A guide to murine fibrinolytic factor structure, function, assays, and genetic alterations


  • This is the third of three review articles originating in the Subcommittee on Animal, Cellular and Molecular Models, Scientific and Standardization Committee, International Society on Thrombosis and Haemostasis.

Gerhard J. Johnson, Professor of Medicine, VA Medical Center and University of Minnesota, One Veterans Drive, Minneapolis, MN 55417, USA.
Tel.: +1 612 467 4135; fax: +1 612 725 2149; e-mail: johns337@umn.edu


Summary.  The components and functions of the murine fibrinolytic system are quite similar to those of humans. Because of these similarities and the adaptability of mice to genetic manipulation, murine fibrinolysis has been studied extensively. These studies have yielded important information regarding the function of the several components of fibrinolysis. This review presents information on the structure, function and assay of mouse fibrinolytic parameters and it discusses the results of the extensive studies of genetically modified mice. It is intended to be a convenient reference resource for investigators of fibrinolysis.

Plasminogen and plasminogen activators

Plasminogen is a single-chain glycoprotein proenzyme that is converted to the active serine proteinase, plasmin, which degrades fibrin into soluble degradation products. Plasminogen activators (PAs) convert plasminogen to plasmin. PAs are classified into two types. One type includes the fibrin-specific PAs, such as tissue-type PA (t-PA) or single-chain urokinase-type PA (scu-PA), and the other is the non-fibrin-specific PA, two-chain urokinase-type PA (tcu-PA, urokinase).


Structure and function The mouse plasminogen gene (Plg) (the gene symbols used in this review are those recommended by the International Committee on Standardized Genetic Nomenclature for Mice, revised January 2006; http://www.informatics.jax.org; accessed 20 February 2007) cDNA encodes a signal peptide of 19 amino acids and a mature protein of 793 amino acids (Table 1). Mouse and human plasminogen are 79% and 76% identical at the protein and DNA level, respectively. The NH2-terminal amino acid is Asp; conversion to plasmin requires cleavage of the Arg562-Val563 peptide bond and the active site triad is composed of His605, Asp648 and Ser743. Murine plasminogen contains functional lysine-binding sites in domains similar to human plasminogen [1].

Table 1.   Summary of murine fibrinolytic components
ProteinGene symbol*Chromosome*AAMr (Da)Plasma levelsHomozygous-deficiencyin mice
Embryonic lethalSpontaneous thrombosis
  1. *Mouse Genome Informatics (http://www.informatics.jax.org); precursor polypeptide from UniProt Knowledgebase (http://www.uniprot.org); data derived from references cited in text; §NR, not reported; urine UPA ∼2 μg mL−1 [23]; **alternative splicing.

PlasminogenPlg17812∼91 00085 μg mL−1NoYes
t-PAPlat8559∼63 0002.5 ng mL−1NoNo
u-PA High MW
Low MW
Plau14433∼48 000
PAI-1Serpine15402∼45 0001–2 ng mL−1NoNo
PAI-2Serpine21397∼44 000NRNoNo
∼35 000
∼25 000**
Annexin IIAnxa29338∼38 000NRNoNo

Assays Plasminogen antigen levels in murine plasma, determined with an enzyme-linked immunosorbent assay (ELISA) based on rabbit polyclonal antibodies calibrated with murine plasminogen, are about 85 μg mL−1 compared with 130–180 μg mL−1 in humans [2]. No information is available on quantitative or qualitative variations in different mouse strains. Plasminogen activity in human plasma is determined by complex formation with excess streptokinase and measurement of the plasminogen–streptokinase complex with a chromogenic substrate for example S-2403.

Plasminogen activity in murine plasma cannot be determined in this way because of poor reactivity of plasminogen with streptokinase and/or inhibition of the plasmin(ogen)–streptokinase complex by plasma protease inhibitors [3]. Plasminogen in murine plasma is also relatively resistant to activation with urokinase [1], but this can be overcome by acidification.

