Polyurethane Arterial Prosthesis" Experimental Evaluation Claude Martin-Mondiere, MD, Philippe David, MD, Daniel Loisance, MD, CrOteil, F r a n c e
Two types of microporous polyurethanes have been evaluated both in vitro and in vivo. In vitro polyurethane disks have been seeded by endothelial cells from bovine aortic origin and from fresh human greater omentum. Various pretreatments permit comparison of six different experimental groups. The benefit of poly L lysine, laminin, fibronectin and previous astrocyte cell seeding is shown. The cell proliferation was assessed daily, using trypan blue in Malassez cells. Cell identification was performed by class I MHC antigen characterization and factor VIII staining. In vivo, 4 mm polyurethane arterial prostheses were implanted as carotid interpositions. Evaluation of polyurethane, both in vitro and in vivo, failed to demonstrate a satisfactory hemocompatibility of the material. However, previous treatment of polyurethane with laminin, fibronectin, and astrocyte cell seeding improves the biologic characteristics of the raw material. (Ann Vasc Surg 1990;4:52-57) KEY WORDS: fibronectin.
Prostheses; polyurethane prostheses; endothelial seeding; laminin;
evaluate two types of polyurethane fabrics as 4 mm 1D arterial substitutes both in vitro and in vivo.
At the present time, and despite extensive experimental work, there is still no satisfactory small arterial prosthesis available for clinical use. Everyone of the 3 to 4 mm internal diameter (ID) conduits which has been evaluated has shown poor results experimentally. The patency rate is unacceptably low, and studies on the healing process have shown a combination of rapid and extensive internal thrombosis, thick subendothelial hyperplasia, and pseudoaneurysmal dilatation. Segmented polyurethane has been proposed and used for various types of vascular, valvular, and ventricular prostheses. Its physical characteristics appear to be optimal for small caliber arterial prostheses [1]. The aim of the present study is to
MATERIALS AND METHODS The polyurethane samples were kindly provided by Matrix* and Medinvent ~. Both of them-presented a microporous structure with a hydrophilic surface. Size of the micropores ranged from 20 to 100 A. Physiological properties were optimal for surgical applications: suturability and creep resistance are better than any other types of synthetic prostheses [2,3]. Studies were performed both in vitro and in vivo. In vitro studies
From the Centre de Recherches Chirurgicales, CNRS, UA 591, Universit( Paris XII, CrOteil, France.
Polyurethane disks, 14 mm diameter and 0.1 mm thickness, were prepared. They were seeded by
Reprint requests: Daniel Y. Loisanee, MD, Centre de Recherches Chirurgicales, Facutt( de M~decine, C H U H. Mondor, 8, rue du G~n(ral Sarrail, 94000 Cr(teil, France.
*White Bridge, Colorado. ~Medinvent, Lausanne, Switzerland. 52
VOLUME 4
No
1 - 1990
POLYURETHANE ARTERIAL PROSTItESIS
endothelial cells from a bovine aortic origin*, and by human endothelial cells obtained from flesh omental tissue using our own technique combining enzymatic and physical dissociation. Dispase, Grade II*, was used for enzymatical dissociation, pore mesh (25 /~) was used for mechanical filtration**, and Percoll ~gradient was used to obtain more than 95% of phenotypic identical cells. A mixture of Dubelco 4.5 mg glucose, minimum essential medium D-Valine, and fetal calf serum** was used as culture nutriment. Culture 5 was incubated at 37°C with 6% CO2 and inspected daily for 15 days. The cells were passaged continuously as previously described [4] and used between passages three and 10 for the in vitro assay. According to different methods of pretreating the polyurethane disk six different groups were defined: group I was a control; group II underwent incubation for 12 hours in isogenic fibronectin* (20 ~g/mi); group III underwent incubation for three hours in poly L lysine '~'~ (10 /zg/ml); group IV underwent incubation for 12 hours in laminin**; group V underwent incubation in poly L lysine for three hours and subsequently, incubation in laminin overnight; group VI underwent incubation in poly L lysine for three hours and then astrocyte cell seeding (1.I05 cells/unit/cm 2) for 48 hours. Astrocytes were purified from the cerebral cortex of young adult brain as previously described [5]. More than 95% of these cells were labeled by antiserum against glial fibrillary acid protein (GFAP), a specific marker for astrocytes in the central nervous system [6]. All of the GFAP-positive cells had a morphological and antigenic phenotype characteristic type I astrocytes [7]. Four experiments were performed in each group twice using each kind of polyurethane. The cultures were evaluated daily by light microscope in Hoffman interference system***. Cell proliferation was evaluated daily using trypan blue for counting in Malassez cell. The cell phase was characterized by anti 5-bromodeoxyuridine staining, endothelial cell lineage by anti factor VIII *t*, and on human cells class I MHC antigen was revealed by monoclonal antibody W6-32 ~~.
