JOURNAL

OF SURGICAL

RESEARCH

51, 33-39 (1991)

Comparison of Immediate Seeding of Endothelial Cells with Culture Lining of Small Diameter ePTFE Carotid Interposition Graftsip SAJJAD M. HUSSAIN, M.D., GRAHAM W. LONG, M.D., RANDALL S. JULEFF, M.D., MARGARET MCKAIN, JOHN L. GLOVER, M.D., PHILLIP J. BENDICK, PH.D.,~ AND LAURACE E. TOWNSEND, PH.D. Department

of Surgery,

William Beaumont

Hospital,

Royal Oak, Michigan

B.S.N.,

48073

Submitted for publication November 20, 1990

oped because it was assumed that an endothelial lining in small caliber grafts would improve patency and allow replacement or bypass of small arteries. In all but one of the experimental studies which support this hypothesis [l-3], antiplatelet medications started prior to surgery were necessary to achieve these results, which is not surprising because it takes days at least, and in some instances weeks, for these linings to develop. Consequently, it seems logical to try to produce completely confluent linings on the grafts prior to implantation using tissue culture methods. Campbell showed that such a technique might obviate the need for antiplatelet agents [4], but no reports in the literature compare the rates of patency and the healing characteristics of these two approaches. This study was undertaken to directly compare the best available techniques of immediate endothelial seeding versus the preoperative development of a confluent lining in tissue culture in small diameter ePTFE grafts placed opposite one another in the carotid position in the same animal.

This study is the first to compare chronic healing characteristics of immediately seeded grafts with those of grafts lined by autogenous venous endothelial cells in tissue culture prior to implantation. Ten mongrel dogs had a segment of external jugular vein excised for enzymatic harvest of endothelial cells. After approximately 2 1 days growth in tissue culture, 4 X 10” cells/ml were inoculated into a 6-cm length of 4 mm i.d. ePTFE for formation of a confluent lining in culture media. The remaining external jugular vein had its endothelial cells enzymatically harvested for immediate seeding of an identical length of preclotted ePTFE. Both grafts were implanted end-to-end in the carotid position and excised after 30 days. In 6 of the 10 dogs, grafts were patent bilaterally; all others were occluded. Planimetric measurements on patent grafts with immediate seeding showed a thrombus-free surface area of 56 + 39% compared to 86 k 15% for culture-lined grafts (P = 0.046). Endothelial coverage was 70 + 24% for immediately seeded grafts and 29 + 21% for culturelined grafts (P = 0.016). We conclude that immediate seeding and culture lining of autogenous endothelial cells in small diameter ePTFE grafts produce equivalent short-term patency. While culture-lined grafts have an initially less thrombogenic luminal surface, subsequent development of a confluent endothelial lining is slower than that with an immediate seeding preparation, and thus would appear to offer no significant clinical benefit, especially in light of the complexity culture lining adds to the procedure. o 1991 Academic Press.

METHODS

Ten mongrel dogs weighing from 18 to 29 kg (mean weight 23 + 4 kg) were anesthetized with intravenous sodium pentobarbital, 25 mg/kg. The animals were intubated and ventilated with room air using a constant volume ventilator at a tidal volume of 15 ml/kg and a rate of lo-13 breaths per min. Anesthesia was maintained with supplemental intravenous doses of sodium pentobarbital as needed. Intravenous lactated Ringer’s solution was infused at a rate of approximately 5 ml/kg/hr, with boluses administered as needed to maintain peripheral blood pressure. For all procedures the animals were placed in a supine position and the neck was shaved and painted with an iodophor solution. Animal care complied with the National Institutes of Health Principles of Laboratory Animal Care and the Guide for the Care and Use of Laboratory Animals [5], and the research protocol was reviewed and approved by the institutional animal care committee. At the initial operation, endothelial cells were harvested for growth in tissue culture. One external jugular

Inc.

INTRODUCTION

The technique of seeding grafts with autogenous endothelial cells just prior to implantation has been devel’ This research was supported by the William Beaumont Hospital Research Institute, RI-87-11. ’ Presented at the Annual Meeting of the Association for Academic Surgery, Houston, TX, November 14-17, 1990. 3 To whom correspondence and reprint requests should be addressed at Department of Surgery, William Beaumont Hospital, 3601 West Thirteen Mile Road, Royal Oak, MI 48073. 33

0022-4804/91$1.50 Copyright 0 1991 by Academic Press, Inc. All rights of reproduction in any form reserved.