Genetic alteration Ploplis et al. [2] inactivated murine Plg by replacing genomic sequences encoding the exons containing the catalytic triad amino acids His605 and Asp648 with a neomycin phosphotransferase expression cassette(Plgtm1Dco). Bugge et al. [4] incorporated a targeting vector in the murine genome by homologous recombination, resulting in a 9-kb deletion within Plg that includes proximal promoter sequences, exon 1 and exon 2 (Plgtm1Jld).

Disruption of the plasminogen gene in mice (Plg−/− mice) causes a severe thrombotic phenotype, but is compatible with development and reproduction [2,4]. Plg−/− mice display a greatly reduced spontaneous lysis of pulmonary plasma clots and young animals develop multiple spontaneous thrombotic lesions. Restoration of normal plasminogen levels in these mice results in normalization of the thrombolytic potential and in removal of endogenous fibrin deposits in the liver, thus establishing conclusively that in vivo fibrin dissolution is critically dependent on the plasminogen/plasmin system [5]. Interestingly, removal of fibrin(ogen) from the extracellular environment (mice with combined deficiency of plasminogen and fibrinogen) alleviates the diverse pathologies associated with plasminogen deficiency.

For some processes, for example, response to vascular injury, contradictory results have been reported in different models. Mechanical or electrical injury models in Plg−/− mice revealed impaired neointima formation [6,7]. In an arterial transplant model Plg−/− mice showed reduced infiltration of inflammatory cells and diminished smooth muscle intimal migration [8]. However, with a venous graft stenosis model neointima formation was similar in wild-type and Plg−/− mice [9]. Furthermore, mice with combined deficiency of plasminogen and apolipoprotein E (ApoE) had accelerated lesion formation compared with mice with single deficiency of ApoE [10].

Plg−/− mice displayed delayed skin [11] and corneal [12,13] wound healing and developed ligneous conjunctivitis, similarly to humans [14]. Plg−/− mice were more resistant to neuronal degeneration [15], but more sensitive to axonal demyelination [16]. Lewis lung carcinoma skin transplants in Plg−/− mice resulted in smaller and less hemorrhagic primary tumors than in wild-type mice, but lung metastases were comparable [17]. However, tumor growth was significantly suppressed in Plg−/− mice when transplanted in the footpad, apparently as a result of the persistence of occlusive thrombi that limit tumor blood supply [18]. In contrast, using the Polyoma middle T antigen model, lung metastases were significantly reduced in Plg−/− mice [19]. A role of plasminogen in in vivo angiogenesis is supported by the findings that in Plg−/− mice corneal angiogenesis [20] and choroidal neovascularization [21] were decreased compared with controls.

The role of plasmin(ogen) in different processes is summarized in Table 2.

Table 2.   (Patho)physiological roles of plasmin(ogen) as studied in gene-deficient mice
ProcessEffect of Plg deficiencyReferences
Physical developmentNone[2,4]
Vascular remodelingReduced[6,7]
Transplant arteriosclerosisReduced[8]
Developmental behaviorNone[87,88]
Stress-induced behavior (grooming)Enhanced[87,88]
Dopamine release (morphine-induced)Reduced[89]
Adipose tissue formationReduced[90]
Rheumatoid arthritisEnhanced[92]
Pulmonary fibrosisEnhanced[38]
Skin wound healingReduced[11]
Corneal wound healingReduced[12,13]
Ligneous conjunctivitisEnhanced[14]
Neuronal degenerationReduced[15]
Axonal demyelinationEnhanced[16]
Tumor metastasis:
 Lewis lung carcinomaNone/reduced[17,18]
 Polyoma middle T antigenReduced[19]
Corneal angiogenesisReduced[20]
Choroidal neovascularizationReduced[21]

Tissue-type plasminogen activator

Structure and function The mouse t-PA gene (Plat) (Table 1) consists of 14 exons separated by 13 introns which code for a 2800 nucleotide transcript. The amino terminal sequence of the protein was originally reported to be Gly-Ala-Arg, but it was later observed to be Ser-Tyr-Arg [22]. This sequence corresponds to deletion of the three terminal amino acids from the sequence originally reported. Apparently this discrepancy is explained by heterogeneity in the N-terminal region of single chain t-PA which results in production of two types of proteins that differ by three amino acids, as observed in human t-PA. Mouse t-PA has 81% sequence identity with its human counterpart.

