Mechanisms of arterial graft failure. II. Chronic endothelial and smooth muscle cell proliferation in healing polytetrafluoroethylene prostheses

Mechanisms of arterial graft failure. II. Chronic endothelial and smooth muscle cell proliferation in healing polytetrafluoroethylene prostheses Alexander W. Clowes, M.D., Thomas R. Kirkman, B.A., and Monika M. Clowes, B.A., Seattle, Wash. In a previous study of arterial bypass grafts (4 mm polytetrafluoroethylene [PTFE]) in baboons we observed that endothelial and smooth muscle cells (SMCs) formed the neointima and were derived from the cut edges of adjacent artery. The purpose of this study was to determine at late times whether endothelial cells would continue to migrate and to proliferate to cover the graft surface and whether the underlying proliferating SMCs would produce a progressively thickened intima, graft stenosis, and eventual thrombosis. At 6 and 12 months after grafts were placed, endothelial coverage by ingrowth from the anastomoses was more advanced than at 3 months, and by 12 months 60% of grafts (7 to 9 cm in length) were covered. Endothelial cells proliferated in association with the growing edge and focally in other regions. Underlying SMCs proliferated in the region of the growing edge of the endothelial cells and also at anastomoses. Intimal crosssectional area was greatest at anastomoses and at late times was principally due to an increase in connective tissue; actual SMC mass remained constant after 3 months. These results demonstrated slow but progressive healing of the grafts by ingrowth of endothelium. There was also an increased turnover rate of SMCs and endothelial cells in estabfished intima at late times, which might be the consequence of chronic endothelial injury. This condition represents a stable state since it does not produce further infimal thickening and accumulation o f SMCs and does not lead to a high rate o f thrombosis. (J VASC SURG 1986; 3:877-84.)

The management of peripheral occlusive disease in small vessels by vein bypass grafting has yielded satisfactory long-term results. In the absence of suitable vein grafts, synthetic prostheses have been used but unfortunately have not exhibited the same durable results as have vein grafts. In part this can be attributed to the thrombogenic character of such materials; in addition, the characteristics of wound healing in these grafts also may have a deleterious effect on long-term results. In a previous study we examined the cellular aspects of early healing in implanted small-diameter arterial polytetrafluoroethylene (PTFE) prostheses (30 ~m internodal distance, Gore-Tex, W. L. Gore From the Department of Surgery and the RegionalPrimate Research Center, Universityof Washington Schoolof Medicine. Supported by Grant Nos. HL 30946, HL 01108, and RR 00166 from the National Institutes of Health, United States Public Health Service. Reprint requests: Alexander W. Clowes, M.D., Department of Surgery, University of Washington, School of Medicine, RF-25, Seattle,WA 98195.

& Associates, Inc., Flagstaff, Ariz.). 1 We found that the intima of healing grafts was formed by the ingrowth of endothelial and smooth muscle cells (SMCs) derived from the cut edge of adjacent artery. We observed that endothelial cells proliferated not only at the growing edge but also focally in areas in which the endothelial layer was fully established. Like endothelium, SMCs proliferated at the growing edge; in addition, SMCs proliferated in thickened intima overlying anastomoses. These observations contrasted sharply with those obtained in animal models of injury-induced intimal thickening in which both endothelial cells and SMCs ceased proliferating once endothelial coverage had occurred. 2 We concluded that intima overlying synthetic grafts must be subjected to chronic injury and that, in time, this would lead to repeated rounds of endothelial repair and further SMC proliferation, further intimal thickening, stenosis of the graft lumen, and eventual graft thrombosis. To test this hypothesis we have now examined the proliferative behavior of endothelial cells and SMCs at late times (6 and 12 months). Our 877

Journal of VASCULAR SURGERY

878 Clowes,Kirkman, and Clowes

Table I. Endothelial ingrowth

Proximal Distal

6 mo

12too

3.18 --- 0.37 (5) 3.32 -+ 0.30 (5)

2.68 + 0.34 (4) 2.55 -+ 0.37 (4)

NOTE: Data expressed as distance (in centimeters) of white-blue boundary (mean -+ SEM) from toe of adjacent anastomosis in partially endothelialized grafts; number in parenthesis is number of grafts evaluated.

