Recognition of Plasminogen Activator Inhibitor Type 1 as the Primary Regulator of Fibrinolysis

Abstract

The fibrinolytic system consists of a balance between rates of plasminogen activation and fibrin degradation, both of which are finely regulated by spatio-temporal mechanisms. Three distinct inhibitors of the fibrinolytic system that differently regulate these two steps are plasminogen activator inhibitor type-1 (PAI-1), α2-antiplasmin, and thrombin activatable fibrinolysis inhibitor (TAFI). In this review, we focus on the mechanisms by which PAI-1 governs total fibrinolytic activity to provide its essential role in many hemostatic disorders, including fibrinolytic shutdown after trauma. PAI-1 is a member of the serine protease inhibitor (SERPIN) superfamily and inhibits the protease activities of plasminogen activators (PAs) by forming complexes with PAs, thereby regulating fibrinolysis. The major PA in the vasculature is tissue-type PA (tPA) which is secreted from vascular endothelial cells (VECs) as an active enzyme and is retained on the surface of VECs. PAI-1, existing in molar excess to tPA in plasma, regulates the amount of free active tPA in plasma and on the surface of VECs by forming a tPA-PAI-1 complex. Thus, high plasma levels of PAI-1 are directly related to attenuated fibrinolysis and increased risk for thrombosis. Since plasma PAI-1 levels are highly elevated under a variety of pathological conditions, including infection and inflammation, the fibrinolytic potential in plasma and on VECs is readily suppressed to induce fibrinolytic shutdown. A congenital deficiency of PAI-1 in humans, in turn, leads to life-threatening bleeding. These considerations support the contention that PAI-1 is the primary regulator of the initial step of fibrinolysis and governs total fibrinolytic activity.

1. INTRODUCTION

The fibrinolytic system is sequentially composed of plasminogen activation and fibrin degradation [1, 2]. Three distinct inhibitors of the fibrinolytic system that differently regulate these two steps are plasminogen activator inhibitor type-1 (PAI-1), α2-antiplasmin (α2AP) and thrombin activatable fibrinolysis inhibitor (TAFI) [3]. PAI-1 limits the amount of free tissue-type plasminogen activator (tPA) both in plasma and on vascular endothelial cells (VECs), and regulates plasminogen activation potential to dissolve fibrin [4]. α2AP inhibits plasmin activity in plasma and also attenuates plasmin-catalyzed fibrin degradation after it is cross-linked to fibrin by the transglutaminase, activated coagulation factor XIII [5]. Activated TAFI, a carboxypeptidase, catalyzes the removal of C-terminal lysines of fibrin, which are essential for both effective plasminogen activation and effective fibrinolysis [6, 7]. These regulators function in a spatio-temporal manner, and enable the fibrinolytic system to quickly dissolve pathological thrombi, while at the same time protecting physiological haemostatic thrombi from premature lysis [3]. Disruption of the spatio-temporal regulation of the fibrinolytic system naturally leads to the development of either lethal bleeding or thrombotic disorders. A comprehension of the general scheme of fibrinolysis is necessary to evaluate the specific function of a particular component in a holistic manner. An accurate assessment of the contribution of each regulator to the global process of fibrinolysis, however, is problematic, since these regulators function differently in different environments. The lack of an apparent haemostatic phenotype in gene-deficient animals of these regulators [4, 8] also makes it difficult to extensively assess the functions of each regulatory component.

In the present review article, we focus on PAI-1 and discuss how this protein regulates the expression of fibrinolytic activity and how modification of this PAI-1 dependent regulation is causally related to either thrombotic or bleeding disorders. Although attention has focused on the pleiotropic functions of PAI-1 for drug development, we stress its principal function in the vasculature as a PA inhibitor.

2. PAI-1 GOVERNS THE “FIBRINOLYTIC POTENTIAL” IN PLASMA

Plasminogen activation in plasma takes place only in the presence of fibrin, and occurs immediately when a thrombus is formed. This is known as a coagulation-associated enhancement of fibrinolysis, in which the assembly of plasminogen and tPA on the fibrin surface occurs [1]. A template mechanism, as well as a conformational change of Gluplasminogen, the mature native plasminogen having glutamic acid in its N-terminus, to an easily activatable form after binding to fibrin through its lysine binding site (LBS) in kringle 5 is at the basis of these considerations [2, 3, 9, 10] (Fig. 1). Thus, fibrin formation is required to assess the full competence of fibrinolysis in plasma. However, this makes it difficult to identify the extent of the contribution of each regulator to the whole process of fibrinolysis in which all the regulators are involved in a complex manner. We here define the potential to express fibrinolytic activity when fibrin is formed as “fibrinolytic potential” and discuss how this is regulated.