34

JOURNAL

OF SURGICAL

RESEARCH:

vein was excised through a lateral longitudinal neck incision on the right side in five dogs and on the left side in the other five dogs. Throughout the dissection, the vein was handled minimally, and side branches were carefully ligated and divided at least 2 mm from the jugular vein itself. A 12- to 13-cm length of vein was excised from each animal. The vein was everted on a metal rod and placed in culture medium M-199, a synthetic cell culture medium with Earle’s balanced salt solution and Hepes (Whittaker Bioproducts, Walkersville, MD). Endothelial cells were harvested by submerging the vein in 0.2% collagenase for 20 min and then rinsing it in phosphate-buffered solution. The solution was centrifuged and the resulting pellet of cells was resuspended in a canine cell culture medium of M-199 to which had been added fetal bovine serum (20%), endothelial cell growth factor (l%), and heparin (0.6%). The presence and purity of an endothelial cell population was confirmed by staining for Factor VIII-related antigen by means of the peroxidase-antiperoxidase method [6]. The cells were first cultured on a 35-mm Petri dish to confluence and split twice using trypsin to a T75 flask. After approximately 21 days of growth in culture to confluence, they were removed from the flask, centrifuged, and resuspended in 4 ml of phosphate-buffered solution at a concentration of lo6 cells/ml. Grafts were lined in culture with endothelial cells following the technique of Kesler et al. which was shown to provide rapid confluent coverage, with enhanced strength of attachment [ 71. The inner surface of a 7-cm length of 4-mm i.d. ePTFE (W.L. Gore and Associates, Flagstaff, AZ) with an internodal distance of 22 pm was coated with a 0.1 mg/ml solution of fibronectin (Boehringer-Mannheim, Indianapolis, IN). One milliliter of the endothelial cell suspension was placed within the graft and both ends were clamped. After 30 min the suspension was decanted, the graft was rotated 90”) and a second l-ml aliquot was placed in the graft for 30 min. This was repeated for the third and fourth quadrants of the graft, with the fourth quadrant cells remaining in place for 60 min. This final aliquot was then removed, and the graft was unclamped and placed in canine cell culture media until the time of implantation the next day. When the culture-lined graft was prepared, the dog corresponding to that line of endothelial cells was returned to the operating room and again anesthetized and prepared as described above. At this operation a midline longitudinal neck incision was made and the opposite external jugular vein was exposed. This vein was excised using the same careful handling procedures as during the first operation, and the endothelial cells were enzymatitally harvested in a similar fashion using 0.2% collagenase. However, following harvest, the vein was rinsed with tissue culture medium M-199, the cell suspension was centrifuged as before, and the resulting pellet of cells was resuspended in 4 ml of M-199. One milli-

VOL.

