- Email: [email protected]
Definitive proof of endothelialization of a dacron arterial prosthesis in a human being
Definitive proof of endothelialization of a Dacron arterial prosthesis in a human being Moses H o n g - D e W u , M D , Q u n Shi, M D , Arlene R. Wechezak, P h D , Alexander W . Clowes, M D , I a n L. G o r d o n , M D , P h D , and Lester R. Sauvage, M D , Seattle, Wash. A 10 mm woven Dacron axillofemoral bypass graft was removed from a 65-year-old patient during redo surgery after an implant period of 26 months, because of a large seroma that surrounded the entire length of the graft. Tissue blocks were taken from representative areas along the entire length of the graft surface and evaluated by light microscopy with hematoxylin and eosin and Masson trichrome staining, scanning electron microscopy, transmission electron microscopy, and immunocytochemical staining. Paraffin-embedded sections were stained with smooth muscle cell oe-actin, which demonstrated smooth muscle cells in the pseudointima, and Ham 56 Stain to identify macrophages. Endothelial factor VIII/yon Willebrand factor and Ulex europaeus agglutinin identified human endothelial cells on the flow surface, in areas far removed from the anastomoses to the native vessels. This is the first definitive proof in a human of endothelialization of a synthetic arterial graft beyond the pannus ingrowth zone. (J VASC SURG 1995;21:862-7.)
Definitive p r o o f o f endothelialization o f a synthetic arterial graft beyond the limited zone ofparmus ingrowth has not been shown in humans. Although our report in 1975 described endothelial-like cells on the flow surface o f a preclotted, knitted Dacron axillofemoral graft implanted for 20 months, sophisticated studies were then unavailable for definite proof o f the nature o f these cells. 1,2 Recently we have had the opporttmity to study a woven Dacron axillofemoral graft, implanted for 26 months. Our studies o f this specimen with definitive immunocytochemical techniques comprise the substance o f this report. CASE R E P O R T
In September 1990, one of the authors (I.L.G.) of the Long Beach Veteran's Administration Hospital implanted a 10 mm Cooley Verisoft woven Dacron graft (Meadox From the Hope Heart Institute, the ProvidenceMedical Center, and the Department of Surgery, University of Washington, Seattle. Reprint requests: Lester R. Sauvage,MD, Medical Director, the Hope Heart Institute, 528 18th Ave., Seattle,WA 98122, Copyright 9 1995 by The Society for Vascular Surgery and InternationalSocietyfor CardiovascularSurgery,North American Chapter. 0741-5214/95/$3.00 + 0 24/4/62968 862
Medicals, Inc., Oakland, N.J.) as a left axillofemoral bypass in a 65-year-old man to relieve severe lower extremity ischemia caused by inflow obstruction, occurring as a result of complications after abdominal aortic aneurysm resection and grafting. Within a few months, a large seroma developed around the entire length of the graft. A cultured aspirate of the fluid was sterile. Because of the size of the seroma, the graft was replaced by a polytetrafluoroethylene prosthesis in November 1992. Material and M e t h o d s Explanation o f specimen. There was no tissue attachment to the prosthesis, which was floating in serorna fluid. The central trtmk (47 cm long) o f the graft was removed, rinsed immediately with saline solution, and fixed in 10% buffered formalin. The specimen was sent to The H o p e Heart Institute as part o f an international study, where it was received 12 days later. E x a m i n a t i o n and evaluation. As soon as the specimen was received, it was removed from the formalin and inspected grossly and under the stereomicroscope. Sets o f tissue samples were taken representing the different gross findings along the entire length o f the graft surface to evaluate at higher magnification (Fig. 1, A; areas 1 to 4). Because
JOURNAL OF VASCULAR SURGERY Volume 21, Number 5
W u et al.
863
.. of gross sp.ecimen. White dots identify areas from which sets of tissue Fig. 1. A, Flow,. . surface samples were taken. B, Clo~eup view of area 2~ ~ith 2 ~ 3turn isled of thin, white tissue.
