surgery-biocompatible - Europe PMC

6 downloads 0 Views 1MB Size Report
Mechem, C., Williams, S., Jarrell, B., et al.: Intravital fluorescence microscopy of endothelial cells on vascular grafts. Monolayer evaluation in the operating room.
156

PERSPECTIVES IN VASCULAR

SURGERY-BIOCOMPATIBLE

VASCULAR SURFACES: THE PAST AND FUTURE ROLE OF ENDOTHELIAL CELLS BRUCE E. JARRELL, M.D., STUART K. WILLIAMS, PH.D., JOHN R. HOCH, M.D., AND R. ANTHONY CARABASI, M.D. Department of Surgery Jefferson Medical College Philadelphia, Pennsylvania

U SE of synthetic vascular grafts as conduits to bypass arterial occlusive lesions has increased significantly during the past few decades. Progress has in part been based on our increased ability to diagnose vascular disease at an earlier stage and to interpose grafts prior to irreversible tissue ischemia. Vascular procedures are not always safe for patients, and surgeons have learned over the years to compare the potential benefits of a vascular graft with its possible complications: the potential benefit of salvaging a limb or revascularizsing an ischemic heart must be compared to the risk to a patient's life posed by the procedure itself. In most cases the decision is based on the patient's overall vital organ status as well as the proved efficacy of the bypass graft intended for use. In general, currently available vascular procedures can be divided into three categories based upon the probability of long-term patency (see table). The success of a given vascular reconstruction largely depends upon the anatomic location and type of graft implanted and the flow rate through the graft. ' In general, high flow rate, large diameter prosthetic grafts perform well in most patients regardless of the material used. When a prosthetic graft is placed in a less favorable location where flow rates are lower, the material used for the bypass becomes critical. Autogenous grafts in this setting continue to provide satisfactory long-term patency. Synthetic grafts in low flow locations tend to occlude frequently because of their inherent throm-

bogenicity. Address for reprint requests: Bruce E. Jarrell, M.D., Department of Surgery, Jefferson Medical College, 1025 Walnut Street, Philadelphia, Pennsylvania 19107

Bull. N.Y. Acad. Med.

ROLE OF ENDOTHELIAL CELLS

ROLE OFEDTEILC

L

S

5

157

SUCCESS RATE OF VASCULAR GRAFTS Moderate but unpredictable Low probability of success High success rate with with less than 25 % patency success with either less greater than 90% one year at one year than 50% patency at one patency and continued patency over 10 years year or a continuing failure rate 1. Dacron aortofemoral bypass grafts 2. Carotid endarterectomy with no graft 3. Femoral-popliteal bypass graft a. For polytetrafluorethylene above knee b. For sapenous vein below knee c. For umbilical vein grafts 4. Aortocoronary bypass using saphenous vein or internal mammary artery

1. Femoral-popliteal bypass grafts below the knee or to the tibial arteries using polytetrafluorethylene 2. Axillo-femoral bypass grafts using dacron or polytetrafluorethylene 3. Dialysis vascular access

1. Dacron femoral popliteal bypass grafts below knee 2. Most venous bypass grafts 3. Any prosthetic graft where the diameter is less than 6 mm or the flow rate is low

grafts using polytetrafluorethylene 4. Aortocoronary bypass graft using polytetrafluorethylene

Graft failure is a complex problem related to platelet and coagulation protein accumulation within the lumen of the graft as well as to cellular proliferation occurring at or near anastomotic areas.2 No unifying hypothesis explains these events, but at least one major factor is the inherent thrombogenicity that exists in most clinical prosthetic grafts. Many novel polymers have been developed that appear to possess nonthrombogenic properties in vitro, but no surface has demonstrated long-term patency in low flow rate conditions that equals the patency seen with native vessel grafts. While many differences exist between thrombogenic polymer surfaces and a native blood vessel surface, one obvious and important difference is the presence of a living endothelial cell lining on the native vessel luminal surfaces. Because we can culture human adult endothelial cells, a new understanding of the integral role of endothelium in blood vessel homeostasis has evolved.3 Although these cells form only a single monolayer separating blood from body tissues, they are not passive cells. Metabolically active, they interact with blood and tissues to control many different processes. Certain endothelial cell functions are integral to the maintenance of a smooth antithrombogenic surface. They have anticoagulant properties in their native state that prevent blood coagulation, yet when injured and in certain Vol. 63, No. 2, March 1987