51, NO. 1, JULY

1991

liter of this suspension was used for evaluation of cell count and cell viability. Factor VIII staining was again used to confirm the presence and purity of an endothelial cell population. The remaining 3-ml aliquot, which contained 5 X 104-2 X lo5 viable cells, was transported to the operating room. During the harvest of endothelial cells, a second 7-cm length of 4-mm i.d. ePTFE was prepared for immediate seeding using established techniques [l, 81. The graft was first preclotted with unheparinized blood by forcing it through the interstices of the graft. Excess clot and blood were removed from the graft 20 min later using a balloon embolectomy catheter. The suspended cells were injected into the graft in four equal volumes, again clamping both ends and keeping the graft stationary for 10 min in each position prior to rotating 90” and injecting the next aliquot. When the two grafts were prepared, both carotid arteries were exposed through the midline incision and the dog was systemically anticoagulated using intravenous sodium heparin (100 U/kg). The side of the earlier harvested external jugular vein determined the position of the culture-lined graft. Both grafts were cut to a length of 6 cm; the excess portion of the culture lined graft was evaluated by Factor VIII staining to confirm the purity of the endothelial lining and the complete confluence of the cells. All culture-lined grafts showed a single-cell thickness luminal layer of endothelial cells only with a confluence from 96-100%. Each graft was then placed in an end-to-end fashion using running 7-O polypropylene suture after excising approximately 5 cm of native carotid artery. Throughout the procedure the inner surface of each graft was kept moist by instilling M-199 into the immediate-seeded graft and canine cell culture media into the culture-lined graft periodically through an open end of the graft. The neck incision was closed in layers using running absorbable polyglactin suture. All dogs were given 80 mg aspirin and 25 mg dipyridamole once per day beginning the day prior to graft implantation and continuing each day throughout the course of the study to ensure the highest achievable patency rates. Each dog also received antibiotic prophylaxis using 1 gm cefotetan intravenously at induction and 500 mg intramuscularly each day for 3 subsequent days. All grafts were left in place for 30 days. At that time, each dog was returned to the operating room and anesthetized, ventilated, and anticoagulated as before. Both grafts were exposed by reopening the midline neck incision. Graft patency was determined by first noting the presence or absence of a distal pulse. The distal native vessel was then partially transected and flow characteristics were observed. After obtaining proximal control, an 18-gauge catheter was placed in the native carotid artery proximal to each graft and 2.5% glutaraldehyde solution in cacodylate buffer was infused at 100 mm Hg pressure to perfusion fix both grafts in situ; the infusion was continued until the effluent was

HUSSAIN

ET AL.:

SEEDING

TECHNIQUES

clear. The grafts were then excised, opened longitudinally, and the luminal surface was photographed. Projections of these color transparencies of each graft were made onto quadrille paper for planimetric measurement of surface area free of grossly visible red thrombus or blood cell deposition. The specimens were immersed in 2.5% glutaraldehyde prior to performing histologic studies. All grafts were bisected longitudinally with half processed for light microscopy and half for scanning electron microscopy. The portion selected for light microscopy was ethanol dehydrated, embedded in JB-4 plastic, sectioned in 6-pm thick slices, and stained with toluidine blue. That portion designated for scanning electron microscopy was also ethanol dehydrated, critical point dried, and coated with gold/palladium. Comparison of patency rates between immediateseeded and culture-lined grafts and the presence or absence of a confluent monolayer of endothelial cells in the midportion of patent grafts was made using Fisher’s exact test. Comparison of thrombus-free surface area and the percentage of graft luminal surface covered by endothelial cells was made using Student’s t test. RESULTS In six of 10 animals, grafts were patent bilaterally. The grafts of the other four dogs were occluded bilaterally, for a patency rate of 60% for both immediateseeded and culture-lined grafts. Three separate patterns of healing were found on the luminal surfaces of the patent grafts by gross examination: (1) clear, thrombus free areas; (2) areas with gross thrombus; and (3) areas with adherent fibrin and platelets, with or without a one- or two-cell thickness layer of red blood cells (Figs. 1 and 2). Planimetric measurements showed that patent grafts with immediate seeding had an average thrombus-free luminal surface area of 56 + 39%, compared to 86 + 15% for patent culture-lined grafts (P = 0.046). Scanning electron microscopy (SEM) evaluation of the midportion of patent grafts identified the presence of an endothelial cell monolayer in five of six (83%) of the immediate-seeded grafts and in only one of six (17%) of the culture-lined grafts (P = 0.040). Scanning electron micrographs taken at randomly sampled sites of the luminal surface of patent grafts confirmed the presence of small clusters of endothelial cells and endothelial cell monolayers by their characteristic cobblestoned pattern, nuclear bulge, and well-defined intercellular borders (Fig. 3). Some areas of each type of graft were covered with fibrin, platelets, and a layer of red blood cells, or with gross thrombus, seen clearly with light microscopy, but unsuited for SEM. Several grafts showed scattered endothelial cells by SEM beneath a thin fibrin and platelet layer with scattered red blood cells (Fig. 4). Some of each showed exposed nodes and fibrils of the

