Expression of the plasminogen activator system in the human
vascular wall
M.Y. Salame *, N.J. Samani, I. Masood, D.P. deBono
Department of Cardiology,Glenfield General Hospital,Leicester,England, UK
Received 15 January 1999; received in revised form 7 October 1999; accepted 25 October 1999
Abstract
In this paper we describe the expression of the tissue plasminogen activator (tPA), urokinase-type plasminogen activator (uPA), plasminogen activator inhibitor-1 (PAI-1) and the uPA receptor (uPAR), in normal and atheromatous human vascular tissue obtained at coronary and peripheral vascular surgery. tPA, uPA, PAI-1 and uPAR antigens were localised by immunohistochem-istry. Vessel homogenates were used to quantitate tPA, uPA and PAI-1 antigens as well as uPA and PAI-1 activities using immunoassay and immunoactivity assays, respectively. Quantitative reverse transcription polymerase chain reaction assays (PAI-1 and uPA) were developed and used to quantify PAI-1 and uPA mRNA. In-situ hybridisation (tPA, uPA and PAI-1) was used to localise mRNA. In normal saphenous vein or internal mammary artery, expression of tPA, uPA and PAI expression is associated with endothelium and with intimal or medial smooth muscle cells, but expression is at a low level. uPAR protein was seen on the endothelium of normal saphenous vein or internal mammary artery but absent on the smooth muscle cells. In complex atheroma tPA, uPA, PAI and uPAR proteins were associated with the endothelium, groups of smooth muscle cells (in the intima and around vascular channels, but not with the media), infiltrating mononuclear cells, and also with acellular areas. PAI-1, tPA and uPA mRNA were demonstrated in atheroma in endothelial cells and smooth muscle cells, as well as in areas rich in macrophages. In stenosing saphenous vein grafts there was strikingly increased tPA and uPA (but not PAI-1) expression in neointimal smooth muscle cells and migrating SMC at the intima/media border. A major difference between complex atheroma and either normal vessel or saphenous vein grafts was greatly increased expression of PAI-1 mRNA associated with smooth muscle cells (SMC) in the former. In spite of the greatly increased PAI-1 mRNA expression in atheromatous lesions, the immunoactivity assay showed PAI-1 activity to be low compared to normal internal mammary artery. Our findings would be compatible with previous reports implicating the plasminogen activator/inhibitor system in the initiation and control of matrix remodelling during normal and pathological vessel growth and repair, but also emphasize the complexity of this process in human vessels. © 2000 Elsevier Science Ireland Ltd. All rights reserved.
Keywords: Atherosclerosis; Plasminogen activator; Urokinase-type plasminogen activator (uPA); Urokinase-type plasminogen activator (uPA) receptor; PAI-1; Saphenous veins
www.elsevier.com/locate/atherosclerosis
1. Introduction
The ability of vessel wall tissue to initiate fibrinolysis, originally demonstrated by Astrup [1] and by Todd [2], was subsequently shown to be due to the production of both tissue type- (tPA) and urokinase type- (uPA) plasminogen activators [3 – 5]. Plasminogen activator
expression by endothelial cell was initially linked teleo-logically to the prevention of vascular thrombosis. However, Clowes and colleagues advanced the hypoth-esis that increased plasminogen activator expression in the vessel wall is also important in vascular remodelling by facilitating vascular smooth muscle cell migration and cell proliferation by initiating a proteolytic cascade of matrix degradation [6]. In keeping with this hypothe-sis, are the findings that vessel wall plasminogen activa-tor activity is increased after experimental vascular injury and also in human atheroma [7 – 9]. It has been suggested that plasminogen activator expression by mononuclear cells may be important in destabilising the * Corresponding author. Present address: Andreas Gruentzig
Car-diovascular Center, Emory University Hospital, 1364 Clifton Road, N.E., Suite F606, Atlanta, GA, USA. Tel.: +1-404-712-7821; fax
+1-404-712-5622.
E-mail address:[email protected] (M.Y. Salame).