158

158

AND OTHERS B. B. JARRELL JARRELL AND

OTHERS~~~~~~~~~~~~~~~~~~-

pathologic states they change to a procoagulant state, resulting in blood coagulation.4 The endothelial cell is an active participant in the transport of materials between blood and tissue,5 many investigators have demonstrated endothelial cell participation in the immune response,6 they play a role in lipid metabolism, and may be a major participant in the development of

atherosclerosis.7 It is sobering to realize the multitude of endothelial cell functions identified during just the past 10 years and the failure of an intensive research effort over the past 30 years even to closely approximate one cellular function, namely, anticoagulant activity. The multitude of endothelial cell functions raises a question whether a passive polymer surface will ever be developed that will successfully reproduce the endothelial anticoagulant function. It is not enough to prevent platelet and fibrin deposition on a surface as the only important characteristic of a prosthetic surface: the surface must also prevent leukocyte and macrophage attachment, bacterial attachment, complement activation and should maintain those qualities indefinitely. As we learn more about the polymer characteristics necessary for long-term patency, major difficulty persists with respect to the ability effectively to evaluate a surface even if identified. Specifically, there is no good human model for blood surface interactions, and thus potential human applications must be extrapolated from less than ideal animal models. ENDOTHELIALIZATION OF PROSTHETIC SURFACES

An alternative approach to creating an artificial nonthrombogenic surface is to develop a graft amenable to endothelial cell lining. While basic research has defined mechanisms that help to explain the antithrombogenic nature of the endothelium, clinical studies have established the importance of an intact endothelial lining for the maintenance of normal blood vessel function. The high success rate with saphenous vein or internal mammary artery grafts indicates the superiority of a native "living"' vessel. Additional studies suggest that, long-term patency may be related to preservation of the endothelium in the saphenous vein after removal.8 Several investigators have shown that poor surgical technique, pressure distention of the vein, and vasospasm may adversely affect endothelial integrity,9 while a cold preservation medium or anticoagulated blood within the graft while it is awaiting implantation probably affords better endothelial preservation than room temperature saline solution. '0 These pieces of evidence suggest that cellular integrity within the vessel wall is important for subsequent normal vascular wall function. "CelBull. N.Y. Acad. Med.

ROLE OF ENDOTHELIAL CELLS

ROLE

OF

ENDOTHELIAL

159

159

lular integrity" is not necessarily limited to the endothelial cell monolayer. Normal function of smooth muscle cells within the vessel wall may have an important regulatory role in endothelial cell function, and thus preservation of all the cellular and extracellular elements of the vessel wall may be the most important end product of proper vessel handling and preservation. A second observation about the response of a native vein when placed in an arterial position is vessel wall hypertrophy in response to arterial pressure. " This may allow the thin venous wall better to accommodate to the higher mural stresses encountered in an artery of equivalent size. Regardless of the mechanism of hypertrophy, this response implies a complex vessel wall structure/function relationship. The net effect of hypertrophy upon endothelial cell function has not been examined, but is undoubtedly fertile ground for study. Another important aspect of native vessels is their innate ability to repair injuries. Endothelial cells are frequently injured during surgery from local trauma and preservation injury. When injury occurs, the cell may lose its anticoagulant function, resulting in either thrombin activation on the injured cell or cell slough with subsequent exposure of the subendothelial surface. When this occurs, nearby cells can migrate into the region and reendothelialize the surface. This ability to repair cellular defects in saphenous veins used for bypass grafts has been documented in humans and laboratory animals. 2'3 Regeneration of an endothelial monolayer may be more rapid when the in situ method of grafting is used when compared to the standard excised and reversed saphenous vein graft. In the insitu procedure, the vein is handled less, which may contribute to the improved long term patency. 14 When large areas of vessel surface have been denuded, the migratory ability of nearby endothelial cells may be inadequate for reendothelialization of the surface. This occurs in long endarterectomy sites and may explain the discontinuance of endarterectomy procedures in locations of lower flow rate such as the superficial femoral artery. Last, the composition and structural integrity of the collagen lattice which underlies and surrounds the cellular components of the vessel wall has been shown to influence endothelial cell function and morphology.'5 Since collagen synthesis and structure may be regulated by such factors as shear rate and synthetic polymer structure, further work must be done to determine the role of collagen in native graft function. Interactions between collagen and vessel wall cells could be important in regulating numerous vessel wall functions. Vol. 63, No. 2, March 1987

