INTRODUCTION

Top Abstract Introduction Conclusions References

TARGETED GENE MANIPULATION in animal models has largely contributed to a better understanding of the molecular mechanisms involved in the formation as well as the pathology of blood vessels. Several of these factors play distinct roles not only during embryonic development but, frequently, also during adult vascular pathology. The vessel wall consists of two major cell types, endothelial and smooth muscle cells, that closely interact with each other. The vascular endothelial growth factor (VEGF) system is essential for endothelial cell biology, and defective signaling results in abnormal formation of endothelial cell-lined channels. In contrast, the initiator of coagulation plays a role in smooth muscle cell function, and loss of its function during embryogenesis results in immature fragile vessels that form microaneurysms and rupture with secondary bleeding. The coagulation and plasminogen systems are essential for proper hemostasis because the former controls bleeding by formation of fibrin clots and the latter guarantees maintenance of vascular patency by removal of fibrin clots. When deregulated, both systems may contribute to thrombotic or hemorrhagic disorders. In addition, these systems have been implicated in tissue remodeling and cellular migration, essential mechanisms for the repair of blood vessels after acute injury (e.g., balloon angioplasty) or chronic inflammation (atherosclerosis). The present overview aims at integrating the pleiotropic roles of the VEGF, coagulation, and plasminogen systems in vascular biology, as deduced from targeted gene manipulation (gene inactivation or gene transfer) studies in the mouse.

	THE VEGF SYSTEM

VEGF-A was purified as a growth factor able to increase vascular permeability and endothelial cell proliferation (45, 49; Fig. 1A). VEGF-A distinguishes itself from previously known angiogenic factors by a unique combination of properties: 1) it is secreted and exerts a direct effect on endothelial cells via interaction with cellular receptors Flk-1 and Flt-1 (45, 49); 2) it is produced by cells in close proximity to endothelial cells, suggesting paracrine regulation of blood vessel formation (10); 3) expression of VEGF-A is highly regulated by hypoxia, providing a physiological feedback mechanism to accommodate insufficient tissue oxygenation by promoting blood vessel formation (95); and 4) VEGF-A is a potent factor, because its over- or underexpression significantly affects blood vessel formation in vivo (52). VEGF-A is transcribed from a single gene but is alternatively processed to various isoforms. The shortest form (VEGF-A¹²⁰) is diffusible in the surrounding extracellular milieu, whereas the longer isoforms (VEGF-A¹⁶⁵, VEGF-A¹⁸⁹, and VEGF-A²⁰⁶) contain basic amino acid residues that mediate increasing binding to heparin-rich extracellular matrix (45, 49). Although these various isoforms exhibit a tissue-specific pattern of expression, their differential (patho)physiological roles in vivo remain largely undetermined.

View larger version (55K): [in this window] [in a new window]   Fig. 1.   A: schematic representation of vascular endothelial growth factor (VEGF) system. VEGF and placental growth factor (PlGF) bind as homo- or heterodimers, with different specificities and affinities to their different cellular receptors, Flt-1, Flk-1, and Flt-4. VEGFR, VEGF receptor. B: formation of blood vessels. Early endothelial stem cells (hemangioblasts) differentiate in situ into endothelial cells (left) and become organized into a primitive vascular plexus consisting of endothelial lined channels (middle). Subsequently, these primitive blood vessels acquire a muscular wall (right) to accommodate increased blood pressure during further embryonic development.

More recently, other VEGF-related factors have been identified (Fig. 1A): VEGF-B, which is expressed primarily in the heart, brain, muscle, and kidney (76), and VEGF-C, which appears to play a role in development of lymphatic vessels (61). Another homolog, placental growth factor (PlGF), was identified in the placenta and, to a lesser extent, in heart, lung, and thyroid (71). Three receptor tyrosine kinases, containing seven immunoglobulin domains, have thus far been identified that bind the VEGF family members with different specificity and affinity; VEGF receptor-1 (or Flt-1) binds VEGF-A and PlGF, VEGF receptor-2 (or Flk-1) binds VEGF-A and VEGF-C, and VEGF receptor-3 (or Flt-4) binds VEGF-C (41, 56, 72). Homo- or heterodimerization of these ligands may determine their biological specificity (44). Receptor activation, via intracellular signaling, results in a pleiotropic pattern of angiogenic activities including proliferation, apoptosis, migration, and permeability of endothelial cells and their production of matrix-degrading proteinases and tissue factor (TF; Refs. 45, 49).