M.Y.Salame et al./Atherosclerosis152 (2000) 19 – 28
20
fibrous caps of atheromatous plaques [10,11]. Plasmino-gen activator inhibitor type 1 (PAI-1) is also present in atheromatous lesions [12,13]. Further information on the role of the plasminogen activator system in the vessel wall has been obtained from genetic manipula-tion studies. uPA gene deficient (but, curiously, not tPA deficient) mice demonstrated impaired healing of arte-rial wall injuries [14]. Over-expression of PAI-1 in transgenic mice carrying human cDNA resulted in thrombotic problems in these animals, whereas me-chanical vessel injury in gene knockout mice deficient in PAI-1 resulted in increased neointimal formation [15]. Given the emerging putative role of the plasminogen activator system in the blood vessel wall, we studied the expression of tPA, uPA, PAI-1 and uPAR, in normal human artery and veins (internal mammary artery and saphenous veins, respectively) as well as in arteries and veins that have undergone atheromatous degeneration development (atheromatous coronary and carotid en-darterectomies) and intimal hyperplasia development (saphenous vein graft stenosis). Immunohistochemistry was used to localise tPA, uPA, PAI-1 and uPAR anti-gens. Immuno(antigen) assays and immuno(activity) as-says were performed on filtered vessel homogenates to quantitate tPA, uPA and PAI-1 antigens as well as uPA and PAI-1 activities. tPA, uPA and PAI-1 mRNA were localised by in-situ hybridisation whereas quantitative reverse transcription polymerase chain reaction (RT-PCR) assays were developed and used to quantify PAI-1 and uPA mRNA.
2. Materials and methods
Saphenous veins, internal mammary arteries, coro-nary and carotid endarterectomies and stenoses of saphenous vein grafts, were obtained from patients undergoing routine coronary, carotid or peripheral vas-cular surgery. Tissues allocated for mRNA in-situ hy-bridisation and immunohistochemistry were fixed in 10% formol saline, dehydrated in graded alcohol and embedded in paraffin wax. Tissue for RTPCR and protein analyses were snap-frozen in liquid N2 and stored at −80°C until analysed.
2.1. Immunohistochemistry
Immunohistochemical studies were performed on
4-mm paraffin sections using mouse monoclonal
antibod-ies to tPA, uPA, PAI-1 and uPAR. Antibodantibod-ies and optimum conditions for the primary antibody incuba-tions were as follows. tPA: PAM3 (Biopool, Sweden) at 1:25 dilution for 1 h at 37°C; uPA: c3689 (American Diagnostics) at 1:25 dilution for 1 h at 37°C; PAI-1: MAI-12 (Biopool, Sweden) at 1:100 dilution for 1 h at 25°C; uPAR: c3936 (American Diagnostics) at 1:25
dilution for 1 h at 37°C. Cell-type characterisation was done using: antibodies to a-actin (M0851, Dako) for
smooth muscle cells; von Willebrand factor (A080, Dako), and CD31 (M0823, Dako) for endothelial cells; CD68 (Dako) for macrophages; and CD45 (Dako) (leukocyte common antigen) for lymphocytes. The sec-ondary antibody was biotin-labelled sheep anti-mouse antibody (Amersham, UK) at 1/200 dilution for all the immunohistochemical analyses except those for vWF where sheep anti-rabbit was used (1:200). ABC (Dako) and 3,3%-diamino-azobenzidine were used for detection
of antigen and sections were counter-stained in Mayer’s haematoxylin. Negative controls used consisted of (i) omission of the primary antibody, and (ii) replacing the primary antibody with non-immune serum.
2.2. RNA extraction from 6essel wall
Total RNA was extracted from tissue using the RNAzol B method [16]. The RNA was quantified by optical density at 260 A, and its integrity assessed by running 5 mg of the RNA on a 0.8% agarose gel.
Extracted RNA were stored as 3-mg aliquots at −80°C
for the RTPCR and used once only after thawing
2.3. Specific mRNA quantification
Quantitative RTPCR was used to quantitate the level of expression of uPA and PAI-1 mRNA in normal and atheromatous vessels. The design of the RTPCR is shown schematically in Fig. 1. The plasmids used are shown in Table 1. The oligonucleotide primers used were: uPA primer A (CTGTAATACGACTCAC-TATAGGGGGCACCGG), uPA primer B (TC-CGGATAGAGATAGTCGGTGTGGTGAGCAAG), uPA primer 1 (GCTCTGTCACCTACGTGTG), uPA primer 2 (TCCGGATAGAGATAGTCGG), PAI-1 primer A (AACCCAGCAGCTAATACGACTCAC-TATAGGGGTTC), PAI-1 primer B (CTGGCCGTT-GAAGTAGAGGCTGGCTCTCTC), PAI-1 primer 1
(GGAACAAGGATGAGATC AGC), and PAI-1
primer CTGGCCGTTGAAGTAGAGG).