160 160

B. JARRELL AND OTHERS B.

JARRELL

SPONTANEOUS MECHANISMS CAPABLE OF GENERATING AN ENDOTHELIAL MONOLAYER

Although endothelial cell interactions with natural surfaces have been extensively examined, interactions with artificial surfaces have not been fully characterized. One area of investigation is the development of a prosthetic surface which stimulates the spontaneous generation of an endothelial cell monolayer on a prosthetic surface. Multiple in vivo mechanisms exist which would allow a spontaneous endothelial monolayer to form on an implanted vascular surface. All involve migration of nearby endothelial cells onto or into the graft. Migration of endothelial cells onto a surface may occur in two possible ways. The first way, termed pannus ingrowth, has been repeatedly observed in both humans and nonhuman species. 16 With this mechanism, endothelial cells on the surface of the native vessel migrate slowly across the suture line and attach and grow to the prosthetic surface. One important aspect of pannus formation is pannus arrest. Once the pannus advances to a certain distance, the endothelial cells cease to continue migration. Understanding the mechanism of pannus arrest may yield information that could improve this method of graft endothelialization. The rate of pannus ingrowth shows some variability between animal species. When porous vascular grafts are implanted into mammals, pannus formation and ingrowth begins at each anastomotic site. In laboratory animals, this process proceeds at a variable rate. Dogs are able to cover approximately 1 cm of each end of the graft by one year while baboons can cover a larger surface (up to 5 cm) in a shorter period of time. In human studies only about 1 cm is covered and this requires one to two years to become complete. Thus, although this mechanism of graft endothelialization has appealing aspects, its major drawbacks are its incompleteness and the amount of time required. A second method of graft endothelialization is transinterstitial ingrowth of microvessel endothelial cells through the graft and onto the luminal surface. Certain nonhuman species are capable of endothelializing porous dacron grafts in the central portion of the graft well before pannus ingrowth could account for coverage in that area. Sauvage proposed that central graft endothelialization occurred from transinterstitial ingrowth of microvessel endothelial cells. 17 Central graft endothelialization could be prevented in his studies by wrapping the porous graft with an impervious material. Additional studies by Clowes and coworkers revealed that a polytetrafluoroethylene graft of large pore size allowed microvessel ingrowth while grafts composed of Bull. N.Y. Acad. Med.

ROLE OF ENDOTHELIAL CELLS

161

161

smaller pore sizes prevented ingrowth. 18 Dye studies demonstrated continuity between external, intramural, and luminal endothelial cells. These studies are compatible with the concept that microvessel endothelial cells are capable of not only spontaneously colonizing a surface but also performing some functions of large vessel endothelial cells. This includes the formation of a cobblestone monolayer that appears antithrombogenic with the ability to withstand arterial shear stresses. This process has not been observed to occur in grafts implanted into humans. Anderson and associates have examined grafts removed from humans and compared them histologically to grafts implanted into animals. 19 He examined the midportion of the graft, therefore excluding pannus growth. An absence of endothelial cells was observed in the human grafts. There was also a marked species difference in the cellular response of the perigraft tissue to the implanted graft. Absence of endothelial cells in the midportion of grafts implanted into humans appears to be the case in most reported series. This is unique to the human species in that multiple studies have demonstrated spontaneous islands of endothelial cells within the midportion of porous grafts implanted into dogs, pigs, rats, and baboons. 17 This disparity in healing response suggests a species specific difference in microvessel endothelial cell response to prosthetic grafts. As a result, in vitro migration studies across graft materials should be designed using human endothelial cells. The effect of different variables might then be more directly relevant to human implantation studies. A third potential method of graft endothelialization is migration of pluripotential blood-borne cells from a distal site to the graft surface where it attaches and subsequently covers the surface. Several investigators have suggested that fibroblasts, monocytes, or other pluripotential cells are capable of transforming into cells with endothelial-like characteristics.20-23 Both vascular endothelial cells and blood cells originate from hemangioblasts of mesodermal origin. Thus, this mechanism has a scientific basis but more convincing data need to be obtained. Endothelial cell seeding. Another mechanism that can expediate graft endothelialization is the process of "seeding" the graft, which involves the procurement of endothelial cells isolated from either an autologous large vessel or microvessel source and placement of these cells upon the graft surface allowing adherence to occur. Early seeding studies by Herring and coworkers utilized the isolation of endothelial cells from a vein by mechanical scraping.24 These endothelial cells were then mixed with blood and passed through the interstices of the graft. Eight to 12 weeks following implantaVol. 63, No. 2, March 1987