VEGF expression has been implicated in vascular development during embryogenesis and postnatally during heart ischemia (5), atherosclerosis (64), diabetic retinopathy (50), tumorigenesis (79), arthritis, and wound healing (54). VEGF gene therapy has been considered to improve blood flow to ischemic limbs and hearts (100) and to reduce stenosis of blood vessels after arterial injury by promoting reendothelialization (4).

	THE COAGULATION SYSTEM

Initiation of the plasma coagulation system is triggered by TF, which functions as a cellular receptor and cofactor for activation of the serine proteinase factor VII to factor VIIa (Refs. 40, 46, 55; Fig. 2A). This complex activates factor X directly or indirectly via activation of factor IX, resulting in the generation of thrombin, which mediates conversion of fibrinogen to fibrin (40, 46, 55). Thrombin and factor Xa produce a positive-feedback stimulation of coagulation by activating factor VIII and factor V, which serve as membrane-bound receptors/cofactors for the proteolytic enzymes factor IXa and factor Xa, respectively (40, 46, 55).

View larger version (18K): [in this window] [in a new window]   Fig. 2.   A: schematic representation of coagulation system. Tissue factor (TF) is primary cellular activator of blood coagulation and via activation of factor VII initiates a cascade activation of factor X and prothrombin resulting in conversion of fibrinogen to fibrin. Anticoagulation is mediated by TF pathway inhibitor (TFPI), thrombomodulin, protein C, protein S, and antithrombin III. See text for details. B: schematic representation of plasminogen (Plg) system. Plg is an inactive proenzyme that can be activated by tissue-type (t-PA) or urokinase-type (u-PA) Plg activator. Their action is inhibited by PA inhibitor-1 (PAI-1). u-PA binds to a cellular receptor, u-PAR, that has been involved in cellular migration. 2-Antiplasmin is principal plasmin inhibitor.

In contrast, when thrombin is bound to its cellular receptor thrombomodulin, it functions as an anticoagulant by activating the protein C anticoagulant system (39, 47). Activated protein C, in the presence of its cofactor protein S, inactivates factor Va and factor VIIIa, thereby reducing thrombin generation (39, 47). Anticoagulation is further provided by antithrombin III, which binds to and inactivates thrombin, factor IXa, and factor Xa in a reaction that is greatly enhanced by heparin (78). Anticoagulation is further secured by TF pathway inhibitor, which inhibits factor Xa and, in a factor Xa-dependent manner, produces feedback inhibition of the factor VIIa/TF catalytic complex (11). A revised hypothesis of coagulation has been suggested in which factor VIIa/TF is responsible for the initiation of coagulation but, because of TF pathway inhibitor-mediated feedback inhibition, amplification of the procoagulant response through the actions of factor VIII, factor IX, and factor XI is required for sustained hemostasis (11). Deficiencies of anticoagulant factors or disturbed expression of procoagulant factors have been implicated in thrombosis during inflammation, sepsis, atherosclerosis, and cancer (7, 46), whereas deficiencies of procoagulant factors have been related to increased bleeding tendencies (9, 60). Evidence has been provided that the coagulation system may also be involved in other functions beyond hemostasis, including cellular migration and proliferation, immune response, angiogenesis, embryonic development, metastasis, and brain function (1, 37, 46). Its precise role and relevance in these processes in vivo remain, however, largely unknown.