The mouse t-PA can be found in a two-chain form which has the proteolytic processing site (Arg279-Ile280) conserved in human t-PA. Also, as observed with human t-PA, the amino acids that form the catalytic site (His326, Asp375 and Ser481) are conserved.

The catalytic efficiencies of autologous t-PA are comparable for mouse and human plasminogen, and in both species, t-PA has low affinity for its substrate (see Table 3 for biochemical properties of mouse and human t-PA). However, the activation of human plasminogen by mouse t-PA exhibits 2- to 3-fold higher catalytic efficiency than that by human t-PA, as a result of a more than tenfold lower Km, but also a 5- to tenfold lower k2. In contrast, the activation of mouse plasminogen by human t-PA revealed forty to sixtyfold lower catalytic efficiency than that by mouse t-PA, mainly as the result of much lower k2. Thus, mouse plasminogen appears to be resistant to activation by human t-PA [22].

Table 3.   Biochemical properties of mouse tissue-type plasminogen activator (t-PA) and human t-PA
 Mouse t-PAHuman t-PA
Plasma concentration2.5 ± 1.0 ng mL−13.4 ± 0.8 ng mL−1
Catalytic efficiency for mouse plasminogenKm = ∼ 800 μm
k2 = 0.29 s−1
k2/Km = 0.0004 μm−1s−1
Km = 92 μm
k2 = 0.0013 s−1
k2/Km = 0.00001 μm−1s−1
Catalytic efficiency for human plasminogenKm = 20 μm
k2 = 0.027 s−1
k2/Km = 0.0013 μm−1s−1
Km = ∼200 μm
k2 = 0.12 s−1
k2/Km = 0.0006 μm−1s−1
Stimulation by CNBr-digested fibrinogen in autologous systemHundredfoldOne hundred and sixtyfold
The second-order rate constant for the inhibition by mouse PAI-14.9 × 107m−1s−13.9 × 107m−1s−1
The second-order rate constant for the inhibition by human PAI-12.0 × 107m−1s−11.8 × 107m−1s−1
The inhibition in mouse plasma (Half-life at concentration of 1 μg mL−1)6.5 min (wild-type mice)
4.2 min (PAI-1-deficient mice)
80 min
The inhibition in human plasma (Half-life at concentration of 1 μg mL−1)8 min67–89 min
Clot lysis in autologous system20–30% in 2 h with 100 nm50% in 2 h with 3.5 nm

Plasminogen activation by t-PA is stimulated one hundred and one hundred and sixtyfold in autologous murine and human systems, respectively, with saturating concentrations of 0.45 and 0.32 μm, respectively, of CNBr-digested fibrinogen (Table 3). At a fibrin concentration of 1.7 mg mL−1, almost quantitative binding (85–90%) of t-PA to fibrin occurs in both autologous and heterologous systems [22].

Both mouse and human t-PA are very rapidly inhibited by autologous or heterologous PAI-1 [22].

Two-chain mouse t-PA is inhibited in normal or PAI-1-deficient murine plasma with half-lives of 6.5 and 4.2 min, respectively, suggesting that murine plasma contains proteinase inhibitors other than PAI-1 which efficiently inhibit autologous t-PA. Furthermore, two-chain mouse t-PA is inhibited in human plasma with a half-life of 8 min, as compared with 67–89 min for human t-PA [22]. Thus, it is suggested that mouse t-PA possesses a higher reactivity with proteinase inhibitors.