162 162

B. JARRELL AND OTHERS

tion into dogs, grafts demonstrated increased endothelial cell coverage compared to control grafts.25-27 This finding has been repeatedly demonstrated in animal models and even recently reported in one human graft.28 A number of uncertainties persist in canine seeding studies, including the observation that spontaneous endothelialization occurs in canine models even in the absence of seeding procedures. In addition, the origin of endothelial cells ultimately forming the monolayer was originally assumed to arise from the seeded endothelial cells. This assumption has recently been challenged by Hollier and associates, who observed that porcine endothelial cells seeded onto grafts implanted into dogs yielded results similar to autologous seeding,29 suggesting that an endothelial cell-derived factor rather than the cells themselves may be responsible for the more rapid endothelialization. This line of investigation should be continued in that it may yield information useful for stimulating transinterstitial endothelial cell ingrowth. A second method of graft seeding involves placing a large number of endothelial cells directly upon the graft surface rather than mixing the cells within a clot and allowing the mixture to jell upon the graft surface. The principal advantage of this method is that it allows a monolayer to form very rapidly. A mature monolayer may be stimulated to form in less than one hour when a sufficient number of endothelial cells is seeded upon a surface receptive to endothelial cell attachment and spreading.30 An immediate monolayer opens the possibility that a surface could be actively antithrombogenic at the time of blood flow restoration through the graft. Many variables affecting monolayer formation must be examined, including endothelial cell source, seeding parameters, monolayer stability in the presence of shear forces, and factors that influence endothelial cell anticoagulant activity. But, once characterized, immediate monolayers offer prosthetic surface applications in areas where current surgical techniques fail. These include not only small diameter low flow arterial grafts, but also venous grafts, other intravascular prostheses and artificial organs. CRITICAL AREAS OF STUDY Variables that affect the interaction between endothelial cells and polymeric surfaces must be examined and understood before routine graft endothelialization can become a clinical reality. DONOR VARIABLES

Since endothelial cells are derived from individual donors, the possibility exists that the cells may demonstrate donor-to-donor variation. Although Bull. N.Y. Acad. Med.

ROLE OF ENDOTHELIAL CELLS

163

163

there are many possible important donor factors, it is probable that two very significant variables may be the donor age and the presence of systemic disease. The effect of in vivo donor age on endothelial cell morphology was suggested by Repin et al. ,31 who observed increasing endothelial cell size with increasing age on human adult arterial vessels. Increasing bovine and human endothelial cell size in vitro has been observed by Levine and associates to be related to increased endothelial cell age in culture, particularly as cellular senescence begins.32 Diabetic patients may also exhibit altered endothelial cell function. We have noted that the yield of microvessel endothelial cells from diabetic patients is significantly less than from nondiabetic patients. These and many other undefined variables suggest that the success rate for endothelializing a surface may be donor dependent. SURFACE VARIABLES The second major determinant for a successful endothelial cell monolayer is the polymeric surface. A major problem in this area is the well documented observation that dacron or expanded polytetrafluoroethylene as supplied by the manufacturers does not support optimal human endothelial cell adherence or growth. As a result, many investigators have resorted to a "coating" placed upon the surface prior to the introduction of the endothelial cells. These coatings have included collagens, fibronectin, laminin, blood-derived plasma, blood, and glycosaminoglycans. One difficulty with this approach is the potential nonuniformity of the coating. We have observed on many occasions that the polymer surface as received from the manufacturer frequently needs extensive degreasing before a uniform surface as determined by endothelial cell adherence occurs. This undoubtedly affects protein adsorption as well as cell adherence and may also affect protein conformation on the surface. A second difficulty with coatings is the possible metabolism of surface coatings by the cell itself. Grinnell has demonstrated fibroblast degradation of fibronectin adsorbed onto polystyrene.33 This might affect adsorbed as well as covalently-linked proteins and could result in cellular detachment. One potential solution to this problem is identification of polymeric surfaces that do not require exogenously added proteins for rapid cellular attachment and spreading to occur. After completion of the steps necessary to allow a monolayer to form, it would be very desirable to know whether a monolayer truly exists. This quality control step is necessary to avoid implanting a graft that fails to endothelialize and therefore that would be doomed to thrombosis and failure. It is exceedingly difficult to visualize living endothelial cells on dacron and Vol. 63, No. 2, March 1987

