Early Loss of Thrombomodulin Expression Impairs Vein Graft Thromboresistance
Implications for Vein Graft Failure
Thrombosis is the major cause of early vein graft failure. Our aim was to determine whether alterations in the expression of the anticoagulant proteins, thrombomodulin (TM) and the endothelial cell protein C receptor (EPCR), impair endothelial thromboresistance that may contribute to vein graft failure. Immunohistochemical staining of autologous rabbit vein graft sections revealed that the expression of TM, but not EPCR, was reduced significantly early after graft implantation. Western blot analysis revealed that TM expression was reduced by >95% during the first 2 weeks after implantation, with gradual but incomplete recovery by 42 days. This resulted in up to a 95% reduction in the capacity of the grafts to activate protein C and was associated with an increase in bound thrombin activity, which peaked on day 7 at 28.7±3.8 mU/cm2 and remained elevated for more than 14 days. Restoration of TM expression using adenovirus vector-mediated gene transfer significantly enhanced the capacity of grafts to activate protein C and reduced bound thrombin activity on day 7 to levels comparable to that of normal veins (5.7±0.4 versus 5.2±1.1 mU/cm2, respectively, P=0.74). Surprisingly, neointima formation was not affected by this inhibition of local thrombin activity. These data suggest that the early loss of TM expression significantly impairs vein graft thromboresistance and results in enhanced local thrombin generation. Although enhanced local thrombin generation may predispose to early vein graft failure due to thrombosis, it does not seem to contribute significantly to late vein graft failure due to neointimal hyperplasia.
Autologous vein grafts are the most frequently used conduits for coronary and peripheral arterial bypass surgery. Compared with arterial grafts, vein grafts suffer a significantly higher failure rate that limits their clinical efficacy. One- and 12-month failure rates for coronary bypass vein grafts approach 12% and 20%, respectively, and are predominantly due to occlusive thrombosis.1,2⇓ Late vein graft failure is due mainly to neointimal hyperplasia and accelerated atherosclerosis. Although surgical trauma and technical factors have often been invoked as possible causes of early graft failure, little is known about changes that may occur to the graft endothelium that might predispose to thrombosis.3,4⇓
Thrombomodulin (TM) and the endothelial cell protein C receptor (EPCR) constitute key components of the protein C pathway and are therefore major contributors to vascular thromboresistance.5,6⇓ Upon binding to TM, thrombin is rendered incapable of enzymatically cleaving fibrinogen or cellular thrombin receptors but acquires the ability to activate protein C. Activated protein C (APC) proteolytically degrades factors Va and VIIIa of the coagulation cascade, thereby inhibiting further thrombin generation. EPCR has recently been found to augment the efficiency of APC generation by binding protein C in solution and “presenting” it to the thrombin/TM complex.7 Evidence suggests that reduced expression of TM and/or EPCR, with the consequent impairment of APC generation, contributes to the thrombotic manifestations of several diseases, including sepsis, transplant rejection, and radiation enteropathy.8–11⇓⇓⇓
We hypothesize that the exposure of veins to an arterial environment adversely affects endothelial cell thromboresistance, thereby predisposing vein grafts to thrombosis and early failure. To characterize this process, we investigated the expression patterns of TM and EPCR in rabbit vein grafts and made correlations to local APC and thrombin activity. We then determined the effect of restoring TM expression using adenovirus vector-mediated gene transfer. Our findings suggest that TM is a major contributor to vein graft thromboresistance and that reductions in its expression facilitate local thrombin generation that may predispose to thrombotic graft occlusion.
Materials and Methods
All animal procedures were approved by the Johns Hopkins University Animal Care and Use Committee. A well-characterized rabbit vein graft model was used with modifications.12 Male New Zealand White rabbits (Robinson Services, Clemmons, NC) weighing 2.5 to 3.5 kg were sedated with 25 to 75 mg of intravenous thiopental, intubated, and anesthetized with 1% to 2% halothane. The left jugular vein and carotid artery were exposed through a midline incision. An approximate 2-cm segment of the external jugular vein was isolated and resected. End-to-side interpositional grafting of the jugular vein segment to the carotid artery was performed using 8-0 Prolene sutures. Rabbits were administered 100 U/kg of intravenous heparin before performing the arteriotomies. To control for the effects of surgical manipulation, a group of 3 rabbits underwent interpositional grafting of vein segments into the contralateral jugular vein in similar fashion. Wounds were closed with 4-0 Vicryl sutures. At harvest, vein grafts were isolated and the proximal and distal carotid segments cannulated. Grafts were flushed gently with saline and either used immediately for in situ coagulation assays or perfusion-fixed under distending pressure with 10% formalin in phosphate-buffered saline for immunohistochemical and morphometric analysis.
