a preclotted graft was inserted, and on the other side a preclotted and endothelial cell seeded graft was placed. Ten milliliters of heparinized saline solution (Pharmacia, Uppsala, Sweden), 10 IU/ml, was injected into each occluded artery; no other antithrombotic medication was administered. Blood flow was reduced to 35 ml/min and checked regularly with a transit time
Endothelial cell seeding of Dacron grafts
531
flowmeter (Transonic TC 101, Transonic Systems Inc., Ithaca, N.Y.).
Endothefial Cell Harvest and Seeding After careful dissection with ligation of all tributaries, approximately 18 cm of the left jugular vein was excised, and small cannulas were introduced and tied at each end. The vein was flushed with phosphate-buffered saline (PBS; Nordvac Medica, Stockholm, Sweden) and then filled with a solution of neutral protease (Dispase II, Boerhinger-Mannheim, Mannheim, Germany), 5 mg/ml in PBS, and incubated for 60 minutes in 37 ° C. The veins were then emptied, filled with PBS, and slightly massaged between fingers to loosen any residual cells. Both cell suspensions were centrifuged at 200 x g for 7 minutes and the pellet was suspended in 1 ml of medium 199 (M-199, Gibco Laboratories, Grand Island, N.Y.), an aliquot was counted, and viability was assessed with the trypan blue exclusion test. Ceils from this experiment were not subcultured, but ceils harvested in an identical manner have been cultured and stained for factor VIII-related antigen and found to be positive. The grafts were preclotted with repeated injections of blood and the blood was allowed to coagulate. The graft was cleared with an embolectomy catheter, and the process was then repeated until the graft was totally leak proof. One of the two preclotted grafts was occluded at one end with a clamp, filled with the cell suspension, topped up with M-199, and occluded at the other end. It was left for 45 minutes at 37 ° C and was rotated 90 degrees every 10 minutes. After the adherence period, the fluid was emptied and the graft inserted.
Separation and Labeling of Leukocytes A total of 50 ml of blood was drawn without anticoagulant. Platelets were removed and the blood was defibfinated by swirling it in a 100 ml flask containing 30 to 40 glass beads (2 to 3 inin in diameter) for i0 minutes. Defibrinated blood was mixed with an equal amount of PBS, and 6 ml of the mixed blood was layered over 3 ml of Lymphoprep (Nyegaard & Co., Oslo, Norway). The tubes were centrifuged for 20 minutes at 600 × g. The pelleted fraction contained erythrocytes and granulocytes. After removing the separation fluid down to 2 m m above the erythrocyte mass, the cells in the pellet were resuspended in 2 rnl of PBS, transferred to
532
Annals of Vascular Surgery
Jensen et al.
other tubes, and an additional 5 ml of PBS was added prior to centrifugation for 7 minutes at 600 x g. The pelleted cells were resuspended three times in 1.8 ml of distilled water followed by 200 ~,1 of PBS in order to lyse the red blood cells. The tubes were filled up to 10 ml with PBS and centrifuged for 7 minutes at 600 x g. The pellets, which now contained granulocytes, were resuspended in 0.9% sodium chloride and combined into one 12 ml tube; 2 x 25 fzl was removed from the suspension to obtain white blood cell and platelet counts. The tube was then centrifuged for 7 minutes at 600 x g, and the pellet was added to 4 ml of technetium-99m-HMPAO, 400 MBq (HM-PAO-Ceretec, Amersham International, Amersham, U.K.). The leukocytes were incubated for 20 minutes, and in order to stop the labeling process, 6 ml of cell-free plasma was added. After centrifugation (600 x g for 5 minutes), the supernate was removed and the leukocyte preparation was resuspended in 5 to 10 ml of cell-free plasma. Before intravenous injection, the radioactivity in the cells and in the supernate was measured, and the labeling efficiency was calculated.
