Chapter 135: Fibrinolysis and Thrombolysis

In response to vascular injury, fibrin, the insoluble end product of the action of thrombin on fibrinogen, is deposited in blood vessels, thus stemming the flow of blood. Once the vessel has healed, the fibrinolytic system is activated, converting fibrin to its soluble degradation products through the action of the serine protease, plasmin (Fig. 135–1A). Fibrinolysis is subject to precise control because of the actions of multiple activators, inhibitors, and cofactors.¹ In addition, receptors expressed by endothelial, monocytoid, and myeloid cells provide specialized, protected environments where plasmin can be generated without compromise by circulating inhibitors (Fig. 135–1B).^2,3 Beyond its more traditional role in fibrin degradation, the fibrinolytic system also supports a variety of tissue remodeling mechanisms. This chapter reviews the fundamental features of plasmin generation, considers the major clinical syndromes resulting from abnormalities in fibrinolysis, and discusses approaches to fibrinolytic and antifibrinolytic therapy.

PLASMINOGEN

Synthesized primarily in the liver,^4,5 plasminogen is a Mr approximately 92,000 single–chain proenzyme that circulates in plasma at a concentration of approximately 1.5 μM⁶ (Table 135–1). The plasma half-life of plasminogen in adults is approximately 2 days.⁷ Its 791 amino acids are crosslinked by 24 disulfide bridges, 16 of which give rise to five homologous triple loop structures called “kringles” (Fig. 135–2).⁸ The first (K1) and fourth (K4) of these 80–amino-acid, Mr approximately 10,000 structures impart high- and low-affinity lysine binding, respectively.⁹ The lysine–binding domains of plasminogen appear to mediate its specific interactions with fibrin, cell surface receptors, and other proteins, including its circulating inhibitor α₂–plasmin inhibitor (α₂-PI).^{10,11,12,13,14}

Posttranslational modification of plasminogen results in two glycosylation variants (forms 1 and 2; see Table 135–1).^15,16,17 O–linked oligosaccharide, consisting of sialic acid, galactose, and galactosamine resident on Thr345, is common to both forms. Only form 2, however, contains N–linked oligosaccharide on Asn 288 that is comprised of sialic acid, galactose, glucosamine, and mannose. The carbohydrate portion of plasminogen appears to regulate its affinity for cellular receptors, and may also specify its physiologic degradation pathway.

Activation of plasminogen results from cleavage of a single Arg–Val peptide bond at position 560–561,⁶ giving rise to the active protease, plasmin (see Table 135–1). Plasmin contains a typical serine protease catalytic triad (His 602, Asp 645, and Ser 740), but exhibits broad substrate specificity when compared to other proteases of this class.¹⁸ The circulating form of plasminogen, aminoterminal glutamic acid plasminogen (Glu–Plg), can be converted by limited proteolysis to several modified forms known collectively as Lys–Plg.^19,20 Hydrolysis of the Lys77–Lys78 peptide bond gives rise to a conformationally modified form of the zymogen that more readily binds fibrin, displays two- to threefold higher avidity for cellular receptors, and is activated 10 to 20 times more rapidly than Glu-Plg^11,21,22 Lys–Plg does not normally circulate in plasma,²¹ but has been identified on cell surfaces.^23,24

Spanning 52.5 kb of DNA on chromosome 6q26–27, the Plg gene consists of 19 exons^25,26 and directs expression of a 2.7-kb mRNA⁸ (see Fig. 135–2). The 5′ upstream region of the Plg gene contains two regulatory elements common to genes for acute–phase reactants (CTGGGA) and six interleukin (IL)-6 response elements.²⁶ Plg gene activity, moreover, is stimulated by the acute-phase-mediator IL-6 both in vitro and in vivo.²⁷ The gene is closely linked and structurally related to that of apolipoprotein(a), an apoprotein associated with the highly atherogenic low density lipoprotein–like particle lipoprotein(a),²⁸ and more distantly related to other kringle-containing proteins such as tissue-type plasminogen activator (t-PA), urokinase-type plasminogen activator (u-PA), macrophage-stimulating protein, and hepatocyte growth factor.^{29,30,31,32,33,34}

The Physiologic Functions of Plasmin(ogen)

Mice made completely deficient in Plg through gene targeting undergo normal embryogenesis and development, are fertile, and survive to adulthood (Table 135–2).^35,36 These animals display runting and ligneous conjunctivitis,³⁷ and harbor spontaneous thrombi in the liver, stomach, colon, rectum, lung, and pancreas, as well as fibrin deposition in the liver and ulcerative lesions in the gastrointestinal tract and rectum. These results suggested that Plg is not strictly required for normal development, but does play a central role in fibrin homeostasis. In humans, Plg deficiency presents most often with ligneous mucositis as a result of fibrin deposition, and is rarely a cause of macrovascular thrombosis (see Fibrinolytic Deficiency and Thrombosis below).

PLASMINOGEN ACTIVATORS

Tissue-Type Plasminogen Activator

One of two major endogenous Plg activators, t-PA consists of 527 amino acids comprising a glycoprotein of Mr approximately 72,000 (see Table 135–1).³⁸ t-PA contains five structural domains including a fibronectin–like “finger,” an epidermal growth factor–like domain, two “kringle” structures homologous to those of Plg, and a serine protease domain (see Fig. 135–2). Cleavage of the Arg275–Ile276 peptide bond by plasmin converts t-PA to a disulfide–linked, two–chain form.³⁸ Although single–chain t-PA is less active than two–chain t-PA in the fluid phase, both forms demonstrate equivalent activity when fibrin–bound.³⁹

The two glycosylation forms of t-PA are distinguishable by the presence (type 1) or absence (type 2) of a complex N–linked oligosaccharide moiety on Asn184 (see Table 135–1).^40,41 Both types, however, contain high mannose carbohydrate on Asn 117, complex oligosaccharide on Asn448, and an O–linked α–fucose residue on Thr61.⁴² The carbohydrate moieties of t-PA may modulate its functional activity, regulate its binding to cell surface receptors, and specify its degradation pathways.

Located on chromosome 8p12–q11.2, the gene for human t-PA is encoded by 14 exons spanning a total of 36.6 kb (see Fig. 135–2).^43,44,45 Although exon 1 encodes a 58-nucleotide mRNA leader sequence, each of the structural domains of t-PA is encoded by one or two of the remaining 13 exons. This arrangement suggests that the t-PA gene arose by an evolutionary process called “exon shuffling,” whereby functionally related genes evolved through rearrangement of exons encoding autonomous domains. Consistent with this hypothesis, deletion of exons encoding the fibronectin–like finger or kringle 2, but not kringle 1, domains of t-PA results in expression of mutants resistant to the cofactor activity of fibrin, while catalytic activity in the absence of fibrin remains intact.⁴⁶

The proximal promoter of the human t-PA gene contains binding sequences for potentially important transcriptional factors including AP1, NF1, SP1, and AP2,^47,48 as well as a potential cyclic adenosine monophosphate (cAMP)–responsive element (CRE).⁴⁹ In vitro, many agents have been shown to exert small effects on the expression of t-PA mRNA, but relatively few enhance t-PA synthesis without augmenting plasminogen activator inhibitor (PAI)–1 synthesis as well. Agents that regulate t-PA gene expression independently of PAI–1 include histamine, butyrate, retinoids, arterial levels of shear stress, and dexamethasone.^{50,51,52,53,54,55} Forskolin, which increases intracellular cAMP levels, has been reported to decrease synthesis of both t-PA and PAI–1.^48,56

In the vascular system, t-PA is synthesized and secreted primarily by endothelial cells belonging to a restricted set of blood vessels. In rodents, t-PA expression appears in 7- to 30-μm diameter precapillary arterioles in the lung, postcapillary venules, and vasa vasorum; much less expression is seen in endothelial cells of the femoral artery, femoral vein, carotid artery, or aorta.⁵⁷ In the mouse lung, bronchial arteriolar endothelial cells express t-PA antigen, especially at branch points, while pulmonary blood vessels are uniformly negative.^51,58,59,60 t-PA has also been detected in sympathetic neurons associated with the blood vessel wall.⁶¹ Release of t-PA is governed by a variety of stimuli such as thrombin, histamine, bradykinin, epinephrine, acetylcholine, arginine vasopressin, gonadotropins, exercise, venous occlusion, and shear stress.^50,51,62,63 Its circulating half-life is exceedingly short (~5 minutes). Alone, t-PA is actually a poor activator of Plg, but, in the presence of fibrin, the catalytic efficiency of t-PA–dependent plasmin generation (k_cat/K_m) increases by at least two orders of magnitude.²² This is the result of a dramatic increase in affinity (decreased K_m) between t-PA and its substrate Plg in the presence of fibrin. Although it is also expressed by extravascular cells, t-PA appears to represent the major intravascular activator of Plg.¹⁸

Urokinase

The second endogenous Plg activator, single–chain u-PA or prourokinase, is a Mr approximately 54,000 glycoprotein consisting of 411 amino acids (see Table 135–1). u-PA possesses an epidermal growth factor–like domain, a single Plg-like “kringle,” and a classical catalytic triad (His204, Asp255, Ser356) within its serine protease domain (see Fig. 135–2).⁶⁴ Cleavage of the Lys158–Ile159 peptide bond by plasmin or kallikrein converts single-chain u-PA to a disulfide–linked two–chain derivative.⁶⁵ Located on chromosome 10, the human u-PA gene is encoded by 11 exons spanning 6.4 kb, and expressed by activated endothelial cells, macrophages, renal epithelial cells, and some tumor cells.^66,67 Its intron–exon structure is closely related to that of the t-PA gene. u-PA expression appears to be induced during neoplastic transformation, possibly through the action of transcription factors AP1 and AP2.⁶⁸ Other in vitro u-PA inducers include hormones, angiogenic growth factors, and cAMP,⁵⁵ as well as tumor necrosis factor and transforming growth factor-β (TGF-β).^69,70,71