*Kindly provided by S. Eskin. Baylor College of Medicine, Houston, Texas. *Boehringer, Mannheim Inc., Mannheim, West Germany. '~St. Gantois, St, Die, France. **Pharmacia, Sweden. '~BRL, Long Island, New York. ~Sigma, LaVerpilliere, France. ***Nikon, Japan. ***Immunotech, Marseilles, France. ~ W e thank Pr. Dr. Charron for his generous gift.
53
In vivo studies
The polyurethane straight arterial prosthesis, 4 mm ID, 5 cm length, was prepared. Experiments were carried out on young pigs (10-15 kg body weight). Under general anesthesia arterial interposition was performed on the carotid artery (mean ID 3 ram), using microsurgical technique and continuous 7-0 Prolene sutures. Neither heparin, nor antiplatelet, nor antiinflammatory drugs were used during surgery or in the postoperative period. The animals were allowed to survive for a period of two weeks after which they were put to death. The prosthesis and the adjacent tissues were explanted for various studies. Macroscopically, the degree of periprosthetic adherence was evaluated. The patency of the prosthesis was assessed directly after arteriotomy. The prosthesis was embedded in me*hacrylate. The sections were stained by toluidine blue, and hemotoxylin-eosin. Finally, the freshly harvested internal wall of the prosthesis was scrapped and cultures were expanded according to the technique previously described. A growth factor (EGF) was added to the culture. Various experimental groups were performed according to the pretreatment of the polyurethane prosthesis. In each group the interposition on the left carotid was used as a control, with the polyurethane prosthesis left untreated. On the right carotid artery pretreatmen* of the polyurethane prosthesis with 24 hours laminin incubation was evaluated (n = 6 Matrix polyurethane, n = 6 Medinvent polyurethane). RESULTS In vitro studies
No healthy endothelial cell could ever be obtained after 48 hours of culture on Medinvent polyurethane (Fig. 1). No treatment was able to promote any attachment of cells to the polyurethane surface: at two weeks only a few ghosts could be found inside the microscope of its Matrix. However, every culture performed on Matrix polyurethane was successful. The influence of the pretreatment is shown in Figure 2. The most homogeneous and rapid growth was observed in the groups where laminin, fibronectin, and astrocytes were used prior to the cell seeding. The best results in terms of morphology were observed in the astrocyte-seeded polyurethane disks (Fig. 3). Immunofluorescence studies confirmed the endothelial cell lineage. More than 10% of the cells were on the S phase. Every cell was W6-32 positive (Fig. 4). In vivo experiments
Results were quite different in the Matrix versus the Medinvent polyurethane groups. In the left
P O L Y U R E T H A N E ARTERIAL PROSTHESIS
54
ANNALS OF VASCULAR SURGERY
Fig. 1. Nonfunctional endothelial cells, seeded on a Medinvent polyurethane disk.
Fig. 3. Endothelial cells growing on previously seeded polyurethane disk, with astrocyte cells.
carotid artery, where a nonpretreated prosthesis was implanted, early thrombosis was observed in every experiment before the end of the surgery when the polyurethane was of Medinvent origin. However, when the prosthesis was a Matrix, thrombosis occurred between the fifth and tenth postoperative day in all experiments. In both cases the thrombus was found attached to the proximal suture line, floating freely in the prosthesis lumen (Fig. 5). In the right carotid artery, where the laminin pretreated prosthesis was implanted, thrombosis occurred on the fifth day in every experiment but one, both in the Matrix and Medinvent polyurethane (Fig. 6). The aspect of the thrombus was identical to the one found early in the untreated Matrix polyurethane prosthesis. A slight yellow color was observed on the inner face of the graft. No drastic differences could be observed morpho-
logically between the nonpretreated and the pretreated prosthesis, the only difference being the delay before occlusion. The only two prostheses patent at two weeks were pretreated laminin grafts. Both suture lines were perfectly free of any thrombus. The inner surface of the prosthesis was yellow colored and quite clean. The prosthesis was surrounded in every experimental group by an inflammatory reaction. The dense periprosthetic tissue was not firmly adherent to the material itself. Histology revealed the lack of any cellular or material deposit on the inner surface of the graft in the thrombosed specimen. In the specimens found patent at 15 days, the suture line was partly covered by an organized subendothelial hyperplasia. This tissue was in the process of covering the adjacent prosthesis. Rarely, ceils were found trapped in the micropores in the vicinity of the inner surface, and the Matrix was impregnated with a glycoproteinlike substance.