Fig. (1).

Glu-plasminogen changes its conformation from a tight to loose after binding to the fibrin surface.

Glu-plasminogen adopts a tight conformation (a) in the fluid phase in Cl−, but changes its conformation to a more relaxed conformation (c) after binding to fibrin through its lysine binding site in kringle 5. In the relaxed form, PAs easily cleave the 561-562 peptide bond in Glu-plasminogen to generate plasmin. Modified from [9].

The fibrinolytic potential of individual plasma samples varies widely. Several different assay methods have been developed to assess global fibrinolytic activity. Spontaneous plasma clot lysis occurs over days, and is impractical as a laboratory test [11]. The tPA-supplemented plasma clot lysis time is frequently used, but it is also not suitable since large amounts of tPA in excess of PAI-1 are required to obtain measurable clot lysis, which largely modifies the innate fibrinolytic potential. Thus, an euglobulin clot lysis time (ECLT) has been developed to more rapidly assess fibrinolytic potential. The euglobulin fraction of plasma, an isoelectric precipitate at acidic pH (pH 5.2-5.9), contains fibrinogen, plasminogen, tPA, and PAI-1 but does not contain major plasmin inhibitors of α2AP and α2-macroglobulin [12]. Thus, its lysis time is reasonably short and suitable for use as a laboratory test.

The ECLT varies to a large extent among plasma samples obtained from different individuals and shows circadian variations [13]. We observed a statistically significant inverse correlation between ECLT and tPA activity, and significant positive correlations between ECLT and both total and free PAI-1 levels in plasma [14–16]. Notably, the calculated amounts of free tPA, based upon the assumption that tPA and PAI-1 form a high molecular weight complex in plasma, showed significant positive correlations with tPA activity and a negative correlation with ECLT (Fig. 2) [15]. These data clearly suggest that the plasma concentration of PAI-1 in excess of tPA would directly govern free active tPA levels in the plasma, as well as the fibrinolytic potential in plasma, which is responsible for triggering plasmin generation when fibrin is formed. Such regulation of the initial step of fibrinolysis is different from those of other steps of the coagulation and fibrinolysis cascades. Whereas most of the other steps in hemostasis are regulated by more complex cascade-type mechanisms, in which the efficacy of the enzyme generation, as well as the balance between the amounts of the generated enzymes and the corresponding SERPINs are involved, the plasminogen activation step is basically reflected by the balance of plasma concentrations of tPA and PAI-1. This is mainly due to a unique characteristic of tPA of having endogenous activity to interact with both plasminogen and PAI-1 even in its single-chain form [17, 18].

Fig. (2).

Balance between PAI-1 and tPA governs the amounts of free tPA, PA activity and fibrinolytic potential in plasma.

Concentrations of tPA (mass) and PAI-1 (total, complex and free form) as well as tPA activity in plasma, and ECLTs were measured using plasma obtained from 25 male normal volunteers. The amounts of free tPA ([tPA]) were calculated by equation (2) based upon an assumption that tPA and PAI-1 interacts according to equation (1) even in plasma milieu. Calculated free tPA level showed a significant strong correlation (Spearman’s rank correlation coefficient) with tPA activity in plasma, and a negative correlation with ECLT. Modified from [15].

3. REGULATION OF THE FIBRINOLYTIC POTENTIAL ON THE SURFACE OF VECs

tPA possesses another unique characteristic. This protein is secreted from VECs by both regulated and constitutive secretory mechanisms [19]. Employing tPA attached to green fluorescence protein (GFP), we analyzed the secretory dynamics of tPA-GFP by total internal reflection-fluorescence (TIR-F) microscopy. After activation and opening of tPA-containing granules, the released tPA-GFP appeared to be retained on the surface of VECs for longer than several seconds [20]. This is unique to tPA since other peptide substances including insulin are secreted within milliseconds after activation of the corresponding granules. This unique characteristic of tPA contributes to the maintenance of a high fibrinolytic potential on VEC surfaces [21], and the retained tPA on the VEC contributes to rapid fibrin dissolution. Modification of affinities of tPA to the surfaces of both VECs and fibrin naturally alter the efficacy of the dissolution of the fibrin clot on the surface of VECs [22].