Two–chain u-PA occurs in both high– (Mr 54,000) and low-molecular-weight (Mr 33,000) forms that differ by the presence or absence, respectively, of a 135–residue aminoterminal fragment released by plasmin cleavage between Lys135 and Lys136.^72,73 Although both forms are capable of activating Plg, only the high-molecular-weight form binds to the u-PA receptor (see Urokinase Plasminogen Activator Receptors below). u-PA has much lower affinity for fibrin than t-PA, and is an effective Plg activator both in the presence and in the absence of fibrin.^74,75

Accessory Plasminogen Activators and Fibrinolysins

Under certain conditions, proteases traditionally classified within the intrinsic arm of the coagulation cascade have been shown to be capable of activating Plg directly. These include kallikrein, factor XIa, and factor XIIa.^76,77,78 These proteases, however, normally account for no more than 15 percent of total plasmin generating activity in plasma.⁷⁹ In addition, the membrane type 1 matrix metalloproteinase (MT1- MMP) appears to exert fibrinolytic activity in the absence of Plg, and may explain the unexpectedly mild phenotype observed in Plg-deficient mice.⁸⁰

Physiologic Function of the Plasminogen Activators

Because there are no clinical examples of complete deficiency of t-PA or u-PA in humans, except for patients with deficient release in the setting of chronic renal disease and hypertension,^81,82,83 the most compelling data regarding the physiologic functions of t-PA and u-PA come from gene disruption analysis in mice.⁸⁴ Both u-PA and t-PA null deletion mice exhibit normal fertility and embryonic development. However, u-PA–/– mice develop rectal prolapse, nonhealing ulcerations of the face and eyelids, and occasional fibrin deposition in tissues. Although they show normal lysis rates of pulmonary clots injected via the jugular vein, endotoxin–induced microvascular thrombus formation is significantly enhanced. t-PA–deficient mice also display a normal spontaneous phenotype, but have a decreased rate of lysis of artificially induced pulmonary thrombi, as well as enhanced thrombus formation, in response to injection of endotoxin. Like Plg–/– mice, mice doubly deficient in t-PA and u-PA (t-PA–/–; u-PA–/–) exhibit rectal prolapse, nonhealing ulceration, runting, and cachexia, with extensive fibrin deposition in liver, intestine, gonads, and lung. Not surprisingly, clot lysis is also markedly impaired.

INHIBITORS OF FIBRINOLYSIS

Plasmin Inhibitors

The action of plasmin is negatively modulated by a family of serine protease inhibitors, called serpins (see Table 135–1).⁸⁵ Serpins form an irreversible complex with the active site serine of their target protease following proteolytic cleavage of the inhibitor by the target protease. Within such a complex, both protease and inhibitor lose their activity.

A single–chain glycoprotein of Mr approximately 70,000, α₂-PI is synthesized primarily in the liver, circulates in plasma at relatively high concentrations (~0.9 μM), and enjoys a plasma half–life of 2.4 days (see Table 135–1).⁸⁶ This serpin contains approximately 13 percent carbohydrate by mass and consists of 452 amino acids with two disulfide bridges.⁸⁷ In humans, the gene is located on chromosome 18 and contains 10 exons distributed over 16 kb of DNA.⁸⁸ The promoter region of the α₂-PI gene contains a hepatitis B–like enhancer element that directs tissue–specific expression in the liver.⁸⁷ α₂-PI is also a constituent of platelet α granules.⁸⁹ Plasmin released into flowing blood or in the vicinity of a platelet–rich thrombus, is immediately neutralized upon forming an irreversible 1:1 stoichiometric, lysine-binding site–dependent complex with α₂-PI. Interaction with plasmin is accompanied by cleavage of the Arg364–Met365 peptide bond, and the resulting covalent complexes are cleared in the liver. Mice globally deficient in α₂-PI display reduced fibrin deposition, following treatment with endotoxin and enhanced lysis of injected plasma clots, but no spontaneous bleeding (see Table 35–2).⁹⁰

Several additional proteins can act as plasmin inhibitors (see Table 135–1). α₂–Macroglobulin is a Mr 725,000 dimeric protein synthesized by endothelial cells and macrophages, and found in platelet α granules. This nonserpin inhibits plasmin with approximately 10 percent of the efficiency exhibited by α₂–PI by forming noncovalent complexes with several distinct serine proteases.⁹¹ C₁-esterase inhibitor can inhibit t-PA in plasma⁹² and protease nexin may function as a noncirculating cell surface inhibitor of trypsin, thrombin, factor Xa, urokinase, or plasmin, resulting in protease–inhibitor complexes that are endocytosed via a specific nexin receptor.^93,94 The purpose of these multiple plasmin inhibitors is to guard against premature plasmin activation and subsequent degradation of fibrinogen, until intravascular fibrin begins to appear/.

Plasminogen Activator Inhibitors

Plasminogen Activator Inhibitor-1 Of the two major Plg activator inhibitors, PAI–1 is the most ubiquitous (see Table 135–1).⁹⁵ This Mr approximately 52,000 single–chain, cysteine–less glycoprotein is released by endothelial cells, monocytes, macrophages, hepatocytes, adipocytes, and platelets.^96,97,98 Release of PAI-1 is stimulated by many cytokines, growth factors, and lipoproteins common to the global inflammatory response.^{69,70,99,100,101} The PAI-1 gene consists of nine exons, spanning 12.2 kb on chromosome 7q21.3–q22.¹⁰² The serpin-reactive site is located at Arg346–Met347, and activity of this labile serpin is stabilized upon complex formation with vitronectin, a component of plasma and pericellular matrix.^103,104,105

Regulation of PAI-1 gene expression is complex.^106,107 The upstream regulatory region of the human PAI-1 gene contains a strong endothelial cell/fibroblast-specific element^108,109 a glucocorticoid–responsive enhancer,¹⁰⁹ and TGF-β responsive elements.¹¹⁰ TGF-β is known to stimulate fos and jun, the two components of the AP1 complex, and an AP1 binding site (GGAGTCA) is located upstream of the PAI–1 cap site.¹¹¹ Agents shown to enhance expression of PAI–1 at the message level, the protein level, or both, without affecting t-PA synthesis, include the inflammatory cytokines lipopolysaccharide, IL–1, tumor necrosis factor–α,^{69,70,99,112,113} TGF-β and basic fibroblast growth factor,^{71,99,110,114} very-low-density lipoprotein and lipoprotein(a),^115,116 angiotensin II,¹¹⁷ thrombin,^118,119 and phorbol esters.¹²⁰ In addition, endothelial cell PAI–1 is downregulated by forskolin⁵⁶ and by endothelial cell growth factor in the presence of heparin.¹²¹

PAI-1 is the most important and rapidly acting physiologic inhibitor of both t-PA and u-PA. Transgenic mice that overexpress PAI–1 exhibit thrombotic occlusion of tail veins and swelling of hind limbs within 2 weeks of birth.¹²² Mice deficient in PAI–1, on the other hand, exhibit normal fertility, viability, tissue histology, and development, and are resistant to endotoxin-induced thrombosis, but show no evidence of overt hemorrhage (see Table 135–2).^123,124 These observations contrast with the moderately severe bleeding disorder observed in a human patient with complete PAI-1 deficiency.¹²⁵

Plasminogen Activator Inhibitor-2 Originally purified from human placenta, PAI-2 is a 393-amino-acid member of the serpin family whose reactive site is the Arg358–Thr359 peptide bond¹²⁶ (see Table 135–1). The gene encoding PAI–2 is located on chromosome 18q21–23, spans 16.5 kb, and contains eight exons.¹²⁷ PAI–2 exists as both a Mr 47,000 nonglycosylated intracellular form and an Mr 60,000 glycosylated form secreted by leukocytes and fibrosarcoma cells. Functionally, PAI–2 inhibits both two–chain t-PA and two–chain u-PA with comparable efficiency (second order rate constants 10⁵ M^–1s^–1). However, it is less effective toward single-chain t-PA (second order rate constant 10³ M^–1s^–1), and does not inhibit prourokinase.