164

164

B. JARRELL AND OTHERS B.

polytetrafluorethylene surfaces because of their opacity at visible light wave lengths and autofluorescence in the 400 to 500 nm wavelength range. Staining with conventional nuclear dyes yields cell numbers, but is destructive and gives little information regarding cell spreading or cell-to-cell contact. Therefore, cell visualization must be performed using either cytoplasmic dyes or dyes that exhibit fluorescent emission in the 500 to 600 nm wavelength range.34 This is perhaps optimally utilized on the implanted graft itself rather than a "control" segment to maximize the chance of graft success. Thus, the dye would need to be nontoxic and acceptable for intraarterial use. ENDOTHELIALIZED SURFACE EVALUATION In many respects, evaluation of an endothelialized surface is similar to evaluations performed on a pure polymer graft. These evaluations include measurements of thrombogenicity, surface stability, compliance and downstream effects such as effects on anastomotic cellular hyperplasia. While nonthrombogenicity has been the goal in polymer prosthesis research, endothelialized grafts should be antithrombogenic, implying active surface properties. These properties include clotting protein degradation and release of such factors as plasminogen activator and prostaglandins. Although this represents a complex in vitro and in vivo series of tests, the major advantage for the endothelialized graft is the standard, which is native vessel endothelium, available for comparison. While the final endpoint of "nonthrombogenicity" is difficult to define for pure polymer surfaces, the antithrombogenic features of an endothelialized graft can be evaluated.

Initial requirements for cell monolayers would include normal morphology of the monolayer and its cellular components, including a mature junctional integrity and production of a full complement of membrane proteins. Further requirements for an endothelial cell monolayer on a prosthetic surface are the ability to degrade thrombin, to release plasminogen activator, to produce prostaglandins, and to demonstrate normal endothelial cell immunological surface markers. Since endothelial cells are known to possess the ABH blood group system and the HLA A, B, and D histocompatibility system, these antigens should be present or inducible on the monolayer.35'36 In addition, the graft endothelial cell should not stimulate autogenous lymphocytes in mixed endothelial cell-lymphocyte coculture differently from native vessel endothelial cells. Any immunological attack upon the monolayer might result in monolayer destruction and loss of the antithrombogenic surface. Finally, there should be evidence that there has been no genetic damage to endothelial cells during the isolation procedure. Bull. N.Y. Acad. Med.

ROLE OF ENDOTHELIAL CELLS

ROLE

OF

ENDOTHELIAL

165

165

Another area of concern is the long-term in vivo stability of the endothelial cell monolayer. Grafts seeded with endothelial cells have been observed to develop a subendothelial network of cells and connective tissue over the course of months following graft implantation.37 These layers have ranged up to several hundred microns in thickness. The precise explanation for these layers is unknown but could relate to a cellular response to wall stress, endothelial cell incompatibility with the prosthetic surface, lack of a smooth muscle cell substrate, or a "healing" response. If subendothelial matrix remains static after a brief period of time, then no further concern is necessary. If the matrix is unstable or there appears to be a significantly elevated baseline mitotic index of the overlying monolayer, then smooth muscle cells and/or a collagen-based graft structure may be necessary for a stable monolayer to occur. Another possible problem with the endothelial cell monolayer may be related to the use of microvessel endothelial cells to establish a monolayer in a large diameter vessel. Microvessel endothelial cells are known to have the capacity to form small tubular structures rather than a cobblestone monolayer.37 This tendency to express multiple cellular morphologies may be important if it affects the antithrombogenicity of the graft. A second cause of vascular graft failure is the development of strictures at the anastomotic junction between native vessels and grafts. The precise mechanism for these strictures is unknown, but cellular hyperplasia is a major component histologically. One factor important in its development may be compliance differences between the graft and vessel. These variables may be important regardless of the presence or absence of endothelium. A second factor in anastomotic stenosis may be the activation of platelets and coagulation factors by the thrombogenic prosthetic surface. Addition of an endothelium could greatly alter these factors and change cellular stimuli occurring at the anastomosis. A related matter is the effect of pulsatile contraction and expansion on endothelial cell monolayer function. Previous investigators have noted changes in endothelial cell prostaglandin production when pulsatile perfusion is added to a rigid static experimental system.38 In summary, either spontaneously generated or seeded endothelial lined prosthetic surfaces will emerge in the clinical arena over the next decade. These surfaces should be evaluated using many of the currently available techniques that have been applied to pure polymer surfaces. Endothelialized surfaces offer the first step in replicating normal vessel wall function. Careful quality control measures must be undertaken to assure that the morphology and function of the endothelial cell monolayer are very similar to the native vessels. Vol. 63, No. 2, March 1987