2.4. Synthesis of the RNA internal standards(mimic)
Fig. 1. Schematic diagram to show the design of the reverse transcriptase polymerase chain reaction with internal controls used to quantitate PAI-1 and uPA mRNA.
2.5. Re6erse transcription
Aliquots (0.5mg) of vessel RNA spiked with known
amounts of internal standard RNA (in serial dilutions) were denatured by heating, quick-chilled on ice and added to the reaction mixture (see below) and incu-bated at 37°C for 1 h. Reaction mixture: 2.0 ml of 5×
buffer; 0.2ml of 0.1 M DTT; 0.1ml of RNAse inhibitor;
0.1ml of MMLV enzyme; 1ml of dNTP (10 mM mix);
10 pmol primer 2; and DEPC water made up to 10.0ml.
2.6. Polymerase chain reaction
Of the resulting cDNA 2.0ml was used for the PCR.
Each tube contained: 1.5ml of NTP mix (2.5 mM each);
2.0 ml of 10× buffer; 0.6 ml of MgCl2 (50 mM) (1.5 mM final concentration), 0.2 ml of Taq polymerase,
15.0 pmol of upstream primer (primer 1), 13.0 pmol of downstream primer (primer 2), water 10.8ml. Optimum
cycle numbers were 26 for PAI-1 and 31 for uPA and an annealing temperature of 60°C was used. Negative controls were (1) water instead of cDNA and (2) water instead of upstream primer.
PCR products were identified on electrophoresis by size and by Southern blotting and hybridising with labelled cDNA probes that were derived from the spe-cific plasmids. Pilot studies were performed to deter-mine the appropriate concentrations of RNA standard, and to confirm exponential phase amplification of cDNA from both tissue mRNA and RNA standard.
Quantitative RTPCR was carried out in the presence of 32P dCTP, and products run on an polyacrylamide gel
which was dried, exposed to film and the resulting autoradiograph scanned. The amount of tissue mRNA was derived from a plot of the intensity of signal of the internal standard/tissue signal (y-axis) versus internal standard (x-axis). The amount of specific mRNA in the RNA extract was derived from the equivalence point, i.e. the amount of internal standard needed to produce the same signal intensity as the tissue signal (corrected for the difference in number of dCTP in standard DNA and tissue cDNA).
2.7. mRNA in-situ hybridization
The paraffin-embedded sections on slides were de-waxed in xylene, re-hydrated, incubated with 150 ml of
Proteinase K (Boehringer Mannheim, 20ml/ml) at 37°C
Table 1
Structure of the plasmids used
Restriction Plasmid cDNA
cDNA Bacteria
fragment enzyme
tPA pBR322 2.0 kilobase Bgl II NM522 pairs
pSP64 1426 base TG1
uPA Hind III
pairs
Eco R1 TG1
2.5 kilobase PUC
PAI-1
M.Y.Salame et al./Atherosclerosis152 (2000) 19 – 28
22
for 1 h and refixed in 0.4% paraformaldehyde in PBS for 20 min at 4°C. An aliquot of salmon sperm DNA (0.5 ml of 10 mg/ml), incubated at 95°C for 10 min was added to 10 ml of solution A (see below) to form the pre-hy-bridisation solution which was applied to the slides for a 3-h incubation at 42°C. Solution A was made up in the following proportions: 6.0ml of 20× SSPE, 12.5ml of
100% formamide, 1.25ml of 100× Denhardt’s solution,
4.50 ml PEG mix (1.25 ml 10% SDS, 2.25 ml DEPC-treated water, 1.5 g 6% PEG 6000). Hybridisation solution (i.e. solution A with labelled probe) was added to the slides, and incubated overnight at 42°C. Slides were then washed twice in 3× SSPE/0.1% SDS, once in 1× SSPE/0.1% SDS, dehydrated in graded alcohols, dried in air, then exposed to photographic emulsion (Kodak K5 Emulsion, Ilford, UK) according to manu-facturer’s instructions. Slides were placed in a light-tight box for 21 days, removed under safe-light and developed using Developer Kodak D19 following the manufactur-er’s instructions and counter-stained in haematoxylin.. Negative control slides for the mRNA hybridisation were treated with RNase (Sigma) in 2× SSC for 1 h at 37°C.