Clot lysis experiments in autologous plasma indicate that the murine plasma fibrinolytic system is more resistant to activation than the human system [22]. These quantitative interactions between purified components of the murine fibrinolytic system appear to be comparable to those of their human counterparts. Thus, the murine fibrinolytic system may be a suitable model to study human disease; however, it is note-worthy that murine plasma clots are greater than thirtyfold more resistant to lysis with autologous t-PA than human plasma clots. The quantitative and qualitative interspecies differences must be carefully considered when implications regarding human fibrinolysis are drawn from studies of murine systems.

Assays Mouse t-PA can be quantitated by ELISA. This assay was developed using monoclonal antibodies raised against mouse t-PA in t-PA gene-inactivated mice. The ELISA assay is linear for t-PA between 5 and 0.1 ng mL−1, with intra-assay, inter-assay and inter-dilution coefficients of variation of 6% to 14%. t-PA antigen in murine plasma is 2.5 ± 1.0 ng mL−1 [23].

The PA activities of both mouse and human t-PA can also be measured by electrophoretic fibrin-enzymography [24].

Genetic alteration Mice with a deficiency of t-PA (Plattm1Mlg) were generated by gene targeting [25]. Plat−/− mice appear normal at birth, are fertile and have a normal life span, suggesting that t-PA is not required for normal embryonic and postnatal development. Although spontaneous fibrin deposits were not observed in Plat−/− mice, these mice have a reduced thrombolytic potential and increased incidence of endotoxin-induced thrombosis [25].

The role of t-PA in thrombus formation and removal in vivo was investigated by comparing Plat−/− mice with wild-type controls. T-PA was observed to play an important role in clot lysis after arterial endothelial injury in mice [26] and other studies of Plat−/− mice indicate that the absence of t-PA significantly reduces the antithrombotic effect of GPIIb/IIIa antagonism [27]. In apoE3-Leiden mice, which have a human-like lipid profile, t-PA deficiency delayed the atherosclerotic process [28].

The role of t-PA in the nervous system has also been investigated in Plat−/− mice. Injection of excitotoxin into the hippocampus induces t-PA expression in neurons and microglia, which degrade laminin and other substrates and result in neural damage [29]. Other mouse studies demonstrate that t-PA is involved in synaptic plasticity that is important in learning and memory [30]. T-PA plays an important role in motor learning [31] and formation of neonatal white matter lesions [32].

Mechanical injury of the sciatic nerve results in induction of t-PA expression by Schwann cells. The substrate for plasmin in this case is fibrin that is detrimental to the nerve. Therefore, degradation of this protein is beneficial and helps axons survive [16].

Plat−/− mice are resistant to retinal cell damage caused by administration of the excitotoxin N-methyl-d-aspartate (NMDA); therefore, endogenous t-PA appears to act as a facilitator of NMDA-induced retinal cell damage [33].

T-PA over-expression in the hippocampus of mice results in enhanced hippocampal-dependent learning [34]. Transient augmentation of clot lysis occurs within 4 h of adenovirus-mediated Plat transfer, but elevated expression levels subside within 7 days [35].

Understanding of the in vivo function of t-PA has been significantly enhanced by the study of mouse models. However, some biochemical properties of mouse t-PA differ from those of human t-PA. Therefore, caution must be exercised in the direct transposition of data derived from murine systems to humans.

Urokinase-type plasminogen activator

Structure and function Characteristics of the urokinase-type plasminogen activator (u-PA) gene (Plau) are recorded in Table 1. U-PA is found in urine and is secreted by various cell types as a single-chain glycoprotein. Its main function is exerted via binding to a specific glycolipid-anchored cell surface receptor, uPA receptor (u-PAR), which increases the efficiency of plasminogen activation by u-PA. Plasmin generated by this system is protected from inhibition by α-2antiplasmin. The cell surface uPA/plasmin activates matrix metalloproteinases as well as growth factors and cytokines.