data confirm the resuks of the earlier experiments and demonstrate that, despite progressive endothelial coverage of the grafts, endothelial cells and SMCs proliferate over a long period of time, particularly in regions overlying anastomoses. This proliferation may represent a stable state of chronic injury and repair since intimal thickening and SMC accumulation were not markedly increased in grafts examined at late times. MATERIAL AND METHODS Male baboons (Papio cynocephalus) weighing approximately 10 kg and 2 years of age were used in this study. Animals were anesthetized with halothane, and 4 mm PTFE grafts (Gore-Tex, 30 ~m mean intemodal distance, W. L. Gore & Associates, Inc., Flagstaff, Ariz.) 7 to 9 cm in length were inserted in the common iliac circulation with standard vascular techniques. End-to-side anastomoses were performed with continuous 6-0 polypropylene suture (Davis & Geck, Danbury, Conn.). Each graft was sewn to the distal aorta and to the external iliac artery; the common iliac artery was then ligated and divided without disturbing the internal iliac artery. All animals received heparin during the operation. Animal care complied with the "Principles of Laboratory Animal Care" and the "Guide for the Care and Use of Laboratory Animals" (NIH Publication No. 80-23, revised 1985). Morphology. The morphologic techniques have been described in detail in a previous publication. In brief, all animals received three doses oftritiated thymidine (0.5 mCi/kg/dose) by intramuscular injection in the 24 hours before death at 6 or 12 months after graft placement. One half hour before fixation, heparin (3000 units) and Evans blue (50 mg/kg) l were given by intravenous injection. Animals were then killed by anesthetic overdose and their arteries fixed by perfusion with 2.5% glutaraldehyde in 0.1 M phosphate buffer (pH 7.4) at 100 mm Hg. The grafts and adjacent normal arteries were retrieved, further fixed by immersion in fixative, and then washed extensively in buffer. The distance between the toe of

the proximal or distal anastomoses and the adjacent blue-white interface was measured and recorded as a measure of extent of endothelial ingrowth. One of each pair of grafts was washed with 0.1 M glycine buffer (pH 7.4), opened longitudinally, and pinned out on Teflon sheets for either en face light microscopic autoradiography or scanning electron microscopy. These specimens were dehydrated through graded alcohol solutions and dried by the critical-point method. One half of each specimen was mounted on studs and sputtercoated so that the luminal surface could be examined by scanning electron microscopy. The remaining half of the specimen was processed for autoradiography as described below. From these specimens the surface morphologic characteristics of the grafts, the extent of endothelial ingrowth, endothelial size, and the thymidine labeling indices could be measured. The other aortoiliac graft was subdivided into rings at 0.5 cm intervals beginning proximally in the aorta and ending distally in the external lilac artery. These samples were then embedded in paraffin for cross sectioning (6 ~m) perpendicular to the long axis of the graft. Histologic sections were used for morphometric measurements of intimal area and SMC autoradiography. Intimal areas were obtained from cross sections along the length of each graft by means of a camera lucida and planimetric measurement with a digitizing pad (Hewlett-Packard Company, Waltham, Mass.). The fraction of intima occupied by SMCs was obtained by reembedding the intima from one paraffin block of each graft (at the site of maximal intimal thickening over the distal anastomosis). The paraffin of the blocks was dissolved in xylene; the intima was then separated from each graft, fixed in osmium tetroxide, and embedded in Epon for cross sectioning. Four photographs of each quadrant were obtained from these sections by transmission electron microscopy (× 2000). These negatives were enlarged (× 5000) and the fraction of intima occupied by SMCs determined by the point-hit method. ~ The actual volume of SMCs in the intima for each centimeter of graft was then calculated by multiplying this fraction by intimal volume (cross-sectional area × length). If the volume of a single cell is assumed to remain constant over time, then this calculation provides a direct estimate of the number of intimal SMCs at the distal anastomosis. For this portion of the long-term study of PTFE graft healing, grafts taken from animals at 1 and 3 months, 1 as well as grafts taken at 6 and 12 months, were examined.

