Matrix metalloproteinase inhibition reduces intimal hyperplasia in a porcine arteriovenous-graft model

BASIC RESEARCH STUDIES

Matrix metalloproteinase inhibition reduces intimal hyperplasia in a porcine arteriovenous-graft model Joris I. Rotmans, MD,a,b Evelyn Velema, BSc,a Hence J. M. Verhagen, MD, PhD,c Jan D. Blankensteijn, MD, PhD,c Dominique P. V. de Kleijn, PhD,a Erik S. G. Stroes, MD, PhD,b and Gerard Pasterkamp, MD, PhD,a Utrecht and Amsterdam, The Netherlands Background: The patency of arteriovenous (AV) polytetrafluoroethylene grafts for hemodialysis is impaired by intimal hyperplasia (IH) at the venous outflow tract. IH mainly consists of vascular smooth muscle cells, fibroblasts, and extracellular matrix proteins. Because matrix metalloproteinases (MMPs) are enzymes able to degrade extracellular matrix proteins such as elastin and collagen and also stimulate migration of vascular smooth muscle cells, we hypothesized that BB2983 (a broad-spectrum MMP inhibitor) could reduce IH in AV grafts. Methods: In 12 pigs, AV grafts were created bilaterally between the carotid artery and the jugular vein. Six pigs received the oral MMP inhibitor (MMPi), and six pigs served as a control. Four weeks after AV shunting, the grafts and adjacent vessels were excised and underwent histologic analysis. Quantification of elastin content was performed on Elastin von Gieson–stained sections. Results: At the venous outflow tract, IH was strongly inhibited after MMPi when compared with the control group (1.02 ⴞ 0.26 mm2 vs 2.14 ⴞ 0.38 mm2; P ⴝ .027). The medial area did not differ significantly. In the control group elastin density decreased compared with nonoperated veins. This decrease was not observed in the MMPi group (nonoperated, 6.3% ⴞ 0.4%; MMPi, 7.2% ⴞ 0.7% vs untreated, 3.6% ⴞ 0.5%; P ⴝ .0004). Outward remodeling of the vein was not influenced by MMP inhibition. Conclusion: MMPi reduces IH formation at the venous outflow tract of AV grafts in pigs, probably by inhibiting elastin degradation. These data suggest that MMP inhibitors might be useful for minimizing IH in AV grafts, thus prolonging patency rates of AV grafts in patients on hemodialysis. (J Vasc Surg 2004;39:432-9.)

Complications at the hemodialysis access site constitute a major cause of morbidity for patients with end-stage renal disease who are undergoing chronic hemodialysis. In the United States approximately 70% of the 250,000 hemodialysis patients use an expanded polytetrafluoroethylene (ePTFE) graft for permanent vascular access.1 The current 1- and 2-year primary patency rates of these ePTFE grafts are 50% and 25%, respectively.2 These data are in line with an incidence rate of complications, which approaches 40% From the Department of Experimental Cardiology,a University Medical Center, Utrecht, Department of Vascular Medicine,b Academic Medical Center, Amsterdam, Department of Vascular Surgery,c University Medical Center, Utrecht. Competition of interest: none. Supported by grant No. CSP 6001 from the Dutch Kidney Foundation. The matrix metalloproteinase inhibitor BB2983 was supplied by British Biotech Pharmaceuticals, Oxford, England. The ePTFE grafts were supplied by W. L. Gore and Associates, Flagstaff, Ariz. Clopidogrel was supplied by Sanofi-Synthelabo, Paris, France. Reprint requests: Gerard Pasterkamp, MD, Department of Experimental Cardiology, University Medical Center, Heidelberglaan 100, 3584 CX Utrecht, The Netherlands (e-mail: [email protected]). 0741-5214/$30.00 Copyright © 2004 by The Society for Vascular Surgery. doi:10.1016/j.jvs.2003.07.009

432

in patients with ePTFE grafts within the first 6 months. The failure of hemodialysis access grafts is predominantly due to the rapid development of an intimal hyperplastic response in the region of anastomosis between the ePTFE graft and the vein (ie, venous outflow tract). The intimal hyperplasia mainly consists of vascular smooth muscle cells (VSMCs) and extracellular matrix proteins.3-5 Despite coordinated attempts to solve this clinically important issue, approaches to reduce graft stenosis have thus far proven unsuccessful. Matrix metalloproteinases (MMPs) belong to a group of zinc-dependent proteases and cause breakdown of extracellular matrix proteins such as collagen and elastin, which is a prerequisite for migration of VSMCs. Flow increase and vascular injury are potent stimuli for activation of MMP-2 and MMP-9.6,7 These enzymes are able to degrade elastic fibers in the vessel wall.8,9 Furthermore, recent studies have emphasized the importance of elastin as a potent regulator of VSMC proliferation and subsequent neointima formation after vascular injury.10-12 The beneficial effects of tissue inhibitor metalloproteinase (TIMP)-213 and TIMP-3 overexpression14 on intimal hyperplasia after vein grafting of carotid arteries have underlined the capacity of MMP inhibition as a potential therapy.