Fig, 2. HistO!ogic findings from area:3 in Figure 1,A; findings are also representative of area 1. A, Light microscopy: (H&E stain: Original magnification x 120.) B, Scanning electron microscopy. (Original magnification X 500.)
surface endothelialization was observed on the area 4 set of tissue samples during the initial studies, an additional two sets next to the area 4 sections were taken 8 weeks later (areas 4A and 4B;. Each set consisted of four tissue blocks from adjacent areas; they were evaluated by light microscopy with hematoxylin and eosin (H&_E) and Masson trichrome staining, scanning electron microscopy, transmission electron microscopy, and immunocytochemistry
studies. For the latter, paraffin-embedded sections were stained with endothelial factor VIII/yon Willebrand factor (code No. M-616; Dako Corp., Carpinteria, Calif.) and Ulex europaeus agglutinin (peroxidase-antiperoxidase) (Vector Laboratories, Inc., Burlingame, Calif.) for identifying human endothelial cells, H a m 56 (code No. M-632; Dako Corp.) for identifying human macrophages, and smooth muscle a-actin for identifying smooth muscle
864
W u et al.
JOURNALOF VASCULARSURGERY May 1995
Fig. 3. Histologic findings from area 2 in Figure 1, A. A, Light microscopy. ( H & E stain. Original magnification • 120.) B, Light microscopy. (Masson trichrome stain. Original magnification • 480.)C, Light microscopy. (Positive smooth muscle ~-actin stain. Original magnification • 480.)
Fig. 4. Endothelial cells identified on area 4 of Figure 1,A. A, Light microscopy. (H&E stain. Original magnification x 120.) B, Light microscopy. (H&E stain. Original magnification x465.) C, Scanning electron microscopy. (Original magnification • D, Scanning electron microscopy. (Original magnification • 2000.)
JOURNAL OF VASCULAR SURGERY Volume 21, Number 5
W u et al.
865
Fig. 4 , E-H, E , Transmission etectron microscopy, (Original magnification • 3050.) F, Confirmation ofendothelium b y positive :factor VIII/von Willebrand factor staining result. (Originaimagnification • iLG, Co~rmation of endothelium by positive Ulex europaeus agglutinin staining result. (Original magnification • H, Confirmation of macrophages by ~ 56 immunocytochemistry staining~ (Original magnification • 485.) cells (code No. M-85t, Dako Corp.). The detailed techniques of the above-mentioned methods were described in previous publications. 3-6
translucent material. These areas were larger and more numerous in the middle of the distal part of the graft (area 4) (Fig. 1, A).
Results
Microscopic observations of areas 1 and 3 (a compact fibrin c o a g u l u m ~ surface). N o external capsule of
Gross findings. There was no perigraft tissue attachment along the external surface of the graft. It was fully patent, and about 85% of its flow surface was free ofthrombus (Fig. 1,A). In the middle of the proximal sections, there were a few narrow longitudinal strips of thrombus (area 1). Most of the thromborac areas consisted of thin, red thrombus deposited in the valleys o f crimps (area 3). There were also some scattered small clumps of white thrombus on the flow surface. In the middle of the graft (area 2) there was a 2 • 3 mm island of thin, white, opaque tissue that was not attached tightly to the graft fabric and tended to detach from the wall when a tissue sample was taken (Fig. 1, B). The areas that were free ofthrombus were covered mainly by a thin,
connective tissue was attached to the graft. In the valleys of some crimps on the external surface there were accumulated red cells and white cells. The crimp depth in this woven Dacron graft was about 500 to 700 ~m, and this wavy configuration of the flow surface had been smoothed and leveled by compact fibrin coagulum. This coagulum layer contained impacted red cells at the bottom of the crimp valleys and a fibrinous layer associated with varying amounts of red cells, macrophages, and neutrophils at the inner area close to the flow surface. There was no microscopic evidence o f organization of surface clot and no tissue in the graft interstices. H & E stain indicated that the interstices were filled with a pink substance that resembled fibrin. The