	THE PLASMINOGEN SYSTEM

The plasminogen system is composed of an inactive proenzyme, plasminogen, that can be converted to plasmin by either of two plasminogen activators (PA), tissue-type PA (t-PA) or urokinase-type PA (u-PA) (34, 105; Fig. 2B). This system is controlled at the level of PAs by PA inhibitors (PAIs), of which PAI-1 is believed to be physiologically the most important (65, 90, 107), and at the level of plasmin by ₂-antiplasmin (34). Because of its fibrin specificity, t-PA is primarily involved in clot dissolution, although it has also been involved in ovulation, bone remodeling, and brain function (34, 105). Cellular receptors for t-PA and plasminogen have been identified that might localize plasmin proteolysis to the cell surface (57, 82). u-PA also binds a cellular receptor, the urokinase receptor (u-PAR), and has been implicated in pericellular proteolysis during cell migration and tissue remodeling in a variety of normal and pathological processes including angiogenesis, atherosclerosis, and restenosis (8, 104). u-PAR binds to vitronectin (106), whereas PAI-1 controls recognition of vitronectin by u-PAR or the _v3-integrin receptor, suggesting a role in coordinating cell adhesion and migration (97). It is presently unclear whether or under what conditions binding of u-PA to u-PAR is required in vivo. Plasmin is able to degrade fibrin and extracellular matrix proteins directly or indirectly via activation of other proteinases (such as the metalloproteinases; Refs. 30, 87). Plasmin can also activate or liberate growth factors from the extracellular matrix, including latent transforming growth factor- (TGF-), basic fibroblast growth factor, and vascular endothelial growth factor (49, 87). Cell-specific clearance of plasminogen activators or of complexes with their inhibitors by low-density lipoprotein receptor-related protein or gp330 may modulate pericellular plasmin proteolysis (2).

	TARGETED MANIPULATION AND ADENOVIRUS-MEDIATED TRANSFER OF GENES IN MICE

Novel gene technologies have allowed the manipulation of the genetic balance of candidate molecules in mice in a controlled manner. Homologous or site-specific recombination in embryonic stem cells allows the study of the consequences of deficiencies, mutations, and conditional or tissue-specific expression of gene products in transgenic mice (15, 73). With a novel embryonic stem cell technology (aggregation of embryonic stem cells with tetraploid embryos), it has become possible to generate completely embryonic stem cell-derived embryos in a single step. In addition, the technology allows the bypass of conventional germ line transmission, to separate extra- from intraembryonic phenotypes and to study homozygous deficient phenotypes of genes that cause embryonic lethality when heterozygously deficient (18, 74). Viral gene transfer can also be used to manipulate the expression of genes, e.g., via implantation of retrovirally transduced cells or via adenovirus-mediated gene transfer in vivo (88). In fact, intravenous administration of a recombinant adenovirus results in expression of target genes to plasma levels >10 µg/ml. Such studies allow us not only to generate but also to rescue disease models and to evaluate possible gene-transfer therapies.

	FORMATION OF BLOOD VESSELS AND HEMOSTASIS

Formation of endothelial cell-lined channels during embryogenesis and wound healing. After initial differentiation of stem cells into endothelial cells and their assembly into endothelial cell-lined channels (vasculogenesis), the embryonic vasculature further develops via sprouting of new channels from preexisting vessels (angiogenesis; Ref. 84, Fig. 1B). This latter process is recapitulated in adulthood during tissue neovascularization. The genetic mechanisms controlling these processes are, however, poorly defined. With the advent of the recently developed gene technology to generate mice with inactivation or mutation of target genes, remarkable progress has been made in the understanding of the molecular processes in blood vessel formation. The following discussion briefly summarizes recent findings on the role of the VEGF-PlGF system in blood vessel development during embryogenesis and wound healing (Fig. 3).

View larger version (37K): [in this window] [in a new window]   Fig. 3.   Overview of VEGF targeting studies. Deficiency of VEGF-A results in haploinsufficient embryonic lethality because of delayed endothelial cell development and abnormal assembly, lumen formation, sprouting, and organization of vascular channels. Deficiency of PlGF results in normal embryonic development and fertility but impairs skin wound healing, possibly via abnormal formation of granulation tissue and neovascularization. Deficiency of Flk-1 results in embryonic lethality caused by abortive endothelial cell development and lack of blood vessels. This phenotype is different from that of VEGF-A deficiency, suggesting presence of other Flk-1 ligand(s). Deficiency of Flt-1 results in embryonic lethality because of abnormal organization of endothelial cells in vascular channels. There appear to be more endothelial cells.

Hemostasis: vessel fragility vs. clot formation. Once the endothelial cells are assembled into vascular channels, they become surrounded by smooth muscle cells/pericytes that may affect maturation of the blood vessels not only by providing the fragile primitive blood vessels the structural support required to accommodate the increased blood pressure but also by controlling endothelial cell proliferation, vascular permeability, and tone (Ref. 75; Fig. 1B). Although a role for vascular smooth muscle cells/pericytes in the pathogenesis of vasculopathies during adulthood, including atherosclerosis, restenosis, and diabetic retinopathy, has been documented, their role during vascular development has been poorly characterized. Surprisingly, targeted gene inactivation of some coagulation factors has revealed their possible implication in blood vessel development. Because these coagulation factors are also expressed during restenosis or atherosclerosis, a better understanding of their role during embryogenesis may help us to understand their function during adult disease processes and to design appropriate therapeutic strategies.