The surgical protocol was modified for experiments involving viral transduction of the grafts. Jugular vein segments were isolated, cannulated, and flushed with M-199 medium (Mediatech Inc). They were then incubated in situ under gentle distending pressure with 400 μL of M-199 medium, containing 1010.5 pfu/mL of the appropriate adenovirus vector. After 1 hour, the viral solution was removed and the transduced vein segments were flushed gently with fresh medium. Segments were then interposed into the carotid circulation as described above.
Construction of Adenovirus Vectors
The plasmid, pUC19TM15, containing the full-length cDNA sequence of the human TM gene, was obtained from the American Type Culture Collection, Manassas, Va (ATTC No. 61348). The EcoRI fragment containing the TM sequence was ligated into the multicloning site of pAdlox, a shuttle plasmid containing the CMV immediate-early promoter and SV-40 virus polyadenylation signal. The first-generation recombinant adenovirus vector, AdTMh5, was generated by cotransfection of pAdTMh5 and purified ψ5 adenovirus DNA into CRE8 cells as previously described.13 pAdlox, the ψ5 adenovirus, and CRE8 cells were a generous gift of Dr Stephen Hardy, Cell Genesys Inc, Foster City, Calif. AdNull-1, a control adenovirus vector containing no transgene, was generated in similar fashion. Recombinant virus was propagated in 293 cells and purified by double cesium chloride centrifugation. Viral stocks were plaque titered on 293 cells and replication incompetence was verified using A549 cells.
Three to seven transverse sections were cut from formalin-fixed and paraffin-embedded vein grafts harvested at the indicated times. Care was taken to avoid sections close to the arteriovenous anastomoses. Adjacent sections were immunostained for TM using a goat anti-rabbit TM polyclonal antibody (No. 236; American Diagnostica, Greenwich, Conn), von Willebrand factor (vWF) using a mouse monoclonal anti-human vWF antibody (F8/86; Dako, Carpeneria, Calif), and EPCR using gt αpgtEPCR, a goat anti-human EPCR polyclonal antibody.14 Sections were incubated with an appropriate biotinylated secondary antibody and then horseradish peroxidase-labeled streptavidin. Peroxidase activity (red reaction product) was revealed by amionethylcarbazole (Dako).
Quantification of TM and vWF expression was performed on the above sections using computerized digital threshold analysis. Microscopic images of immunostained sections were captured using a Spot RT digital camera (Diagnostic Instruments Inc) mounted on an Olympus BX60 microscope (Olympus Optical Co). Images were normalized for luminosity and the intima masked to exclude nonspecific background staining. The degree of red spectrum-specific intensity reaching a prespecified threshold was quantified using SigmaScan Pro 5.0 software (SPSS). The number of pixels reaching threshold in the masked intima layer was divided by the lumen perimeter. The normalized values for TM and vWF expression in each section were used to calculate a vessel mean. The ratio of TM to vWF staining was determined for each vessel and normalized to the average value for ungrafted jugular veins.
Western Blot Analysis
Vein grafts and control vessels were harvested at the indicated times, minced, and sonicated in lysis buffer containing 1% Triton X-100, 100 μg/mL PMSF, and 0.1 mol/L NaCl in 20 mmol/L Tris-Hcl (pH 7.5). After centrifugation, lysates were assayed for total protein content using a BCA protein assay kit (Pierce). One microgram of each sample was electrophoresed through a 10% to 20% gradient SDS-polyacrylamide gel (Bio-Rad) and then transferred to an Immobilon-P membrane (Millipore) using a Trans-Blot Electrophoretic Transfer Cell (Bio-Rad), according to the manufacturer’s instructions. After overnight blocking, duplicate blots were incubated with 1:1000 dilutions of either a goat anti-rabbit TM polyclonal antibody (No. 236; American Diagnostica) or a mouse anti-human CD31 monoclonal antibody (JC/70; Dako) followed by incubation with a 1:25 000 dilution of an appropriate horseradish peroxidase-labeled secondary antibody (Amersham). TM and CD31 bands were detected by autoradiography using enhanced chemiluminescence (ECL-Plus; Amersham) and quantified by densitometric analysis using UN-SCAN-IT software (Silk Scientific).