JOURNAL OF VASCULAR SURGERY
866 Wu et al.
histologic findings from areas 1 and 3 were similar, except that the surface fibrinous coagulum on area 1 was thicker and contained more blood cells (Fig. 2, A, and B). Microscopic observations on area 2: (presence of smooth muscle cells on the inner surface of the graft). Light microscopic studies with both H & E and Masson trichrome-stained slides taken from area 2 (Fig. 1, A, and B), where a small area of white, opaque tissue coverage was observed grossly, are shown in Fig. 3, A, and B. These studies revealed a limited "pseudointimal" area 400 to 550 Ixm thick, consisting mainly o f collagen with many fusiform cells dispersed throughout. Smooth muscle e~-actin staining demonstrated that the majority were smooth muscle a-actin-positive cells (Fig. 3, C). H & E and Masson trichrome staining demonstrated that the surface lacked endothelium and was composed of collagen and fibrin. Microscopic observations on area 4: (presence of en&thelial cells on the flow surface). H&E-Stained histologic slides taken from area 4, as indicated in Fig. 1, A, showed a continuous single layer of endothelial-like cells on the flow surface (Fig. 4, A and B). Scanning electron microscopic study demonstrated that these ceils had typical endothelial cell morphology with microvilli on the cellular surface (Fig. 4, C and D). With transmission electron microscopic study, although the ultramicrocellular structure of the endothelial cells was not well demonstrated (perhaps caused by the fixative used during initial preservation of the specimen), endothelial cell junctions could be observed and the ceils had endothelial morphology (Fig. 4, E). The cells on the flow surface of tissue samples from area 4 were confirmed as endothelium by immunoperoxidase staining for endothelial factor VIII/von Willebrand factor and Ulex europaeus agglutinin (Fig. 4, F and G). The H&E-stained histologic slides also showed that the graft interstices contained no tissue ingrowth (Fig. 4, A). These endothelial cells were located on a 50 to 150 ~,m thick layer with macrophages and neutrophils (Fig. 4, A and B). The identity of the macrophages was confirmed by staining with Ham 56 (Fig. 4, H). Similar studies done on tissue samples taken from area 4B showed findings similar to those observed on the proximal neighboring section of area 4, with endothelial cells proved by factor VIII/yon Willebrand factor staining on the flow surface. However, on area 4A, which adjoined the distal neighboring section of area 4, I-I&E staining showed occasional scattered endothelial-like cells on the surface but no clear positive results for endothelial factor VIII/von Willebrand factor.
May 1995 DISCUSSION Although this graft had extensive seroma formation with no perigraft tissue attachment or tissue growth into the graft, it functioned well as a bypass for 26 months. The most important finding was the unequivocal demonstration that an area of the flow surface far beyond the zone of pannus ingrowth was covered by endothelium. However, the microscopic examination method we used did not a~low us to quantify the extent of the flow surface that was covered by these cells. In 1975 we reported the presence of endothelial-like cells, on the basis of silver nitrate staining, of 32% of the flow surface of a preclotted, 8 mm diameter knitted Dacron, left axillofemoral graft implanted for 20 months. 1,2 Several additional aspects of that specimen are worthy of emphasis: (1) It was removed in the operating room within minutes of the patient's death and immediately silver stained; (2) the outer capsule of the graft was thin, compliant, and well attached to the wall; and (3) standard microscopy revealed fibrous-appearing tissue throughout the interstices of the graft wall. It was subsequently commented that silver nitrate staining was not specific proof that the surface cells we observed were endothelium. 7 We readily acknowledge the validity of this statement but consider the findings suggestive that the cells were endothelium. 7 The paucity of reports in the literature concerning endothelial cells on clinical graft flow surfaces may be due to the fragility of the endothelium and the loss of cells in postmortem specimens. There is also a lack of references concerning the actual survival time of endothelial cells on a prosthesis after the circulation has stopped, but generally it is thought that any postmortem or postexplant clot on the flow surface will greatly diminish the possibility of finding endothelium. Perhaps the increasing number of redo cases will enable more investigation of this topic, as in this report. There appear to be thre e possible sources for flow surface endothelialization of a porous synthetic graft: (1) transanastomotic pannus ingrowth, (2) full-wall transinterstitial ingrowth of microvessels across the prosthetic wall, and (3) deposition of circulating endothelial or multipotentia!, precursor cells on the flow surface, with subsequent proliferation and migration. 8~2 In this case transanastomotic pannus ingrowth apparently can be excluded, because the endothelialized area was located far beyond the distance over which pannus can extend. There is small likelihood of a contribution from interstitial tissue ingrowth because there was no perigraft tissue attachment and no microscopic evidence of tissue in
JOURNAL OF VASCULARSURGERY Volume 21, Number 5