View larger version (56K): [in this window] [in a new window]   Fig. 4.   Targeting studies of coagulation system. A: schematic representation of yolk sac connected to embryo via vitelloembryonic blood vessels, mediating transfer of maternally derived nutrients to the embryo proper. B: schematic representation of a transversal section through blood vessels in wild-type (TF+/+) and TF-deficient (TF/) yolk sacs, aligned by inset in A. Vascular defects are induced by TF deficiency; smooth muscle cell (SMC)/pericyte-like mesenchymal cells accumulate and differentiate around endothelial cell-lined capillaries in yolk sac from a TF+/+ embryo beyond 9 days of gestation to sustain increased blood pressure load. In TF/ yolk sac, these mesenchymally derived SMC/pericyte-like cells fail to accumulate and differentiate around endothelial cells. As a result, visceral endoderm and mesothelial cell layers detach (arrowheads, left), and fragile capillaries rupture and form microaneurysms and "blood lakes" that ultimately leak blood into the (extravascular) extracelomic cavity (arrows, right). As a result of vascular defects, blood circulation from yolk sac to the rapidly growing embryo is compromised, inducing wasting of embryo because of deprivation of essential nutrients. Notably, leakage of blood from defective vessels ("vascular" bleeding) only occurs at a final stage of necrosis and deterioration. C: targeting studies of coagulation system. TF is a membrane receptor, the extracellular domain of which is necessary and sufficient for initiation of coagulation cascade and fibrin formation. The intracellular domain has been implicated in intracellular signaling. It remains to be determined whether TF interacts with factor VII or a novel, unidentified factor to mediate its morphogenic properties during embryogenesis and whether this involves intracellular signal transduction. Indeed, factor VII-deficient embryos develop normally until birth, after which they die because of massive bleeding, suggesting that, in these embryos, TF either interacts with another factor or acts independently. Alternatively, placental transfer of factor VII might rescue factor VII-deficient embryogenesis, but transfer of human factor VIIa across mouse placenta was undetectable (see text). Reproduced with permission from Ref. 17.

	THROMBOSIS AND THROMBOLYSIS

Fibrin deposits and pulmonary plasma clot lysis in transgenic mice. Deficient fibrinolytic activity, e.g., resulting from increased plasma PAI-1 levels or reduced plasma t-PA or plasminogen levels, might participate in the development of thrombotic events (3, 65, 90, 107). Fibrin surveillance in different knockout mice was analyzed in quiescent conditions and after challenge (Fig. 5). In unstressed conditions, u-PA-deficient mice developed occasional minor fibrin deposits in liver and intestines and excessive fibrin deposition in chronic, nonhealing skin ulcerations, whereas in t-PA-deficient mice, no spontaneous fibrin deposits were observed (16, 27). Mice with a single deficiency of plasminogen or a combined deficiency of t-PA and u-PA, however, revealed extensive intravascular and extravascular fibrin deposits in several organs (Refs. 13, 16, 27, and 81 and unpublished observations). Interestingly, mice with a combined deficiency of t-PA and u-PAR did not display such excessive fibrin deposits, suggesting that sufficient plasmin proteolysis can occur in the absence of u-PA binding to u-PAR (Fig. 5B; Refs. 12, 42). Loss of both PAs or of plasminogen severely affected general health and caused a multiorgan dysfunction syndrome characterized by dyspnea, anemia, sterility, cachexia, and premature death (Fig. 5A).

View larger version (54K): [in this window] [in a new window]   Fig. 5.   Role of Plg system in fibrin surveillance. Fibrin deposits in Plg system knockout mice before and after experimental challenges. A: fibrin deposits during unchallenged conditions were occasionally observed in u-PA-deficient mice and more extensively in Plg- or combined t-PA + u-PA-deficient mice, indicating that both PAs cooperate in prevention of fibrin deposition. Inflammatory and/or traumatic challenges including local injection of proinflammatory endotoxin in footpad, skin wounding, experimental glomerulonephritis, lung hypoxia, and arterial injury induce extravascular fibrin deposits but also intravasular thrombosis in veins or capillaries (endotoxin, lung hypoxia, skin wounding, glomerulonephritis) and arteries (arterial injury). B: combined deficiency of t-PA and u-PA but not of t-PA and u-PAR results in widespread fibrin deposition, reduced fertility, multiorgan dysfunction with dyspnea, lethargia, and cachexia, ultimately leading to premature death, indicating that sufficient pericellular plasmin proteolysis occurs in the absence of binding of u-PA to its cellular receptor. Reproduced with permission from Ref. 17.