The fibrinolytic potential on VEC surfaces is also governed by PAI-1. PAI-1 facilitates the detachment of tPA from VEC surfaces after forming the tPA-PAI-1 complex [20]. When the PAI-1 concentration is elevated in plasma, the amounts of tPA and its activity on VECs are suppressed, and the levels of the tPA-PAI-1 complex increase in plasma. This is similar to the amounts of free tPA and its activity in plasma both of which are principally determined by the balance of tPA and PAI-1 [14, 15]. All of this is governed by the unique characteristic of tPA which possess endogenous activity as a single chain form [17, 23].

4. SUPPRESSED FIBRINOLYTIC POTENTIAL BY PAI-1 IS RELATED TO THROMBOGENESIS

An increase in PAI-1 in plasma increases the risk of thrombosis by lowering the free tPA concentration and fibrinolytic potential in plasma [24]. High plasma levels of tPA antigen have been also reported as a risk factor for stroke and mitral infarction [25, 26]. An elevated level of plasma tPA antigen, in the form of PA-PAI-1 complex, indicates suppression of fibrinolytic potential on the surface of VECs. This results from the dissociation of retained tPA as a result of the formation of the tPA-PAI-1 complex [20]. The inner surface of the vascular wall is the site for thrombus formation, and thus rapid lysis after thrombus formation through highly maintained fibrinolytic potential is essential to maintain vascular patency. The dissociation of free active tPA from VEC surfaces by PAI-1 naturally increases the risk of thrombosis.

PAI-1 contained in platelets also plays an important role under physiological and pathological conditions [27, 28]. Although the regulatory mechanism of its secretion has not yet been clarified, PAI-1 originating in platelets contributes to thrombogenesis under shear stress [29–31], which is related to arterial thrombogenicity.

Fluctuations in fibrinolytic activity due to gender difference, aging, circadian regulation, exercise loading, lipid metabolism disorders, and infection/inflammation, have been shown to be caused by changes in plasma PAI-1 levels due to its modified gene expression [24, 32]. Several cytokines are known to enhance PAI-1 gene expression and to increase plasma PAI-1 concentrations as much as several hundredfold, which is a major mechanism for thrombogenesis and/or multiple organ failure in infection/inflammation [33].

5. INCREASED PAI-1 CAUSES FIBRINOLYSIS SHUTDOWN AFTER TRAUMA

Trauma-induced coagulopathy (TIC) is a well-known phenomenon in which uncontrollable bleeding occurs immediately after severe brain injury. This has been postulated to be associated with inflammation, dysfunctional vascular endothelial cells, shock, and organ failure [34]. Effective reduction of blood loss after tranexamic acid infusion, and/or blood component transfusion are typically seen [35]. TIC patients also show an increase in coagulation markers, including soluble fibrin monomer and the thrombin-antithrombin complex, as well as fibrinolysis markers, including the plasmin-α2AP complex in plasma [34]. In animal models, brain-derived microparticles bearing tissue factor on their surface has been shown to exist in circulation, which was then rapidly transferred to lung tissue [36]. These data suggest that the pathogenesis of the uncontrollable bleeding in TIC is considered to be consumptive coagulopathy associated with enhanced fibrinolysis, which is typically seen in disseminated intravascular coagulation (DIC). The loss of platelet function observed in TIC [37], however, seems an exception since this has not been reported in DIC. It is possible that TIC could be considered as a form of DIC which develops a wide range of phenotypes depending on the pathogenesis of the underlying disease [38]. A fibrinolysis-dominant phenotype of DIC, typically seen in abdominal aortic aneurysms having large thrombi [39], and in vascular intimal carcinogenesis, having high expression of annexin A2 [40], seems similar to TIC. Possible impairment of platelet function in these DIC patients should be analyzed in order to assess whether a highly elevated fibrinolytic activity is involved in the shedding of platelets’ receptors.

Recently, a stage of suppressed fibrinolysis termed “fibrinolytic shutdown” has received attention. This is typically seen in TIC following a hyperfibrinolytic stage, and contributes to the development of organ failure through microthrombi accumulation [41, 42]. A sudden increase in PAI-1 in plasma, as an acute phase reactant corresponding to systemic inflammation, is considered responsible for this phenomenon [43].