the interstices. Therefore deposition o f cells f r o m the bloodstream appears to be the m o s t likely source for the endothelial cells f o u n d o n the flow surface o f this graft. Several animal studies have suggested surface endothelialization resulting f r o m fallout sources in the arterial system. 11,~2I n recent separate studies, Shi et al. ~3,14 and Scott et al. ~s p r o v e d fallout endothelialization in b o t h the arterial and venous systems in the d o g ) s-is Experimentally, complete healing o f a p o r o u s synthetic graft wall consists o f fibrous tissue invasion o f the interstices o f the prosthesis f r o m the perigraft tissue response, with the flow surface formed o f a confluent monolaver o f endothelial cells coveting a layer o f s m o o t h muscle cells with an extracellular matrix o f varying degrees o f thickness? ,Is H o w e v e r , m this case ~the single layer o f endothelial cells rested o n a collection o f macrophages and neutrophils (Fig. 4). M o s t investigators believe that these cells are attracted very early to implanted vascular grafts, s Their role in this specimen is U n k n o w n , T h e i m m u n o c v t o c h e m i c a l and histochemical studies also demonstrated the presence o f s m o o t h muscle ceils and collagen as the extracellular matrix in the small, white, opaque area o n the graft surface (Fig. 3). The histologic composition o f this tissue was similar to that observed beneath the endothelial fining o f the healed neointima o f p o r o u s synthetic arterial grafts i n the dog, ~'ls but n o endothelial cells were f o u n d in this area. A l t h o u g h we did n o t identify the collagen types, the presence o f s m o o t h muscle cells in this area suggests that types I, I I I , and I V could be present, because these typesare expressed by s m o o t h muscle ceils. 16'17 T h e origin o f these cells is obscure. This study has demonstrated that neoendothelialization can take place o n a p o r o u s synthetic arterial prosthesis implanted in a h u m a n being. M a n y questions remain concerning the mechanisms involved. Investigation o f individual h u m a n specimens recovered in surgery m i g h t lead to additional understanding o f this mechanism and methods that w o u l d p r o m o t e complete healing and endothelialization o f vascular prostheses implanted in h u m a n beings. We appreciate the assistance of Dorothy Mungin, H T (ASCP), David Criss, Medical Photographer, Mary Ann Sedgwick Harvey, Medical Editor, and Mau-Ann Nelson, Medical Illustrator, for their contributions to this manuscript. REFERENCES 1. Sauvage LR, Berger I(, Beilin LB, Smith JC, Wood SJ, Mansfield PB. Presence of endothelium in an axillary-femoral
W u et al.
2.
3. 4.
5.
6. 7. 8. 9. 10.
li. I2. 13.
14.
15. 16.
17.
867
graft of knitted Dacron with an external velour surface. Aim Surg 1975;182:749-53. Sauvage LR, Berger K, Wood SJ, et al. The USCI-Sauvage filamentous vascular prosthesis: rationale, clinical results, and healing in man. In: Sawyer PN, Kaplitt MJ, eds. Vascular grafts. New York: Appleton-Century-Crofts, 1977:18596. Mathisen SR, Wu H-D, Sauvage LR, Usui Y, Walker MW. An experimental study of eight current arterial prostheses. J VASCSURG i986;4:33-4I. Clowes AW, Gown AM, Hanson SR, Reidy MA. Mechanisms of arterial graft failure, I: role of cellular proliferation in early healing of PTFE prostheses. Am J Pathol I985;118:4354. Gown AM, Vogel AM. Monoclonal antibodies to human intermediate filament proteins, IX: distribution of filament proteins m normal human tissues. Am J Pathol 1984;i14: 30%21. Gown AM, Vogel AM, Gordon D, Lu PL. A smooth muscle-specific monoclonal antibody recognizes smooth muscle actin isozymes, l Cell Biol 1985;100:807-13. Bauman FG, Geun-Eun K, Imparato AM. Letter. Ann Surg i976;184:652-4. Greisler HP. New biologic and synthetic vascular prostheses. Austin: RG Landes, i991:2-19. Poole JCE~ Sabiston DC Jr, Florey HW, Allison PR. Growth of enc[o~ellum in arterial prosthetic grafts and following endarterectomy. Surg Forum 1962;13:225-7. Flore~[ ~ , Greer SJ, Kiser I, Poole JCF, Telander R, Wertherssen NT. The development of the pseudointima fining fabric grafts of the aorta. Br J Exp Pathol 1962;43: 655-60. Stump MM, lordan GL Jr, DeBakey ME, Halpert B. Endothelium grown from circulating blood on isolated intravasCular Dacron hub. Am I Pathol 1963;43:361-7. Mackenzie JR, Hackett M, Topuzlu C, Tibbs DJ. Origin of arterial prosthesis lining from circulating blood cells. Arch Surg 1968;97:879-85. Shi Q, Wu H-D, Hayashida N, Wechezak AR, Clowes AW, Sauvage LR. Neoendothelialization of isolated Dacron grafts from endothelial cells in the circulation: observations in a canine model [Abstract]. I992:95. Shi Q, Wil M H-D, Hayashida N, WechezakAR, Clowes AW. Proof of fallout endothelialization of impervious Dacron grafts ifi the aorta and inferior vena cava of the dog. J VASC SURG 1994;20:546-57. Scott SM. Barth MG, Gaddy LR, Ahl ET. The role of circulating cells in the healing of vascular prostheses. J VASC SuRe 1994;19:585-93. Tan EM, Dodge GR, Sorger T, et al. Modulation of extracellular matrix gene expression by heparin and endothelial cell growth factor in human smooth muscle ceils. Lab Invest 199i;64:474-82. Miller EJ, Furuto DK, Narkates AJ. Quantitation of type I, III, and V collagens in human tissue samples by highperformance liquid chromatography of selected cyanogen bromide peptides. Anal Biochem I99i;196:54-60.