JOURNAL OF VASCULAR SURGERY Volume 39, Number 2

Rotmans et al 433

In the present study, we hypothesized that BB2983, an oral broad-spectrum MMP inhibitor, could reduce extracellular matrix breakdown and thus reduce the intimal hyperplastic response at the venous outflow tract of arteriovenous (AV) ePTFE-grafts in pigs. MATERIAL AND METHODS Animals. Twelve female Landrace pigs weighing 52 kg ⫾ 4.6 kg were studied. ePTFE AV grafts were created bilaterally between the carotid artery and the internal jugular vein (n ⫽ 24 grafts). Six pigs received the oral MMP inhibitor (MMPi), and six pigs served as controls. All animals were killed 28 days after surgery. The study protocol was approved by the Ethical Committee on Animal Experimentation of the University Medical Center, Utrecht, and animal care followed established guidelines.15 Anesthesia. Before operation and termination, the animals were deprived of food and water overnight and premedicated with intramuscular ketamine hydrochloride 10 mg/kg, midazolam 0.4 mg/kg, atropin 0.5 mg, and intravenous thiopental sodium 4 mg/kg. They were then intubated and ventilated with a mixture of O2 and air (1:2) and halothane (0.6%). An ear vein was used for continuous administration of 0.3 mg/kg per hour midazolam and 2.5 ␮g/kg per hour sufentanil. Operative procedures. Starting 6 days before the operation, the pigs received acetylsalicylic acid 80 mg/dd. Clopidogrel 225 mg was given 1 day prior to the operation and was continued at a dose of 75 mg/d until termination. Through a longitudinal incision in the midline of the neck, the common carotid artery and the internal jugular vein were dissected bilaterally. Heparin 100 IU/kg intravenously was given before manipulation of the vessels. All the AV grafts were created by the same experienced vascular surgeon (H.V.). The artery was clamped with atraumatic clamps, and an arteriotomy was made (8 mm). An end-toside anastomosis was created at an angle of 45° by use of a continuous suture of 8-0 polypropylene (Prolene; Ethicon, Somerville, NJ). All ringed ePTFE grafts were 5 mm in diameter and 7 cm in length (W. L. Gore and Associates, Flagstaff, Ariz). The venous anastomosis was created in a similar fashion (Fig 1). The blood flow through the artery (proximal and distal to the anastomosis) and vein (proximal and distal) was measured several minutes after completing the anastomosis with a 4-mm perivascular flow probe (Transonic Systems, Maastricht, The Netherlands). Graft flow was calculated as flow of the proximal carotid artery minus the flow of the distal artery. The external diameter of the jugular vein was measured with a microscope with an internal raster (40⫻ magnification), 1 cm from the anastomosis (proximal and distal). After 4 weeks, pigs were anesthetized as described previously. After we measured flow, the animals were killed. Because of extensive formation of granulation tissue, we refrained from measuring the external diameter of the veins in view of the risk of tearing. The proximal carotid artery supplying the graft was cannulated and ligated beyond the cannula. The grafts and adjacent vessels were perfused with

Fig 1. ePTFE graft between the carotid artery and the internal jugular vein.

saline for 3 minutes. Subsequently, the grafts were perfused with formalin at physiologic pressure (100 mmHg). Two minutes later, both sides of the arteries and the veins were ligated to pressure-fixate the vessels. Grafts and adjacent vessels were then excised and immersed in formalin for a minimum of 24 hours. MMP activity. To study MMP activity in our model, portions of the grafted jugular vein of control pigs, 3 cm proximal and distal to the anastomosis, were fresh frozen in liquid nitrogen at the time of death. The femoral vein served as an unoperated control. Zymography was performed on each sample with electrophoresis through a 10% sodium dodecyl sulfate⫺polyacrylamide gel electrophoresis (SDS-PAGE) gel containing 1% gelatin, as described previously.7 MMP inhibition. The oral nonspecific synthetic MMP inhibitor BB2893 was supplied by British Biotech Pharmaceuticals Limited, Oxford, UK. This drug inhibits, in the nanomolar range, all known classes of MMPs. Pigs in the MMP inhibitor group were treated until death with 10 mg/kg given orally twice a day, starting 1 day after the operation. An unpublished dosimetry study in Landrace pigs in our lab established that a dose of 10 mg/kg of the comparable compound BB2516 twice a day would give exposure of 100 to 200 ng/mL plasma, a concentration which has been shown to be effective in animal models of cancer. To demonstrate bioavalability of the MMP inhibitor in the vascular tissue of the treated pigs, a collagenase activity assay was performed (Molecular Probes, The Netherlands). Thirty fresh-frozen sections (5 ␮m each) of the femoral arteries of four pigs were dissolved in dimethyl sulfoxide (DMSO). These extracts were tested on their ability to inhibit collagenase activity. As a positive control, BB2983 was diluted in DMSO. Collagen from bovine skin (100 ␮g/mL) was diluted in reaction buffer and added to the different tissue extracts. Next, collagenase type IV (0.5 U/mL) was added to start the proteolytic reaction.