In Situ Protein C and Thrombin Activity Assays
Measurement of APC-generating capacity was performed on freshly excised grafts and control vessels mounted in a modified template as previously described.15 Because of the small size of rabbit carotid arteries, segments of aorta were used as arterial controls. The template was constructed by removing the bottom of a polystyrene 96-well plate. Vessels were positioned in the template such that the bottom of the well was formed by the luminal surface. Vessels were incubated at 37°C with 40 nmol/L human α-thrombin (Boehringer Mannheim, Indianapolis, Ind) and 1 μmol/L human protein C (American Diagnostica) in 250 μL Hanks’ buffered salt solution (HBSS; Life Technologies). After 90 minutes, the thrombin was neutralized by the addition of 5 μL of 50 mg/mL lepirudin (Hoechst Marion Roussel) and 150-μL aliquots were removed and incubated at room temperature with 25 μL of a 3 mmol/L solution of the chromogenic substrate, S-2366 (Chromogenix). The rate of conversion of the substrate was determined spectrophotometrically using a Vmax Kinetic Microplate Reader (Molecular Devices), and the amount of protein C activation was determined by comparison to a standard curve of human APC (American Diagnostica).
Measurement of bound thrombin activity was determined using modifications of previously described methods.16,17⇓ Freshly excised vessels were washed with HBSS and placed in the template as described above. Because of the technical inability of placing rabbit jugular vein segments into the template reliably without exposing their adventitial surface to the substrate, rabbit inferior vena cavae (IVC) were used as venous controls. Vessels were incubated at 37°C with 333 μmol/L of the chromogenic substrate, S-2238 (Chromogenix), in 250 μL of 50 mmol/L Tris-HCl, 175 nmol/L NaCl, and 2 mmol/L CaCl2 (pH 7.8). After 30 minutes, the supernatants were removed and the conversion of substrate was determined spectrophotometrically. Vessel segments were then incubated for 5 minutes with excess lepirudin, washed with HBSS, and incubated a second time with S-2238. The difference in absorbance at 405 nm before and after lepirudin treatment was taken to represent thrombin-specific conversion of the substrate. Bound thrombin activity was quantified by comparison to a human α-thrombin standard curve. Because of differences between the activities of soluble and bound thrombin, unit values reported are only approximately equivalent to NIH units.17
In Vivo Human TM Expression
Untransduced and adenovirus-transduced grafts were harvested at the indicated times, minced, and sonicated in lysis buffer containing 0.2% Triton X-100 and 100 mmol/L K2HPO4 (pH 7.8). After centrifugation, lysates were assayed for human TM (hTM) using an Immubind Soluble Thrombomodulin ELISA kit (American Diagnostica) and normalized to total protein. The Immubind ELISA specifically recognized hTM but not rabbit TM, with a lower limit of detection equal to 0.625 ng/mL.
Measurement of Neointima Formation
Untransduced and adenovirus-transduced grafts were harvested at the indicated times, perfusion-fixed, and embedded in paraffin. Three to seven transverse sections were cut as described above and stained with Movat’s stain. Microscopic images were captured and calibrated for distance and area. The lumen, neointima, and media areas were masked and measured using SigmaScan Pro 5.0 software. Average neointima thickness was calculated by this formula:
To compensate for variations in graft size, neointimal area was normalized to the maximal area bounded by the internal elastic lamina (IEL) calculated from measurements of IEL perimeter. Mean values for each graft were determined from the separate sections and used to calculate a group mean.
All data are presented as mean±SEM. Where indicated, comparisons between two groups were by two-tailed t test and comparisons between multiple groups by one-way ANOVA. A P<0.05 was considered statistically significant.