Significant levels of PAI–2 are found in human plasma primarily during pregnancy. The gene’s 5′–untranslated region contains a potent silencer, the PAUSE-1 element, which may be responsible for its low level of expression in nonpregnant individuals.^127,128 The 3′–downstream sequences include the TTATTTAT motif which has been identified with inflammatory mediators.^129,130 In macrophages in vitro, secretion of PAI-2 is enhanced by endotoxin and phorbol esters^130,131 and dexamethasone decreases PAI–2 expression in HT–1080 cells.⁵⁵

Thrombin-Activatable Fibrinolysis Inhibitor

Thrombin-activatable fibrinolysis inhibitor (TAFI) is a plasma carboxypeptidase with specificity for carboxy terminal arginine and lysine residues.¹³² The action of TAFI eliminates binding sites for Plg and t-PA on fibrin.¹³³ This single-chain Mr 60,000 polypeptide circulates in plasma at concentrations of approximately 75 nM, and undergoes limited proteolysis in the presence of thrombin, which leads to its activation.^134,135,136 The profibrinolytic effect of activated protein C in plasma is a result of its ability to inactivate coagulation factors Va and VIIIa, which reduces activation of thrombin, the primary activator of TAFI.¹³² The profibrinolytic effect of activated protein C in an in vitro plasma-based system was TAFI-dependent,¹³⁷ and, in a system of purified components, TAFI has been shown to downregulate t-PA–induced fibrinolysis half-maximally at a concentration of approximately 1 nM, which is 2 percent of its concentration in plasma.¹³⁸ Inhibition of either the intrinsic pathway of coagulation or TAFI itself results in a doubling of endogenous clot lysis in an in vivo rabbit jugular vein model of thrombolysis.^139,140 TAFI-deficient mice display increased lysis of plasma clots and reduced injury-induced venous thrombosis (see Table 135–2).^141,142 In plasma, TAFI may regulate Plg binding to both cell surface receptors and to fibrin.¹⁴³

CELLULAR RECEPTORS

A large number of structurally diverse fibrinolytic “activation” and “clearance” receptors have been described. Here, we focus on endothelial cell activation receptors that are likely to contribute to homeostatic control of plasmin activity (see Table 135–1).² Clearance receptors eliminate plasmin and Plg activators from the blood or focal microenvironments.

Activation Receptors

Plasminogen Receptors Proposed Plg receptors include α–enolase, glycoprotein IIb/IIIa complex, the Heymann nephritis antigen, amphoterin, the annexin A2/S100A10 complex, histone H2B, and plasminogen receptor-KT (Plg-R_KT)^2,3; these are expressed on a wide spectrum of cells, including monocytoid cells, platelets, renal epithelial cells, neuroblastoma cells, endothelial cells, and tumor cells.^{144,145,146,147,148,149,150,151} Typically, Plg receptors interact with the kringle structures of Plg through carboxyl–terminal lysine residues that are either present on the native protein, or generated by limited proteolysis.¹⁴⁴

Urokinase Plasminogen Activator Receptor The u-PA receptor (uPAR) is expressed on monocytes, macrophages, fibroblasts, endothelial cells, and many tumor cells (see Table 135–1).^152,153 uPAR complementary DNA (cDNA) was cloned and sequenced from a human fibroblast cDNA library¹⁵⁴ and encodes a protein of 313 amino acids with a 21–residue signal peptide. The gene consists of seven exons distributed over 23 kb of genomic DNA, and places this glycoprotein within the Ly-1/elapid venom toxin superfamily of cysteine rich proteins.^155,156 uPAR is anchored to the plasma membrane through glycosylphosphatidylinositol linkages.¹⁵⁷ u-PA bound to its receptor maintains its activity and susceptibility to the physiologic inhibitor, PAI–1.¹⁵⁸ Formation of u-PA–PAI-1 complexes hastens clearance of u-PA by hepatic or monocytoid cells.^{158,159,160,161}

Although originally thought to function only as a means of localizing Plg activation to the cell surface, uPAR now appears to play a central role in cellular signaling and adhesion events.^152,162 The uPAR-deficient mouse has normal development and fertility, and unimpaired fibrin clot lysis (see Table 135–2).^163,164 uPAR binds the adhesive glycoprotein vitronectin at a site distinct from the u-PA binding domain^165,166 and u-PA transfected renal epithelial cells acquire enhanced adhesion to vitronectin while they lose their adhesion to fibronectin.¹⁶⁷ uPAR, furthermore, colocalizes with integrins in focal contacts and at the leading edge of migrating cells,¹⁶⁸ and also associates with caveolin, a major component of caveolae, structures abundant in endothelial cells and thought to participate in signaling events.^169,170,171 In addition, cleaved and soluble forms of uPAR have recently been detected in the sera of patients with cancer, and these modified forms are thought to regulate the activity of several receptors involved in inflammatory and angiogenic responses.¹⁵³

The Annexin A2-S100a10 System Annexin A2, a Mr 36,000, 339-amino-acid member of the annexin superfamily of calcium-dependent, phospholipid-binding proteins, forms a heterotetramer with the S100 family protein, S100A10 (see Table 135–1).^172,173,174 It is highly conserved, and abundantly expressed on endothelial cells,^{175,176,177,178} monocyte/macrophages,^179,180 early myeloid cells,¹⁸¹ developing neuronal cells,¹⁸² and some tumor cells.^183,184,185 All of the more than 60 annexin family members have in common a conserved membrane-binding C-terminal “core” region and a more variable N-terminal “tail.”¹⁸⁶ The human annexin A2 gene consists of 13 exons distributed over 40 kb of genomic DNA on chromosome 15 (15q21).¹⁸⁷

Annexin A2 is unique among fibrinolytic receptors in that it possesses binding affinity for both Plg (Kd 114 nM)¹⁴⁸ and t-PA (Kd 30 nM), but not u-PA.¹⁴⁹ In a fluid phase system of purified proteins, native human annexin A2 stimulates the catalytic efficiency of t-PA–dependent Plg activation by 60-fold.¹⁸⁸ This effect is completely inhibited in the presence of lysine analogues or upon treatment of annexin A2 with carboxypeptidase B, an agent that removes basic carboxyl-terminal amino acids. Although it lacks a classical signal peptide, annexin A2 is constitutively translocated to the endothelial cell surface within 16 hours of its biosynthesis. This translocation event can be stimulated either by thrombin or by heat stress, in a process that requires phosphorylation of annexin A2 at Tyr23, the action of a Src family kinase, and the presence of the annexin A2 binding protein p11 (S100A10).¹⁸⁹

At the cell surface, A2 binds phospholipid via core repeat 2, which contains the linear amino acid sequence KGLGT and downstream aspartate residue (Asp 161); together these moieties constitute a classical “annexin” motif.¹⁹⁰ The annexin A2 heterotetramer, which consists of two A2 monomers and two protein p11 subunits and constitutes the cell surface form of A2, appears to have even greater stimulatory effects on t-PA–dependent plasmin generation.¹⁷⁷ Interestingly, A2 regulates endogenous levels of protein p11 in the endothelial cell by masking a polyubiquitination site on p11, which otherwise directs p11 to the proteasome where it is rapidly degraded.¹⁹¹

Plg and t-PA appear to bind to distinct domains. Lys307 appears to be crucial for the effective interaction of Plg with annexin A2, and may be revealed upon limited proteolysis of the parent protein.¹⁸⁸ The atherogenic low-density lipoprotein (LDL)-like particle, lipoprotein(a), competes with Plg for binding to annexin A2 in vitro,¹⁹² thereby reducing cell surface plasmin generation. t-PA binding to annexin A2 requires a domain consisting of residues 8 to 13 (LCKLSL) within the receptor’s amino terminal “tail” domain.¹⁹³ This region is a target for homocysteine (HC), a thiol-containing amino acid that accumulates in association with nutritional deficiencies of vitamin B₆, vitamin B₁₂, or folic acid, or in inherited abnormalities of cystathionine β-synthase, methylenetetrahydrofolate reductase, or methionine synthase,¹⁹⁴ and is associated with atherothrombotic disease.^195,196 In vitro, HC impairs t-PA–dependent plasmin generation at the endothelial cell surface by approximately 50 percent¹⁹⁷ by forming a covalent derivative with Cys,¹⁹⁷ and mice with diet-induced hyperhomocysteinemia have deficient annexin A2 function¹⁹⁸ The half-maximal dose of HC for inhibition of t-PA binding to annexin A2 is approximately 11 μM HC, a value close to the upper limit of normal for HC in plasma (12 μM).

The important role of S100A10 in fibrin balance has recently been underscored. S100A10–/– mice display increased deposition of fibrin in the vasculature and reduced clearance of batroxobin-induced vascular thrombi, and S100A10-deficient endothelial cells demonstrate a 40 percent reduction in Plg binding and plasmin generation in vitro (see Table 135–2).¹⁹⁹ S100A10 also appears to contribute to Plg-dependent macrophage invasion in vitro by enhancing plasmin-dependent activation of matrix metalloporetinase-9.²⁰⁰

Several studies suggest a physiologic role for the annexin A2 system in fibrin homeostasis. First, blast cells from human patients with acute promyelocytic leukemia overexpress annexin A2 in proportion to their degree of hyperfibrinolytic coagulopathy¹⁸¹; S100A10 also appears to be upregulated by the PML-RAR-α oncoprotein,²⁰¹ and both annexin A2 and S100A10 are downregulated by treatment with all-trans-retinoic acid. Second, in rats, arterial thrombosis can be significantly attenuated by pretreatment with intravenous annexin A2.²⁰² Third, the prevalence of high-titer anti–annexin A2 antibodies correlates with a history of severe thrombosis in humans with antiphospholipid syndrome and in a cohort of individuals with cerebral venous thrombosis.^203,204 Finally, mice with total deficiency of annexin A2 display impaired clearance of artificial arterial thrombi, fibrin deposition in the microvasculature, and angiogenic defects in a variety of tissues (see Table 135–2).²⁰⁵