Separation and Labeling of Platelets The labeling procedure was modified from that described by Thakur et al. ~5 Whole blood (51 ml) was drawn and mixed with 9 ml of acid citrate dextrose solution. The mixture was centrifuged at 200 × g for 15 minutes. The supernate (plateletrich plasma [PRP] ) was decanted and centrifuged at 1000 x g for 10 minutes, and the supernate (platelet-poor plasma [PPP]) was removed and saved. The platelet buttons were resuspended in 2 ml of modified Tyrode's solution and combined into one tube and centrifuged once again at 1000 × g for i0 minutes. The platelet button was then resuspended in 4 ml of modified Tyrode's solution, samples were taken to obtain platelet and white blood cell counts, and the mixture was incubated in 12.5 MBq indium-111--oxine (DRN 4908 indium oxinate, Mallinckrodt Diagnostica BV, Petten, The Netherlands) for 15 minutes. To remove any indium not bound to the platelets, 2 ml of PPP was added to the labeled platelet suspension followed by centrifugation at 1000 × g and removal of the supemate. The platelet button was resuspended in PPP and before reinfusion, the radioactivity in the cells and the supernate was measured for calculation of the labeling efficiency.
Separation and Labeling of Fibrinogen Homologous fibrinogen was prepared prior to the experiments. From one sheep approximately 1000 ml of plasma was obtained. Fibrinogen was purified according to the method of Blomb~ck and Blomb~ck, 16 and labeling was performed using iodine-125 according to the method described by McFarlane~7; aliquots containing 0.2 mg of fibrinogen in each were frozen for later use. Iodine125 has a half-life of 60 days, which makes this prefabrication possible.
Radioactivity Measurements and Data Processing Four g a m m a radiation sodium iodide detectors (NaI IT1], model 0.91 × M.510/.75 BLP-X, Bicron Corp. Inc., Newbury, Ohio) placed in excavations in two lead collimators, two detectors in each colimator, were used simultaneously. The detector was connected to a linear amplifier and the signals were analyzed using a multichannel analyzer (model 918 A, 476-8, EG & G Ortec Nuclear Systems, Oak Ridge, Tenn.), which was convected to a computor (Vectra 45940 A, Hewlett-Packard, Sunnyvale, Calif.). Since measurements were carried out simultaneously with four detectors, a multichannel buffer and a multiplexer/router were used to handle the signals from the detectors before they were analyzed by means of the multichannel analyzer. The count rates of the radionuclides were corrected to allow for the physical decay, the influence of scattered radiation, and the limited energy resolution of the detectors. After correction of the count rates, the changes in these rates were representative of the changes in the a m o u n t of platelets, leukocytes, and fibrinogen. During the first measurement period (60 seconds × 17) all detectors were placed over the intact arteries. Measurements were carried out to detect any radionuclide contamination. Thereafter the radiolabeled leukocytes, platelets, and fibrinogen were injected and activity in the native artery was measured. These activity measurements were used as a baseline. After the grafts were inserted, blood flow was reduced to 35 ml/min by a distal clamp. One detector was placed proximally over the artery and one over the graft. Measurements were then carried out at 5-minute periods for 4 hours. The arteries as well as the grafts were wrapped in plastic film, and physiologic saline solution was added underneath to keep them moist.
Vol. 10, No. 6 1996
Endothelial cellseeding of Dacrongrafts
Statistics
A lead a p r o n w a s u s e d to shield t h e s h e e p ' s b o d y to reduce r a d i a t i o n scatter. After the m e a s u r e m e n t period, the grafts w e r e excised, o p e n e d lengthwise, a n d t h e t h r o m b u s free surface w a s assessed. Excess liquid b l o o d w a s a b s o r b e d f r o m t h e a d h e r e n t t h r o m b u s u s i n g blotting paper, a n d the film of t h r o m b u s w a s r e m o v e d w i t h a scalpel a n d w e i g h e d . The a n i m a l s w e r e t h e n killed.
The values p r e s e n t e d in Figs. 1 to 3 are b a s e d o n the corrected c o u n t rates a n d c o r r e s p o n d e d to c o u n t r a t e s in t h e n a t i v e arteries as m e a s u r e d before i n s e r t i o n of t h e grafts. The results are p r e s e n t e d as m e d i a n s a n d interquartfle ranges. For analysis t h e M a n n - W h i t n e y U test a n d the W i l c o x o n test for paired o b s e r v a t i o n s w e r e used; p < 0.05 w a s c o n s i d e r e d significant.
10 ] Endothelialcell seeded grafts ] Non-seeded grafts .= 7,5-
tTw[1 T
,~
5-
2:3
E
~ 2,5-
I
I
I
I
I
!