Submitted Aug. 25. 1994; accepted Dec. 21. 1994.
Definitive p r o o f o f endothelialization o f a synthetic arterial graft beyond the limited zone ofparmus ingrowth has not been shown in humans. Although our report in 1975 described endothelial-like cells on the flow surface o f a preclotted, knitted Dacron axillofemoral graft implanted for 20 months, sophisticated studies were then unavailable for definite proof o f the nature o f these cells. 1,2 Recently we have had the opporttmity to study a woven Dacron axillofemoral graft, implanted for 26 months. Our studies o f this specimen with definitive immunocytochemical techniques comprise the substance o f this report. CASE R E P O R T
In September 1990, one of the authors (I.L.G.) of the Long Beach Veteran's Administration Hospital implanted a 10 mm Cooley Verisoft woven Dacron graft (Meadox From the Hope Heart Institute, the ProvidenceMedical Center, and the Department of Surgery, University of Washington, Seattle. Reprint requests: Lester R. Sauvage,MD, Medical Director, the Hope Heart Institute, 528 18th Ave., Seattle,WA 98122, Copyright 9 1995 by The Society for Vascular Surgery and InternationalSocietyfor CardiovascularSurgery,North American Chapter. 0741-5214/95/$3.00 + 0 24/4/62968 862
Medicals, Inc., Oakland, N.J.) as a left axillofemoral bypass in a 65-year-old man to relieve severe lower extremity ischemia caused by inflow obstruction, occurring as a result of complications after abdominal aortic aneurysm resection and grafting. Within a few months, a large seroma developed around the entire length of the graft. A cultured aspirate of the fluid was sterile. Because of the size of the seroma, the graft was replaced by a polytetrafluoroethylene prosthesis in November 1992. Material and M e t h o d s Explanation o f specimen. There was no tissue attachment to the prosthesis, which was floating in serorna fluid. The central trtmk (47 cm long) o f the graft was removed, rinsed immediately with saline solution, and fixed in 10% buffered formalin. The specimen was sent to The H o p e Heart Institute as part o f an international study, where it was received 12 days later. E x a m i n a t i o n and evaluation. As soon as the specimen was received, it was removed from the formalin and inspected grossly and under the stereomicroscope. Sets o f tissue samples were taken representing the different gross findings along the entire length o f the graft surface to evaluate at higher magnification (Fig. 1, A; areas 1 to 4). Because
JOURNAL OF VASCULAR SURGERY Volume 21, Number 5
W u et al.
863
.. of gross sp.ecimen. White dots identify areas from which sets of tissue Fig. 1. A, Flow,. . surface samples were taken. B, Clo~eup view of area 2~ ~ith 2 ~ 3turn isled of thin, white tissue.
Fig, 2. HistO!ogic findings from area:3 in Figure 1,A; findings are also representative of area 1. A, Light microscopy: (H&E stain: Original magnification x 120.) B, Scanning electron microscopy. (Original magnification X 500.)
surface endothelialization was observed on the area 4 set of tissue samples during the initial studies, an additional two sets next to the area 4 sections were taken 8 weeks later (areas 4A and 4B;. Each set consisted of four tissue blocks from adjacent areas; they were evaluated by light microscopy with hematoxylin and eosin (H&_E) and Masson trichrome staining, scanning electron microscopy, transmission electron microscopy, and immunocytochemistry
studies. For the latter, paraffin-embedded sections were stained with endothelial factor VIII/yon Willebrand factor (code No. M-616; Dako Corp., Carpinteria, Calif.) and Ulex europaeus agglutinin (peroxidase-antiperoxidase) (Vector Laboratories, Inc., Burlingame, Calif.) for identifying human endothelial cells, H a m 56 (code No. M-632; Dako Corp.) for identifying human macrophages, and smooth muscle a-actin for identifying smooth muscle
864
W u et al.