TM and EPCR Protein Expression After Vein Graft Implantation
Alterations in the expression of endothelial cell TM and EPCR were assessed in vein grafts after implantation. Formalin-fixed autologous rabbit jugular vein grafts, harvested between 1 and 42 days after implantation, were subjected to immunohistochemical analysis with specific TM and EPCR antibodies. TM expression by the luminal endothelium was visibly reduced in grafts harvested on day 3 compared with those harvested on day 42 and ungrafted jugular vein and aorta controls (Figure 1). In contrast, EPCR expression did not seem to change appreciably over time. Preserved endothelial cell integrity at all time points was verified by immunostaining for vWF in adjacent sections, as well as by scanning electron microscopy and immunostaining for CD31 of selected cryofixed grafts (data not shown).
To quantify changes in TM expression, Western blot analysis was performed on whole vessel lysates (Figure 2). TM protein expression declined precipitously early after implantation, reaching a nadir between 3 and 7 days after implantation to <3% that of ungrafted jugular veins and carotid arteries (P<0.02). In contrast, vein grafts reimplanted into the venous circulation for 3 days (VVG) expressed 30-fold more TM than day-3 arterial grafts (P<0.01), indicating that the surgical procedure itself was not a major contributor to TM loss. TM expression in day-42 grafts was ≈60% that of ungrafted jugular veins, although the difference did not reach statistical significance (P=0.06). Similar quantification of EPCR expression was not performed because of the large amount of soluble protein uptake by inflammatory cells adherent to or infiltrating the vessel wall.18
Several weeks after implantation, vein grafts were observed to develop adventitial capillaries that express TM (Figure 3A). To differentiate TM expression by the luminal endothelium from that of the adventitial capillaries, digital threshold analysis of the intima layer was performed on graft sections immunostained for TM and vWF (Figure 3B). By this method, intimal TM expression was also reduced early after implantation compared with ungrafted jugular vein and carotid artery controls. Similar to Western blot analysis, TM expression in day-42 grafts remained substantially below that of ungrafted jugular vein and carotid artery controls.
Loss of TM Expression Is Associated With Decreased Graft Thromboresistance
We investigated the functional consequences of diminished TM expression on vein graft thromboresistance. This was accomplished by in situ measurements of the APC-generating capacity of the luminal endothelium and of the amount of thrombin activity bound to the vessel wall, a reflection of local thrombin activation (Figure 4). Similar to TM protein expression, the APC-generating capacity of vein grafts was markedly reduced early after implantation, with gradual, but incomplete, return at later time points. APC generation by day-7 grafts was only 5% that of jugular vein controls (P<0.001). Even after 42 days, the capacity of grafts to generate APC was only 46% and 68% that of aorta and jugular vein controls, respectively (P<0.01 for each). Early loss of APC-generating capacity was associated with a marked increase in bound thrombin activity appearing on the graft luminal surface, indicative of increased local thrombin generation. Bound thrombin activity peaked on day 7 at 28.7±3.8 mU/cm2 of graft surface area, with gradual return to baseline levels by day 28. This degree of bound thrombin activity is similar to that observed after acute balloon injury of rabbit aortas.16,17⇓
Restoration of TM Expression Preserves Graft Thromboresistance
To determine a causal relationship between loss of APC-generating capacity and local thrombin generation, TM expression was restored using adenovirus-mediated gene transfer. A first-generation adenovirus vector, AdTMh5, was constructed that contained the full-length cDNA encoding for hTM. Transduction of vascular cells in vitro with AdTMh5 resulted in a dose-dependent increase in both the expression of hTM antigen and capacity to generate APC (data not shown). Preliminary in vivo dosing studies, using an identical vector expressing the nuclear-targeted β-galactosidase gene, revealed that the optimal dose for gene transfer to rabbit vein grafts using our protocol was 1010.5 pfu/mL. At this dose, there is substantial transduction of graft endothelial cells with moderate transduction of medial smooth muscle cells (data not shown).
The expression of hTM in vein graft lysates was quantified using an ELISA that specifically detects human, but not rabbit, TM. Transduction of rabbit vein grafts with AdTMh5 at the time of implantation resulted in a maximal expression on day 7 of 257±49 ng hTM/mg total graft protein (Figure 5). Significant hTM protein expression persisted for more than 14 days and was detectable in low quantities even after 42 days. Untransduced grafts and grafts transduced with AdNull-1, an identical adenovirus vector expressing no transgene, had no detectable hTM at any time point.