JOURNAL OF VASCULAR SURGERY February 2004

434 Rotmans et al

Intima and media areas of Elastin von Gieson–stained sections obtained from the venous anastomosis

Fig 2. A, Frontal view. V-1 is the jugular vein located 1 cm proximal of the anastomosis. V-2 is the center of the anastomosis. V-3 is the jugular vein 1 cm distal of the anastomosis. B, Cross section of the anastomosis (V-2). Intimal hyperplasia is situated at the facing vein (1 ⫽ cushioning region) and at the edge of the ePTFE graft (2 ⫽ shoulder region).

Tissue preparation and morphometric analysis. The fixated veins were cut in 5-mm blocks and embedded in paraffin. Sections (5 ␮m thick) were prepared of the vein 1 cm proximal and distal to the anastomosis and at the center of the anastomosis (Fig 2). To minimize the variability of sectioning the anastomosis, we selected the sections in which half of the circumference is composed of the ePTFE graft and the other half of the recipient vein. (Fig 2, B). The standardized angle of the anastomosis (45°) was used to be sure that sections were cut perpendicular to the vein. For morphometric analysis, sections were stained for general morphology (Elastin von Gieson) and collagen (picro-Sirius red). All sections were stained by one technician, in three sessions. Although the intensity of the staining can slightly differ between sessions, the area measurements could be performed reliably because, above a certain threshold, they do not depend on the intensity of the stainings. Area measurements were performed on captured images obtained with a color camera attached to a light microscope with the use of computer-assisted morphometry. The area and thickness of the intima and media were manually traced by Elastin von Gieson–stained section. The intima was defined as the area above the internal elastic lamina. Interrupted elastic laminae were connected by manual tracing. The intima could be easily distinguished from the media by the disorganized structure of intimal cells. Thrombus formation was discriminated from intimal hyperplasia by using hematoxylin-eosin– and Elastin von Gieson–stained sections. In the sections of the venous anastomosis, two parts of the intima area were distinguished: (1) the intima covering the graft (shoulder region) and (2) the intima located at the venous part of the anastomosis (cushioning region) (Fig 2, B). The venous diameter was traced in the sections that were located 1 cm proximal and distal to the anastomosis. This diameter is a valid measure for outward remodeling because the veins were pressure-fixated. Quantification of elastin content was performed on Elastin von Gieson–stained sections of the grafted vein, 1

Venous anastomosis

Control pigs

BB2983treated pigs

P

Total intima area Shoulder region Cushioning region Media area Intima/media ratio

2.14 ⫾ 0.38 1.08 ⫾ 0.23 1.07 ⫾ 0.31 1.71 ⫾ 0.22 1.32 ⫾ 0.21

1.02 ⫾ 0.26 0.32 ⫾ 0.12 0.70 ⫾ 0.23 2.10 ⫾ 0.38 0.51 ⫾ 0.11

0.027 0.009 0.57 0.45 0.004

Intima and media values are expressed in square millimeters and are the mean area ⫾ SEM.

cm proximal and distal to the anastomosis and at the center of the venous anastomosis. Only the elastin-positive pixels of captured images were selected by using a color threshold mask focused on the black pixels. Elastin area was expressed as a percentage of the vessel wall by dividing the elastinpositive pixels by the medial and adventitial cross-sectional area. Quantification of collagen content of the grafted veins was performed by using picro-Sirius red-stained sections and digital image microscopy with circularly polarized light as described by Perree et al.16 Collagen density was calculated by dividing the collagen content by the vessel wall cross-sectional area. Statistical evaluation. Data are presented as mean ⫾ SEM. SPSS 11.0 (Chicago, Ill) was used for all statistical calculations. We performed the Mann-Whitney test to ascertain the significance of differences among the two groups. P ⬍ .05 was considered significant. RESULTS Characteristics of animal model All grafts except one in the MMP inhibitor group were patent at the time of harvest. Patent grafts (n ⫽ 23) were included for further analysis. The proximal vein of the occluded graft was completely thrombosed, most likely because of local infection. The increase in weight during follow-up was 10.3 ⫾ 1.6 kg in the MMPi-treated group and 8.6 ⫾ 0.9 kg in the control group. None of the pigs showed side effects like diarrhea or other signs of toxicity. No differences in wound healing were observed between the groups. Baseline flow through the jugular vein was 42 ⫾ 6 mL/min. AV graft implantation resulted in a graft flow of 829 ⫾ 71 mL/min (proximal vein, 448 ⫾ 55 mL/min; distal vein, 381 ⫾ 42 mL/min). On day 28, flow measurements of the grafts revealed no significant differences between the controls and BB2983-treated animals (961 ⫾ 91 mL/min vs 956 ⫾ 96 mL/min, respectively). Gelatin zymographic analysis of protein extracts from the jugular vein revealed three bands of metalloproteinase activity with molecular weights of approximately 68, 72, and 95 kDa, corresponding with pro-MMP2, active MMP-2, and MMP-9, respectively. At 28 days, protein