FOR SMALL-DIAMETER

GRAFTS

35

ePTFE surface with no surface coverage. Systematic sampling of appropriate sites free of gross thrombus by SEM showed the percentage of luminal surface covered by endothelial cells in patent grafts to be 70 + 24% for the immediate-seeded group and 29 & 21% for the culture-lined grafts (P = 0.016). DISCUSSION The lack of difference in patency rates at 1 month (both 60%) between immediate-seeded and culturelined grafts suggests that these two methods produce equivalent patency in small diameter prosthetic grafts. Other investigators, studying the two techniques separately in the dog model, have found similar overall patency rates. Campbell et al. [4] reported a patency of 58% at 1 month in their study of immediate-seeded 4-mm i.d., 60-mm long ePTFE grafts in the carotid position. Douville et al. [9] studied immediate-seeded 4-mm internal diameter, 70-mm long, highly porous experimental ePTFE grafts in the carotid position, in which mongrel dogs on anti-platelet medications were sacrificed at intervals between 12 and 52 weeks. The patency rates of seeded grafts in that study at 12 weeks was 63%, although the 8 week difference in postoperative experimental interval may make comparison of our results with these questionable. The patency rates of unseeded control grafts in these two studies were 10 and 32%, respectively, demonstrating the benefits of endothelial seeding on overall patency. Shindo et aZ. [3] studied 4 mm X 150 mm Dacron culture-lined grafts in the femoral position in mongrel dogs. At 1 month their overall patency was 67%, which approximates that of our ePTFE culture-lined grafts. However, they reported that three of their occluded grafts were Weaveknit (Meadox Medical) prostheses that were found to be kinked at explanation. Therefore, they reported the results of the other type of Dacron graft, Microvel DoubleVelour (Meadox Medical), separately. In this subgroup of grafts the patency rate was 86%, higher than that found in our study, but not significantly so; small numbers of experimental animals and the difference in graft material may account for this difference. There was a significant difference in thrombus-free surface area between the immediate-seeded (56%) and culture-lined (86%) grafts. Theoretically, the culturelined grafts have a higher number of endothelial cells attached to their lumens, confluent at the time of implantation. On the other hand, the immediate-seeded endothelial cells would be lower in number and most of these would still be in their rolled-up position on the preclotted luminal surface at implantation. Thus preclotting and immediate-seeding technique would expose greater areas of fibronectin to the blood stream and would be expected to attract larger numbers of platelets, leading to a decreased thrombus-free surface area. Campbell et al. [4] reported a 56% thrombus-free surface

36

JOURNAL

OF SURGICAL

FIG. 1. Specimen photograph of an immediate-seeded free area with endothelial healing.

IEARCH: RESEARCH:

VOL.

51, NO. 1, JULY

graft showing areas of gross thrombus

area in their study of immediate-seeded grafts, duplicating our own result. The thrombus-free surface area of immediate-seeded grafts of highly porous ePTFE in the study by Douville et al. [9] was 83%, but the experimental interval was 12 weeks. Shindo et al. [3] did not report thrombus-free surface area in their study of culturelined Dacron grafts. To our knowledge no studies using ePTFE culture-lined grafts in uiuo have been reported. The midportion of each graft was examined by scanning electron microscopy for the presence of an endothelial cell monolayer, since this is the area of the graft unlikely to develop a monolayer by pannus ingrowth from the anastomotic areas. The significant difference between the immediate-seeded and culture-lined grafts,

1991

on the luminal

surface as well as a thrombus-

83% versus 17%, was surprising for two reasons. First, there were theoretically more confluent endothelial cells present in the culture-lined grafts at the time of implantation, which should make the presence of an endothelial cell monolayer in the midgraft more likely than in the immediate-seeded grafts; and second, the culturelined grafts showed a significantly higher thrombus-free surface area than the immediate-seeded grafts. Many factors other than the presence or absence of a cellular lining may affect thrombus-free surface, however, and studies of the attrition of endothelial linings, while common in immediately seeded grafts, have not been reported in culture lined grafts. The significant difference in percentage of graft lu-

FIG. 2. Specimen photograph of a culture-lined graft with a virtual 100% thrombus-free surface area. Scanning electron microscopy, however, showed little endothelialization with a predominantly fibrin layer and microscopic areas of bare ePTFE graft exposed.

HUSSAIN

ET AL.:

FIG. 3. Scanning electron micrograph (original magnification 630X).