Transduction of vein grafts with AdTMh5 restored their capacity to generate APC (Figure 6). APC generation paralleled recombinant hTM expression, peaking on day 7 to levels that were 17 and 40 times higher than AdNull-1-transduced and untransduced grafts, respectively (P<0.01 for each). At all time points beyond 3 days, APC generation was significantly higher for AdTMh5-transduced grafts than for control grafts (P≤0.03 by ANOVA). Restoration of graft APC-generating capacity significantly reduced local thrombin generation. Transduction of grafts with AdTMh5 reduced bound thrombin activity on day 7 by 75% to 80% (P<0.03 for AdTMh5 versus each control group) to levels found in control veins (Figure 7). Bound thrombin activity was uniformly low in AdTMh5-transduced grafts harvested earlier on day 3 (8.3±1.9 mU/cm2) and later between 14 and 42 days (range 3.9±0.4 to 8.3±2.0 mU/cm2), demonstrating the consistency and persistence of this effect.
Effect of Thrombin Inhibition on Neointima Formation
Inhibition of local thrombin activity has been demonstrated to reduce neointimal hyperplasia after acute arterial injury in several animal models.19 Given that restoration of TM expression significantly reduced local thrombin activity in vein grafts, we investigated whether this might also reduce neointima formation. Average neointimal thickness and normalized neointimal area were measured using digital morphometric analysis of grafts that were transduced with no virus, AdNull-1, or AdTMh5 and harvested after 42 days (Figure 8). Despite inhibiting local thrombin generation, there were no reductions in neointimal thickness or area. There was also no difference in neointima to media area ratios between groups, although difficulties in accurately identifying the external elastic lamina in rabbit vein grafts make this a relatively unreliable measurement of neointimal hyperplasia (data not shown).
The major findings of the present study are (1) the expression of TM, but not EPCR, is dramatically reduced early after vein graft implantation and exhibits only partial recovery after 6 weeks; (2) loss of TM expression diminishes the capacity of grafts to generate APC, which facilitates unopposed local thrombin generation; (3) TM expression and activity can be restored using adenovirus-mediated gene transfer; and (4) local thrombin generation does not seem to promote vein graft neointima formation, as it does after arterial injury.
Local Thrombin Generation in Vein Grafts
Thrombosis is the major cause of early vein graft failure, yet little is known from in vivo studies about factors that might predispose vein grafts to thrombus formation. Acute arterial injury is known to cause tissue factor-mediated activation of the coagulation system resulting in local thrombin generation.20 Measurements of thrombin activity bound to the wall of balloon-injured rabbit arteries reveals that local thrombin generation peaks within the first 24 to 48 hours after acute injury but may persist at low levels for more than 7 days.16,17,21⇓⇓ Similar measurements in vein grafts have not previously been reported. In the present study, the extent of local thrombin generation in vein grafts was found to be of the same magnitude as in injured arteries but with a later peak and of longer duration. This suggests that vein grafts are subjected to a significant but more chronic form of injury.
Detailed morphological studies of rabbit vein grafts demonstrate significant inflammation and evidence of endothelial injury, but not of substantial endothelial cell loss, during the first 2 weeks after implantation.12,22⇓ Scanning electron micrographs performed in our laboratory confirm these findings and also reveal a carpeting of platelets and fibrin over the graft luminal surface during this period (data not shown). Channon et al23 found a marked increase in tissue factor expression within the intima of rabbit vein grafts harvested early on day 3, but not in grafts harvested later on days 14 or 28. Interestingly, tissue factor expression was found to colocalize to infiltrating leukocytes, not to graft endothelial cells. These findings suggest the critical importance of the inflammatory response in initiating the shift toward a local procoagulant environment.
Although nascent tissue factor expression may initiate local thrombin generation, alterations in endothelial cell thromboresistance are likely to contribute to its magnitude and persistence. Two elements from the present study support this concept. The first is that local thrombin generation continues well beyond 14 days, correlating more closely with loss of APC-generating capacity than with tissue factor expression. The second is that restoration of TM expression with adenovirus vector-mediated gene transfer was able to completely blunt local thrombin generation. It clearly would have been desirable to determine whether the restoration of TM could reduce acute graft thrombosis. Unfortunately, the model used for the present study is inadequate for this purpose and no widely accepted models of acute vein graft thrombosis currently exist.