Assays Tissue u-PA expression has been studied by immunocytochemical techniques. Monoclonal mouse antimurine and polyclonal rabbit antimurine, goat antihuman, and rabbit antihuman u-PA antibodies have been used equally effectively to identify u-PA in murine tissues [36]. U-PA mRNA has been identified by in situ hybridization with [35S]-labeled cRNA probes [37]. The functional consequences of u-PA conversion of plasminogen to plasmin may be evaluated by measurement of matrix metalloproteinase (MMP) activation by zymography [37,38]. Normal murine urinary u-PA antigen is 1.8 ± 1.9 μg mL−1 [23].

Genetic alteration Mouse models with a targeted disruption of Plau (Plautm1Mlg) have demonstrated roles for u-PA in physiological fibrinolysis [25], the response of the arterial wall to injury [36], susceptibility to infection [39] and ovulation [40].

As Plau−/− mice have greatly impaired inflammatory cell recruitment [39], u-PA appears to be required for cell migration [41,42]. Although u-PA is not required for neutrophil recruitment to the lung in bacterial infection [43], neutrophil activation by lipopolysaccharide is greatly accentuated by u-PA [44].

A study of the resolution of experimental pulmonary emboli, performed in Plau−/− and Plat−/− mice found t-PA, but not u-PA, to be important for clot removal [25]. In contrast, venous thrombus resolution was dependent on u-PA, through recruitment of monocytes into the thrombus and their fibrinolytic activity, but was unaffected by the absence of t-PA [42].

In a model of acute myocardial infarction, the absence of u-PA reduces the migration of leukocytes, in particular monocytes, into tissue, prevents cardiac rupture but causes cardiac failure as a result of impaired angiogenesis [41]. Deficiency of t-PA had no similar effect in this model [41]. Furthermore, it is reported that u-PA is induced in skeletal muscle regeneration, and that its absence causes severe dystrophic changes with persistent fibrin deposition and decreased recruitment of monocytes in experimentally damaged skeletal muscle [45].

The roles of u-PA and u-PAR in neovascularization have been studied in mouse models, and trials to prevent or suppress the process by inhibiting u-PA/u-PAR have been conducted. T-PA is more important than u-PA in mouse models of angiogenesis and arthritis [46], and a u-PAR-independent role of u-PA during arteriogenesis has been reported [47]. In a model of luminal stenosis after arterial injury, T-PA functions primarily in intravascular clot lysis while u-PA modulates vessel wall cell migration [48]. Furthermore, u-PA is essential for angiotensin II-induced abdominal aortic aneurysm formation in ApoE-deficient mice [37].

In experimental autoimmune encephalomyelitis, a mouse model of multiple sclerosis, the induction of u-PA in the areas of the inflammatory damage suggests a role for u-PA in inflammation and demyelination. Enhanced production of u-PA is also observed in another model of brain injury, but u-PA deficiency does not affect the size of ischemic cerebral infarction [49]. However, delayed functional recovery and attenuated MMP2 and MMP9 activities are observed in Plau−/− mice after sciatic nerve crush [50].

U-PA-plasmin activates latent forms of hepatocyte growth factor (HGF) and transforming growth factor (TGF)-ß in mouse models of hepatic regeneration and fibrosis[51]. In u-PA-deficient mice, retarded regeneration occurs after hepatic injury or partial hepatectomy, probably as a result of impaired activation of HGF [51].

Retarded tumor progression and metastasis have been demonstrated in Plau−/− mice using a fibrosarcoma tumor model [52]. Soluble u-PAR inhibits cancer growth and invasion by trapping u-PA and by u-PA-independent inhibition of the signaling activity of membrane anchored u-PAR.

Albumin-u-PA/SCID transgenic mice have severe hepatic necrosis. This model has been used to establish chimeric human liver models for study of hepatitis C as well as liver formation and regeneration [53].