JOURNALOF VASCULARSURGERY May 1995
Fig. 3. Histologic findings from area 2 in Figure 1, A. A, Light microscopy. ( H & E stain. Original magnification • 120.) B, Light microscopy. (Masson trichrome stain. Original magnification • 480.)C, Light microscopy. (Positive smooth muscle ~-actin stain. Original magnification • 480.)
Fig. 4. Endothelial cells identified on area 4 of Figure 1,A. A, Light microscopy. (H&E stain. Original magnification x 120.) B, Light microscopy. (H&E stain. Original magnification x465.) C, Scanning electron microscopy. (Original magnification • D, Scanning electron microscopy. (Original magnification • 2000.)
JOURNAL OF VASCULAR SURGERY Volume 21, Number 5
W u et al.
865
Fig. 4 , E-H, E , Transmission etectron microscopy, (Original magnification • 3050.) F, Confirmation ofendothelium b y positive :factor VIII/von Willebrand factor staining result. (Originaimagnification • iLG, Co~rmation of endothelium by positive Ulex europaeus agglutinin staining result. (Original magnification • H, Confirmation of macrophages by ~ 56 immunocytochemistry staining~ (Original magnification • 485.) cells (code No. M-85t, Dako Corp.). The detailed techniques of the above-mentioned methods were described in previous publications. 3-6
translucent material. These areas were larger and more numerous in the middle of the distal part of the graft (area 4) (Fig. 1, A).
Results
Microscopic observations of areas 1 and 3 (a compact fibrin c o a g u l u m ~ surface). N o external capsule of
Gross findings. There was no perigraft tissue attachment along the external surface of the graft. It was fully patent, and about 85% of its flow surface was free ofthrombus (Fig. 1,A). In the middle of the proximal sections, there were a few narrow longitudinal strips of thrombus (area 1). Most of the thromborac areas consisted of thin, red thrombus deposited in the valleys o f crimps (area 3). There were also some scattered small clumps of white thrombus on the flow surface. In the middle of the graft (area 2) there was a 2 • 3 mm island of thin, white, opaque tissue that was not attached tightly to the graft fabric and tended to detach from the wall when a tissue sample was taken (Fig. 1, B). The areas that were free ofthrombus were covered mainly by a thin,
connective tissue was attached to the graft. In the valleys of some crimps on the external surface there were accumulated red cells and white cells. The crimp depth in this woven Dacron graft was about 500 to 700 ~m, and this wavy configuration of the flow surface had been smoothed and leveled by compact fibrin coagulum. This coagulum layer contained impacted red cells at the bottom of the crimp valleys and a fibrinous layer associated with varying amounts of red cells, macrophages, and neutrophils at the inner area close to the flow surface. There was no microscopic evidence o f organization of surface clot and no tissue in the graft interstices. H & E stain indicated that the interstices were filled with a pink substance that resembled fibrin. The
JOURNAL OF VASCULAR SURGERY
866 Wu et al.
histologic findings from areas 1 and 3 were similar, except that the surface fibrinous coagulum on area 1 was thicker and contained more blood cells (Fig. 2, A, and B). Microscopic observations on area 2: (presence of smooth muscle cells on the inner surface of the graft). Light microscopic studies with both H & E and Masson trichrome-stained slides taken from area 2 (Fig. 1, A, and B), where a small area of white, opaque tissue coverage was observed grossly, are shown in Fig. 3, A, and B. These studies revealed a limited "pseudointimal" area 400 to 550 Ixm thick, consisting mainly o f collagen with many fusiform cells dispersed throughout. Smooth muscle e~-actin staining demonstrated that the majority were smooth muscle a-actin-positive cells (Fig. 3, C). H & E and Masson trichrome staining demonstrated that the surface lacked endothelium and was composed of collagen and fibrin. Microscopic observations on area 4: (presence of en&thelial cells on the flow surface). H&E-Stained histologic slides taken from area 4, as indicated in Fig. 1, A, showed a continuous single layer of endothelial-like cells on the flow surface (Fig. 4, A and B). Scanning electron microscopic study demonstrated that these ceils had typical endothelial cell morphology with microvilli on the cellular surface (Fig. 4, C and D). With transmission electron microscopic study, although the ultramicrocellular structure of the endothelial cells was not well demonstrated (perhaps caused by the fixative used during initial preservation of the specimen), endothelial cell junctions could be observed and the ceils had endothelial morphology (Fig. 4, E). The cells on the flow surface of tissue samples from area 4 were confirmed as endothelium by immunoperoxidase staining for endothelial factor VIII/von Willebrand factor and Ulex europaeus agglutinin (Fig. 4, F and G). The H&E-stained histologic slides also showed that the graft interstices contained no tissue ingrowth (Fig. 4, A). These endothelial cells were located on a 50 to 150 ~,m thick layer with macrophages and neutrophils (Fig. 4, A and B). The identity of the macrophages was confirmed by staining with Ham 56 (Fig. 4, H). Similar studies done on tissue samples taken from area 4B showed findings similar to those observed on the proximal neighboring section of area 4, with endothelial cells proved by factor VIII/yon Willebrand factor staining on the flow surface. However, on area 4A, which adjoined the distal neighboring section of area 4, I-I&E staining showed occasional scattered endothelial-like cells on the surface but no clear positive results for endothelial factor VIII/von Willebrand factor.