Although TM seems to be an important contributor to vein graft thromboresistance, alterations in the expression or activity of other molecules may also predispose grafts to early thrombosis. For example, Mann et al24 found that stimulated nitric oxide production by rabbit vein grafts harvested at 28 days was <50% that of ungrafted jugular veins and <30% that of carotid arteries. This study did not specifically address how the loss of nitric oxide production might contribute to vein graft failure. Nevertheless, it is consistent with our data in suggesting that a prolonged period of endothelial dysfunction exists after vein graft implantation. Further studies are needed to characterize other molecular changes occurring as part of this maladaptive process and to determine when, if ever, vein grafts become fully “arterialized.”
Mechanism of Thrombomodulin Loss
The mechanism responsible for the loss of TM protein expression after vein graft implantation is likely to be complex and multifactorial. Circumstantial evidence suggests that inflammation may play an important role. It is known, for example, that TM can be cleaved and shed from the surface of endothelial cells by granulocyte elastase that is secreted by activated neutrophils.25 Inflammatory cytokines, particularly tumor necrosis factor-α (TNF-α), are known to be potent inhibitors of TM gene expression.26,27⇓ Recent in vivo studies have also demonstrated that TNF-α potentiates neutrophil-mediated endothelial cell injury and shedding of surface TM.28 These findings are particularly relevant given that rat vein grafts explanted at 4 weeks and placed in organ culture elaborated nearly 10-fold more TNF-α than do control veins or vein grafts reimplanted into the venous circulation.29
TM expression may be altered by the exposure of vein grafts to arterial pressure and flow. Endothelial cells in culture are known to downregulate TM gene expression in response to elevated shear stress or cyclical strain.30,31⇓ Gosling et al31a found that placing freshly excised human saphenous vein segments under arterial flow conditions for 90 minutes using an ex vivo perfusion system resulted in a 40% reduction in TM protein expression as detected by quantitative immunohistochemical analysis. Although changes in TM gene expression were not measured, there was no detectable increase in soluble TM in the graft perfusate to suggest significant shedding of the surface protein. More sophisticated studies isolating the effects of mechanical forces from those of inflammation will be required to determine the relative contributions to the loss of in vivo TM expression.
Effect of Thrombin Generation on Neointima Formation
Neointimal hyperplasia is a major cause of late vein graft failure.1,2⇓ Local thrombin generation contributes to neointima formation after acute arterial injury19 and is often assumed to have similar effects in vein grafts. Recent in vitro studies seem to support this concept by demonstrating that thrombin is significantly more mitogenic for smooth muscle cells derived from human saphenous veins than from internal mammary arteries.32 It was therefore surprising that restoring TM expression, with documented blunting of local thrombin activation, did not reduce neointima formation. These results contrast the findings in balloon-injured rabbit arteries where local adenovirus-mediated TM expression did result in a 38% reduction in the average neointima to media area ratio.33 Although it is conceivable that soluble fragments of TM or small amounts of residual thrombin activity may promote neointimal growth,34 a more plausible explanation for our results is that local thrombin generation is not a significant stimulus for neointimal hyperplasia in vein grafts, as it is in arteries after acute injury. Experiments using alternative methods of inhibiting local thrombin activation, ie, with a direct thrombin inhibitor, might help to confirm this finding.
Early vein graft thrombosis remains a significant cause of morbidity and mortality in patients undergoing peripheral and cardiac bypass surgery. The present study demonstrates that TM is an important contributor to vein graft thromboresistance and that its loss leads to enhanced local thrombin generation. The mechanism of TM loss is likely complex and will require further investigation to be more clearly defined. Strategies that either prevent the loss of TM expression or restore its functional activity may hold promise in the treatment of patients with vein graft disease.
This work was supported by an American Heart Association Scientist Development Grant (J.J.R.), a grant from the W.W. Smith Charitable Trust (J.J.R.), a National Research Service Award (HL-10250-01; J.L.S.), and generous gifts from the families of Peter Belfer and John Maguire. Special thanks goes to Melissa Haggerty for her technical assistance with the animal surgery.
This manuscript was sent to Donald D. Heistad, Consulting Editor, for review by expert referees, editorial decision, and final disposition.
Original received July 19, 2001; resubmission received November 13, 2001; accepted December 20, 2001.
- 2.↵Motwani JG, Topol EJ. Aortocoronary saphenous vein graft disease. Circulation. 1998; 97: 916–931.