The functional roles of u-PA defined by studies of gene-deficient mice are summarized in Table 4.

Table 4.   (Patho)physiological roles of u-PA as studied in gene-deficient mice
ProcessEffect of u-PA deficiencyReferences
Physical developmentNone[25]
 Fibrin depositionOccasionally increased[25]
 A flow model of venous thrombosisImpaired[42]
Susceptibility to infection:Increased[39]
 Monocyte recruitmentReduced[41,42]
 Neutrophil recruitmentUnchanged[43]
 Neutrophil activationReduced[44]
Platelet survivalUnchanged[93]
Response to arterial injury:Reduced[36,94]
 Cardiac ruptureDisappeared[41]
 AAA formationReduced[37]
Skeletal muscle regenerationReduced[45,95]
Neuronal injury:
 Focal cerebral ischemic infarctionUnchanged[49]
 Functional recovery after sciatic nerve crushDelayed[50]
HGF activationReduced[51]
Tumor developmentRetarded[52]

Plasminogen activator inhibitors

Plasminogen activator inhibitor-1 (PAI-1) and plasminogen activator inhibitor 2 (PAI-2) have been identified in plasma. A third form, initially identified in urine and called PAI-3, was later shown to be identical to the activated protein C inhibitor.

Plasminogen activator inhibitor-1

Structure and function The murine PAI-1 gene (Serpine 1; Table 1) cDNA encodes a protein which is 78% identical to human PAI-1; the Arg-Met P1-P1′ peptide bond is conserved [54].

PAI-1 is the main physiological inhibitor of t-PA and u-PA. Increased levels of plasma PAI-1 have been associated with thrombosis in mice [55].

Assays PAI-1 antigen levels can be measured by ELISA in mouse plasma [23]. PAI activity in acidified plasma can be measured by addition of t-PA and measurement of residual t-PA after a short incubation time.

In mice, PAI-1 activity is ∼2 ng mL−1 [23], and about 5-fold lower than in man [56].

Genetic alteration Studies with transgenic mice (Serpine1tm1Mlg) have confirmed many of the observations in man, but also revealed functional roles for PAI-1 in several other phenomena. It is not always established if these functions depend on the antiproteolytic activity of PAI-1 or on its interference with migration or matrix binding of cells. Transgenic mice overexpressing murine PAI-1 or stable human PAI-1 do not show evidence of arterial or venous thrombosis up to 4 months of age [55], whereas 90% of transgenic mice older than 6 months develop spontaneous occlusions [55].

Serpine1−/− mice are viable, fertile and develop normally [56]. These animals exhibit no overt bleeding tendency, but show a faster lysis of pulmonary clots and develop venous thrombi less frequently than their wild-type counterparts. In contrast to delayed rebleeding seen in man, PAI-1-deficient mice do not display spontaneous bleeding or delayed rebleeding. This difference may be because of the ∼5-fold lower basal plasma levels of active PAI-1 in wild-type mice, or to species-dependent differences.

Study of PAI-1-deficient mice suggests that PAI-1 inhibits vascular wound healing and arterial neointima formation after mechanical or electrical injury, mainly by affecting cellular migration [57]. In contrast, in murine models of vascular injury induced by ferric chloride, rose bengal or copper, a positive correlation was observed between PAI-1 levels and neointima formation [58,59]. A critical feature in these studies appears to be the presence or absence of thrombus/fibrin, raising the hypothesis that PAI-1 may inhibit neointima formation in the absence of fibrin (mechanical or electrical injury with transient thrombosis), but enhance it in the presence of fibrin (chemical injury with persistent thrombus). In atherosclerosis-prone mice, PAI-1 also promotes neointima formation after oxidative vascular injury [60]. PAI-1 deficiency in mice aggravates intimal proliferation after allogeneic transplantation, suggesting that PAI-1 plays a role in limiting the early phase of allograft vascular disease [61].