I un
I
t
xJD
I
I
1
I~
Time (min.) Fig. 1. Comparison of platelet uptake of indium during a 4-hour period in endothelial cell seeded and unseeded Dacron grafts.
10 [] Endothelialcell seeded grafts 7,5-
Non-seeded grafts
[]
c.)
[
O
5-
ml
8
I
Cq
533
I
I
I
I
I
rch
Time (rain.) Fig. 2. Comparison of leukocyte uptake of technetium during a 4-hour period in endothelial cell seeded and unseeded Dacron grafts.
534
Annals of Vascular Surgery-
Jensen et aL
30
' ' " ""="
25~>+ -~
[]
Endothelial cell seeded grafts
[]
Non-seeded grafts
20-
T
TI T
T T
t
a
:
r,
I
|
z
I
I
I
_
Time (min.)
Fig. 3. Comparison of fibrinogen uptake of iodine during a 4-hour period in endothelial cell seeded and unseeded Dacron grafts.
RESULTS A median of 4.8 x l0 s endothelial cells were harvested (interquartile range 2 to 6.4 x i05). The cell harvest allowed a seeding density of 4.2 x 104 cells/cm2 of graft surface (interquartile range 1.7 to 5.6 x 104 cells/cm2). A median of 54 x 106 leukocytes were used for labeling (interquartile range 40 to 61 x 106) of which 62% were polymorphonuclear cells (interquartile range 55% to 70%) that had a median value for contamination of platelets of 19 x 106 (interquartile range 1.4 to 27 x 106). Platelet contamination varied among experiments and was greater t h a n 10% in eight experiments and less t h a n 10% in six. A median of 640 MBq 99m-technetium was used and 76% of the activity was found to be located intracellularly (interquartile range 63% to 89%). Figs. 1 to 3 show the attachments of the various thrombus components as activity increased compared to baseline values. No differences could be demonstrated in the attachment of platelets, leukocytes, and fibrinogen. With regard to platelets (Fig. 1), there was a great deal of interindividual variability in platelet attachment, but there was less variation within each experiment. During the 4 hours of measurement a slight increase in platelet attachment was noted, which was found to be approximately 6-fold greater t h a n the baseline value by the end of the measurement period. As for leukocytes (Fig. 2), there was only a slight increase during the measurement period. The m a x i m u m activity was approximately 3.5
times that of the baseline value. Both platelets and leukocytes showed an immediate attachment. Fibrinogen uptake (Fig. 3) was initially low and showed a continuous increase with a maxim u m activity approximately 10 to 15 times that of baseline measurements. One of the 19 endothelial cell seeded grafts occluded after 115 minutes compared to three unseeded grafts (68, 178, and 194 minutes) (not significant). The median thrombus weight was 311 mg (interquartile range 218 to 408 mg) in the endothelial cell seeded group and 299 mg (interquartile range 210 to 395 mg) (not significant) in the unseeded group.
DISCUSSION In animal experiments, endothelial cell seeding shows a beneficial effect after 3 to 5 weeks/As but a longer healing time may diminish the differences. 19'2° The possible effects, beneficial or negative, of seeded endothelial cells during the first hours or days have received very little attention. When seeding endarterectomized baboon aortas with a very high density of endothelial cells (6 X l0 s cells/cm2), Schneider et al. 2' noted a decrease in platelet deposition during the first hour after seeding and also demonstrated the spreading out of endothelial cells at this time. This shows that seeded cells can regain their functional properties very soon after seeding. In comparison, our seeding densities were lower. Another difference was the seeding base, preclot-
Vol. 10, No. 6 1996
ted Dacron, which differed from endarterectomized artery. Enzymatic methods for harvesting endothelial cells could induce alterations in endothelial cell function. 6 Enzymatic harvesting from different species has been shown to be fairly efficient but in cell harvesting from h u m a n veins this efficiency is suboptimal. We have achieved the best harvesting efficiency with h u m a n saphenous vein using Dispase II. Various collagenases have been less effective in our hands. Another potential drawback of collagenase-harvested endothelial cells is the possibility of functional alterations, which have been reported. 