JOURNAL OF VASCULAR SURGERY Volume 39, Number 2

Rotmans et al 435

Fig 3. Representative gelatin zymographic analysis of protein extracts obtained from jugular and femoral veins of control pigs at 28 days. Lane one, recombinant MMP-9; lane two, recombinant MMP-2; lane three, tissue extract femoral vein; lane four, tissue extract left jugular vein at 28 days; lane five, tissue extract right jugular vein at 28 days. Note, the molecular weights of porcine MMP-2 and MMP-9 are slightly smaller then the molecular weights of human recombinant MMP-2 and MMP-9.

extracts from the grafted jugular veins revealed augmented gelatinolytic activity of all three bands, compared with the femoral veins (Fig 3). Morphometric analysis Intima and media areas. At 4 weeks, vein grafts in control animals showed extensive intimal hyperplasia. In the untreated group, more intimal hyperplasia was formed compared with the MMPi-treated animals (Fig 4, A-B). ␣–Smooth muscle actin staining confirmed that the intimal hyperplasia predominantly consisted of VSMCs (Fig 4, E). Data of the intima and media area measurements are summarized (Table). A 53% reduction in intima formation at the venous anastomosis was achieved in the group treated with BB2983 for 28 days. The reduction of the intima covering the graft (70%) and the facing vein (34%) both contributed to this inhibitory effect of BB2983. At 1 cm proximal and distal to the anastomosis, the intimal and medial areas did not differ significantly between the BB2983-treated animals and the control group. Elastin area and collagen density. Densitometric analysis of elastin area and collagen density of the grafted veins is shown in Fig 5. At all three locations the elastin area was larger in the BB2983-treated group, compared with the untreated pigs (anastomosis, 7.2% ⫾ 0.7% vs 3.6% ⫾ 0.5%; P ⬍ .001; proximal, 7.0% ⫾ 0.7% vs 4.9% ⫾ 1.1%; P ⫽ .04; distal, 6.7% ⫾ 0.7% vs 3.8% ⫾ 0.4%; P ⫽ .03, respectively). Linear regression analysis showed a significant relationship between the elastin area and intimal hyperplasia (IH) at the venous anastomosis (R ⫽ 0.47, P⫽ .02). Representative Elastin von Gieson–stained sections at large magnification are shown in Fig 4, C-D,; the elastic fibers are black. Collagen density was not significantly different between control and BB2983-treated animals, except for a modest difference in the distal vein (26.7 ⫾ 3.3 gray values/mm2 vs 36.7 ⫾ 4.3 gray values/mm2, respectively, P ⫽ .05). An example of a picro-Sirius–stained section is shown in Fig 4, F. Venous diameters. Immediately after creation of the AV graft, the mean external diameter of the proximal and distal jugular vein did not differ among control and

BB2983-treated animals (proximal, 3.91 ⫾ 0.29 mm vs 3.37 ⫾ 0.25 mm; P ⫽.26; distal, 3.90 ⫾ 0.21 mm vs 3.94 ⫾ 0.22 mm; P ⫽.42, respectively). At 28 days, histologically determined venous diameters did not differ between the untreated pigs and the BB2983-treated pigs. (proximal, 4.15 ⫾ 0.37 mm vs 3.61 ⫾ 0.40; distal, 5.06 ⫾ 0.55 vs 5.37 ⫾ 0.44 mm, respectively). MMP inhibition in tissue extracts. A 30% reduction in collagenase activity was observed in tissue extracts of the arteries treated with MMP inhibitor (n ⫽ 2), when compared with control arteries, implying that MMP inhibitory compound is actually present within arterial tissue (untreated, 195 arbitrary units ⫾ 9 [SEM] vs MMPi-treated, 137 arbitrary units ⫾ 11 [SEM]; BB2983 in DMSO 1 ␮g/mL, 120 arbitrary units ⫾ 8). DISCUSSION In the present study we show that neointimal formation at the venous outflow tract of ePTFE AV grafts is significantly reduced after MMP inhibition. The elastin content in operated vessels was preserved in the MMPi group, whereas a sharp decrease in elastin content was observed in the control group. Finally, venous outward remodeling was not negatively influenced by MMP inhibition. MMP inhibition and venous intimal hyperplasia. MMPs are key enzymes during rebuilding of the vessel wall matrix, which occurs in the process of vascular remodeling.17,18 Surgical preparation of the venous wall19 and flow changes (ie, shear stress alterations),20-23 both characteristics of the AV grafting procedure, are potent stimuli for increasing MMP activity. Because MMPi has been shown to decrease matrix breakdown and VSMC migration,24-26 MMPi can be expected to exert beneficial effects on intimal hyperplasia, particularly in an ePTFE AV-graft model. Indeed, in the present study we show a reduction in intimal hyperplasia at the venous outflow tract in a porcine AV-graft model. These findings concur with previous data, showing a reduction of neointimal formation in grafted veins during TIMP overexpression.13,14,27 In the study of Porter et al,27 doxycycline reduced intima formation in cultured human saphenous veins by inhibiting MMP-9 activity. Those studies