Clearance Receptors

Clearance of serpin-enzyme complexes, such as t-PA–PAI–1 and u-PA–PAI-1, occurs mainly in the liver, and is mediated by a large two–chain receptor called the LDL receptor–related protein 1 (LRP1).^206,207 LRP1 binds a large number of serpin-protease complexes and other ligands, indicating a multifunctional role in mammalian physiology. An additional Mr 39,000 “receptor associated protein” copurifies with LRP1 and appears to regulate the binding and uptake of LRP1 ligands.²⁰⁸ Interestingly, LRP1 “knockout” embryos undergo developmental arrest by 13.5 days after conception, suggesting that regulation of serine protease activity may be crucial for early embryogenesis.^209,210 Although PAI–1–independent clearance pathways for t-PA have been proposed involving the mannose receptor,²¹¹ or an α–fucose–specific receptor,²¹² in vivo studies in mice suggest that LRP1 and the mannose receptor play a dominant role in t-PA clearance.²¹³

DEGRADATION OF FIBRINOGEN AND FIBRIN

Fibrinogen

Plasmin releases carboxyl–terminal Aα and N–terminal fibrinopeptide B moieties from fibrinogen (Fig. 135–3). This reaction is distinct from the proteolytic cleavage of fibrinogen by thrombin, which releases fibrinopeptide A, exposing the Gly–Pro–Arg tripeptide sequence and allowing fibrinogen to polymerize and form insoluble fibrin.²¹⁴ Plasmin cleavage of fibrinogen (Mr 340,000) initially produces carboxyl–terminal fragments from the α chain within the D domain of fibrinogen (Aα fragment).^{205,206,207,208,215,216,217,218}. Simultaneously, but more slowly, the N–terminal segments of the β chains are cleaved, releasing a peptide containing fibrinopeptide B. The resulting Mr approximately 250,000 molecule is termed fragment X and represents a clottable form of fibrinogen. Additional cleavage events may release the Bβ fragment from the β chain’s carboxyl–terminus, and, in a series of subsequent reactions, plasmin cleaves the three polypeptide chains that connect the D and E domains giving rise to free D domain (Mr ~100,000) plus the binodular D–E fragment known as fragment Y (Mr ~150,000). Finally, domains D and E are separated from each other, and some of the N–terminal fibrinopeptide A sites on domain E are also modified. Although fragment X can be converted to fibrin by thrombin, the fragments Y, D, and E are all nonclottable, and, in fact, may inhibit polymerization of fibrinogen.²¹⁹

Fibrin

Plasmin degradation of fibrin leads to a distinct set of molecular products (see Fig. 135–3).²²⁰ Species similar to fragments Y, D, and E, but lacking fibrinopeptide sites, are released from noncrosslinked fibrin. If fibrin has been extensively crosslinked by factor XIII, however, the resulting D fragments are crosslinked to an E domain fragment. Assay of crosslinked D–dimer fragments is employed clinically to identify disseminated intravascular coagulation–like states associated with excessive plasmin–mediated fibrinolysis. Several biologic activities, including inhibition of platelet function,²²¹ potentiation of the hypotensive effects of bradykinin,²²² chemotaxis,²²³ and immune modulation,²²⁴ have been ascribed to fibrin breakdown products.

TISSUE-TYPE PLASMINOGEN ACTIVATOR –MEDIATED PLASMINOGEN ACTIVATION

With or without fibrin, t-PA–mediated activation of Plg follows Michaelis–Menten kinetics.²² In the absence of fibrin, t-PA is a weak activator of Plg. However, in the presence of fibrin, the catalytic efficiency (k_cat/K_m) of t-PA–dependent Plg activation is enhanced by approximately 500-fold. This is the basis for its specificity as a lytic agent in the treatment of thrombosis. The affinity between t-PA and Plg in the absence of fibrin is low (K_m 65 μM), but increases significantly in its presence (K_m 0.16 μM), even though the catalytic rate constant remains essentially unchanged (kcat ~0.05 sec^–1). When plasmin forms on the fibrin surface, both its lysine binding sites and its active site are occupied. Thus, it is relatively protected from its physiologic inhibitor, α₂-PI.²²⁵

The interaction of t-PA with fibrin is probably initiated by its “finger” domain. However, once fibrin is modified by plasmin, carboxy-terminal lysine residues are generated, and these become binding sites for “kringle” 2 of t-PA and “kringles” 1 and 4 of Plg.²²⁶ Therefore, fibrin accelerates its own destruction by (1) enhancing the catalytic efficiency of plasmin formation by t-PA, (2) protecting plasmin from its physiologic inhibitor, α₂-PI, and (3) providing new binding sites for Plg and t-PA once its degradation has begun.

UROKINASE-TYPE PLASMINOGEN ACTIVATOR–MEDIATED PLASMIN GENERATION

For the activation of Glu–Plg by u-PA in a fibrin–free system, reported Michaelis constants (K_m) vary from 1.4 to 200 μM, while catalytic rate constants (kcat) range from 0.26 to 1.48 sec^–1.¹ Interestingly, activation of Glu–Plg by two–chain u-PA is increased in the presence of fibrin by approximately 10–fold even though u-PA does not bind to fibrin.²²⁷ In contrast, single-chain u-PA has considerable fibrin–specificity. This may reflect neutralization by fibrin of components in plasma that impair Plg⁷⁴ also reflect a conformational change in Plg upon binding to fibrin.²²⁸ It is important to recognize, however, that the intrinsic Plg activating potential of single-chain u-PA is less than 1 percent of that of two-chain u-PA. Two–chain u-PA has been used effectively as a thrombolytic agent for many years.²²⁹

PLASMIN AS A TISSUE REMODELER

A large number of in vitro studies suggest a role for plasmin in tissue remodeling. Basement membrane proteins such as thrombospondin,²³⁰ laminin,²³¹ fibronectin,²³² and fibrinogen,²³³ are readily degraded by plasmin in vitro, suggesting possible roles in inflammation,²³⁴ tumor cell invasion,²³⁵ embryogenesis,²³⁶ ovulation,²³⁷ neurodevelopment,^238,239 and prohormone activation.^240,241 Plasmin also activates MMPs 3 and 13 in the mouse, thereby facilitating the degradation of matrix proteins such as the collagens, laminin, fibronectin, vitronectin, elastin, aggrecan, and tenascin C.²⁴² On the other hand, activation of other MMPs apparently proceeds in the absence of Plg, possibly providing the basis for the mild phenotype observed in Plg-null homozygote animals.⁸⁰

Roles for plasmin in tissue remodeling and host defense mechanisms are further supported by in vivo observations in Plg-deficient mice (see Table 135–2). Impaired wound healing is observed in the Plg “knockout,”²⁴³ and is reversed upon simultaneous deletion of fibrinogen.²⁴⁴ Plg-deficient mice also display diminished recruitment of monocytes in response to intraperitoneal thioglycolate,²⁴⁵ and impaired neointima formation following electrical injury to blood vessels.²⁴⁶ In studies involving Borrelia burgdorferi, the agent of Lyme disease, dissemination of the spirochete within its arthropod vector Ixodes dammini is absolutely dependent upon host Plg even though the deer tick contains no fibrin.²⁴⁷ Furthermore, kainate-induced excitotoxicity and attendant neuronal cell dropout in the hippocampus is not observed in Plg knockout mice but does occur in fibrinogen-deficient animals.²⁴⁸ The latter two studies may define new roles for plasmin, which appear to be unrelated to degradation of fibrin.

In the lung, the fibrinolytic system mediates lung matrix remodeling, through mechanisms that appear to be independent of fibrin degradation.²⁴⁹ In mice, deficiency of fibrinogen has no effect on the development of bleomycin-induced pulmonary fibrosis.²⁵⁰ Mice lacking either PAI-1 or TAFI are protected from lung fibrosis in the same model,^251,252,253 whereas inducible expression of u-PA within alveoli abrogates the fibrotic response.²⁵⁴

Plasmin may play a role in the activation of growth factors. TGF-β is a Mr 25,000 homodimeric polypeptide that regulates vascular cell responses and epithelial-mesenchymal transformation in development and in tissue fibrosis.^255,256 In culture, cell–associated plasmin appears to convert latent TGF-β to its physiologically relevant active state. Inhibition of wound healing in this system was dependent upon active TGF-β, and activation of this agent could be blocked in the presence of plasmin inhibitors such as aprotinin or α₂-PI. Activation of TGF-β by plasmin may reflect alteration of its tertiary structure upon cleavage of an aminoterminal glycopeptide.²⁵⁷ Once activated by plasmin, TGF-β can stimulate production of PAI–1, thus impairing further activation of Plg.