2.8. Immuno(antigen) assays and immuno(acti6ity)
assays
Tissues were homogenised in 1 ml of acid acetate extraction buffer (75 mM acetic acid, 225 mM NaCl, 75 mM KCl, 10 mM EDTA, 100 mM arginine, and 0.25% (vol/vol) Triton X-100, pH 4.2) and the homogenate centrifuged at 4°C, the supernatant filtered using 0.45-mm
filters and stored at−80°C until needed. Antigens (tPA, uPA and PAI-1) were measured by enzyme linked immunosorbent as says (ELISA) using Imulyse tPA (Biopool), TintElize uPA (Biopool) and TintElize PAI-1 (Biopool), respectively, all according to the manufactur-er’s instructions using the double antibody principle [17]. uPA and PAI-1 activities of the filtered vessel ho-mogenates were measured using Chromolyse uPA®
and Spectrolyse®
pL PAI-1 (Biopool), respectively.After ini-tial titration experiments, 10ml of filtered homogenate
was required for the tPA and PAI assays, whereas 30ml
was needed for uPA. The quantities of antigens (tPA, uPA and PAI-1) and activities (uPA and PAI-1) in the vessel wall were expressed per wet weight (g or mg) of tissue.
2.9. Statistical analysis
Data are presented as mean9S.E.M. Statistical com-parisons between the different types of vessels were made using Student’st-tests, or when assumptions of normal distribution or equal variance were violated, Mann – Whitney rank sum tests were performed. An alpha level of PB0.05 was considered to indicate a significant difference.
3. Results
3.1. Histology immunohistology and in-situ hybridisation
Normal saphenous veins (Fig. 2) have an intima of variable thickness covered by endothelium and sepa-rated from the media by a rudimentary internal elastic lamina (elastin is stained blue in Fig. 2b). The media consists of multiple layers of a-actin positive smooth
muscle cells, separated by bundles of collagen, ground substance and occasional short elastic fibres (Fig. 2b). tPA, uPA and PAI-1 antigens were associated with the endothelium of the main lumen (Fig. 2d, e, j) and vasa vasorum (not shown) as well as with occasional smooth muscle cells of the media and intima. Compared to undistended veins, veins that had been surgically dis-tended had greatly increased immunostaining (Fig. 2 c – j). PAI-1 mRNA (but not tPA or uPA mRNA) was just detectable in endothelial and smooth muscle cells (Fig. 2k).
Internal mammary arteries (Fig. 3) have a well-defined internal elastic lamina separating the media from the intima that was thin and largely devoid of smooth muscle cells (Fig. 3a, e). In places there is duplication of the internal elastic lamina (Fig. 3b). The spindle shapes of the smooth muscle cells are less apparent in the media due to the abundance of elastic tissue with consequent recoil (Fig. 3a). The spindle-shaped cells are immunopositive (brown staining) for
a-actin confirming that they are smooth muscle cells
(Fig. 3b). In normal internal mammary arteries, tPA, uPA and PAI-1 proteins were associated with some smooth muscle cells of the media and vasa vasorum as well as the endothelium of the vessel lumen (Fig. 3d, e) and the vasa vasorum (not shown); however, the level of immunostaining at these sites was not intense. uPAR protein was largely absent within the walls of internal mammary arteries (Fig. 3f). Using the in-situ hybridisa-tion condihybridisa-tions described, there was an absence of signal in microscopically normal internal mammary arteries for tPA, uPA and PAI-1 mRNA (not shown). Endarterectomy specimens from either coronary or carotid arteries had an outer rim ofa-actin
mam-Fig. 2. Immunohistochemistry and in-situ hybridisation, normal saphenous vein. Original magnification×40. The vessels are orientated with the endothelial lining at the top of the photographs with the intima, superficial and deeper aspects of the media underneath and the adventitia not shown. The boundary between the superficial and deep layers of the media is indicated by the symbol *. Brown staining signifies immunopositivity to the specific antigens as labelled. (b) Orientation of thea-actin positive SMCs. Surgically distending the veins clearly and consistently resulted
in the increase in immunostaining for tPA, uPA, PAI-1 and uPAR. (k) Signal (punctate dots) from32P-labelled cDNA probes used in the in-situ hybridisation.