436 Rotmans et al

JOURNAL OF VASCULAR SURGERY February 2004

Fig 4. A, Representative Elastin von Gieson-stained 5-␮m section of a venous anastomosis of a BB2983-treated pig and an untreated pig (B) (original magnification, 12.5⫻). C, Detail of Elastin von Gieson-stained section of a BB2983-treated pig and an untreated pig (D), showing elastic fibers in black (original magnification, 100⫻). E, Detail of ␣-actin smooth muscle cell-stained section (original magnification, 200⫻). F, Picro-Sirius-stained section of venous anastomosis, visualized with polarized light (original magnification, 12.5⫻).

consistently show a beneficial effect of MMP inhibition on venous intimal hyperplasia. The intimal proliferation at the venous outflow tract of AV grafts has characteristic localizations, both at the edge of the ePTFE graft (shoulder region) and at the facing vein (cushioning region). Whereas stretch injury is thought to be of major importance for the proliferation at the shoulder region, the absence of shear (no/low-flow zone) may be of particular relevance at the cushioning region.28-30 Interest-

ingly, we observed that BB2983 inhibits intimal formation in both these regions. The inhibitory effect of the MMPi on intimal hyperplasia was limited to the anastomotic region (V-2 in Fig 2). A possible explanation for this local effect might be that the stimuli for MMP activation (ie, vascular injury and flow changes) differ among regions. In this respect, it should be noted that the tissue extracts for zymography were not obtained from the center of the anastomosis.

JOURNAL OF VASCULAR SURGERY Volume 39, Number 2

Rotmans et al 437

Fig 5. A, Densitometric analysis of elastin area of grafted veins from untreated pigs (n ⫽ 12) and BB2983-treated pigs (n ⫽ 11). Control veins are jugular veins of pigs without AV graft. B, Densitometric analysis of collagen density of grafted veins from untreated pigs (n ⫽ 12) and BB2983-treated pigs (n ⫽ 11).

MMP inhibition and elastin area. A higher elastin density in the vessel wall was found in the animals treated with MMP inhibitor. These results are in accordance with the study of Forough et al31 that demonstrated elastin accumulation after endogenous MMPi by TIMP-1 overexpression. Interestingly, perivascular release of elastase after venous and arterial injury limits neointimal hyperplasia by directionally guiding VSMC responses, implying a direct inhibitory effect of the intact elastic lamina on neointima formation.32,33 The results of the present study confirm an important role for MMPs in elastin breakdown occurring after vascular injury. The inverse correlation between elastin content and intimal hyperplasia indicates that reduced elastic fiber breakdown in the MMPi animals may have contributed to a decreased neointimal area. Besides the guiding function of elastin in VSMC responses, this inverted relationship might also be explained by the ability of MMPs to attract and activate inflammatory cells, which has been described in abdominal aortic aneurysms.34 However, the lack of effect of MMP inhibition on macrophage infiltration in bal-

loon-dilated porcine arteries does not support this hypothesis.35 MMP inhibition and collagen content. In contrast to the elastin density, no difference in collagen content was observed between the groups. Different studies have emphasized the role of MMPs and MMP inhibitors in collagen turnover after vascular injury. MMPs are not only involved in collagen degradation, but they also affect collagen synthesis and cross-linking.36-38 Collagen synthesis and collagen breakdown are both up-regulated after vascular injury.36 Inhibition of collagen degradation by the MMP inhibitor might decrease collagen synthesis by a feedback loop.39,40 Furthermore, the MMP inhibitor could inhibit the shedding of receptors, such as tumor necrosis factor ␣ (TNF-␣),41 that affect collagen synthesis.42 When both collagen degradation and synthesis are reduced, then total collagen content does not change. The net effect of MMPi on collagen content may also vary between different MMP inhibitors and animal models. In addition to the influence on matrix composition of MMPi, increased apoptosis of VSMCs, causing a reduction