Role of t-PA in stroke. It has been recognized for years that several components of the plasminogen system are expressed in the brain, but functional proof for their in vivo role was lacking. Recently, t-PA- and plasminogen-deficient mice were found to be significantly protected against neuronal cell damage in the hippocampus induced by administration of neuroexcitatory toxins, suggesting that t-PA-mediated plasmin proteolysis is involved in neuronal cell destruction (102). In addition, intracerebral injection of t-PA restored the sensitivity of t-PA-deficient mice to neuronal insult, whereas intracerebral injection of PAI-1 protected wild-type mice against cell death. Taken together, these studies suggest that t-PA might play a role in stroke beyond fibrin clot dissolution and warrant further analysis.

Adenovirus-mediated transfer of t-PA or PAI-1. More recently, we have used adenovirus-mediated transfer of fibrinolytic system components in these knockout mice in an attempt to revert their phenotypes. Intravenous injection of adenoviruses expressing a recombinant PAI-1-resistant human t-PA (rt-PA) gene in t-PA-deficient mice increased plasma rt-PA levels 100- to 1,000-fold above normal and restored their impaired thrombolytic potential in a dose-related manner (28). Conversely, adenovirus-mediated transfer of recombinant human PAI-1 in PAI-1-deficient mice resulted in 100- to 1,000-fold increase in plasma PAI-1 levels above normal and efficiently reduced the increased thrombolytic potential of PAI-1-deficient mice (unpublished observations).

	NEOINTIMA FORMATION

Vascular interventions for the treatment of atherothrombosis induce "restenosis" of the vessel within 3-6 mo in 30-50% of treated patients (33, 67). This may result from remodeling of the vessel wall (such as occurs predominantly after balloon angioplasty) and/or accumulation of cells and extracellular matrix in the intimal layer (such as occurs predominantly after intraluminal stent application). Proteinases may participate in proliferation and migration of smooth muscle and/or endothelial cells and in matrix remodeling during this wound healing response (103). Two proteinase systems have been implicated, the plasminogen (or fibrinolytic) system and the metalloproteinase system, which in concert can degrade most extracellular matrix proteins. In contrast to the constitutive expression of t-PA by quiescent endothelial cells (34, 105) and of PAI-1 by uninjured vascular smooth muscle cells (90, 96), u-PA and t-PA activity in the vessel wall are significantly increased after injury, coincident with the time of smooth muscle cell proliferation and migration (32). This increase in plasmin proteolysis is counterbalanced by increased expression of PAI-1 in injured smooth muscle and endothelial cells and by its release from accumulating platelets (48).

u-PA-mediated plasmin proteolysis promotes arterial neointima formation. Two experimental models of arterial injury have been used, one based on application of an electric current (26) and the other based on an intraluminal guide wire (23, 24), to examine the molecular mechanisms of neointima formation in mice deficient in fibrinolytic system components. The electric current injury model differs from mechanical injury models in that it induces a more severe injury across the vessel wall, resulting in necrosis of all smooth muscle cells. This necessitates wound healing to initiate from the adjacent uninjured borders and to progress into the central necrotic region. Microscopic and morphometric analysis revealed that the rate and degree of neointima formation and the neointimal cell accumulation after injury were similar in wild-type, t-PA-deficient, and u-PAR-deficient arteries (Refs. 22 and 23 and unpublished data). However, neointima formation in PAI-1-deficient arteries occurred earlier after injury (24). In contrast, both the degree and the rate of arterial neointima formation in u-PA-, plasminogen-, and combined t-PA + u-PA-deficient arteries was significantly reduced until 6 wk after injury (22, 23, 25). Infiltration of the media by leukocytes was also significantly reduced in plasminogen-deficient mice (25). Similar genotypic differences were obtained after mechanical injury (23, 24), which more closely mimics the injury in patients.