May 1995 DISCUSSION Although this graft had extensive seroma formation with no perigraft tissue attachment or tissue growth into the graft, it functioned well as a bypass for 26 months. The most important finding was the unequivocal demonstration that an area of the flow surface far beyond the zone of pannus ingrowth was covered by endothelium. However, the microscopic examination method we used did not a~low us to quantify the extent of the flow surface that was covered by these cells. In 1975 we reported the presence of endothelial-like cells, on the basis of silver nitrate staining, of 32% of the flow surface of a preclotted, 8 mm diameter knitted Dacron, left axillofemoral graft implanted for 20 months. 1,2 Several additional aspects of that specimen are worthy of emphasis: (1) It was removed in the operating room within minutes of the patient's death and immediately silver stained; (2) the outer capsule of the graft was thin, compliant, and well attached to the wall; and (3) standard microscopy revealed fibrous-appearing tissue throughout the interstices of the graft wall. It was subsequently commented that silver nitrate staining was not specific proof that the surface cells we observed were endothelium. 7 We readily acknowledge the validity of this statement but consider the findings suggestive that the cells were endothelium. 7 The paucity of reports in the literature concerning endothelial cells on clinical graft flow surfaces may be due to the fragility of the endothelium and the loss of cells in postmortem specimens. There is also a lack of references concerning the actual survival time of endothelial cells on a prosthesis after the circulation has stopped, but generally it is thought that any postmortem or postexplant clot on the flow surface will greatly diminish the possibility of finding endothelium. Perhaps the increasing number of redo cases will enable more investigation of this topic, as in this report. There appear to be thre e possible sources for flow surface endothelialization of a porous synthetic graft: (1) transanastomotic pannus ingrowth, (2) full-wall transinterstitial ingrowth of microvessels across the prosthetic wall, and (3) deposition of circulating endothelial or multipotentia!, precursor cells on the flow surface, with subsequent proliferation and migration. 8~2 In this case transanastomotic pannus ingrowth apparently can be excluded, because the endothelialized area was located far beyond the distance over which pannus can extend. There is small likelihood of a contribution from interstitial tissue ingrowth because there was no perigraft tissue attachment and no microscopic evidence of tissue in
JOURNAL OF VASCULARSURGERY Volume 21, Number 5
the interstices. Therefore deposition o f cells f r o m the bloodstream appears to be the m o s t likely source for the endothelial cells f o u n d o n the flow surface o f this graft. Several animal studies have suggested surface endothelialization resulting f r o m fallout sources in the arterial system. 11,~2I n recent separate studies, Shi et al. ~3,14 and Scott et al. ~s p r o v e d fallout endothelialization in b o t h the arterial and venous systems in the d o g ) s-is Experimentally, complete healing o f a p o r o u s synthetic graft wall consists o f fibrous tissue invasion o f the interstices o f the prosthesis f r o m the perigraft tissue response, with the flow surface formed o f a confluent monolaver o f endothelial cells coveting a layer o f s m o o t h muscle cells with an extracellular matrix o f varying degrees o f thickness? ,Is H o w e v e r , m this case ~the single layer o f endothelial cells rested o n a collection o f macrophages and neutrophils (Fig. 4). M o s t investigators believe that these cells are attracted very early to implanted vascular grafts, s Their role in this specimen is U n k n o w n , T h e i m m u n o c v t o c h e m i c a l and histochemical studies also demonstrated the presence o f s m o o t h muscle ceils and collagen as the extracellular matrix in the small, white, opaque area o n the graft surface (Fig. 3). The histologic composition o f this tissue was similar to that observed beneath the endothelial fining o f the healed neointima o f p o r o u s synthetic arterial grafts i n the dog, ~'ls but n o endothelial cells were f o u n d in this area. A l t h o u g h we did n o t identify the collagen types, the presence o f s m o o t h muscle cells in this area suggests that types I, I I I , and I V could be present, because these typesare expressed by s m o o t h muscle ceils. 16'17 T h e origin o f these cells is obscure. This study has demonstrated that neoendothelialization can take place o n a p o r o u s synthetic arterial prosthesis implanted in a h u m a n being. M a n y questions remain concerning the mechanisms involved. Investigation o f individual h u m a n specimens recovered in surgery m i g h t lead to additional understanding o f this mechanism and methods that w o u l d p r o m o t e complete healing and endothelialization o f vascular prostheses implanted in h u m a n beings. We appreciate the assistance of Dorothy Mungin, H T (ASCP), David Criss, Medical Photographer, Mary Ann Sedgwick Harvey, Medical Editor, and Mau-Ann Nelson, Medical Illustrator, for their contributions to this manuscript. REFERENCES 1. Sauvage LR, Berger I(, Beilin LB, Smith JC, Wood SJ, Mansfield PB. Presence of endothelium in an axillary-femoral
W u et al.