Cells number (xlO~ 1000
.....~" / I~/" //.-" V/"
......7--_L-" / .-....;~-.c~ I
500
/
l
10( ............ g . . - ;. ..... .... . . . . . . .
0 24 hrs
48 hrs
day 3
,. 1
day 5
~> day 8
Time
Fig. 2. Effect of various types of polyurethane pretreatment on endothelial cells multiplication, in vitro. Various experimental groups are as follows: I is control; II fibronectin; III poly L lysine; IV laminin; V poly L lysine and laminin; Vi astrocyte cells seeding.
Fig. 4. Endothelial cells identified by W632 monoclonal antibody.
VOLUME 4 N o 1 - 1990
POLYURETHANE ARTERIAL PROSTHESIS
55
Porc L M I du 9 au 17-12-86
Fig. 5. Macroscopic aspect of thrombosed laminin pretreated polyurethane showing attachment of fresh clot on proximal suture line. Thrombus is floating free into prosthesis lumen.
Results of culturing the deposits obtained by scrapping the inner face of the prosthesis were negative in every experiment but one, the one from the Matrix polyurethane, which was found patent at 15 days. In this case cultures on defined medium with an addition of growth factor led to the development of partly confluent cells. Immunofluorescence studies of this culture after 10 days revealed a factor VIII presence.
DISCUSSION The recent developments in reconstructive and small artery surgery have drastically increased the need for an adequate "off the shelf" small arterial substitute. The optimal characteristics of such a prosthesis have been clearly defined [8] and various options have been evaluated in a great variety of animal models. Prostheses of biological origin have been for years one of the most promising alternatives for arterial replacement. Clinical use of preserved heterograft has demonstrated however, the risk of dilation and aneurysm formation, a consequence of the active proteolytic process, imperfectly blocked by the preservation procedure. Optimizing the collagen crosslinking on the other hand leads to too stiff a structure and an unsatisfactory prosthesis in surgical terms [9]. The hemocompatibility of the biological material, which is quite unsatisfactory after the crosslinking process is another problem which appears quite difficult to solve, even by incorporating covalence links of heparin molecules [10]. The last difficulty observed in the evaluation of the biological prosthesis is the persistence of an immunological stimulant in the host after subcuta-
G
Fig. 6. Macroscopic aspect of polyurethane arterial prosthesis, interposed on carotid arteries. On left (G) an unpretreated polyurethane prosthesis, which is found thrombosed. On right (D) laminin coated polyurethane which remains patent.