Volume 3 Number 6 June 1986

Arterial graft failure 879

Fig. 1. Scanning electron micrographs demonstrate endothelium over graft (A) and adjacent artery (B) at 6 months (bar = 50 ~m). Autoradiography. Unstained deparaffinized histologic cross sections were dipped in emulsion (Kodak NTB-2), allowed to dry, and were kept at 4 ° C in light-tight boxes for 2 weeks. Slides were developed in Kodak D-19 and fixed. The sections were then stained with hematoxylin and evaluated at high power ( x 100) under oil. All cell nuclei that showed five or more overlying silver grains were considered labeled. The thymidine labeling index was obtained by dividing the number of labeled nuclei by the total number of nuclei and multiplying by 100. One cross section was evaluated from each block. These autoradiographic preparations were useful principally for the assessment of cellular proliferation within the intima, the graft, and the surrounding adventitia. For the measurement of proliferation of cells on

the luminal surface, specimens were pinned out flat on Teflon sheets and critical-point dried as described earlier. Dilute emulsion (Kodak NTB-2 emulsion diluted 3 parts to 7 parts water) was dripped onto the luminal surface and allowed to dry. Preparations were developed as described earlier at 2 weeks. These samples were viewed in a light microscope (× 25) and the number of labeled cells per field determined at various locations along the graft. Measurement of the total cell number per unit area was obtained from scanning electron micrographs of specimens not coated with emulsion. Thymidine labeling indices were determined as described earlier on the basis of at least 50 fields (five along the graft by 10 across) from each anatomic region (adjacent artery anastomosis and midzone between anastomoses).

Journal of VASCULAR 880

Clowes, K i r k m a n , a n d Cloves

SURGERY

Fig. 2. Scanning electron micrograph of endothelial growing edge (arrows) at 6 months (bar = 50 ixm).

Table II. Endothelial area Artery Graft

Table III. Thymidine indices--endothelium

6 mo

12 mo

526 _+ 15 697 _+ 65

464 -+ 45 701 + 31

NOTE: Data expressed as cell area (in Ixm2 [mean z SEM]) for endothelial cells overlying graft and adjacent artery. There were five grafts in each group (minimum of six micrographs for each data point). At each time endothelial cells over graft were significantly larger (p < 0.05, paired t test) compared with those over artery.

RESULTS Endothelial coverage. Ten grafts in five animals were available for study at each time point. In each group thrombus had developed in one graft. On gross inspection three of nine grafts (33%) at 6 months and five of nine grafts (56%) at 12 months were entirely white (indicating full endothelial coverage). In the partially covered grafts, the white-blue boundary (endothelial growing edge) was 3.2 -+ 0.2 cm (SEM) from the toe of the anastomosis at 6 months and 2.6 __ 0.2 cm at 12 months (Table I). The grafts exhibiting complete endothelial coverage were 7.1 _+ 0.2 cm (n = 8) in length, whereas those grafts lacking full coverage were 8.4-_-0.3 cm (n = 10); these results suggest that the ability of the endothelial growing edge to progress may slow down substantially with time and that full coverage was obtained in some grafts because they were somewhat shorter than the rest. By scanning microscopy the surface cells exhib-

Artery Graft Proximal anastomosis Distal anastomosis Midzone

6 mo

12 mo

<0.01

<0.01

0.37 + 0.17 0.17 _+ 0.04 0.26 _+ 0.09

0.15 -+ 0.07 0.25 --_ 0.12 0.24 _+ 0.18

NOTE: Thymidine labeling indices expressed in percentages (mean _+ SEM); there were five grafts in each group.

ited the characteristic features of endothelium and formed a continuous monolayer without apparent gaps. Endothelial ceils located over the graft tended to be larger than cells overlying adjacent artery (Fig. 1 and Table II). In grafts lacking full coverage, an endothelial growing edge could readily be defined corresponding closely to the white-blue boundary (Fig. 2). The denuded region contained primarily clusters of platelets and fibrin. The exact location and amount of endothelial proliferation (thymidine labeling index) was defined by autoradiography on specimens taken from along the entire length of four grafts at each time point. Several observations were made. (1) Endothelial proliferation was undetectable (less than 0.01%) over proximal aorta and distal iliac artery in regions more than five high-power fields from the anastomoses. (2) Proliferation was observed focally along the entire graft in regions covered by endothelium (Figs. 3 and 4) (Table III) as well as over anastomoses and growing