View larger version (33K): [in this window] [in a new window]   Fig. 6.   Receptor-independent u-PA promotes neointima formation. SMC is surrounded by an extracellular matrix (ECM) that must be proteolytically degraded to allow cellular migration. In wild-type (WT) SMC, u-PA is bound to u-PAR, mediating Plg activation and plasmin degradation of extracellular matrix such that cells can migrate. In u-PAR-deficient (u-PAR/) SMC, u-PA becomes localized to cell surface, possibly via interaction with other matrix molecules (denoted as "X?"), allowing sufficient pericellular plasmin proteolysis for cells to migrate. u-PA might also accumulate to increased levels because of deficient u-PAR-mediated clearance. In contrast, SMC that lack u-PA (u-PA/) have reduced pericellular plasmin proteolysis and fail to migrate efficiently, resulting in reduced neointima formation. Proposed impairment of SMC migration in mice lacking u-PA-mediated Plg activation is suggested by observations that proliferation of u-PA-deficient cells was similar to WT and that SMC migrated over a shorter distance in the Plg-deficient than in WT arteries. Reproduced with permission from Ref. 17.

Inhibition of neointima formation by adenovirus-mediated PAI-1 gene transfer. The involvement of plasmin proteolysis in neointima formation was supported by intravenous injection in PAI-1-deficient mice of a replication-defective adenovirus that expresses human PAI-1, which resulted in >100- to 1,000-fold increased plasma PAI-1 levels and in a similar degree of inhibition of neointima formation as observed in u-PA-deficient mice (97). Proteinase inhibitors have been suggested as antirestenosis drugs. Our studies suggest that strategies aimed at reducing u-PA-mediated plasmin proteolysis may reduce intimal thickening. However, antifibrinolytic strategies should be aimed at inhibiting plasmin proteolysis and not at preventing the interaction of u-PA with its receptor.

Transplant atherosclerosis in plasminogen-deficient mice. More recently, we (in a collaboration with V. Shi and E. Haber, Harvard Univ., Boston, MA) have started to analyze the role of the plasminogen system in a mouse model of transplant arteriosclerosis that mimics in many ways the accelerated arteriosclerosis in coronary arteries of transplanted cardiac allografts in humans (94). In this model, host-derived leukocytes infiltrate beneath the endothelium and form a predominantly leukocyte-rich neointima within 15 days after transplantation, whereas at later times, smooth muscle cells derived from the donor graft accumulate in the neointima. Because previous targeting studies have shown that migration of leukocytes and smooth muscle cells is dependent on plasmin proteolysis (22-25, 80), carotid arteries from B10.A(2R) wild-type mice were transplanted in C57BL6:129 plasminogen-deficient mice. Initial analysis suggests that neointima formation within 45 days after transplantation is reduced in these mice, suggesting a significant role for plasmin proteolysis in this process. Whether cellular migration, proliferation, or matrix remodeling is affected remains to be determined.

	ATHEROSCLEROSIS

Epidemiologic, genetic, and molecular evidence suggests that impaired fibrinolysis resulting from increased PAI-1 or reduced t-PA expression or from inhibition of plasminogen activation may contribute to the development and/or progression of atherosclerosis (59, 62, 91), presumably by promoting thrombosis or matrix deposition. Indeed, plasma levels of PAI-1 are elevated in patients with ischemic heart disease, angina pectoris, and recurrent myocardial infarction (58). Recent genetic analyses revealed a link between polymorphisms in the PAI-1 promoter and the susceptibility of atherothrombosis (59). A possible role for increased plasmin proteolysis in atherosclerosis is, however, suggested by the enhanced expression of t-PA and u-PA in plaques (69, 89). Plasmin proteolysis might indeed participate in plaque neovascularization, induction of plaque rupture, or ulceration and formation of aneurysms (69, 89). A causative role of the plasminogen system in these processes has, however, not been conclusively demonstrated.