2.
3. 4.
5.
6. 7. 8. 9. 10.
li. I2. 13.
14.
15. 16.
17.
867
graft of knitted Dacron with an external velour surface. Aim Surg 1975;182:749-53. Sauvage LR, Berger K, Wood SJ, et al. The USCI-Sauvage filamentous vascular prosthesis: rationale, clinical results, and healing in man. In: Sawyer PN, Kaplitt MJ, eds. Vascular grafts. New York: Appleton-Century-Crofts, 1977:18596. Mathisen SR, Wu H-D, Sauvage LR, Usui Y, Walker MW. An experimental study of eight current arterial prostheses. J VASCSURG i986;4:33-4I. Clowes AW, Gown AM, Hanson SR, Reidy MA. Mechanisms of arterial graft failure, I: role of cellular proliferation in early healing of PTFE prostheses. Am J Pathol I985;118:4354. Gown AM, Vogel AM. Monoclonal antibodies to human intermediate filament proteins, IX: distribution of filament proteins m normal human tissues. Am J Pathol 1984;i14: 30%21. Gown AM, Vogel AM, Gordon D, Lu PL. A smooth muscle-specific monoclonal antibody recognizes smooth muscle actin isozymes, l Cell Biol 1985;100:807-13. Bauman FG, Geun-Eun K, Imparato AM. Letter. Ann Surg i976;184:652-4. Greisler HP. New biologic and synthetic vascular prostheses. Austin: RG Landes, i991:2-19. Poole JCE~ Sabiston DC Jr, Florey HW, Allison PR. Growth of enc[o~ellum in arterial prosthetic grafts and following endarterectomy. Surg Forum 1962;13:225-7. Flore~[ ~ , Greer SJ, Kiser I, Poole JCF, Telander R, Wertherssen NT. The development of the pseudointima fining fabric grafts of the aorta. Br J Exp Pathol 1962;43: 655-60. Stump MM, lordan GL Jr, DeBakey ME, Halpert B. Endothelium grown from circulating blood on isolated intravasCular Dacron hub. Am I Pathol 1963;43:361-7. Mackenzie JR, Hackett M, Topuzlu C, Tibbs DJ. Origin of arterial prosthesis lining from circulating blood cells. Arch Surg 1968;97:879-85. Shi Q, Wu H-D, Hayashida N, Wechezak AR, Clowes AW, Sauvage LR. Neoendothelialization of isolated Dacron grafts from endothelial cells in the circulation: observations in a canine model [Abstract]. I992:95. Shi Q, Wil M H-D, Hayashida N, WechezakAR, Clowes AW. Proof of fallout endothelialization of impervious Dacron grafts ifi the aorta and inferior vena cava of the dog. J VASC SURG 1994;20:546-57. Scott SM. Barth MG, Gaddy LR, Ahl ET. The role of circulating cells in the healing of vascular prostheses. J VASC SuRe 1994;19:585-93. Tan EM, Dodge GR, Sorger T, et al. Modulation of extracellular matrix gene expression by heparin and endothelial cell growth factor in human smooth muscle ceils. Lab Invest 199i;64:474-82. Miller EJ, Furuto DK, Narkates AJ. Quantitation of type I, III, and V collagens in human tissue samples by highperformance liquid chromatography of selected cyanogen bromide peptides. Anal Biochem I99i;196:54-60.
Submitted Aug. 25. 1994; accepted Dec. 21. 1994.