Volume 3 Number 6 June 1986

Arterial graft failure 881

Fig. 3. Light micrograph of en face autoradiographic preparation demonstrates two labeled endothelial cells (arrows) (bar -- 100 I~m). edges (in unhealed grafts [n -- 6], the labeling index at the growing edge was 0.21% _+ 0.09%). Formation o f intima. Intimal cross-sectional area was measured along the entire length of four grafts at 6 and at 12 months. Although there was considerable variability, the maximal values were associated with the anastomoses and were not significantly greater than anastomotic intimal areas at 3 months (Fig. 5) (Table IV). 1 There were no significant differences between proximal and distal anastomoses. Transmission electron microscopic examination of intima overlying anastomoses showed accumulations of SMCs and connective tissue as previously described.~ The actual mass of SMCs at 6 and 12 months (obtained by mukiplying the SMC volume fraction by intimal volume) was the same when compared with values at 3 months (Table V). Autoradiography showed proliferation of SMCs underneath the endothelial growing edge in grafts lacking a full endothelial coveting (Table VI). In all grafts SMCs proliferated in intima overlying anastomoses. Proliferation of SMCs in adjacent artery was undetectable (thymidine index, less than 0.01%). The thymidine labeling index of intimal SMCs was the same at 6 and 12 months as at 3 months. 1 DISCUSSION

Failure of synthetic arterial bypass grafts is frequently associated with the development of flowreducing intimal lesions at anastomoses, but the biologic process leading to the formation of anastomotic intimal thickening remains poorly defined. We

Table IV. Intimal area

Proximal anastomosis Midzone Distal anastomosis

6too

12 mo

2.60 -+ 0.43 2.43 + 0.10 3.22 + 0.10

2.98 -+ 1.16 3.17 -+ 0.98 3.72 -+ 0.94

NOTE: Data expressed as intimal area (in m m ~ [mean _+ SEMI); there were four grafts in each group.

have suggested that, in part, intimal thickening may be the result of normal healing processes; in certain persons with risk factors such as cigarette smoking, diabetes mellitus, hypertension, or hypercholesterolemia, the wound healing response may be particularly exuberant. In a previous study we showed that endothelium, SMCs, and fibroblasts all participate in the early healing process of PTFE (Gore-Tex) grafts. ~ Endothelium and SMCs are derived from the cut edges of adjacent artery and form the neointima within the graft; fibroblasts infiltrate the matrix of the graft but do not participate in the development of the intima. At late times the cellular organization within the graft represents a recapitulation of the organization within normal artery; invariably the intima is composed of SMCs covered by a luminal monolayer of endothelial cells. In turn, fibroblasts are found in perigraft tissue and graft matrix external to SMCs. Endothelial cells grow from the anastomoses toward the center of the graft.~ This study demonstrates that at late times this process continues, although at

Journal of VASCULAR SURGERY

882 Clowes,Kirkman, and Clowes

0.90.80.70.6-

~- 0.5- 0,4qa

0.3~o.2017 vvv

Fig. 4. Graph depicts distribution of sites of endothelial proliferation (thymidine indices) along graft. Note that although endothelial cells proliferate all along graft much of the proliferation is associated with the growing edges and the anastomoses.

Table V. Cellular component of distal anastomatic intimal thickening 1~ Intimal area (ram 2) Intimal smooth muscle cell volume fraction (%) Total intimal smooth muscle cell volume (mm3/cm graft)