22 It may be desirable to evaluate the grafts from a morphologic point of view, but after 4 hours under reduced flow, the Dacron surface is covered by a thin film of thrombus. The seeded endothelial cells are hidden underneath. It is extremely difficult to evaluate the surface by means of electron microscopy, given these conditions. We chose preclotted Dacron grafts because of their higher cell attachment rate, about 50%, which is approximately five times higher than what we achieved with expanded polytetrafluoroethylene (PTFE) grafts. 9 After 2 hours of flow, about 50% of seeded cells remain on the graft, but most of the cell loss takes place during the first 20 minutes) ° We also know that the same seeding technology produces measurable improved endothelialization after 3 weeks. ~9 The time allowed for endothelial cells to adhere and spread out was limited in this experimental series to reflect the clinical situation. There are several possible explanations for the lack of measurable effects. The most likely one is that the endothelial cells are in a proliferative state and display fewer nonthrombogenetic properties. T M This raises the question of whether immediate seeding can be of benefit in the clinical situation. Will a nonfunctioning endothelium on a graft lining induce the same reparative process as on a nonseeded graft? No benefit is derived from a nonfunctioning endothelium unless the reparative process is altered. Supraconfluent density seeding 2~ might shorten the time during which cell function is interrupted. Endothelial cells have a distinct contact inhibition and as a result functional variations m a y occur. The incidence of contamination during platelet harvest was low, but during leukocyte harvest the proportion of mononuclear cells was high and platelet contamination was high. This caused some uncertainty regarding the uptake of leukocytes onto grafts. Although there is a possibility
Endothelial cell seeding of Dacron grafts
535
that determinations of leukocyte uptake were inaccurate, the uptake curves were similar w h e n only experiments with low platelet contamination were analyzed. Platelet attachment is used as a quantifiable marker of thrombus formation. Leukocytes have not been used, partly because of methodologic difficulties. Their behavior in relation to platelets, synthetic grafts, and endothelium does warrant some attention. Leukocyte-depleted dogs have a higher rate of endothelial cell retention than control dogs, which suggests that leukocytes cause increased endothelial sloughing. 23 In our study there was no increased attachment of leukocytes to the seeded grafts. Whether the platelet uptake in this experimental model corresponds to conditions in humans is unknown. Without a doubt there are differences in the behavior of platelets between species and sheep platelets do react differently to epinephrine in comparison to h u m a n platelets; adhesion is lower and thromboxane production is almost absent in sheep. On the other hand, this difference is more pronounced in other species such as rabbits and goats where platelet reactivity is lower than it is in humans and sheep. Platelet reactivity is higher in monkeys than in humans. Thus there is no ideal experimental animal. 24 In this series a substantial increase in fibrinogen uptake was noted. This increased uptake is greater than what we have found w h e n blood flow is reduced to mimic a poor runoff state in PTFE grafts. 25 In sheep such a large increase in the uptake of fibrinogen can be the result of reduced fibrinolytic activity. It can also depend on differences in the type of graft material. The reason for the difference between platelet and leukocyte uptake on the one hand and fibrinogen uptake on the other is unclear.
CONCLUSION The multichannel analyzer setup, which uses several radionuclides in parallel, has led to the development of the present-day experimental setring for graft thrombogenicity studies. A conventional gamma camera measures the count rate from one and occasionally two radionuclides, whereas the multichannel analyzer can analyze the gamma count rate from up to three radionuclides simultaneously with good accuracy for prolonged periods of time. The gamma camera uses one detector but the multichannel analyzer uses four. Because each detector obtains a complete energy spectrum and each detector covers one
536
Jensen et al.