JOURNAL OF VASCULAR SURGERY February 2004

438 Rotmans et al

in intimal VSMCs, should be considered. However, MMP inhibition might also result in decreased apoptosis of VSMCs, because apoptosis is regulated by cell-matrix interactions.43 Furthermore, MMPs may also directly promote apoptosis by releasing both soluble Fas ligand44 and membrane-bound TNF-␣ from cells.45 In our study, one infected graft was observed in the MMPi-treated group. The question remains whether MMP inhibition reduced graft incorporation, making it more susceptible to infection and thus graft thrombosis. The role of MMPs in wound and anastomotic healing has been investigated by Balcom et al. 46 In that study, MMP inhibition did not impair wound or enteric healing in a rat model of laparotomy and small bowel resection.46 On the basis of our data, we cannot exclude that MMP inhibition increases the risk of graft infection. However, the limited incidence of graft infection does not support such causal relation. Future clinical application. The first available oral MMP inhibitors have entered clinical trials in the field of oncology.52 With long-term treatment, reversible joint and muscle pain were reported frequently. However, it is likely that more selective MMP inhibitors, which are currently being manufactured and tested, will have fewer side effects. These specific MMP inhibitors will probably result in more knowledge of the role of MMPs in cardiovascular disease, with subsequent accurate targeting of the disease and reduction of possible side effects. STUDY LIMITATIONS Bendeck et al47 reported that the decrease in VSMC migration after MMPi in a model of arterial injury is followed by an increase in VSMC replication and a subsequent “catch up” of lesion size. On the basis of the results of our study, we cannot exclude the occurrence of this catch-up phenomenon. However, this phenomenon is described to occur from 14 days after intervention, whereas our study continued up to a later time point of 4 weeks. Venous outward remodeling is essential for adequate maturation of the vein distal to the graft. Different studies emphasized the role of MMP-2 in flow-mediated arterial enlargement.6,48 Hu et al13 reported a significant reduction of the venous diameter after endogenous MMPi on TIMP-2 gene transfer after 8 weeks. In our study, venous outward remodeling was not impaired after 4 weeks of MMP inhibition. Therefore, we cannot exclude that our follow-up period of 4 weeks has been too short to detect a negative effect of MMP inhibition on venous outward remodeling. Cardiovascular disease is the major cause of death in patients with end-stage renal disease.49 A multitude of risk factors, including the uremic milieu and perhaps the hemodialysis procedure itself, is likely to contribute to premature cardiovascular diseases.50,51 These risk factors, not present in our model, might also play a role in AV-graft failure in hemodialysis patients.

CONCLUSIONS MMP inhibition reduces neointimal formation at the venous outflow tract of AV grafts in pigs. The protective effect of MMPi on elastic fiber breakdown may have contributed to the decreased neointimal area. These data suggest that MMP inhibitors might be useful for minimizing intimal hyperplasia in human AV grafts, thus prolonging patency rates of AV grafts in patients receiving dialysis. Machiel Nagtegaal is acknowledged for excellent assistance with the pathology work-up. REFERENCES 1. Pisoni RL, Young EW, Dykstra DM, Greenwood RN, Hecking E, Gillespie B, et al. Vascular access use in Europe and the United States: results from the DOPPS. Kidney Int 2002;61:305-16. 2. Schwab SJ, Harrington JT, Singh A, Roher R, Shohaib SA, Perrone RD, et al. Vascular access for hemodialysis. Kidney Int 1999;55:2078-90. 3. Roy-Chaudhury P, Kelly BS, Miller MA, Reaves A, Armstrong J, Nanayakkara N, et al. Venous neointimal hyperplasia in polytetrafluoroethylene dialysis grafts. Kidney Int 2001;59:2325-34. 4. Shi Y, O’Brien JEJ, Mannion JD, Morrison RC, Chung W, Ford A, et al. Remodeling of autologous saphenous vein grafts. The role of perivascular myofibroblasts. Circulation 1997;95:2684-93. 5. Hu Y, Mayr M, Metzler B, Erdel M, Davison F, Xu Q, et al. Both donor and recipient origins of smooth muscle cells in vein graft atherosclerotic lesions. Circ Res 2002;91:e13-e20. 6. de Kleijn DP, Sluijter JP, Smit J, Velema E, Richard W, Schoneveld AH, et al. Furin and membrane type-1 metalloproteinase mRNA levels and activation of metalloproteinase-2 are associated with arterial remodeling. FEBS Lett 2001;501:37-41. 7. Southgate KM, Mehta D, Izzat MB, Newby AC, Angelini GD, et al. Increased secretion of basement membrane-degrading metalloproteinases in pig saphenous vein into carotid artery interposition grafts. Arterioscler Thromb Vasc Biol 1999;19:1640-9. 8. Senior RM, Griffin GL, Fliszar CJ, Shapiro SD, Goldberg GI, Welgus HG, et al. Human 92- and 72-kilodalton type IV collagenases are elastases. J Biol Chem 1991;266:7870-5. 9. Nikkari ST, Hoyhtya M, Isola J, Nikkari T. Macrophages contain 92-kd gelatinase (MMP-9) at the site of degenerated internal elastic lamina in temporal arteritis. Am J Pathol 1996;149:1427-33. 10. Karnik SK, Brooke BS, Bayes-Genis A, Sorensen L, Wythe JD, Schwartz RS, et al. A critical role for elastin signaling in vascular morphogenesis and disease. Development 2003;130:411-23. 11. Barolet AW, Nili N, Cheema A, Robinson R, Natarajan MK, O-Blenes S, et al. Arterial elastase activity after balloon angioplasty and effects of elafin, an elastase inhibitor. Arterioscler Thromb Vasc Biol 2001;21: 1269-74. 12. Waugh JM, Li-Hawkins J, Yuksel E, Citra PN, Amabile PG, Hilfiker PR, et al. Therapeutic elastase inhibition by alpha-1-antitrypsin gene transfer limits neointima formation in normal rabbits. J Vasc Interv Radiol 2001;12:1203-9. 13. Hu Y, Baker AH, Zou Y, Newby AC, Xu Q. Local gene transfer of tissue inhibitor of metalloproteinase-2 influences vein graft remodeling in a mouse model. Arterioscler Thromb Vasc Biol 2001;21:1275-80. 14. George SJ, Lloyd CT, Angelini GD, Newby AC, Baker AH. Inhibition of late vein graft neointima formation in human and porcine models by adenovirus-mediated overexpression of tissue inhibitor of metalloproteinase-3. Circulation 2000;101:296-304. 15. National Institutes of Health. Guide for the Care and Use of Laboratory Animals. Rockville, Md: National Institutes of Health (NIH Publication No. 85-23, 1996). 16. Perree J, van Leeuwen T, Velema E, Smeets M, de Kleijn D, Borst C, et al. UVB-activated psoralen reduces luminal narrowing after balloon dilation because of inhibition of constrictive remodeling. Photochem Photobiol 2002;75:68-75.