166

B. JARRELL AND OTHERS

REFERENCES 1. Reichle, F.A. and Taylor, R.R.: Comparison- of long-term results of 364 femeropopliteal or femerotibial bypasses for revascularization of severely ischemic lower extremities. Ann. Surg. 182:449, 1975. 2. Sottiurai, V.S., Yao, J.S.T., Flinn, W.R., and Batson, R.C.: Intimal hyperplasia and neointima: An ultrastructional analysis of thrombosed grafts in humans. Surgery 68:610, 1970. 3. Jarrell, B.E., Shapiro, S., Williams, S., et al.: Human adult endothelial cell growth in culture. J. Vasc. Surg. 1:757, 1984. 4. Bevilacqua, M.P., Pober, J.S., Majeau, G.R., et al.: IL-1 induces biosynthesis and cell surface expression of procoagulant activity in human vascular endothelial cells. J. Exp. Med. 160:618-23, 1984. 5. Landis, E.M. and Pappenheimer, R.: Exchange of Substances Through the Capillary Walls. In: Handbook of Physiology, Section 2, Circulation Vol. 2, Hamilton, W.F. and Dow, P., editors. Baltimore, Williams & Wilkins, 1963, pp. 961-1034. 6. Pober, J.S., Gimbrone, M.A., Cotran, R.S., et al.: Ia expression by vascular endothelium is inducible by activated T cells and by human gamma-interferon. J. Exp. Med. 157:1339-59, 1983. 7. Bondjers, G. and Bjorkerud, S.: Cholesterol accumulation and contention regions with defined endothelial integrity in the normal rabbit aorta. Atherosclerosis 11:451-62, 1970. 8. Adcock, O.T., Adcock, G.L.D., Wheeler, J.R., et al.: Optimal techniques for harvesting and preparation of reversed autogenous vein grafts as arterial substitutes: A review. Surgery 96:886, 1984. 9. LoGerfo, F.W., Quist, W.C., Crenshaw, A.M., and Haudenchild, C.C.: An improved technique for preservation of end6thelial morphology in vein grafts. Surgery 90:1015-24, 1981. 10. Stanley, J.C., Sottiurai, V.I., Fry, R.E., and Fry, L.J.: Comparative evaluation of

11.

12.

13.

14.

15.

16.

17.

18.

19.

vascular graft preparation media: Electro and light microscopic studies. J. Surg. Res. 18:235-46, 1975. Szilagyi, D.E., Elliot, J.P., Hageman, J.H., et al.: Biologic fate of autogenerous vein implants or arterial substitutes: Clinical, angiographic and histiopathic observations in femero-popliteal operations for atherosclerosis. Ann. Surg. 178:232-46, 1973. Fonkalsrudew, S. and Zerubavel, R.: Morphological evaluation of canine autologous vein grafts in the arterial circulation. Surgery 84:253-64, 1978. Buchbinder, D., Sing, J.K., Karmody, A.M., et al.: Comparison of patency rate and structural changes of in situ and reversed vein arterial bypass. J. Surg. Res. 30:213-22, 1981. Leather, R.P. and Karmody, A.M.: The In situ saphenous vein for arterial bypass. In: Biological and Synthetic Vascular Prosthesis, Stanley, J. C., editor. New York, Grune & Stratton, 1982, pp. 351-64. Baker, K.S., Williams, S.K., Jarrell, B.E., et al.: Endothelialization of human collagen surfaces with human adult endothelial cells. Am. J. Surg. 150:197200, 1985. Clowes, A.W., Gown, A.M., Hanson, S.R., and Reidy, M.A.: Mechanisms of arterial graft failure I: Role of cellular proliferation in early healing of PTFE prostheses. Am. J. Path. 118:43-54, 1985. Sauvage, L., Gerger, K., Wood, S., et al.: Interspecies healing of porous arterial prosthesis. Arch. Surg. 109:698705, 1974. Clowes, A.W., Kirkman, T.R., and Reidy, M.A.: Mechanisms of arterial graft healing. Rapid transmural capillary ingrowth provides a source of intimal endothelium and smooth muscle in porous PTFE prostheses. Am. J. Path. 123:22030, 1986. Anderson, J.M., Abbuhl, M.F., Hering, T., and Johnston, K.H.: Immunohistochemical identification of components in the healing response of human Bull. N.Y. Acad. Med.