region of interest, two grafts and native vessels can be examined simultaneously. Our experimental investigation, however, was unable to demonstrate any reduction in uptake of platelets, leukocytes, or fibrinogen in grafts immediately seeded with endothelial ceils compared to unseeded control grafts during the first 4 hours after graft placement. Immediate seeding does not initially produce any reduction in graft thrombogenicity. REFERENCES 1. Stanley JC, Burkel WE, Graham LM, et aL Endothelial cell seeding of synthetic vasollar prosthesis. Acta Chir Scand 1985;529( Suppl): 17-28. 2. Jensen N, Lindblad B, Bergqvist D. Endothelial cell seeded Dacron aortobifurcated grafts: Platelet deposition and longterm follow-up. J Cardiovasc Surg 1994;35:425-429. 3. Ortenwall E Wadenvik H, Risberg B. Reduced platelet deposition on seeded versus unseeded segments of expanded polytetrafluoroethylene grafts: Clinical observations after a 6-month follow-up. J Vasc Surg 1989;10:374-380. 4. Zilla E Deutsch M, Meinhart J, et al. Clinical in vitro endothelialization of femoropopliteal bypass grafts: An actuarial follow-up over three years. J Vasc Surg 1994;19:540-548. 5. Magometschnigg H, Kadletz M, Vodrazka M, et al. Changes following in vitro endothelial cell lining of ePTFE prostheses: Late morphologic evaluation of six failed grafts. Eur J Vasc Surg 1994;8:502-507. 6. Ford JW, Burkel WE. Isolation of adult canine venous endothelium for tissue culture. In Vitro Cell Dev Biol Anim 1981;17:44-50. 7. Engfeldt E A m e r E Ostman J. Nature of the inhibitory effect of collagenase on phosphodlesterase activity. J Lipid Res 1985;26:977-981. 8. Cesarone CE Fugasa E, Gallo G, et al. Collagenase perfusion of rat liver induces DNA damage and repair in hepatocytes. Mutation Res 1984;141:113-116. 9. Jensen N, Lindblad B, Bergqvist D, In vitro attachment of endothelial cells to different graft materials, Eur Surg Res 1996;28:49-54. 10. Jensen N, Lindblad B, Leide S, et ak Loss of seeded endothelial cells in vivo. A study of Dacron grafts under different flow conditions. Eur J Vasc Surg 1994;8:690-693. 11. DiCorletto PE. Cultured endothelial cells produce multiple growth factors for connective tissue cells. Exp Cell Res 1984;153:167-172.
Annals of Vascular Surgery
12. Barrett TB, Gajdusek CM, Schwartz SM, et al. Expression of the sis gene by endothelial cells in culture and in vivo. Proc Natl Acad Sci USA 1984;81:6772-6774. 13. Collins T, Ginsburg D, Boss JM, et al. Cultured human endothelial cells express platelet derived growth factor B chain: cDNA cloning and structural analysis. Nature 1985; 316:748-750. 14. Clowes AW, Gown AM, Hanson SR, et aL Mechanisms of arterial graft failure: 1. Role of cellular proliferation in early healing of PTFE prostheses. Am J Pathol 1985;118:43-54. 15. Thakur ML, Walsh L, Malech HL, et al. Indium-ll 1--labeled human platelets: Improved method, efficacy and evaluation. J Nucl Med 1981;22:381-385. 16. Blomb/ick B, Blomb~ick M. Purification of human and bovine fibrinogen. Arkiv Kemi 1956;10:415-443. 17. McFarlane AS. Efficient trace labeling of proteins with iodine. Nature 1958;182:53. 18. Jensen N, Brunkwall J, F~ilt K, et al. Prostacyclins are produced from endothelial cell-seeded grafts: An experimental study in sheep. Eur J Vasc Surg 1992;6:499-504. 19. Jensen N, Brunkwall J, F~ilt K, et al. Recovery of endothelial cells and prostanoid production in endothelial cell seeded grafts. Eur J Vasc Endovasc Surg (in press). 20. Burkel WE, Ford JW, Vinter DW, et al. Fate of knitted Dacron velour vascular grafts seeded with enzymatically derived autologous canine endothelium. Trans Am Soc Artif Intern Organs 1982;28:178-182. 21. Schneider PA, Hanson SR, Price TM, et al. Confluent durable endothellalization of endarterectomized baboon aorta by early attachment of cultured endothelial cells. J Vasc Surg t990; 11:365-372. 22. Sharefkii1 JB, Fairchild KD, Albus RA, et al. The cytotoxic effect of surgical glove powder particles on adult human vascular endothelial cell cultures: Implications for clinical uses of tissue culture techniques. J Surg Res 1986;41:463472. 23. Emerick S, Herring M, Amold M, et al. Leukocyte depletion enhances cultured endothelial retention on vascular prostheses. J Vasc Surg 1987;5:342-347. 24. Didisheim P. Comparative hematology in the human, call sheep, and goat: Relevance to implantable blood pump evaluation. ASAIO J 1985;8:123-127. 25. Lundell A, Bergqvist D, Lindblad B. The uptake of platelets, fibrinogen and leukocytes in eFrFE vascular grafts in relation to blood f l o w - A n experimental study in the sheep. Eur J Vasc Surg 1993;7:698-703.