Atherosclerosis was studied in mice deficient in apolipoprotein E (apoE; Ref. 83) and in t-PA, u-PA, or PAI-1 and fed a cholesterol-rich diet for 5-25 wk (P. Carmeliet, L. Moons, R. Lijnen, J. Crawley, V. Lemaïtre, P. Tipping, A. Drew, Y. Eeckhout, S. Shapiro, F. Lupu, and D. Collen, unpublished observations). No differences in the size or predilection site of plaques were observed between mice with a single deficiency of apoE or with a combined deficiency of apoE and t-PA or of apoE and u-PA. However, significant genotypic differences were observed in the destruction of the media with resultant erosion, transmedial ulceration, medial smooth muscle cell loss, dilatation of the vessel wall, and microaneurysm formation. At the ultrastructural level, elastin fibers were eroded, fragmented, and completely degraded, whereas collagen bundles and glycoprotein-rich matrix were disorganized and scattered. Although both apoE- and apoE + t-PA-deficient mice developed severe media destruction, apoE + u-PA-deficient mice were virtually completely protected. Plaque macrophages expressed abundant amounts of u-PA mRNA, antigen, and activity at the base of the plaque and in the media, similar to the atherosclerotic, aneurysmatic arteries in patients (69, 89). Macrophages crossed the elastin fibers but only after proteolytic digestion of the elastin, a process that was remarkably enhanced by u-PA. u-PA-generated plasmin may mediate degradation of glycoproteins, surrounding elastin fibers in the aortic wall, thereby exposing the highly insoluble elastin to elastases and facilitating elastolysis in vivo (31). In addition, plasmin promoted degradation of elastin and collagen via activation of matrix metalloproteinases (MMP) such as stromelysin-1 (MMP-3), gelatinase-B (MMP-9), collagenase-3 (MMP-13), and the macrophage metalloelastase (MMP-12) (unpublished observations and Refs. 30, 92). The expression of several of these metalloproteinases was induced in situ in advanced atherosclerotic plaques by macrophages that also expressed increased amounts of u-PA. Taken together, these results implicate an important role of u-PA in the structural integrity of the atherosclerotic vessel wall by triggering activation of metalloproteinases (Fig. 7).

View larger version (13K): [in this window] [in a new window]   Fig. 7.   Hypothetical role of u-PA in aneurysm formation. u-PA-mediated plasmin proteolysis might be involved in media destruction by degradation of extracellular matrix (including insoluble elastin) directly by degradation of fibrin and glycoproteins surrounding elastin fibrils and/or indirectly by activation of other matrix degrading proteinases, including of elastolytic metalloproteinases (gelatinases and macrophage metalloelastase). By mediating infiltration in atherosclerotic vessel wall of cells that produce these matrix degrading proteinases, u-PA may further amplify destruction of media.

In contrast, mice with a combined deficiency of apoE and PAI-1 developed normal fatty streak lesions but subsequently revealed a transient, delayed progression to fibroproliferative plaques. Whether the increased plasmin proteolytic balance in these mice might prevent matrix accumulation and, consequently, delay plaque progression, or whether more abundant plasmin increased activation of latent TGF- with its pleiotropic role on smooth muscle cell function and matrix accumulation, remains to be determined. Taken together, these targeting studies identify a specific role for u-PA in the destruction of the media that may precede aneurysm formation and for PAI-1 in plaque progression, possibly by promoting matrix deposition.

	CONCLUSIONS

Top Abstract Introduction Conclusions References

Gene-targeting studies are useful to obtain novel insights into the role and relevance of a gene during normal or pathological biological processes in vivo. New insights into the role of VEGF/PlGF, the fibrinolytic system, and/or the coagulation system in the formation of a normal blood vessel and in its pathologic progression to disorders such as thrombosis, restenosis, and atherosclerosis have recently been obtained.

Several coagulation factors appear to play an unanticipated role in embryogenesis, presumably in the formation of the primitive muscular wall of blood vessels, indicating a role for these coagulation factors beyond mere control of hemostasis. Whether a molecule such as TF mediates these morphogenic properties via interaction with factor VII or, possibly, with a novel (yet unidentified) factor remains to be evaluated. This might open possible future therapeutic avenues for a specific differential inhibition of hemostasis or morphogenesis in processes such as restenosis or atherosclerosis.

Urokinase-mediated plasmin proteolysis appears to play a significant role in migration of smooth muscle cells during neointima formation after vascular injury and in the destruction of the media and aneurysm formation during atherosclerosis. Such insights have initiated studies aimed at preventing neointima formation. Whether inhibition of plasmin proteolysis is a feasible means to prevent aneurysm formation during atherosclerosis remains to be determined. Ongoing research will determine the role of the plasminogen system in other cardiovascular disorders such as myocardial infarction and transplant atherosclerosis.