M.Y.Salame et al./Atherosclerosis152 (2000) 19 – 28
mary arteries or undistended saphenous veins, atheroma-tous lesions demonstrated prominent immunostaining for tPA, uPA, PAI-1 and uPAR proteins. tPA was proteins. tPA was predominantly associated with en-dothelial cells and periluminal smooth muscle cells (Fig. 4f). PAI-1, uPA and uPAR proteins were also associated with endothelial cells (Fig. 4e, h) and periluminal smooth muscle cells (Fig. 4e, g, h) but in addition, were also found in areas rich in macrophages (Fig. 4d). Unlike normal internal mammary arteries, atheromatous lesions demonstrated tPA, uPA and PAI-1 mRNA on in-situ hybridisation (Figs 4, j – l). This expression was present in both endothelial cells (not shown) and smooth muscle cells that were periluminal or in pockets within the neointima. PAI-1 and uPA mRNA signals were also strongly observed in the macrophage-rich areas within the neointima.
Stenotic saphenous vein grafts were seen to have a hyperplastic neointima principally composed of smooth muscle cells with stellate- or spindle-shaped nuclei (Fig. 5). The lumen is lined with (CD31-immunopositive) endothelial cells. Only a few macrophages were found in the intimal hyperplasia of saphenous vein graft stenotic lesions, and these were found deep in the vessel wall (not shown). Both the smooth muscle cells within the neointima as well as the endothelial cells lining the lumen were immunopositive for tPA, uPA and uPAR (Fig. 5 c, d, f). The peri-medial part of the neointima is seen to contain smooth muscle cells with their long axes perpen-dicular to the endothelium strongly immunopositive for tPA, uPA (Fig. 5d) and uPAR. PAI-1 immunostaining however, was strongly present on the endothelium but largely absent from the neointima (Fig. 5e). The medial smooth muscle cells were consistently immunonegative for all four proteins (not shown). Unlike undistended saphenous veins, grafted saphenous veins that had under-gone graft stenosis demonstrated mRNA signal (as black punctate dots) for tPA (Fig. 5h) and uPA (Fig. 5i) in both
Table 2
Quantity of PAI-1 and uPA mRNA in vessel wall (mean9S.E.M.)
UPA mRNA content PAI 1 mRNA content
(attomol/pg total (attomol/pg total RNA)
smooth muscle cells and endothelial cells. Strong PAI-1 mRNA signal was associated with endothelial cells (Fig. 5g) with very little signal being associated with the neointimal smooth muscle cells unlike that seen in atheromatous arteries.
3.2. uPA and PAI-1 mRNA in 6essel wall measured by
RTPCR
Average yield of total RNA9S.E.M. was 1791.5
mg/100 mg of tissue for saphenous vein, 19.692.4 for internal mammary artery and 7.991.1 for coronary endarterectomy specimens. Specific mRNA contents for PAI-1 and uPA mRNA9S.E.M. are shown in Table 2. Atheromatous coronary endarterectomies contained sig-nificantly higher levels of both PAI-1 and uPA mRNA expression compared to normal internal mammary artery (P=0.02 andPB0.0001, respectively). Surgical disten-sion of veins resulted in significant falls in PAI-1 and uPA mRNA (P=0.004 and 0.04, respectively). PAI-1 expres-sion was significantly higher than uPA in all three tissues (PB0.05). When normal saphenous veins were com-pared to internal mammary arteries, there was no signifi-cant difference in the level of PAI-1 mRNA but there was
Fig. 4. Immunohistochemistry and in-situ hybridization, complex atheroma (coronary endarterectomy). Brown staining signifies immunopositivity to the specific antigens as labelled. (a) Low magnification showing the rim of brown-stained a-actin positive SMCs at the periphery of the
endarterectomy specimen (at the bottom of photograph) where the media was sheared during extraction. On higher magnification, SMCs were seen to be present in sheets in the neointima as well as in the periluminal areas. (b) Same specimen,×40 view, showing vWF immunostaining of the endothelium at the top of the photograph. (c) Area rich in macrophages surrounding a vascular channel deep in the substance of the atheromatous lesion. (d) The same area as in (c) but stained for uPA, showing co-localisation with increased uPA protein. (e) The same area as in (b) with old media at the bottom and the lumen at the top. Note the increased uPA immunopositivity at the border of the media and the neointima, as well as in spindle-shaped cells in the neointima including periluminally. (f) tPA on periluminal SMCs around small vascular channel. (g) uPAR antigen in the neointima (×100) associated with SMCs and extracellular spaces. (h) PAI-1 antigen on endothelium (at top) and SMC in periluminal area. (i) Heavy staining of PAI-1 deep in the neointima associated with neointimal cells and extracellular matrix. (j – l) In-situ hybridisation for specific mRNA in the neointima.