PAI-1-deficient mice exhibit accelerated skin wound healing [62] and bone remodeling [63,64]. Several studies have highlighted a role of PAI-1 in tumor angiogenesis [65,66] and in atherosclerosis [67,68]. Although high levels of PAI-1 in human breast carcinomas correlate with short survival, tumor growth and metastases in Serpin1−/− transgenic breast cancer mice was not affected [69].

A potential role of PAI-1 in development of obesity is supported by studies of PAI-1-deficient mice [70]. Transgenic mice overexpressing a stable human PAI-1 variant had virtually no intraperitoneal fat, but disruption of the PAI-1 gene in genetically obese and diabetic ob/ob mice reduced adiposity and improved the metabolic profile [71]. PAI-1-deficient mice on a high fat diet developed less obesity and insulin resistance than wild-type controls, and downregulation of PAI-1 in wild-type mice ameliorated diet-induced obesity [72]. Some of these differences may be explained by the different genetic background of the mice. PAI-1-deficient mice are protected against chronic airway inflammation [73], rheumatoid arthritis [74], vascular fibrosis [75] and glomerulonephritis [76].

The functional roles of PAI-1 defined by studies of gene-deficient mice are summarized in Table 5.

Table 5.   (Patho)physiological roles of PAI-1 as studied in gene-deficient mice
ProcessEffect of PAI-1 deficiencyReferences
Physical developmentNone[56]
Vascular remodeling:
 Electric/mechanical injuryEnhanced[57]
 Ferric chloride, rose bengalReduced[58,59]
 Oxidative injuryReduced[60]
 Allograft vascular diseaseEnhanced[61]
Skin wound healingAccelerated[62]
Bone remodelingEnhanced[63,64]
Tumor angiogenesisReduced[65,66]
 81% C57Bl/6:19% 129SVEnhanced[70]
Chronic airway inflammationReduced[73]
Rheumatoid arthritisReduced[74]

Plasminogen activator inhibitor-2

Structure and function The murine PAI-2 gene (Serpine2; Table 1) has intron/exon boundaries identical to those of the human gene (GenBank accession nos. AF069683AF069695). Serpine2 expression can be induced by multiple agents [77]. In mice infection, inflammation, tumor promoting phorbol ester and brain neurotrophic factor stimulate Serpine2 expression. Murine Serpine2 is down-regulated in transgenic mouse ovaries overexpressing human lysophosphatidic acid [78].

PAI-2 has an intracellular function in preventing cell death as well as an extracellular function in inhibition of t-PA and u-PA [77].

Increased levels of PAI-2 during the third trimester of human pregnancy suggest that it may have a role in maintenance of the placenta or in embryonic development [77]. However, PAI-2-deficient mice revealed normal development, survival and fertility, and PAI-2 has been reported to be absent from the mouse placenta [77].

Assays PAI-2 levels in plasma measured by ELISA or radioimmunoassay are very low (< 1.0 μg L−1), but are drastically elevated during human pregnancy. Comparable mouse data are not available, but PAI-2 elevation in pregnancy is unlikely (see PAI-2 Structure and function).

PAI-2 expression in murine tissues can be assayed by immunohistochemistry, Western and Northern blots, in situ hybridization and polymerase chain reaction [69,79].

Genetic alterationSerpine2-deficient mice (Serpinb2tm1Dgi) develop and survive normally, are fertile and respond as wild-type controls to bacterial infection or endotoxin infusion. Monocyte recruitment and epidermal wound healing are also similar to wild-type mice. Interestingly, crossing Serpinb2tm1Dgi with Serpine1−/− mice reveals no functional overlap between these inhibitors [79].

Overexpression of murine PAI-2 in the epidermis of mice promotes the development and progression of epidermal papillomas likely as a result of inhibition of apoptosis [80].