For the uncontrollable bleeding in TIC, treatment by tranexamic acid (TXA) is recommended and successful reduction in the amounts of bleeding has been reported [35]. Tranexamic acid suppresses plasminogen activation on the fibrin surface by inhibition of plasminogen binding to C-terminal Lys residues on the fibrin surface [2]. Success in retarding bleeding by TXA further confirms that the uncontrollable bleeding in TIC is caused by elevated fibrinolytic activity. Based on the success in TIC, TXA is now used in the treatment of other uncontrollable bleeding states, including postpartum hemorrhage [44]. However, considering that a stage of “fibrinolytic shutdown” would follow a “hyperfibrinolysis” stage, a question naturally arises as to whether tranexamic acid treatment is necessary and/or safe in TIC [42, 45]. TXA treatment should be considered in underlying diseases only regarding patients whose fibrinolytic activity is highly enhanced. For the safe use of TXA, rapid and accurate laboratory tests are required to assess fibrinolytic activity. Further, understanding the pathogenesis of the sequential modification of fibrinolytic activity after severe insults is prerequisite.

6. CONGENITAL PAI-1 DEFICIENCIES IN HUMAN DEVELOP LIFE-THREATENING BLEEDING

A genetically confirmed PAI-1 deficiency was first reported in an Amish family, the proband of which showed repeated severe bleeding episodes [46]. Two distinct Japanese cases also showed repeated severe bleeding after tooth extractions, surgery, and obstetric complications [4, 47, 48]. Delayed bleeding is the phenotype frequently seen in fibrinolytic disorders [4, 49]. The observations that plasma levels of fibrinogen and plasminogen are maintained in these patients when they are not exposed to insults [47, 48] suggest that fibrinolytic activity is expressed at low levels if fibrin is not formed, even in the absence of PAI-1. In turn, when fibrin is formed, fibrinolytic activity is highly expressed resulting from premature lysis of haemostatic thrombi and delayed bleeding. Pregnancy is such an occasion in which uncontrollable bleeding can occur. In order to maintain pregnancy, highly intensive interventions may be required although the amounts of bleeding are still extraordinary [4, 47, 48].

While PAI-1-deficient animals did not show a bleeding phenotype, these phenotypes observed in human homozygote cases of PAI-1 deficiency clearly demonstrate its principal role in fibrinolysis. Further, these results revealed the limitations of some animal models in the analyses of human pathology [4].

7. PLEIOTROPIC FUNCTION OF PAI-1 AND PROSPECTS OF PAI-1 INHIBITORS

In addition to its primary function to regulate PAs’ activity, PAI-1 has a wide range of pleiotropic activities [50]. Modifications of cellular functions including mobility and binding capacity to matrix proteins are well known, which are related to tumor metastasis, migration of inflammatory cells, angiogenesis, hematopoiesis, etc. These may also explain impaired wound healing in the PAI-1 deficient patients [4, 47, 48]. Underlying mechanisms are explained by the specific function of PAI-1 to alter cellular binding capacity to vitronectin, either through integrins or urokinase-type PA (uPA) / uPA receptor (uPAR) [50]. Modification of cellular binding to matrix proteins naturally alters the associated signal transductions and cellular function as well [51, 52]. Inhibition of the activity of uPAR-bound uPA by PAI-1 also modulates cellular functions [51]. Another intriguing function of PAI-1 is related to aging which was evidenced by a longer leukocyte telomere length in heterozygotes of a congenital PAI-1 deficient family [53]. Retardation of the early senescence found in Klotho-deficient mice by either a simultaneous PAI-1 gene knock-out or a treatment by PAI-1 inhibitor [54] also suggests that PAI-1 plays a role in senescence.

To reduce the risk for thrombosis caused by increased levels of plasma PAI-1, many pharmaceutical companies have developed PAI-1 inhibitors [55, 56], but none of these have been employed in the clinic as anti-thrombotic drugs at this time. To facilitate mobilization of hematopoietic stem cells, however, a phase-1 clinical trial for one of the small molecular weight PAI-1 inhibitors, TM5614, has been completed [57]. Interest in the clinical use of PAI-1 inhibitors is expanding for targeting a wide range of pleiotropic functions of PAI-1.

CONCLUSION

In this review, we considered how PAI-1 directly governs fibrinolytic potential in plasma, as well as on VECs, and why increased PAI-1 levels in the plasma directly induce thrombogenesis. The implication of an elevated PAI-1 level over that of tPA is totally different from the modification of other SERPIN concentrations over those of their target proteases. We believe that such knowledge is mandatory to understand the underlying mechanisms of hemostasis, as well as to establish novel treatments for fibrinolytic shutdown in which appropriate use of PAI-1 inhibitors under development would be one of the choices.