neous implantation [11]. These less than favorable and acceptable experimental observations are confirmed by the results obtained clinically. There is a general agreement today on the superiority of any fresh venous autograft over any prosthesis prepared from animal tissues [8]. This still negative experience with conduits of a biological origin has led us to reconsider synthetic prostheses. Extensive chemical studies on polyurethane elastomers have permitted manufacture of various prostheses for a wide range of uses requiring motion and blood contact. Arterial grafts with optimal surgical characteristics such as suturability, compliance, and flexibility have been manufactured [3, 12-14]. Porosity and compliance of the prosthesis, which have been considered to be key factors in the healing of a prosthesis, may be easily adjusted in the manufacturing process. Furthermore, results obtained both in vitro, in hemocompatibility testing [15], and in vivo, with ventricular prostheses have been extremely encouraging. Finally, the prelimi-
56
POLYURETHANE ARTERIAL PROSTHESIS
nary results of experimental evaluation of small caliber polyurethane tubes have been presented as favorable [3, 13, 16, 17]. A more detailed approach, using various animal models [18] and the present study failed to confirm the initially promising observations. Further unpublished data obtained by our group have demonstrated that these negative results are not the consequence of the model itself, but related to the unsatisfactory nature of the product. As a matter of fact, late patency (over three months) of larger diameter polyurethane arterial prostheses could not be obtained, despite implantation in a favorable model, in terms of hemodynamics, such as an infrarenal aortic interposition or an arteriovenous shunt. The cause of late occlusion (over two months) appeared to be related to fresh thrombus deposits resulting in a thick hyperplastic reaction covering the suture line. Similar information on the poor evolution of the prosthesis in vivo, can also be found in experiments by others [2, 12, 18]. Polyurethane prostheses, identical to the one used in the experiment, failed to remain patent after four weeks of implantation, as femoral grafts or as femorofemoral arteriovenous loops. The lack of biological material inside the microporous wall structure of the prosthesis is pointed out in these studies. This has been viewed as a favorable reaction, possibly reducing potential lodging of bacteria. It could, however, also lead to the conclusion that the lack of cellular infiltration of the matrix prevents adequate healing of the structure. Finally, a strong immunochemical reaction appears to take place on the surface of the graft, suggesting that clinical use is probably premature [ 18]. The possibility of improvement of in vitro and in vivo cellular compatibility of the polyurethane material is suggested by the present experimental results. Few data are available in the literature on the beneficial effects of laminin. This glycoprotein, which is part of the extra cellular matrix, has the property of a cellular growth factor [19]. Furthermore, it facilitates cellular adherence, especially in the case of endothelial cells. Finally, it favors cellular differentiation. These properties may be considered useful in facilitating endothelial healing on synthetic structures. The present data confirm this hypothesis. Further studies however are necessary to confirm the present results. Fibronectin might be another cell attachment protein that plays a major role in the interaction of cells with the extracellular matrix [20]. Seeding the raw material with astrocytes appeared to be another alternative. No identical approach can be found in the literature. In vitro data suggest that growing of endothelial cells on polyurethane previously seeded with astrocytes is more rapid and more confluent as compared to untreated material. Beneficial effects could be the result of factors secreted by astrocytes. This hypothesis is reinforced by the recent demonstration of the in-
ANNALS OF" VASCULAR SURGERY
ductions of blood brain barrier properties in endothelial cells by astrocytes [21]. The relationship between astrocytes and vascular endothelial cells is presently under evaluation. If a vascular endothelium growth factor could be obtained from astrocytes, optimization of the polyurethane hemocompatibility could perhaps be demonstrated. The data collected in the present protocol suggest a tess than satisfactory hemocompatibility of the polyurethane surface, nor do they confirm its reduced tendency to stimulate platelet adhesion and aggregation shown by Lyman [12]. However, our data point out the fact that thrombus formation occurs mostly on the suture line and not on the polyurethane surface itself, which suggest that optimization of the polyurethane cylinder's natural artery junction would improve the overall patency rate of the prosthesis. Preliminary studies in our group failed to demonstrate the beneficial effect of various microsurgical techniques such as separate stitches, everting sutures, or use of biodegradable suture material. The final mechanism of this precise localization of the thrombus formation on the suture line remains unclear. Many factors might be involved such as microturbulence, the discrepancy between the respective compliance of the natural and synthetic structure, or the activation of the native endothelial ceils in the immediate vicinity of the suture line [18].
CONCLUSIONS Evaluation both in vitro and in vivo of polyurethane fabrics which presented optimal characteristics for small arterial grafting failed to demonstrate positive results. However, in vitro studies permitted demonstration of the positive effects of laminin, fibronectin, and previous astrocyte cell seeding. These benefits of pretreatment of the row polyurethane were confirmed by in vivo evaluation.
ACKNOWLEDGMENTS Work supported by the Institut H. Beaufour, and the Action concertde CNRS Organes Artificiels.
REFERENCES 1. GILDING DK, REED AM, ASKILL IN, BRIANA S. Mitrathane: a new polyetherurethane urea for critical medical applications. Trans Am Soc Artif Intern Organs 1984; 30:587-577. 2. IVES CL, ZAMORA IL, ESKIN SG, WEILBAECHOR DG, GAO ZR, NOON GP, DeBAKEYME. In vivo investigation of a new elastomeric vascular graft (Mitrathane). Trans Am Soc Artif lntern Organs 1984;30:587-591. 3. GILDINGDK, REEDAM. Significantvariableswhichmust
VOLUME 4 N o 1 - 1990
POL YURETtIANE ARTERIAL PROSTHESIS
be controlled in the design and synthesis of polyurethanes for use in medicine. Life Support Systems 1985;5:1%24. 4. JARRELL BE, WILLIAMS SK, STOKES G, HUBBARD FL, CARABASI A, KOOLPE E, GREENER D, PRATT K, MORITZ MJ, RADOMSKI J, SPEICHER L. Use of freshly isolated capillary endothelial cells for the immediate establishment of a monolayer on a vascular graft at surgery.