3~

6~

12~

1.4 -+ 0.5 (3) 30 _+ 7

2.6 + 0.4 (4) 23 + 2

3.2 _+ 0.1 (4) 18 +- 2

3.7 + 0.9 (4) 17 _+ 3

3.8 -+ 1.0

5.8 -+ 0.5

5.8 -+ 0.5

6.5 -+ 1.9

NOTE: Data expressed as maximal intimal cross-sectional area over distal anastomosis (mean - SEM). Number in parentheses is number o f grafts in each group. The 1- and 3-month samples were obtained from animals used in a previous study. ~

a reduced rate. That only the shorter grafts were fully healed suggests to us that we may have selected a graft of critical length. Longer grafts might never develop a complete endothelial layer. A similar pattern of endothelial coverage would be expected in human PTFE grafts but the rate and extent of ingrowth would probably vary; unfortunately no precise, noninvasive way of quantitating endothelial coverage in human grafts is presently available to document this important aspect of vascular wound healing. Since excessive intimal thickening correlates with eventual graft failure and SMCs appear to be the principal cell type found within the neointima, it is of some importance to define at which point along the graft and to what extent over time SMCs prolif-

erate. Furthermore, it is important to determine how SMC proliferation relates to overlying endothelial repair of the denuded graft surface. In this regard, our thinking has been greatly influenced by previous studies of endothelial and SMC proliferation in injured arteries. In healing PTFE grafts, SMCs grow in from the cut edges of adjacent artery to form the new intima; SMCs proliferate at the growing edge and at late times continue to proliferate over anastomoses at levels significantly above background. This anastomotic proliferation eventually leads to an increase in SMC mass and concomitant accumulation of connective tissue. Of note, this region is covered early by a complete layer of endothelium; in other vascular structures undergoing repair (e.g., in injured vessels),

Volume 3 Number 6 June 1986

Arterial graft failure

Fig. 5. Histologic cross-section o f anastomosis at 12 months

intimal SMC proliferation ceases underneath regenerated endothelium.2 Why SMCs continue to proliferate at anastomoses despite an intact overlying endothelial layer cannot, as yet, be explained. We have investigated the possibility that the endothelial layer has small defects undetectable by scanning microscopy to which indium 111-labeled platelets might adhere. We have also examined the possibility that cells on the surface identified by microscopy as endothelium are in fact modified SMCs. However, we have been able to demonstrate that the surface layer is made up of only factor VIII antigen-positive cells and that there are no significant regions of chronic denudation (unpublished results). Since graft endothelium continues to exhibit focally a high thymidine labeling index, endothelial cells must be dying and turning over at an increased rate compared with endothelium over adjacent artery. This must represent a chronic state of ongoing endothelial injury without frank denudation. Although growth factors from platelets may be involved in SMC proliferation after extensive arterial injury,4-6platelets and mitogens contained in platelets are probably not present in the endothelialized intima of vascular grafts at late times and therefore are not likely to be responsible for the late anastomotic SMC proliferation. What other sources of mitogen are available? One possibility not previously considered is that either chronically injured endothelial cells or SMCs themselves generate growth factors that act on neighboring SMCs (paracrine stimulation), z-~° It is now known that both endothelial cells and SMCs can synthesize platelet-derived growth factor (PDGF)-Iike molecules and that RNA for the on-

(bar

883

= 1 ~m).

Table VI. Thymidine indices--smooth muscle cells

Healedgrafts Unhealedgrafts Artery Graft Anastomosis Midzone between anastomosis and growing edge Growing edge

(n =4)

(n =4)

0

0.01 -+ 0.01

0.38 -+ 0.19 0.06 +- 0.01

0.15 _+ 0.12 0.03 -+ 0.01

--

0.40 -+ 0.22

NOTE: Maximum thymidine index expressed as percentage (mean _+ SEM) for each region; differences between midzone and anastomosis or growing edge are significant (p < 0.05) by the Wilcoxon paired-sample test.

cogene c-sis, which codes for a molecule apparently identical to PDGF, 11'~2 can be identified in human endothelial cells in vitro and in atherosclerotic plaques but not in normal artery. ~3,~ It is not known what conditions induce endothelium to synthesize growth factors. It is possible that proliferating or injured endothelial cells exhibit this capacity. On the other hand, the studies of Castellot et al. is suggest that quiescent endothelium may synthesize a heparin-like inhibitor of SMC growth. The factors controlling the balance between endothelial production of growth-promoting and growthinhibiting substances has not been defined. In view of the foregoing considerations, we hypothesize that endothelial cells and perhaps intimal SMCs in synthetic vascular grafts are subjected to abnormal stresses and therefore turn over chronically at a high rate. Because the endothelial cells remain continuously in the growth state and do not revert to qui-