The role of the fibrinolytic system in vascular remodeling during atherosclerosis appears to be complex.²⁵⁸ In the evolution of an injury to the endothelial cell lining of blood vessels, deposition of intravascular fibrin and organization of a thrombus occurs.²⁵⁹ As the injury resolves, fibrin participates in plaque growth and luminal narrowing. Evidence of the importance of fibrinolytic balance in this process is that, in the absence of PAI-1, there is less neointima formation and reduced luminal stenosis, possibly because of more rapid resolution of fibrin.²⁶⁰ In areas of the vasculature where injury is not associated with fibrin deposition, however, absence of PAI-1 may lead to enhanced lesion formation, as cells that invade the developing plaque may require plasmin activity for their directed migration.²⁶¹

FIBRINOLYSIS AND ANGIOGENESIS

Although the fibrinolytic system has generally been assumed to be proangiogenic by virtue of its ability to promote “tunneling” of endothelial cells through fibrin-containing matrices, its effect, in actuality, appears to be context specific.^262,263 PAI-1 deficiency in mice, for example, seems to prevent tumor vascularization in a malignant keratinocyte model.²⁶⁴ The same mice are also resistant to laser-induced neovascularization of the choroid.^265,266 The paradoxical proangiogenic effect of PAI-1 in some settings may relate to its ability to protect endothelial cells from apoptosis mediated by FasL, which is activated by plasmin.²⁶⁷

In the mouse cornea, absence of t-PA, u-PA, or TAFI, had no effect on neovascularization, whereas loss of Plg, PAI-1, or annexin A2 significantly diminished this response.^205,268 Within the atherosclerotic plaque, moreover, expression of a truncated form of PAI-1 (rPAI-1₂₃) was antiangiogenic, inhibiting the proliferation of vasa vasorum, and reducing overall plaque area and plaque cholesterol in the descending aorta.²⁶⁹ As a gene product that is transcriptionally upregulated by hypoxia, annexin A2 is required for the normal corneal angiogenic response to growth factor stimulation, and also for hypoxia-induced retinal angiogenesis.^198,205,270

FIBRINOLYTIC DEFICIENCY AND THROMBOSIS

Although partial human Plg deficiency was first described in a young man with a history of venous thrombosis and pulmonary embolism,²⁷¹ there is currently little evidence that hypoplasminogenemia alone is a significant cause of deep venous thrombosis.²⁷² In a study of 23 consecutive patients with thrombophilia, the prevalence of Plg deficiency was only 1.9 percent.²⁷³ Approximately half of these individuals had other risk factors such as deficiency of antithrombin, protein C, or protein S, or resistance to activated protein C. Among 93 patients with type I Plg deficiency, the prevalence of thrombosis was 24 percent, or 9 percent when the propositi were excluded.²⁷⁴ Two additional epidemiologic studies concluded, moreover, that isolated hypoplasminogenemia is not a risk factor for thrombosis.^275,276

Although there are no reported cases of complete absence of Plg in humans, a large number of Plg polymorphisms and dysplasminogenemias have been reported.²⁷² Congenital Plg deficiency has been classified into two types: in type I the concentration of immunoreactive Plg is reduced in parallel with functional activity,²⁷⁷ whereas in type II (dysplasminogenemia), immunoreactive protein is normal while functional activity is reduced.²⁷⁸ Patients with type I Plg deficiency are most likely to present with ligneous conjunctivitis, which resolves completely upon infusion of Lys-Plg.^279,280 In a study of a Japanese cohort, approximately 27 percent of individuals with type II deficiency had a clinical history of thrombosis, but it is not clear whether there were other explanations for thrombophilia in these individuals.²⁸¹ Acquired Plg deficiency may occur in liver disease, sepsis, and Argentine hemorrhagic fever due to decreased synthesis and/or increased catabolism,²⁸² but associated thrombosis may be due to abnormalities in other hemostatic factors in these very ill patients. Similarly, there are no reported cases of complete t-PA or u-PA deficiency in humans, and no mutations or polymorphisms in these genes have so far been clinically linked to thrombophilia. Defects in Plg activator release, as well as increased inhibition of t-PA by PAI-1, have been reported in associated with thrombosis,^283,284 and with chronic renal disease and hypertension.^81,83

Global deficiency in fibrinolytic function, moreover, is associated with increased risk for venous thrombosis, as well as first myocardial infarction in young men.^285,286 Increased circulating PAI-1 appears to represent an independent risk factor for vascular reocclusion in young survivors of myocardial infarction.²⁸⁷ In addition, increased levels of PAI-1 have been associated with deep vein thrombosis in patients undergoing hip replacement surgery²⁸⁸ and in individuals with insulin resistance.²⁸⁹ Although a 4G versus 5G polymorphism in the PAI-1 promoter has been reported, the 4G form being associated with higher PAI-1 plasma levels, it is not yet established as to whether this allele correlates with elevated thrombotic risk.^290,291 With regard to such studies, one should bear in mind that PAI-1 is itself an acute phase reactant, and thus may not be directly responsible for the observed prothrombotic tendency.²⁹²

ENHANCED FIBRINOLYSIS AND BLEEDING

Enhanced fibrinolysis resulting from congenital or acquired loss of fibrinolytic inhibitor activity may be associated with a bleeding diathesis.²⁹³ Patients with congenital deficiency of α₂-PI may present with a severe hemorrhagic disorder as a result of impaired inactivation of plasmin and premature lysis of the hemostatic plug.²⁹⁴ Acquired α₂-PI deficiency may be seen in patients with severe liver disease from decreased synthesis, disseminated intravascular coagulation from consumption, nephrotic syndrome from urinary losses, or during thrombolytic therapy, which induces excessive utilization of the inhibitor.²⁹⁴ TAFI levels are markedly reduced in liver cirrhosis, correlating with enhanced plasma fibrinolysis, and serving as an independent predictor of mortality.²⁹⁵

Patients with acute promyelocytic leukemia demonstrate excessive expression of annexin A2 on their developmentally arrested promyelocytes. Bleeding in this disorder is accompanied by evidence of high levels of plasmin generation and depletion of α₂-PI. Bleeding resolves upon initiation of all-trans-retinoic acid therapy, which eliminates expression of promyelocyte annexin A2, probably through a transcriptional mechanism.¹⁸¹ In this setting, A2 acts most likely in concert with S100A10, which is also upregulated in an acute promyelocytic leukemia (APL) cell line.^201,296

Complete loss of PAI-1 expression resulting in hemorrhage in a 9-year-old child was associated with severe hemorrhage in the setting of trauma or surgery.²⁹⁷ This autosomal recessive trait reflected a frameshift mutation within exon 4 that induced a premature stop codon. This case demonstrates that PAI-1 is a central regulator of fibrinolysis in humans.

DEVELOPMENTAL REGULATION OF THE FIBRINOLYTIC SYSTEM

In the resting, nonstressed state, the plasmin–generating potential in the newborn is significantly less than that of the adult.^298,299 Although the amino acid composition and apparent molecular mass of neonatal Plg are indistinguishable from those of the adult protein,^300,301 plasma concentrations of Plg in the neonate are approximately 50 to 75 percent of those observed in adults.^300,302,303 On the other hand, levels of histidine-rich glycoprotein, a carrier protein that may limit Plg’s interaction with fibrin, are also reduced by 50 to 80 percent in healthy, term newborns.³⁰⁴ Neonatal Plg is heavily glycosylated, less readily activated by t-PA, and only weakly bound to the endothelial cell surface.³⁰¹ Throughout childhood, global plasma fibrinolytic activity and plasmin generation are decreased in comparison to adults, and this relative deficiency may contribute to the high frequency of thrombosis associated with central venous line placement, Kawasaki disease, and Henoch-Schönlein purpura in this age group.³⁰⁵

Although t-PA antigen and activity levels are reduced by 50 to 75 percent compared with adult values throughout childhood,³⁰³ stressed infants, such as those with severe congenital heart disease or respiratory distress syndrome, may have t-PA antigen levels that are increased by up to eightfold because of the t-PA release response.^306,307 In contrast, the principal plasmin inhibitors undergo only minimal change from birth to adulthood.^{302,308,309,310} Thus, reduced fibrinolytic activity may contribute to the thrombogenic state commonly observed in the newborn,³¹¹ but this predilection may be reversed under conditions of physiologic stress.

FIBRINOLYTIC ACTIVITY DURING PREGNANCY AND PUERPERIUM

Pregnancy is a hypofibrinolytic state.^312,313,314 Both Plg and fibrinogen levels in plasma increase by 50 to 60 percent in the third trimester. However, overall fibrinolytic activity, as reflected in euglobulin lysis activity, is reduced, and increased fibrin deposition is suggested by increasing D-dimer levels throughout pregnancy.³¹⁵ Between the 20th week of pregnancy and term, PAI-1 levels increase to three times their normal level, while PAI-2 levels rise to 25 times their level in early pregnancy.³¹² Less dramatic increases in both u-PA and t-PA levels are also observed. Within 1 hour of delivery, however, concentrations of both PAI-1 and PAI-2 begin to decrease, and return to normal within 3 to 5 days.³¹²

In preeclampsia, the hemostatic and fibrinolytic imbalances seen in pregnancy are further exaggerated.³¹⁶ Circulating PAI-1 levels exceed those in normal pregnancy, and fibrin deposition is seen in the glomerular capillaries and spiral arteries of the placenta. Interestingly, levels of PAI-2, a marker of placental function, are reduced during preeclampsia compared with normal pregnancy, and this decrease correlates with intrauterine growth retardation of the fetus. Elevated TAFI levels may be a cause of fibrin deposition and occlusion of placental vessels in preeclampsia.³¹⁷

The goal of thrombolytic therapy is rapid restoration of flow to an occluded vessel achieved by accelerating fibrinolytic proteolysis of the thrombus.³¹⁸ The fibrinolytic system functions physiologically to remove fibrin deposits through the action of plasmin, but this is often too slow to prevent tissue injury following acute vascular occlusion. Because arterial thrombosis immediately renders distal tissue ischemic with rapid onset of dysfunction and necrosis, a critical problem is minimizing time to restoration of flow. Thrombolytic therapy should be viewed as one part of an overall antithrombotic plan that frequently includes anticoagulants, antiplatelet agents, and mechanical approaches, all designed to rapidly restore flow, prevent reocclusion, and promote healing. Here we review thrombolytic approaches to stroke and peripheral vascular disease. Thrombolytic therapy for deep vein thrombosis, pulmonary embolism, and myocardial infarction are discussed elsewhere.