M.Y.Salame et al./Atherosclerosis152 (2000) 19 – 28
26
significantly more uPA mRNA in saphenous veins (PB0.05).
3.3. Quantity of tPA, uPA and PAI-1 antigens in the
6essel wall measured by immunoassay
Atheromatous coronary endarterectomies contained significantly greater amounts of tPA, uPA and PAI-1 antigen compared to normal internal mammary arteries, 652999 versus 141937 ng/g of tissue,P=0.001; 16.49
2.4 versus 3.990.6 ng/g of tissue, P=0.0003; and 272984 versus 16.392.4 ng/g of tissue, P=0.0004, respectively. Saphenous veins contained significantly more tPA and PAI-1 antigens compared to internal mammary arteries; 311938 versus 141937 ng/g of tissue,P=0.03; and 72911.9 versus 16.392.5 ng/g of tissue PB0.0001, respectively.
3.4. uPA and PAI-1 acti6ities in the6essel wall
Atheromatous endarterectomy specimens (n=7) con-tained higher levels of PAI-1 activity compared to internal mammary arteries (n=10), 6.590.9 versus 1.190.2 U/mg tissue, respectively, P=0.0002. However, uPA activity was not significantly different between atheroma-tous endarterectomy 35.5911.0 ng/g tissue (n=7) and normal internal mammary artery 46.697.0 (n=11),
P=0.29. In comparing conduits used for coronary artery bypass grafting (distended saphenous veins (n=15) and internal mammary arteries (n=10)), PAI-1 activity was significantly lower in the distended saphenous veins, 1.590.2 versus 6.590.9 m/mg tissue,PB0.0001. uPA
activity did not differ between distended saphenous veins (n=16), 34.993.1 and internal mammary arteries (n=
11), 46.697.0 ng/g tissue,P=0.25.
4. Discussion
Histologically, the saphenous vein and internal mam-mary artery specimens represent relatively normal adult vessels, whilst the coronary and carotid endarterectomies show classic histological features of complex atheroma, and the saphenous vein grafts an exaggerated fibroprolif-erative response to vascular damage.
Immunohistochemical localisation of tPA, uPA, and PAI-1 antigens accords with previous studies in showing an association of these antigens not only with endothelial cells of the major lumen and vasa vasorum, but also with smooth muscle cells and extracellular matrix [4,5,12,13]. The level of staining associated with smooth muscle cells is variable, and there is an association between more intense staining and smooth muscle cell activation, most clearly seen when comparing the original media and the intima-medial junction of stenosing saphenous vein grafts. uPAR staining, absent in undamaged saphenous
vein and internal mammary artery, was present in saphenous vein grafts associated with proliferating/ mi-grating smooth muscle cells at the intima-medial junction, and in complex atheroma associated with smooth muscle cells and focal collections of lymphocytes and macrophages. The latter also stained for uPA and PAI-1, but not tPA.