- 7.↵Stearns-Kurosawa DJ, Kurosawa S, Mollica JS, Ferrell GL, Esmon CT. The endothelial cell protein C receptor augments protein C activation by the thrombin-thrombomodulin complex. Proc Natl Acad Sci U S A. 1996; 93: 10212–10216.
- 9.↵Taylor FB, Stearns-Kurosawa DJ, Kurosawa S, Ferrell G, Chang ACK, Laszik Z, Kosanke S, Peer G, Esmon CT. The endothelial cell protein C receptor aids in host defense against Escherichia coli sepsis. Blood. 2000; 95: 1680–1686.
- 13.↵Hardy S, Kitamura M, Harris-Stansil T, Dai Y, Phipps ML. Construction of adenovirus vectors through Cre-lox recombination. J Virol. 1997; 71: 1842–1849.
- 16.↵Ghigliotti G, Waissbluth AR, Speidel C, Abendschein DR, Eisenberg PR. Prolonged activation of prothrombin on the wall after vascular injury. Arterioscler Thromb Vasc Biol. 1998; 18: 250–257.
- 18.↵Kurosawa S, Esmon CT, Stearns-Kurosawa DJ. The soluble endothelial cell protein C receptor binds to activated neutrophils: Involvement of proteinase-3 and CD11b/CD18. J Immunol. 2000; 165: 4697–4703.
- 19.↵Patterson C, Stouffer GA, Madamanchi N, Runge MS. New tricks for old dogs:5 nonthrombotic effects of thrombin in vessel wall biology. Circ Res. 2001; 88: 987–997.
- 20.↵Speidel CM, Eisenberg PR, Ruf W, Edgington TS, Abendschein DR. Tissue factor mediates prolonged procoagulant activity on the luminal surface of balloon-injured aortas in rabbits. Circulation. 1995; 92: 3323–3330.
- 21.↵Barry WL, Gimple LW, Humphries JE, Powers ER, McCoy KW, Sanders JM, Owens GK, Sarembock IJ. Arterial thrombin activity after angioplasty in an atherosclerotic rabbit model. Circulation. 1996; 94: 88–93.
- 23.↵Channon KM, Fulton GJ, Davies MG, Peters KG, Ezekowitz MD, Hagen P-O, Annex BH. Modulation of tissue factor protein expression in experimental venous bypass grafts. Arterioscler Thromb Vasc Biol. 1997; 17: 1313–1319.
- 27.↵Moore KL, Esmon CT, Esmon NL. Tumor necrosis factor leads to the internalization and degradation of thrombomodulin from the surface of bovine aortic endothelial cells in culture. Blood. 1989; 73: 159–165.
- 30.↵Malek AM, Jackman RW, Rosenberg RD, Izumo S. Endothelial expression of thrombomodulin is reversibly regulated by fluid shear stress. Circ Res. 1994; 74: 852–860.
- 31.↵Feng Y, Yang J-H, Huang H, Kennedy SP, Turi TG, Thompson JF, Libby P, Lee RT. Transcriptional profile of mechanically induced genes in human vascular smooth muscle cells. Circ Res. 1999; 85: 1118–1123.
- 31A.↵Gosling M, Golledge J, Turner RJ, Powell JT. Arterial flow conditions downregulate thrombomodulin on saphenous vein endothelium. Circulation. 1999; 99: 1047–1053.
- 32.↵Yang Z, Ruschitzka F, Rabelink TJ, Noll G, Julmy F, Joch H, Gafner V, Aleksic I, Althaus U, Luscher TF. Different effects of thrombin receptor activation on endothelium and smooth muscle cells of human coronary bypass vessels. Circulation. 1997; 95: 1870–1876.
- 33.↵Waugh JM, Li-Hawkins J, Yuksel E, Kuo MD, Cifra PN, Hilfiker PR, Geske R, Chawla M, Thomas J, Shenaq SM, Dake MD, Woo SLC. Thrombomodulin overexpression to limit neointima formation. Circulation. 2000; 102: 332–337.
- 34.↵Tohda G, Oida K, Okada Y, Kosaka S, Okada E, Takahashi S, Ishii H, Miyamori I. Expression of thrombomodulin in atherosclerotic lesions and mitogenic activity of recombinant thrombomodulin in vascular smooth muscle cells. Arterioscler Thromb Vasc Biol. 1998; 18: 1861–1869.