884

Clowes,Kirkman, and Clowes

escence they might be incapable of synthesizing hcparin-like inhibitors of SMC growth and might, in fact, synthesize PDGF-like molecules or other mitogens that would continue to stimulate underlying SMCs to proliferate and perhaps to produce connective tissue. Our data indicate that in most situations this process would be expected to reach an equilibrium with endothelial and SMC proliferation being matched by cell death. On the other hand, in some persons (e.g., cigarette smokers), continued high levels of proliferative activity might eventually lead to marked intimal thickening, luminal narrowing, and graft thrombosis. We thank Charles LaChapelle for his assistance in the preparation of the manuscript.

REFERENCES 1. 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 1985; 118:43-54. 2. Clowes AW, Reidy MA, Clowes MM. Kinetics of cellular proliferation after arterial injury. I. Smooth muscle growth in the absence ofendothelium. Lab Invest 1983; 49:327-33. 3. Clowes AW, Reidy MA, Clowes MM. Mechanisms of stenosis after arterial injury. Lab Invest 1983; 49:208-15. 4. Ross R. Atherosderosis: A question of endothelial integrity and growth control of smooth muscle. Harvey Ixct 1983; 77:161-82. 5. Friedman RJ, Stemerman MB, Wenz B, Moore S, Gauldie J, Gent M, Tiell ML, Spaeth TH. The effect of thrombocytopenia on experimental arteriosclerotic lesion formation in rabbits. Smooth muscle cell proliferation and re-endothelialization. J Clin Invest 1977; 60:1191-1201.

Journal of VASCULAR SURGERY

6. Moore S, Friedman RJ, Singal DP, Gauldie J, Blajchman MA, Roberts RS. Inhibition of injury-induced thromboathcrosclerotic lesions by antiplatelet serum in rabbits. Thromb Haemost 1976; 35:70-81. 7. Gajdusek CM, DiCorleto P, Ross R, Schwartz S. An endothelial cell-derived growth factor. J Cell Biol 1980; 85: 467-72. 8. Fox PL, DiCorleto PE. Regulation of production of a platelet-derived growth factor-like protein by cultured bovine aortic endothelial cells. J Cell Physiol 1984; 121:298-308. 9. Seifert RA, Schwartz SM, Bowen-Pope DF. Developmentally regulated production of platelet-derived growth factor-like molecules. Nature 1984; 311:669-71. 10. Walker LN, Bowen-Pope DF, Reidy MA. Secretion of platelet-derived growth factor (PDGF)-Iike activity by arterial smooth muscle cells is induced as a response to injury (abstract). J Cell Biol 1984; 99:416. 11. Doolittle RF, Hunkapiller MW, Hood LE, Devare SG, Robbins KC, Aaronson SA, Antoniades HN. Simian sarcoma virus one gene, v-sis, is derived from the gene (or genes) encoding a platelet-derived growth factor. Science 1983; 221:275-7. 12. Waterfield MD, Scrace GT, Whittle N, Stroobant P, Johnson A, Wasteson A, Westermark B, Heldin CH, Huang JS, Deuel TF. Platelet-derived growth factor is structurally related to the putative transforming protein p28 s~ of simian sarcoma virus. Nature 1983; 304:35-9. 13. Barrett TB, Gajdusek CM, Schwartz SM, McDougalt JK, Benditt EP. Expression of the sis gene by endothelial cells in culture and in vivo. Proc Nat] Acad Sci USA 1984; 81:67724. 14. Barrett TB, Benditt EP. The sis gene is expressed in human atherosclerotic lesions suggesting atherosclerosis may be an autocrine/paracrine disorder (abstract). Fed Proc 1985; 44:737. 15. Castellot Jr JJ, Favreau LV, Karnovsky MJ, Rosenberg RD. Inhibition of vascular smooth muscle cell growth by endothelial cell--derived heparin. Possible role ofa platelet endoglycosidese. J Biol Chem 1982; 257:11256-60.