The increased immunostaining seen in distended veins compared to normal veins was a consistent but unex-pected finding. Whilst our study does not provide a definitive answer for the cause of this, it shows that it is not due to increased mRNA synthesis; vein distension was associated with a decrease in PAI-1 and uPA mRNA. This apparent anomaly could have been caused by the effect of routine forceful distension using heparinised blood in saline either by way of extravasation of these antigens into the vessel wall or possibly by alterations in antigen accessibility to the antibody associated with tissue stretch. In any case, this should caution against the use of immunostaining intensity alone as a measure of expres-sion in vessels forcefully distended with heparinised blood. Other limitations of the information provided by the immunostaining are the antibodies used do not distinguish between free tPA or uPA and that bound to inhibitor, or conversely between free or bound PAI-1. mRNA expression of tPA, uPA or PAI-1 was not detected using in-situ hybridisation in internal mammary artery whereas in saphenous veins only PAI-1 at low levels was detected. RTPCR however did confirm the presence of both PAI-1 and uPA in homogenates of these vessel walls, perhaps reflecting the greater sensitivity of PCR. The fall in PAI-1 and uPA mRNA associated with forceful vein distension could in theory be related to the loss of cells, including endothelial cells, from the vessel wall during the distension process or/and increased RNA degradation. In stenosing vein grafts the striking feature was the dissociation between PAI-1 mRNA, which almost entirely localised to endothelium, and both tPA and uPA, which predominantly localised to proliferating/
4.1. The significance of plasminogen acti6ator
expression
The role of plasminogen activators and their in-hibitors in vessel wall biology has been extensively debated. Mice with tPA uPA, uPAR or PAI-1 gene knockouts do not show any major abnormality in vessel structure, although uPA or PAI-1 deficient mice show altered responses to vascular injury [14,15,18,19]. Clowes and colleagues observed increased tPA activity associated with smooth muscle proliferation and migra-tion after intimal injury in a rat model, and it has been suggested that plasminogen activation is the initial trig-ger to the activation of a chain of metalloproteinases that leads to matrix degradation, thus allowing smooth muscle cells to respond to proliferative and migratory stimuli [19]. Our observation of increased uPA and tPA mRNA expression by smooth muscle cells in the neointima of stenosing saphenous vein grafts would accord with this model. Presumably the co-localisation of PAI-1 antigen, but not PAI-1 mRNA, reflects bind-ing of exogenous PAI-1. It is interestbind-ing that stenosbind-ing saphenous vein grafts, at least at the stage they were examined in the present study, offer an example of a sustained increase in plasminogen activator expression, accompanying sustained SMC proliferation and migra-tion. There is no evidence for a concomitant increase in PAI-1 expression, and this constitutes a major differ-ence from complex atheroma. The association of in-creased proteolytic activity with inflammatory infiltrates in complex atheroma has also been described, and linked to plaque destabilisation leading to plaque rup-ture [10,11,19]. Our observation of uPA and uPAR associated with collections of inflammatory cells would again be in keeping with this model.
4.2. Plasminogen acti6ator inhibitor
The role of PAI-1 in the vessel wall is more contro-versial. Loskutoff and colleagues showed that athero-matous lesions contained PAI-1 mRNA, and in the present study we have both confirmed this and localised at least some of its production to smooth muscle cells. Intuitively, excess active PAI-1 in the vessel wall after vascular injury would be expected to increase the risk of thrombosis, but at the same time inhibit tPA and uPA effects on SMC proliferation and migration. In-deed, evidence from PAI-1 deficient mice demonstrates increased vascular smooth muscle proliferation after arterial injury [15].
It seems clear, both from our own results and those in the literature, that smooth muscle cells in complex atheroma behave differently from quiescent arterial or venous smooth muscle cells, or from smooth muscle in fibroproliferative lesions such as stenosing vein grafts, in that they produce substantial amounts of PAI-1.
Several of the factors controlling PAI-1 expression have been reviewed [20], but in an observational study such as this, one can only speculate as to which might be operating. Conceivably, one factor could be the influ-ence of the inflammatory environment, prominent in complex atheroma but much less so in fibroproliferative lesions. Tilting the balance towards increased tPA and uPA production may favour the development of fibro-proliferative lesions as seen in the vein grafts. One of the paradoxes about the plasminogen activator system in complex atheroma is that despite the large amounts of PAI-1 associated with plaques, the net balance of activator/inhibitor activity is tilted towards activation of fibrinolysis. Our present results suggest that one reason for this may be reduced activity of plaque associated PAI-1, most of which may be in a latent or inactive form.
From a clinical perspective, it would be interesting to know how tPA/uPA production was induced in migrat-ing smooth muscle cells, and whether control of this could influence the development of neointimal hyper-plasia. Conversely we need to know whether increased PAI-1 production is a potential homeostatic mecha-nism, or an independent and potentially harmful re-sponse to inflammation.