13. 14.
Surgery 1986;100:392-399. 5. RAFF MC, ABNEY ER, FOK-SEANG J. Reconstitution of a
t5.
developmental clock in vitro: a critical role for astrocytes in the timing of oligodendrocyte differentiation. Cell 1985;42:61-69. 6. B I G N A M I A , D A H L D. Astrocyte-specific protein and neurogliat differentiation. An immunofluorescence study with antibodies to the glial fibrillary acidic protein. J Comp
16.
57
The effects of chemical structure and surface properties of synthetic polymers on the coagulation of blood. Trans Am Soc Artif lntern Organs 1975;21:4%59. ANNIS D. Polyether-urethane elastomers for small diameter arterial prosthesis. Life Support Systems 1987;5:47-52. GOGOLEWSKI S, GALETTI G. Microporous compliant vascular prosthesis of adjustable rate of degradation. Life Support Systems t984;2(suppl. 1):324-33 t. LYMAN DJ, ALBO D, JACKSON R, KNUSTON K. Development of small diameter vascular prosthesis. Trans Am Soc Artif Intern Organs 1977;23:253-261. GALETT1 G, USSIA R, FARRUGIA F, GIARDINO R, GOGOLEWSKI S. Arterial replacement with a degradable vascular prosthesis. Life Support Systems 1986;4(suppl.
H):74-76.
Neurol 1974;153:27-38. 7. tL~FF MC, ABNEY ER, COHEN J, LINDSAY R, NOBLE M. Two types of astrocytes in cultures of developing rat white matter: differences in morphology, surface gangliosides, and growth characteristics. J Neurosci 1983;3:128%1300. 8. LOISANCE DY. A brief review of the current state of small arterial grafts. Life Support Systems 1984;2:41-44. 9. MOCZAR M, DAVID Ph, LO1SANCE D. Evolution of neointimal hyperplasia on vascular prosthesis from human placenta arteries. Pro Int Syrup Vasc Prostheses NLimegen 1984. In: CRAWFORD N, TAYLOR DM (eds). interaction of cells with natural and foreign surfaces. Plenum Press, New York 1986, pp. 137-141. 10. MOCZAR M, LEANDRI J, DAVID Ph, LOISANCE D. Vascular bioprostheses from porcine carotids. Artif Org Proc ISIAO. Chicago 1985, pp. 1005-1009. 11. MOCZAR M, DAVID Ph, LO1SANCE DY, Bioprosthesis from human umbilical cord vein. 6 Life Support Systems (abstract) 1986; p. 68. 12. LYMAN DJ, KNUTSON K, McNEILL B, SHIBATINI K.
mmm
17. QAYUMI K, JAMICSON WRE, VAN DEN BROCK J, QUENVILLE NF, LINDSAY D, JANUSZ MT, JURALA AJ. Comparison of graft materials for arterial substitute. Life
Support Systems 1986;4(Suppl. 11):91-93. 18. MARTZ H, GUIDOIN R, PAYNTER R, FOREST JC, DOWN S. Microporous hydrophilic polyurethane vascular grafts as substitutes in abdominal aorta of dogs. Life Support
Systems 1986;4(suppl. 11):85-89. 19. KLEIMAN HK, CANNON FB, LAURIE GW, HASSEL IR, AUMAILLEY M, TERRANOVA VP, MARTIN GR, DUBOIS-DALCQ M. Biological activities of laminin. J Cell Biochem 1985;27:317-325. 20. YAMADA KM, AK1YAMA SK, HASEGAWA T, HASEGAWA E, HUMPHRIES, JM, KENNEDY DW, NAGATA K, URUSHIHARA H, OLDEN K, CHEN WT. Recent advances in research on fibronectin and other cell attachment proteins, J Cell Biochem 1985;28:79-97. 21. JANZER RC, RAFF MC. Astrocytes induce blood-brain barrier properties in endothelial cells. Nature 1987;325:253-257.