PRINCIPLES OF THERAPY

The basic principle of all fibrinolytic therapy is administration of sufficient Plg activator to achieve a high local concentration at the site of the thrombus, thereby accelerating conversion of Plg to plasmin, and increasing the rate of fibrin dissolution. However, if large amounts of Plg activator overwhelm the natural regulatory systems, plasmin may be formed in the blood, resulting in degradation of susceptible proteins, the “lytic state.”³¹⁹ In addition, if high concentrations of activator reach fibrin deposits at sites of injury, bleeding, often exacerbated by plasmic proteolysis of other proteins in the blood may ensue.

Several therapeutic agents, from both recombinant and natural sources, are available and approved for thrombolytic use (Table 135–3). The degree of “fibrin specificity,” is critical in determining the intensity of action at the site of a thrombus. The plasma half-life of most agents is short, ranging, for example, from 5 to 70 minutes for t-PA and anistreplase, respectively. Decisions to administer by bolus versus continuous infusion, as well as the duration of therapy, are determined by the agent’s half-life and the condition being treated. Regarding site of delivery, systemic therapy via peripheral vein is simpler and does not require specialized facilities, but results in greater systemic complications. Regional delivery with a catheter placed close to the proximal end of the thrombus can provide a high local concentration with a smaller total dose, thereby increasing the local effect and limiting systemic exposure. Fibrinolytic therapy is often administered in combination with an anticoagulant to block fibrin formation and with an antiplatelet agent to limit continued platelet deposition. Anticoagulant therapy is routinely continued after completion of fibrinolytic therapy to prevent reocclusion. In addition, mechanical approaches such as percutaneous coronary intervention often play a vital role in removing the underlying cause of thrombosis.

The activation of plasmin has effects beyond the thrombus, including a reduction in fibrinogen, increase in fibrinogen degradation products, and depletion of Plg and α₂-plasmin inhibitor. Screening coagulation tests, including the activated partial thromboplastin time (aPTT), prothrombin time (PT), and thrombin clotting time, will be prolonged depending on the intensity of the lytic state. Tests reflecting Plg activation, such as the euglobulin clot lysis time, will be abnormal. Platelet membrane proteins may also be degraded, resulting in abnormal platelet function.^320,321,322 Overall, these effects contribute to a hypocoagulable lytic state that may be beneficial for vessel patency, but may also exacerbate a bleeding complication. High doses of a nonspecific activator, such as streptokinase, will cause a more marked lytic state, compared to that seen with a fibrin-specific agent such as reteplase.

Patient selection for fibrinolytic therapy depends on careful consideration of risks and benefits (Table 135–4). For patients with acute myocardial infarction or stroke there is a higher tolerance of bleeding complications, because lytic therapy can be lifesaving and limit disability. Timing of treatment is also critical, with greater benefit achieved with earlier administration. Whereas fibrinolytic therapy for acute pulmonary embolism may be lifesaving, the potential benefits for venous disease are less clear and more likely to be associated with bleeding problems.

THROMBOLYTIC THERAPY FOR STROKE

Stroke is the third leading cause of death, and the leading cause of serious disability in the United States.³²³ Its incidence has been declining in recent years due to control of risk factors, but total numbers are increasing as a consequence of aging of the population. Although aspirin and anticoagulants may be useful in prevention, thrombolytic therapy is the only available intervention during the acute stage.

The appropriate use of thrombolytic therapy for stroke is based on an understanding of its pathogenesis. Ischemic stroke is most commonly caused by rupture of an atherosclerotic plaque within a large or medium-sized artery in the neck or cranium. In addition, transient ischemic attacks (TIAs) and strokes involving small arteries can result from embolization of platelet-fibrin thrombi that form on atherosclerotic vessels in the neck and ascending aorta, or from embolization of thrombi that form in the heart in association with atrial fibrillation, valve dysfunction, artificial valves, or endocardial thrombi. Up to 30 percent of strokes have no defined etiology.

Current approaches to thrombolytic therapy for stroke are based on imaging to define the etiology, results of clinical trials, and the experience with thrombolysis for acute myocardial infarction. Modern computed tomography (CT) imaging and magnetic resonance imaging (MRI) can identify ischemic areas and localize areas of hemorrhage quite early. Additionally, arteriography can identify obstructed vessels and follow the course of recanalization during thrombolytic therapy. Clinical studies have generally followed the successful designs used for myocardial infarction that demonstrated the critical pathologic role of the occluded vessel, the importance of early recanalization in preserving myocardium, and the impressive decrease in morbidity and mortality resulting from early reperfusion. They have also characterized the bleeding risk.

The experience with thrombolytic treatment for stroke also highlights important differences from myocardial infarction. The arterial anatomy of the brain is more complex, the time from onset of ischemia to irreversible necrosis is shorter, the risk and consequences of bleeding are greater, and there is more variability in the thrombo(embolic) occluding lesion. Further, the occlusive platelet-fibrin thrombus that precipitates a myocardial infarction is quite small, whereas the occlusive lesion causing ischemic stroke may be a large in situ thrombus, small platelet-fibrin embolus, or large embolus of varying age and composition originating from the left atrium. Thrombolysis has had a smaller impact for stroke than it has for myocardial infarction, based largely on these differences.

Early Thrombolytic Studies

The current therapeutic approach began with small, open-label studies that used intravenous or intraarterial streptokinase, urokinase, and t-PA to determine dose, recanalization rate, hemorrhagic potential, and clinical predictors of response.^324–339 These studies demonstrated that recanalization could be achieved, that early treatment was essential, and that the rate of intracranial hemorrhage and hemorrhagic transformation within the ischemic area was high. Phase II studies defined the optimum dosage and time window for intravenous t-PA and served as the basis for larger phase III trials that led to the current t-PA-based approach to thrombolytic therapy for stroke (Table 135–5). At present, the only FDA-approved therapy for acute stroke is intravenous alteplase (recombinant t-PA) given within 3 hours of symptom onset.

Tissue Plasminogen Activator Therapy

The National Institute of Neurological Disorders and Stroke (NINDS) Study was a two-part randomized, double-blind, placebo-controlled study³⁴⁰ to test whether t-PA improved clinical outcome at 24 hours and 3 months. All patients were treated within 3 hours of symptom onset with a total dose of 0.9 mg/kg of t-PA. The combined results showed a 30 percent improvement in clinical outcomes at 3 months and the benefit persisted at 12 months, despite a 10-fold increase in early symptomatic intracranial hemorrhage. At 3 months, there was no difference in mortality between the groups. This study formed the basis of the approval by the FDA of intravenous t-PA for stroke in 1996.

Early randomized trials of IV t-PA did not show clear benefit for patients treated beyond 3 hours after stroke onset. In the European Cooperative Acute Stroke Study (ECASS) Study, subjects with moderate to severe symptoms were randomized to placebo or t-PA within 6 hours of symptom onset³⁴¹; results showed no significant difference in either the primary end point of functional status at 90 days or in 30-day mortality. In the ECASS II, in which patients were randomized and stratified for presentation up to 3 hours after symptom onset or between 3 and 6 hours, there was no significant benefit of thrombolytic therapy using the primary end point of functional capacity of 90 days.³⁴² The Alteplase Thrombolysis for Acute Noninterventional Therapy in Ischemic Stroke (ATLANTIS) Study evaluated the safety of recombinant t-PA (rt-PA) in a double-blind, placebo controlled study with administration of drug between 3 and 5 hours after symptom onset,³⁴³ and the primary end point of excellent neurological recovery was observed in 32 percent of placebo and 34 percent of rt-PA–treated patients. Early symptomatic intracranial hemorrhage occurred in 1.1 percent of control and 7.0 percent of rt-PA–treated patients. There was a nonsignificant trend toward increased mortality with rt-PA treatment at 90 days (6.9 percent vs. 11.0 percent, p = 0.09). A meta-analysis pooling data from NINDS, ATLANTIS, and ECASS II patients who received either alteplase or placebo within 6 hours showed that the odds of a favorable 3-month outcome decreased as the interval from stroke onset to the start of alteplase treatment increased. Furthermore, the study alluded to the potential benefit of extending the treatment window to 4.5 hours with favorable but decreasing odds ratio for alteplase treatment beyond 3 hours.³⁴⁴

The benefit of IV t-PA for treatment beyond the 3-hour window was established by the ECASS III trial.³⁴⁵ This study showed that IV t-PA treatment initiated at 3 to 4.5 hours after ischemic stroke onset led to a modest improvement in the 3-month outcome. More patients had a favorable outcome with t-PA than with placebo (52.4 percent vs. 45.2 percent; odds ratio: 1.34; 95 percent CI 1.02 to 1.76). While the incidence of intracranial hemorrhage was higher with t-PA treatment (2.4 percent vs. 0.2 percent; p = 0.008), there was no difference in mortality between the two groups.