SEEDING

TECHNIQUES

from the midportion

FOR SMALL-DIAMETER

of an immediate-seeded

minal surface area covered by endothelial cells in the immediate-seeded and culture-lined grafts, 70% versus 29%, respectively, contradicts the theoretical advantage of a confluent cellular lining at the time of implantation using a culture-lining technique, and no other equivalent studies are available for comparison. Hussain et al. [8] reported a 39% endothelial cell coverage in 6-mm i.d., lo-cm long immediate-seeded ePTFE grafts, substantially lower than that seen in our study. However, the grafts in that report were placed end-to-side in the aorto-iliac position, which exposed them to a different

37

GRAFTS

graft shows a confluent

monolayer

of endothelial

cells

set of hemodynamics than the carotid interposition grafts used in this study. In the study by Shindo et al. [3], the extent of cell coverage was 78%, which was higher than in our culture-lined grafts, but parallels our own results for immediate-seeded grafts. The percentage endothelial coverage of the luminal surfaces correlated well with the presence or absence of an endothelial cell monolayer in the midportion of the grafts. More of the immediate-seeded grafts contained a midgraft endothelial cell monolayer (83% versus 17%), and the monolayers in these grafts covered the graft lu-

38

JOURNAL

OF SURGICAL

RESEARCH:

VOL.

FIG. 4. Scanning electron micrograph from the midportion of a culture-lined layer; small areas of exposed ePTFE graft can also be seen (original magnification

minal surface to a greater extent (70% versus 29%). However, an unexpected finding was that a greater degree of endothelial coverage did not correlate with a greater thrombus-free surface area, as proposed by Campbell et al. [4]. Furthermore, neither the thrombus-free surface area nor the degree of endothelialization seemed predictive of an increased patency rate in this study, since the latter was the same for both groups of grafts. Interval follow-up studies of graft patency using such techniques as angiography or duplex ultrasound, however, were not performed, so the precise time of graft thrombosis cannot be fixed. If these were early failures, following some

51, NO. 1, JULY

graft showing minimal 1000X).

1991

endothelialization

beneath a fibrin

degree of hemodynamically induced cell desquamation but prior to initiation of natural healing or reparative processes, that might explain a lack of correlation with late findings in patent grafts. The only variable between these two groups, other than the one stated in the purpose for this study, was the technique used to line the grafts prior to endothelial cell placement. A fibronectin solution, 0.1 mg/ml, was used for the culture-lined grafts because that method has been shown to produce the best retention of cells [7]. Whole blood was used to preclot the immediate-seeded grafts because that technique is the reported standard

HUSSAIN

ET AL.:

SEEDING

TECHNIQUES

method used in “immediate seeding” of grafts just prior to implantation [ 1, 81. This study was a comparison of the results of these techniques as they are currently used, and the difference in substrates may have accounted for some of the differences between the two groups. For example, fibronectin may have denatured in tissue culture and caused some decrease in adhesion of the cultured cells. In addition, the thrombin present in the whole blood used to preclot immediate-seeded grafts may have stimulated the endothelial cells to produce platelet-activating factor [lo] which, in turn, would have increased platelet deposition, resulting in the lower thrombus-free surface on the immediate-seeded grafts. It is known that several substances will cause endothelial cells to synthesize PAF. Some of these include thrombin, interleukin-1, and tumor necrosis factor. Once bound to the endothelial cell surface membrane, PAF acts as an initiator of localized platelet aggregation and binds polymorphonuclear leukocytes to the endotheha1 cell. It is thought to be an initiator of both the thrombotic and inflammatory processes, localizing them to a particular area of injury [lo]. In the case of the immediate-seeded grafts in our study, the thrombin in the whole blood preclot could certainly activate the seeded endothelial cells to synthesize PAF. Once bound to the endothelial cell membrane, the PAF, along with the fibronectin in the whole blood preclot, would increase local platelet deposition on the graft. This would then lead to a lower thrombus-free surface area on the luminal surface of the immediateseeded grafts. Furthermore, the polymorphonuclear leukocytes that attach to the membrane-bound PAF molecules would initiate a localized inflammatory response that would destroy the seeded endothelial cells, as shown previously in studies by Emerick et al. [ll] and Weinstein et al. [12]. The preceding hypothesis may explain the contradictory results in our study, The greater thrombus-free surface area seen in the culture-lined grafts might have been due to a decreased level of platelet attraction, inflammation, and subsequent thrombus formation. The prelining using a solution of only fibronectin would incite a much lower level of PAF synthesis by the culturelined endothelial cells than in the immediate-seeded grafts with the whole blood prelining. Furthermore, the fibronectin inner lining of the culture-lined grafts would not be deemed “natural enough” by the body to incite the inflammatory and subsequent reparative processes of rebuilding the endothelial lining to the extent that they would had a whole blood inner lining been used. This would result in a relatively high thrombus-free surface area but a smaller extent of endothelial cell cover-