JOURNAL OF VASCULAR SURGERY Volume 39, Number 2

17. Newby AC, Southgate KM, Davies M. Extracellular matrix degrading metalloproteinases in the pathogenesis of arteriosclerosis. Basic Res 1994;89(suppl 1):59-70. 18. Galis ZS, Khatri JJ. Matrix metalloproteinases in vascular remodeling and atherogenesis: the good, the bad, and the ugly. Circ Res 2002;90: 251-62. 19. George SJ, Zaltsman AB, Newby AC. Surgical preparative injury and neointima formation increase MMP-9 expression and MMP-2 activation in human saphenous vein. Cardiovasc Res 1997;33:447-59. 20. Patterson MA, Leville CD, Hower CD, Jean-Claude JM, Seabrook GR, Towne JB, et al. Shear force regulates matrix metalloproteinase activity in human saphenous vein organ culture. J Surg Res 2001;95:67-72. 21. Loth F, Jones SA, Zarins CK, Giddens DP, Nassar RF, Glagov S, et al. Relative contribution of wall shear stress and injury in experimental intimal thickening at PTFE end-to-side arterial anastomoses. J Biomech Eng 2002;124:44-51. 22. Chesler NC, Ku DN, Galis ZS. Transmural pressure induces matrixdegrading activity in porcine arteries ex vivo. Am J Physiol 1999;277: H2002-H2009. 23. Bassiouny HS, Song RH, Hong XF, Singh A, Kocharyan H, Glagov S. Flow regulation of 72-kD collagenase IV (MMP-2) after experimental arterial injury. Circulation 1998;98:157-63. 24. Bendeck MP, Conte M, Zhang M, Nili N, Strauss BH, Farwell SM, et al. Doxycycline modulates smooth muscle cell growth, migration, and matrix remodeling after arterial injury. Am J Pathol 2002;160:1089-95. 25. Forough R, Koyama N, Hasenstab D, Lea H, Clowes M, Nikkari ST, et al. Overexpression of tissue inhibitor of matrix metalloproteinase-1 inhibits vascular smooth muscle cell functions in vitro and in vivo. Circ Res 1996;79:812-20. 26. Cheng L, Mantile G, Pauly R, Nater C, Felici A, Monticone R, et al. Adenovirus-mediated gene transfer of the human tissue inhibitor of metalloproteinase-2 blocks vascular smooth muscle cell invasiveness in vitro and modulates neointimal development in vivo. Circulation 1998; 98:2195-201. 27. Porter KE, Thompson MM, Loftus IM, McDermott E, Jones L, Crowther M, et al. Production and inhibition of the gelatinolytic matrix metalloproteinases in a human model of vein graft stenosis. Eur J Vasc Endovasc Surg 1999;17:404-12. 28. Perktold K, Leuprecht A, Prosi M, Berk T, Czerny M, Trubel W, et al. Fluid dynamics, wall mechanics, and oxygen transfer in peripheral bypass anastomoses. Ann Biomed Eng 2002;30:447-60. 29. Piorko D, Knez P, Nelson K, Schmitz-Rixen T. Compliance in anastomoses with and without vein cuff interposition. Eur J Vasc Endovasc Surg 2001;21:461-6. 30. Pappas PJ, Hobson RW II, Meyers MG, Jamil Z, Lee BC, Silva MB, et al. Patency of infrainguinal polytetrafluoroethylene bypass grafts with distal interposition vein cuffs. Cardiovasc Surg 1998;6:19-26. 31. Forough R, Lea H, Starcher B, Allaire E, Clowes M, Hasenstab D, et al. Metalloproteinase blockade by local overexpression of TIMP-1 increases elastin accumulation in rat carotid artery intima. Arterioscler Thromb Vasc Biol 1998;18:803-7. 32. Amabile PG, Wong H, Uy M, Bronmand S, Elkins CJ, Yuksel E, et al. In vivo vascular engineering of vein grafts: directed migration of smooth muscle cells by perivascular release of elastase limits neointimal proliferation. J Vasc Interv Radiol 2002;13:709-15. 33. Wong AH, Waugh JM, Amabile PG, Yuksel E, Dake MD. In vivo vascular engineering: directed migration of smooth muscle cells to limit neointima. Tissue Eng 2002;8:189-99. 34. Hance KA, Tataria M, Ziporin SJ, Lee JK, Thompson RW. Monocyte chemotactic activity in human abdominal aortic aneurysms: role of elastin degradation peptides and the 67-kD cell surface elastin receptor. J Vasc Surg 2002;35:254-61. 35. de Smet BJ, de Kleijn D, Hanemaaijer R, Verheijen JH, Robertus L, van der Helm YJ, et al. Metalloproteinase inhibition reduces constrictive

Rotmans et al 439

36.