The effect of alteplase given beyond 3 hours after stroke on infarct growth and reperfusion was studied in the Echoplanar Imaging Thrombolytic Evaluation Trial (EPITHET) trial.³⁴⁶ Alteplase was shown to be significantly associated with increased reperfusion in patients who had mismatch at baseline (p = 0.001), better neurologic outcome (p <0.0001), and better functional outcome (p = 0.010). The observational Safe Implementation of Treatment in Stroke International Stroke Thrombolysis Register (SITS-ISTR) study further supported the safety of administering IV t-PA between 3 and 4.5 hours after acute ischemic stroke.^347,348 Compared to patients treated within less than 3 hours (n = 11,865), those treated at 3 to 4.5 hours (n = 664) had similar rates of independence, symptomatic intracranial hemorrhage, and mortality. An updated pooled analysis of ECASS, ATLANTIS, NINDS, and EPITHET trials continues to demonstrate that patients with ischemic stroke, selected by clinical symptoms and CT, benefit from intravenous alteplase when treated no later than 4.5 hours.³⁴⁹

Streptokinase Therapy

Streptokinase has been evaluated in three large stroke trials. The Australian Streptokinase (ASK) study showed an increase in death rate at 90 days in streptokinase-treated patients, and the study was prematurely terminated.³⁵⁰ The Multicentre Acute Stroke Trial–Italy (MAST-I) study examined benefits and risks of streptokinase treatment with or without aspirin in patients with acute ischemic stroke who presented within 6 hours of symptom onset.³⁵¹ An interim analysis resulted in early termination because streptokinase treatment was associated with a 2.7-fold increase in fatality at 10 days among patients receiving both streptokinase and aspirin. In the Multicenter Acute Stroke Trial-Europe (MAST-E) trial, the mortality rate at 10 days was higher in patients who received streptokinase (34.0 percent) compared with placebo (18.2 percent, p <0.02) primarily because of hemorrhagic transformation of infarcts.³⁵²

Tenecteplase Therapy

Tenecteplase is a genetically modified and genetically engineered recombinant t-PA; it has a higher fibrin specificity and greater resistance to inactivation by its endogenous inhibitor (PAI-1) compared to native t-PA. In a phase IIB study, there were no significant differences in intracranial bleeding or other serious adverse events in patients receiving alteplase versus tenecteplase.³⁵³ However, the two tenecteplase groups had greater reperfusion (p = 0.004) and clinical improvement (p <0.001) at 24 hours compared with the alteplase group. The study outcome supports ongoing phase II–III trials of tenecteplase versus alteplase in the time window that is currently approved for stroke thrombolysis (NCT01472926 and NCT01949948).

Intraarterial Thrombolysis

Intraarterial administration allows delivery of a high concentration of a Plg activator in proximity to the thrombus, more accurate anatomic diagnosis, the ability to observe the course of recanalization, and lower total doses of drug that might reduce intracranial hemorrhage. On the other hand, this approach requires specialized facilities and experienced personnel to perform arteriography and selective catheterization, which may delay treatment. Several small open-label trials observed a high rate of recanalization and apparent clinical benefit with intraarterial therapy using urokinase, streptokinase or t-PA, but hemorrhagic transformation was a frequent problem.^{327,331,334,337,354,355,356,357,358,359,360}

The Prolyse in Acute Cerebral Thromboembolism (PROACT) and PROACT II trials evaluated recombinant human prourokinase by catheter-directed intraarterial administration.^361,362 In the PROACT trial, a significantly higher recanalization rate was observed with prourokinase treatment with no increase in intracranial hemorrhage. This led to the larger PROACT II trial, which revealed a significantly higher recanalization rate with prourokinase (66 percent vs. 18 percent, p <0.001), and superior functional improvement at 90 days.^363,364 Symptomatic intracranial hemorrhage occurred in 10 percent of patients treated with prourokinase and 2 percent of controls. Although promising, these results did not lead to FDA approval of intraarterial prourokinase for treatment of stroke.

A third study, the Middle Cerebral Artery Embolism Local Fibrinolytic Intervention Trial (MELT), was underpowered because of premature study closure.³⁶⁵ A favorable, but not statistically significant, outcome at 90 days was more likely with intraarterial urokinase compared with placebo. The proportion of patients with an excellent functional outcome was significantly better in the intraarterial urokinase group (42 percent vs. 23 percent, p = 0.045). Intracerebral hemorrhage within 24 hours of treatment occurred in 9 percent and 2 percent, respectively (p = 0.206). This study suggested that intraarterial fibrinolysis has the potential to increase the likelihood of excellent functional outcome in appropriate clinical settings.

Overall, these studies show that treatment of acute stroke with thrombolytic therapy can lead to recanalization of the occluded artery and improvement in clinical outcomes. The need for early treatment, which improves outcome, is currently the single largest limitation to greater application of thrombolytic therapy for stroke, and less than 5 percent of stroke patients currently receive t-PA treatment, indicating the need for focused community educational efforts.^366,367,368 Randomized studies with rt-PA have shown that intravenous thrombolytic therapy can be safely extended to 4.5 hours after symptom onset in selected patients, whereas streptokinase was associated with an unacceptably high rate of intracranial hemorrhage.^344,345,347 In addition, intracranial hemorrhage can be reduced by identifying patients at greatest risk using MRI diffusion–perfusion mismatch to identify reversible ischemia.^{346,369,370,371} The combination of potent antiplatelet therapy using a glycoprotein IIb/IIIa antagonist with a lower dose of a thrombolytic agent may improve results.^{372,373,374,375,376,377}

In summary, current recommendations limit thrombolytic therapy for stroke to patients presenting within 3 hours of symptom onset.^378,379,380 The approved therapy is with 0.9 mg/kg (maximum: 90 mg) of t-PA administered intravenously with 10 percent as an initial bolus and the remainder infused over 60 minutes. The best results are obtained in patients who meet strict eligibility requirements (Table 135–6). Patients should be closely monitored for bleeding complications, especially intracranial hemorrhage, and careful attention should be paid to blood pressure and other comorbidities.

PERIPHERAL VASCULAR DISEASE

Acute peripheral arterial occlusion presents with the sudden onset of new, severe leg symptoms or acute worsening of chronic ischemia, and often involves embolic or thrombotic occlusion of leg arteries. The goals of treatment are to preserve limb function through restoration of flow. Anticoagulation is useful to prevent thrombus extension, while thrombolytic therapy or surgery can restore perfusion.

Early approaches to acute peripheral arterial occlusion involved streptokinase. Several small studies demonstrated reperfusion in approximately 40 percent of patients, with greatest success when occlusions were recent; bleeding complications occurred in up to one-third of subjects.³⁸¹ Following the report in 1974 by Dotter³⁸² of successful thrombolysis in peripheral arterial occlusion using locally administered thrombolysis, practice moved progressively to the nearly exclusive use of local intraarterially administered treatment. Advantages include delivery of a high concentration of drug directly to the site of thrombosis, the ability to follow the course of treatment using the treatment catheter, and identification of local vascular lesions requiring endovascular or surgical treatment after recanalization.

Treatment involves arterial access from a remote site followed by fluoroscopic guidance of the catheter to administer drug directly into the thrombus. Therapy is delivered by continuous infusion over hours to days and requires close monitoring and a large dose of thrombolytic agent. Successful reperfusion occurs in approximately three-quarters of cases.³⁸³ Ouriel and colleagues³⁸⁴ reported that thrombolytic therapy resulted in a 70 percent recanalization rate, and a frequency of limb salvage that mirrored that of operative intervention. There was, however, a survival advantage in patients receiving primary thrombolytic therapy resulting primarily from a decrease in the occurrence of in-hospital complications. The Surgery versus Thrombolysis for Ischemia of the Lower Extremity (STILE) trial, which compared the optimal surgical procedure to catheter directed thrombolysis with either t-PA or urokinase, was terminated prematurely because of ongoing or recurrent ischemia at 30 days in surgically treated patients.³⁸⁵ More than half of patients receiving thrombolysis had a decrease in the magnitude of the surgical procedure eventually required, with significant reductions in the 1-year rate of major amputation. In addition, there was no difference in outcome with t-PA versus urokinase.

The Thrombolysis or Peripheral Arterial Surgery (TOPAS) I study compared recombinant urokinase or surgery for initial therapy of acute lower-extremity ischemia of less than 14 days duration.³⁸⁶ The 1-year mortality and amputation-free survival were similar in the urokinase and surgery groups. There was a significant reduction in the frequency and magnitude of surgical interventions eventually required in patients randomized to initial thrombolysis. The larger TOPAS II study showed recanalization in 80 percent of patients who received urokinase.³⁸⁷ Amputation-free survival at 1 year was not significantly different between the surgical and thrombolysis groups, 70 percent and 65 percent, respectively. Major hemorrhagic complications were significantly more frequent with urokinase (13 percent) compared to 6 percent with surgery (p = 0.005).