FOR SMALL-DIAMETER

GRAFTS

39

age and less frequent presence of an endothelial cell monolayer present in the midgraft portion of the culture-lined grafts relative to the immediate-seeded grafts as we found in this study. Future studies limiting the prelining solution to either fibronectin or whole blood for both types of seeding processes will help clarify these events. At present, however, this study suggests that present techniques of fully lining a graft with endothelial cells to confluence in tissue culture prior to implantation offers no significant clinical advantages over the faster, simpler single stage technique of immediate seeding. REFERENCES 1.

Herring, M., Gardner, A., and Glover, J. L. A single-staged technique for seeding vascular grafts with autogenous endothelium.

Surgery84:498,1978. 2. Stanley, J. C., Burkel, W. E., Ford, J. W., Vinter, D. W., Kahn,

3.

4.

5. 6.

R. H., Whitehouse, W. M., and Graham, L. M. Enhanced patency of small-diameter, externally supported Dacron iliofemoral grafts seeded with endothelial cells. Surgery 92: 994, 1982. Shindo, S., Takagi, A., and Whittemore, A. D. Improved patency of collagen-impregnated grafts after in vitro autogenous endotheha1 cell seeding. J. Vusc. Surg. 6: 325, 1987. Campbell, J. B., Glover, J. L., and Herring, M. B. The influence of endothelial seeding on patency of small artery grafts. Eur. J. Vast. Surg. 2: 365, 1988. Guide for the Cure and Use of Laboratory Animals, NIH Publication 85-23, Department of Health and Human Services, 1985, Baughman, S., Herring, M. B., Glover, J. L., and Dilley, R. S. The peroxidase-antiperoxidase staining of Factor VIII related antigen on cultured endothelial cells. J. Biomed. Mater. Res. 18:

561,1984. I.

8.

9.

10.

11.

12.

Kesler, K. A., Herring, M. B., Arnold, M. P., Glover, J. L., Park, H. M., Helmus, M. H., and Bendick, P. J. Enhanced strength of endothelial attachment on polyester elastomer and polytetraAuoroethylene graft surfaces with fibronectin substrate. J. Vast. Surg. 3: 58, 1986. Hussain, S. M., Glover, J. L., Augelli, N., Bendick, P. J., Maupin, D., and McKain, M. M. Host response to autologous endothelial seeding. J. Vast. Surg. 9: 656, 1989. Douville, E. C., Kempczinski, R. F., Birinyi, L. K., and Ramalanjaona, G. R. Impact of endothelial cell seeding on long term patency and subendothelial proliferation in a small-caliber highly porous polytetrafluoroethylene graft. J. Vast. Surg. 5: 544,1987. Zimmerman, G. A., McIntyre, T. M., Mehra, M., and Prescott, S. M. Endothelial cell-associated platelet-activating factor: A novel mechanism for signalling intercellular adhesion. J. Cell Biol. 110: 529, 1990. Emerick, S., Herring, M. B., Arnold, M. P., Baughman, S., Reilly, K. A., and Glover, J. L. Leukocyte depletion enhances cultured endothelial retention on vascular prostheses. J. Vast. Surg. 5: 342, 1987. Weinstein, T. S., Herring, M. B., and Nixon, J. C. Improved endotheiial cell retention and graft patency in leukopenic dogs: A study on prostheses endothelialized in tissue culture. In Proceedings, 13th Annual Meeting of the Midwest Vascular Surgical Society, Chicago, Sept. 30, 1989. [Abstract]

Comparison of immediate seeding of endothelial cells with culture lining of small diameter ePTFE carotid interposition grafts.

This study is the first to compare chronic healing characteristics of immediately seeded grafts with those of grafts lined by autogenous venous endoth...
4MB Sizes 0 Downloads 0 Views