37.

38.

39.

40.

41.

42.

43.

44.

45.

46.

47.

48.

49.

50.

51.

52.

arterial remodeling after balloon angioplasty: a study in the atherosclerotic Yucatan micropig. Circulation 2000;101:2962-7. Strauss BH, Robinson R, Batchelor WB, Chisholm RJ, Ravi G, Natarajan MK, et al. In vivo collagen turnover following experimental balloon angioplasty injury and the role of matrix metalloproteinases. Circ Res 1996;79:541-50. Kuzuya M, Asai T, Kanda S, Maeda K, Cheng XW, Iguchi A. Glycation cross-links inhibit matrix metalloproteinase-2 activation in vascular smooth muscle cells cultured on collagen lattice. Diabetologia 2001; 44:433-6. Uzel MI, Scott IC, Babakhanlou-Chase H, Palamakumbura AH, Pappano WN, Hong HH, et al. Multiple bone morphogenetic protein 1-related mammalian metalloproteinases process pro-lysyl oxidase at the correct physiological site and control lysyl oxidase activation in mouse embryo fibroblast cultures. J Biol Chem 2001;276:22537-43. Amano S, Akutsu N, Matsunaga Y, Nishiyama T, Champliaud MF, Burgeson RE, et al. Importance of balance between extracellular matrix synthesis and degradation in basement membrane formation. Exp Cell Res 2001;271:249-62. Martin J, Yung S, Robson RL, Steadman R, Davies M. Production and regulation of matrix metalloproteinases and their inhibitors by human peritoneal mesothelial cells. Perit Dial Int 2000;20:524-33. Lombard MA, Wallace TL, Kubicek MF, Petzold GL, Mitchell MA, Hendges SK, et al. Synthetic matrix metalloproteinase inhibitors and tissue inhibitor of metalloproteinase (TIMP)-2, but not TIMP-1, inhibit shedding of tumor necrosis factor-alpha receptors in a human colon adenocarcinoma (Colo 205) cell line. Cancer Res 1998;58: 4001-7. Yao J, Bone RC, Sawhney RS. Differential effects of tumor necrosis factor-alpha on the expression of fibronectin and collagen genes in cultured bovine endothelial cells. Cell Mol Biol Res 1995;41:17-28. Bennett MR. Apoptosis of vascular smooth muscle cells in vascular remodelling and atherosclerotic plaque rupture. Cardiovasc Res 1999; 41:361-8. Kayagaki N, Kawasaki A, Ebata T, Ohmoto H, Ikeda S, Inoue S, et al. Metalloproteinase-mediated release of human Fas ligand. J Exp Med 1995;182:1777-83. Black R, Rauch C, Kozlosky C, Peschon JJ, Slack JL, Wolfson MF, et al. A metalloproteinase disintegrin that releases tumour-necrosis factoralpha from cells. Nature 1997;385:729-33. Balcom JH, Keck T, Warshaw AL, et al. Perioperative matrix metalloproteinase inhibition therapy does not impair wound or anastomotic healing. J Gastrointest Surg 2002;6:488-95. Bendeck MP, Irvin C, Reidy MA. Inhibition of matrix metalloproteinase activity inhibits smooth muscle cell migration but not neointimal thickening after arterial injury. Circ Res 1996;78:38-43. Sho E, Sho M, Singh TM, Nanjo H, Komatsu M, Xu C, et al. Arterial enlargement in response to high flow requires early expression of matrix metalloproteinases to degrade extracellular matrix. Exp Mol Pathol 2002;73:142-53. U.S. Renal Data System. USRDS 2002 Annual Data Report: atlas of end-stage renal disease in the United States. Bethesda, Md: National Institutes of Health, National Institute of Diabetes and Digestive and Kidney Diseases; 2002. Kennedy R, Case C, Fathi R, Johnson D, Isbel N, Marwick TH, et al. Does renal failure cause an atherosclerotic milieu in patients with end-stage renal disease? Am J Med 2001;110:198-204. Cheung AK, Sarnak MJ, Yan G, Swyer JT, Heyka RJ, RoccoMV, et al. Atherosclerotic cardiovascular disease risks in chronic hemodialysis patients. Kidney Int 2000;58:353-62. Steward WP, Thomas Al. Marimastat: the clinical development of a matrix metalloproteinase inhibitor. Expert Opin Investig Drugs 2000; 9:2913-22.

Submitted May 12, 2003; accepted Jul 29, 2003.