In other studies, reteplase appears to be equally effective as t-PA or urokinase with comparable recanalization rates, and clinical outcomes and bleeding complications.^388,389 Prourokinase also gave similar overall results to urokinase in a phase II study.³⁹⁰ In an open-label trial, staphylokinase, a highly fibrin-specific Plg activator, resulted in revascularization in 83 percent of subjects with occluded arteries.³⁹¹ Occasional allergic reactions occurred, and severe bleeding complications were comparable to those with other agents. The addition of abciximab, a glycoprotein IIb/IIIa, inhibitor, to urokinase resulted in more rapid clot like lysis in a randomized study,³⁹² and good results were also reported with reteplase and abciximab.³⁹³ Intraoperative thrombolysis during thromboembolectomy has been used successfully to improve clearance of distal thromboemboli with or without adjunctive mechanical thrombectomy.^{394,395,396,397,398}

Thrombolysis should be viewed as one part of a combined, comprehensive management approach to peripheral arterial occlusion. Key points include early, accurate angiographic diagnosis, appropriate intrathrombic catheter positioning, and, in some cases, definitive endovascular or surgical procedures.^{399,400,401,402} Evidence favors mechanical thromboembolectomy as adjunctive therapy for acute limb ischemia resulting from peripheral arterial occlusion.⁴⁰³

OTHER INDICATIONS

Thrombolytic therapy has been useful in treating acute venous and arterial occlusions in a wide variety of sites. Reports document successful treatment of intraabdominal thrombosis including Budd-Chiari Syndrome,⁴⁰⁴ portal vein thrombosis,^405,406,407 and mesenteric vein thrombosis.^407,408,409 Thrombolytic agents are frequently used to open thrombosed central venous catheters,^{410,411,412,413} as well as access devices for hemodialysis.^{414,415,416,417,418}

MANAGEMENT OF BLEEDING COMPLICATIONS

Bleeding complications are more frequent with fibrinolytic than with anticoagulant therapy and require rapid diagnosis and management. The most serious complication, intracranial hemorrhage, occurs in approximately 1 percent of patients and is associated with a high mortality and serious disability in survivors. Risk factors for intracranial hemorrhage, including prior stroke, serious head trauma, intracranial surgery, tumor or vascular disease such as aneurysms or arteriovenous malformation and uncontrolled hypertension, are strong contraindications to fibrinolytic therapy.⁴¹⁹ Bleeding is most common at sites of invasive vascular procedures or preexisting gastrointestinal or genitourinary lesions, and should not interrupt therapy if it can be managed with local pressure or other simple measures.

Treatment of bleeding involves local measures as well as correction of the systemic hypocoagulable state resulting from proteolysis of plasma proteins and platelets (Table 135–7).⁴²⁰ The fibrinolytic agent should be discontinued, and most will be cleared rapidly, because of the short half-life. For serious bleeding, an antifibrinolytic agent such as epsilon aminocaproic acid can be administered, but will be effective only if the fibrinolytic agent remains in the blood. Replacement of fibrinogen and other hemostatic proteins can be accomplished with cryoprecipitate and fresh frozen plasma, respectively; treatment should be monitored with repeated coagulation tests. Administration of platelet concentrates may also be useful because fibrinolytic therapy results in platelet dysfunction from proteolysis of surface proteins. Heparin can be reversed by administration of protamine sulfate, and 1-deamino-8-D-arginine vasopressin (DDAVP) may have some value in reversing platelet dysfunction.

Pharmacologic agents can be used to inhibit fibrinolytic bleeding, but care must be exercised given the risk of thrombosis (Table 135–8). For example, in patients with consumption coagulopathies there may be excessive activation of both the coagulation and fibrinolytic systems, resulting in clinical manifestations of both bleeding and thrombosis. In this situation, inhibiting fibrinolysis to treat bleeding can precipitate or worsen thrombosis.

Both ε-aminocaproic acid and tranexamic acid are synthetic lysine analogues. These agents inhibit fibrinolysis by competitively blocking binding of Plg to lysine residues on fibrin.^{421,422,423,424} Both can be administered orally or intravenously, have rapid absorption after oral administration and are excreted primarily through the kidneys. Only ε-aminocaproic acid is approved for use in the United States, with the exception that tranexamic acid can be used for treatment of menorrhagia. Pharmacologically, tranexamic acid is approximately 10-fold more potent than ε-aminocaproic acid because of its higher binding affinity. Both drugs have a short half-life of 2 to 4 hours and must, therefore, be administered frequently. ε-Aminocaproic acid can be administered intravenously with a loading dose of approximately 100 mg/kg over 30 to 60 minutes followed by a continuous infusion of up to 1 g/h, or the dose can be divided for intermittent administration. For oral treatment, the same loading dose can be administered followed by a maximum dose of 24 g/day in divided doses given every 1 to 6 hours as indicated. The use of tranexamic acid follows similar principles. The intravenous dose is 10 mg/kg followed by 10 mg/kg every 2 to 6 hours as needed. It can also be administered orally in a dose of 25 mg/kg given three or four times daily. Both ε-aminocaproic acid and tranexamic acid are generally well tolerated, but patients must be observed for possible thrombotic complications. Additionally, thrombotic ureteral obstruction can occur in patients with upper urinary tract bleeding, and such patients should be treated only after careful consideration. The risks of ureteral obstruction can be decreased by insuring high urine flow. Thrombotic complications can occur in patients with hypercoagulability, and thrombotic events can be precipitated or worsened in patients with disseminated intravascular coagulation (DIC). Myonecrosis is a rare complication. Minor complications, including rash, abdominal discomfort, nausea, and vomiting, are reported.

Aprotinin is a naturally occurring, broad-spectrum, proteinase inhibitor derived from bovine lung.^425,426,427 It has both antiinflammatory and antifibrinolytic properties. Until recently, aprotinin was used in the United States for reducing perioperative blood loss and blood transfusions in patients undergoing cardiopulmonary bypass. However, its use is associated with an increased risk of postoperative renal dysfunction, cardiac and cerebral events,^428,429 and of increased short- and long-term mortality in patients who received aprotinin compared to ε-aminocaproic acid, tranexamic acid, or placebo. In a retrospective analysis of electronic records from 33,517 aprotinin recipients and 44,682 ε-aminocaproic acid recipients, the unadjusted risk of death within the first 7 days after coronary artery bypass graft was 4.5 percent for aprotinin recipients compared to 2.5 percent for ε-aminocaproic acid recipients. The relative risk of death was significantly increased in the aprotinin group (relative risk: 1.64; 95 percent CI 1.50 to 1.78).⁴³⁰ Another retrospective study found that use of aprotinin was associated with both a significantly increased mortality risk at 1 year, and a larger risk-adjusted increase in serum creatinine (p <0.001).⁴³¹ The prospective Blood Conservation Using Antifibrinolytics in a Randomized Controlled Trial (BART), which was designed to randomize a total of 3000 patients to either aprotinin, aminocaproic acid or tranexamic acid to further assess the safety of aprotinin, was terminated early because of a significantly higher death rate from any cause at 30 days in the aprotinin recipients.⁴³² Based on these studies, aprotinin was removed from U.S. market in May 2008, and its access is limited to investigational use.

Excessive systemic fibrinolytic activation can lead to bleeding and may result in a shortened euglobulin clot lysis time, decreased Plg, decreased α₂-plasmin inhibitor, increased plasmin-antiplasmin complexes, decreased fibrinogen and increased fibrinogen degradation products. Screening tests including the PT and aPTT may be prolonged. It may be difficult to distinguish between abnormal hemostasis caused by DIC versus systemic fibrinolysis. Useful features include a more prominent decrease in fibrinogen and increase in fibrinogen degradation products and relatively less thrombocytopenia, and elevation of D-dimer with primary fibrinolysis. Homozygous deficiencies of either α₂-plasmin inhibitor or of PAI-1 can cause a lifelong bleeding disorder and have been treated effectively with antifibrinolytic agents.^{297,433,434,435,436}

APL is often associated with a severe bleeding disorder that may have elements of both DIC and systemic fibrinolysis in addition to thrombocytopenia. Administration of ε-aminocaproic acid to inhibit fibrinolysis can be useful, but must be given with care to avoid thrombosis.^{437,438,439,440} In severe liver disease, fibrinolysis caused by reduced inhibitor synthesis can contribute to bleeding and may occasionally be the primary abnormality.^441,442,443 During orthotopic liver transplantation, accelerated fibrinolysis often contributes to bleeding, particularly during the anhepatic phase. Treatment with antifibrinolytic agents can improve bleeding complications and decrease blood loss.^{444,445,446,447}

Primary fibrinolysis with bleeding may rarely occur with some malignant tumors including prostatic carcinoma,^444–453 and also with heat stroke.⁴⁵⁴ Fibrinolytic activation routinely occurs as a compensatory mechanism in consumption coagulopathy. If fibrinolytic activation is prominent in DIC and other measures do not control bleeding, use of antifibrinolytic therapy can be helpful, but must be used with caution, to avoid exacerbation of underlying thrombotic events.

The contact system is activated during cardiopulmonary bypass, resulting in alterations in the coagulation, fibrinolytic and complement systems^455,456 and both postoperative bleeding and the need for large transfusion volumes can be a major problems. Several trials of antifibrinolytic therapy have established that total blood loss and transfusion requirements can be reduced, with aminocaproic acid and tranexamic acid often used for this purpose.^{450,451,457,458,459,460} Antifibrinolytic therapy can also be useful in treating bleeding associated with some snakebites and following administration of fibrinolytic therapy.

In hemophilia or von Willebrand disease, bleeding associated with a local lesion such as dental extraction may also respond to antifibrinolytic therapy. Both the oral and urinary mucosas are rich in fibrinolytic activity, and inhibition of normal fibrinolysis can prevent local bleeding, such as after prostatectomy.^461,462,463 Similarly, endometrial fibrinolysis contributes to menstrual bleeding, and antifibrinolytic therapy can be useful in treating menorrhagia.^464,465 Antifibrinolytic therapy may also be useful in rare cases of Kasabach-Merritt syndrome in which a giant hemangioma is associated with consumption coagulopathy.^466,467 Antifibrinolytic therapy has been used in treating gastrointestinal or genitourinary bleeding in patients with severe thrombocytopenia, ulcerative colitis, hereditary hemorrhagic telangiectasia, traumatic hyphema, following tonsillectomy, and with subarachnoid hemorrhage. However, caution is advised in the latter condition, as rebleeding may be decreased with antifibrinolytic therapy, but vasospasm and distal ischemia may worsen.⁴⁶⁷