Plasminogen activator receptors

Urokinase-type plasminogen activator receptor

Structure and function The cell surface receptor for urokinase-type plasminogen activator (u-PAR) is a high-affinity (Kd 10−9 to 10−10), glycosyl phosphatidylinositol-anchored protein. The mouse u-PAR gene (Plaur; Table 1) is organized in seven exons like human Plaur. The mouse and human proteins share 62% amino acid identity [81], glycosylation pattern, GPI-anchor and ligand binding properties – both bind active urokinase (uPA) and pro-u-PA. Despite these similarities, binding of u-PA to u-PAR is species specific [82]. A soluble form of human u-PAR, which lacks the amino-terminal domain, has been identified in cell lines, tumors and in some disease states but not reported in mice.

Assays Several assays commonly used to detect u-PAR in both mouse and human tissues and cells employ: (i) radioligand binding, (ii) immunohistochemistry, (iii) in situ hybridization and (iv) ELISA. Radioligand binding assays have been performed using mouse peritoneal macrophages coated on microtiter wells and an 125I-labeled fragment consisting of amino acids 1–48 of mouse u-PA (mu-PA 1–48) [83]. A modification of the standard binding assay has also been reported in which an 125I-labeled amino terminal fragment of u-PA is covalently cross-linked to the receptor [84].

Genetic alterationPlaur−/− mice were generated by homologous recombination by two groups (Plaurtm1Jld; Plaurtm1Mlg) [83,84]. Selected characteristics of these mice are listed in Table 6. Mice with combined deficiency of Plaur and Plat have normal growth and reproduction, but they have spontaneous tissue fibrin deposits, primarily in liver sinusoids [85].

Table 6.   (Patho)physiological roles of plasminogen activator receptors as studied in gene-deficient mice
ProcessEffect of Deficiency
u-PARReferencesAnnexin IIReferences
  1. *NR, not reported.

Physical developmentNone[83,84]10–13% weight reduction[86]
Spontaneous thrombosisNone[83,84]Increased tissue fibrin deposits[86]
ThrombolysisNR* Reduced[86]
Vascular remodelingNone[96]NR 
AtherosclerosisNR NR 
Myocarial infarctionNone[41]NR 
Hypoxia-induced pulmonary hypertensionNone[97]NR 
Platelet survivalReduced[93]NR 
Inflammatory cell responseReduced[39,43,98]NR 

Tissue-type plasminogen activator receptor

Structure and function The receptor for t-PA is annexin II, a member of the annexin family of calcium- and phospholipid-binding proteins. The annexins are highly conserved across species and widely expressed on multiple cells. The murine annexin II gene (Anxa2; Table 1) contains 12 exons. Annexin II is comprised of a Mr ∼34 000 core domain and a short Mr ∼3400 amino terminal ‘tail’ domain.

Assays The highly conserved amino acid sequence means that essentially all of the antibodies prepared against human annexin II will cross-react with the mouse protein.

Genetic alteration Annexin II-deficient mice (Anxa2tm1Kah) have impaired t-PA mediated cell surface plasmin generation and consequently reduced MMP activity [86]. Selected characteristics of these mice are presented in Table 6.


The authors thank G. J. Johnson, Minneapolis, for assistance in editing the manuscript. They acknowledge the contribution of P. Jagadeeswaran, Chairman, Subcommittee on Animal, Cellular and Molecular Models, Scientific and Standardization Committee, ISTH when this work was initiated. They also recognize the contributions of multiple investigators to the knowledge of murine hemostatic parameters. They sincerely regret that space limitations prevent citation of all relevant literature. This work was supported by ‘High-Tech Research Center’ Project for Private Universities: matching fund subsidy from MEXT (Ministry of Education, Culture, Sports, Science and Technology), 2002–2006 (O. Matsuo), and NIH HL070304, HL074219, HL078663 (S. S. Smyth).

Disclosure of Conflict of Interests

The authors state that they have no conflict of interest.