ROLE OF ENDOTHELIAL CELLS

20.

21.

22.

23.

24.

25.

26.

27.

28.

29.

vascular grafts. J. Am. Soc. Artif. Intern. Organs 8:79-85, 1984. Stump, M.M., Jordan G.L., DeBakey, M.E., and Halpert, B.: Endothelium gron for circulating blood on isolated intravascular dacron HUB. Am. J. Path. 43:361-67, 1963. Mackenzie, J.R., Huckett, M., Topuzlu, C., and Tibbs, D.J.: Origin of arterial prosthesis lining from circulating blood cells. Arch. Surg. 97:879-85, 1968. Brieler, H.S., Thiede, A., and Beck, C.: Monocytogenic endothelialization in dacron grafts. Experimental studies on rats. J. Cardiovasc. Surg. 23:487-89, 1982. Robertson, A.L.: Metabolism and ultrastricture of the arterial wall in atherosclerosis. Cleveland Clin. Quart. 32:97, 1965. Herring, M., Glover, A., and Glover, J.: A single stage technique for seeding vascular grafts with autogenous endothelium. Surgery 84:498-504, 1978. Burkel, W.E., Ford, J.W., Vinter, D.W., et al.: Fate of knitted dacron velour vascular grafts seeded with enzymatically derived autologous canine endothelium. Trans. Am. Soc. Artif. Intern. Organs 28:178-82, 1982. Graham, L., Burkel, W., Ford, W., et al.: Immediate seeding of enzymatically derived autologous canine endothelium in dacron velour vascular grafts. Arch. Surg. 115:1298-94, 1980. Belder, T.A., Schmidt, S.P., Falkow, L.J., and Sharp, W.L.: Endothelial cell seeding of small-diameter vascular grafts. Trans. Am. Soc. Artif. Intern. Organs 28:173-77, 1982. Herring, M., Baughman, J., and Glover, J.: Endothelium develops on seeded human arterial prosthesis: A brief clinical note. J. Vasc. Surg. 2:727-30, 1985. Pennell, R.C., Hollier, L.H., Solis, E.,

Vol. 63, No. 2, March 1987

30.

31.

32.

33.

34.

36.

37.

38.

167

and Kaye, M.P.: Xenograft seeding of dacron grafts in dogs. J. Surg. Res. 40:332-39, 1986. Jarrell, B.E., Williams, S.K., Carabasi, R.A., et al.: Use of an endothelial monolayer on a vascular graft prior to implantation. Temporal dynamics and compatibility with operating room. Ann. Surg. 203:671-78, 1986. Repin, V., Dolgov, V., Zaikina, O., et al.: Heterogeneity of endothelium in human aorta. A quantitative analysis by scanning electron microscopy. Atherosclerosis 50:35-52, 1984. Levine, E.M., Mueller, S.N., Grinspan, J.B., et al.: Endothelial Cell Senescence and the Aetiology of Age-Related Disease. In: Biochemical Interaction at the Endothelium, Cryer, A., editor. Amsterdam, Elsevier Biomedical Press, 1983. Grinnel, F. and Phan, T.: Platelet attachment and spreading on polystyrene surfaces. Dependence on fibronectin and plasma concentration, Thrombosis Res. 39:165-71, 1985. Mechem, C., Williams, S., Jarrell, B., et al.: Intravital fluorescence microscopy of endothelial cells on vascular grafts. Monolayer evaluation in the operating room. J. Surg. Res. In press. Gibofsky, A., Jaffe, E., Fotino, M., and Becker, C.: The identification of ABH antigens on fresh and cultured human endothelial cells. J. Immunol.. 115:73033, 1975. Madri, J.A. and Williams, S.K.: Capillary endothelial cell cultures: phenotyping modulation by matrix components. J. Cell. Biol. 97:153-65, 1983. Frangos, J.A., Eskin, S.G., McIntire, L.V., and Ives, C.L.: Flow effects on prostacyclic production by cultured human endothelial cells. Science 227:147779, 1985.