Acknowledgements
This work was supported by the British Heart Foundation.
References
[1] Astrup T. The biological significance of fibrinolysis. Lancet 1956;25:569.
[2] Todd AS. Histological localisation of plasminogen activator. J Pathol Bacteriol 1959;78:281 – 7.
[3] Kwaan HC, Astrup T. Fibrinolytic activity in human atheroscle-rotic coronary arteries. Circ Res 1967;21:799 – 804.
[4] Lupu F, Heim DA, Bachmann F, Hurni M, Kakkar W, Kruithof EKO. Plasminogen activator expression in human ad-vanced atherosclerotic lesions. Arterioscler Thromb Vasc Biol 1995;15:1444 – 55.
[5] Robbie LA, Booth NA, Brown AJ, Bennett B. Inhibitors of fibrinolysis are elevated in atherosclerotic plaque. Arterioscler Thromb Vasc Biol 1996;16:539 – 45.
[6] Clowes AW, Clowes MM, Au YPT, Reidy MA, Belin D. Smooth muscle cells express urokinase during mitogenesis and tissue-type plasminogen activator during migration in injured rat carotid artery. Circ Res 1990;67:61 – 7.
[7] Jackson CL, Reidy MA. The role of plasminogen activation in smooth muscle migration after arterial injury. Ann NY Acad Sci 1992;667:141 – 50.
M.Y.Salame et al./Atherosclerosis152 (2000) 19 – 28
28
[9] Underwood MJ, deBono DP. Increased fibrinolytic activity in the intima of atheromatous coronary arteries: protection at a price. Cardiovasc Res 1993;27:882 – 5.
[10] Lee RT, Libby P. The unstable atheroma. Arterioscler Thromb Vasc Biol 1997;17:1859 – 67.
[11] Felton CV, Crook D, Davies MJ, Oliver MF. Relation of plaque lipid composition and morphology to the stability of human aortic plaques. Arterioscler Thromb Vasc Biol 1997;17:1337 – 45. [12] Schneiderman J, Sawdrey MS, Keeton MR, Bordin GM, Bern-stein EF, Dilley RB, Loskutoff D. Increased plasminogen activa-tor inhibiactiva-tor type 1 gene expression in atherosclerotic human arteries. Proc Natl Acad Sci USA 1992;89:6998 – 7002. [13] Grancha S, Estelles A, Falco C, Chirivella M, Espana F, Aznar
J. Detailed localization of type-1 plasminogen activator inhibitor mRNA expression and antigen in atherosclerotic plaque on human coronary artery. Fibrinolysis Proteolysis 1998;12:53 – 60. [14] Carmeliet P, Moons L, Herbert JM, Lupu F, Lijuen R, Collen D. Urokinase- but not tissue-type plasminogen activators medi-ate arterial neointima formation in mice. Circ Res 1997;81:829 – 39.
[15] Carmeliet P, Moons L, Lijnen R, Janssens S, Lupu F, Collen D,
Gerard RD. Role of plasminogen activator inhibitor-1 in vascu-lar wound healing and arterial neointima formation. A gene targeting and gene transfer study in mice. Circulation 1997;96:3180 – 91.
[16] Chomczynski P, Sacchi N. Single-step method of RNA isolation by guanidium thiocyanate-phenol-chloroform extraction. Anal Biochem 1987;162:156 – 9.
[17] Bergsdorf N, Nilsson T, Wallen P. An enzyme linked immuno-sorbent assay for determination of tissue plasminogen activator applied to patients with thromboembolic disease. Thromb Haemost 1983;50(3):740 – 4.
[18] Carmeliet P, Collen D. Role of plasminogen/plasmin system in thrombosis, hemostasis, restenosis and atherosclerosis evaluation in transgenic animals. Trends Cardiovasc Med 1995;5:117 – 22. [19] Lijnen RH, Van Hoef B, Lupu F, Moons L, Carmeliet P, Collen
D. Function of the plasminogen/plasmin and matrix metallo-proteinase systems after vascular injury in mice with targeted inactivation of fibrinolytic system genes. Arterioscler Thromb Vasc Biol 1998;18:1035 – 45.
[20] Loskutoff DJ. Regulation of PAI-1 gene expression. Fibrinolysis 1991;5:197 – 206.
