Plasminogen activator receptor assemblies in cell signaling, innate immunity, and inflammation

Abstract

Tissue-type plasminogen activator (tPA) and urokinase-type plasminogen activator (uPA) are serine proteases and major activators of fibrinolysis in mammalian systems. Because fibrinolysis is an essential component of the response to tissue injury, diverse cells, including cells that participate in the response to injury, have evolved receptor systems to detect tPA and uPA and initiate appropriate cell-signaling responses. Formation of functional receptor systems for the plasminogen activators requires assembly of diverse plasma membrane proteins, including but not limited to: the urokinase receptor (uPAR); integrins; N-formyl peptide receptor-2 (FPR2), receptor tyrosine kinases (RTKs), the N-methyl-d-aspartate receptor (NMDA-R), and low-density lipoprotein receptor-related protein-1 (LRP1). The cell-signaling responses elicited by tPA and uPA impact diverse aspects of cell physiology. This review describes rapidly evolving knowledge regarding the structure and function of plasminogen activator receptor assemblies. How these receptor assemblies regulate innate immunity and inflammation is then considered.

INTRODUCTION

The response to tissue injury is an important process in evolution, impacting the ability of organisms to survive and reproduce. To understand wound repair at the cellular and molecular levels, this process is frequently divided into four temporally overlapping phases, including: 1) hemostasis; 2) inflammation; 3) the proliferative phase; and 4) tissue remodeling (1–3). The purpose of hemostasis is to terminate blood loss by the eventual formation of a provisional fibrin-based matrix. Hemostasis begins immediately after wound formation and is followed by the inflammatory phase, during which microorganisms are cleared from the injury site together with foreign materials and damaged tissue. This process is initially accomplished by neutrophils and subsequently, by macrophages and other inflammatory cells. The proliferative phase encompasses a variety of processes that lead to tissue regeneration and/or scarring. Fibroblasts play a central role in the proliferative phase, depositing extracellular matrix (ECM) proteins that form granulation tissue and eventually, fibrillar collagen. Other cellular elements also are important, including endothelial cells, which initiate angiogenesis and revascularization, and epithelial cells, which grow into injured tissue from wound edges. Finally, during the remodeling phase, collagen matrices mature through reorganization and cross-linking, accompanied by apoptosis of various cellular elements to form mature regenerated tissue and/or scar tissue.

Like wound healing, hemostasis may be divided into three-to-four overlapping phases, which have been discussed in depth in excellent reviews (4–9). The earliest event that occurs in hemostasis, when the injury affects blood vessels with smooth muscle cells, is vasoconstriction, which results in blood vessel narrowing. Vasoconstriction limits but does not stop blood loss. Primary hemostasis is achieved by platelets, small cellular fragments that adhere to injured tissue surfaces, aggregate with one another, and then release factors that support subsequent steps in hemostasis and the response to tissue injury. The next major system activated is coagulation, in which cascades of protease activation events result in thrombin generation. Thrombin cleaves fibrinopeptides A and B from the N-termini of the fibrinogen α and β chains to drive fibrin polymerization. Although primary hemostasis terminates blood loss, at least temporarily, coagulation is necessary to stabilize the platelet plug. In the final stage of hemostasis, fibrin clots are dissolved and removed by the fibrinolysis system to allow optimal tissue regeneration during the proliferative phase.

Cell surfaces regulate various steps of hemostasis and wound repair. This is particularly relevant to endothelial cells in hemostasis (10, 11). To assure that blood clots do not propagate unnecessarily at a distance from a blood vessel injury, uninjured endothelial cells display cell-surface macromolecules with anticoagulant activity. For example, the thrombomodulin/endothelial protein C receptor system facilitates thrombin-mediated activation of protein C, which proteolytically inactivates factors Va and VIIIa (12, 13). Uninjured endothelial cells also express heparan sulfate proteoglycans, which bind and activate the anticoagulant Serpin, antithrombin III (11, 14).

Approximate relationships between the four phases of wound repair and three phases of hemostasis are shown in Fig. 1. Overlap between the inflammatory phase of wound repair and the time when fibrinolysis is active is evident. It is thus not surprising that pathways have evolved whereby factors in the fibrinolysis system regulate innate immunity and inflammation. This review begins by briefly describing the fibrinolysis system as it functions independently of cells. Cell-surface receptors that bind plasminogen and amplify fibrinolysis are then described. Finally, two multicomponent receptor systems that recognize and bind the major mammalian plasminogen activators, urokinase-type plasminogen activator (uPA) and tissue-type plasminogen activator (tPA), are discussed. Plasminogen activator receptor assemblies are robust regulators of cell signaling. Rapidly evolving knowledge regarding the activity of these receptor assemblies in innate immunity and inflammation is considered.

Figure 1.

Overlap between the four phases of the response to injury (wounding) (color coded in green) and the three major phases of hemostasis (color coded in blue).

THE FIBRINOLYSIS SYSTEM

The serine protease that is principally responsible for fibrin degradation in mammalian systems is plasmin, which circulates as a 92-kDa zymogen called plasminogen (7, 15, 16). Cleavage of a single Arg-Val peptide bond in the structure of plasminogen converts it into its active, two chain form (plasmin). In mammalian systems, plasminogen activation is catalyzed by uPA and tPA. tPA is expressed by endothelial cells and smooth muscle cells in normal blood vessels whereas uPA is expressed only by macrophages in blood vessels in which atherosclerotic lesions have formed (17). It is, thus, not surprising that fibrinolysis within the vascular system is typically attributed mainly to tPA. However, at least in mice, when the gene encoding tPA (Plat) is deleted, uPA prevents massive, multiorgan fibrin deposition (18). Thus, both plasminogen activators may contribute to intravascular fibrinolysis.

Once plasmin is generated, it degrades fibrin and fibrinogen in a predictable manner (19, 20), generating soluble products of defined size and sequence. The products of fibrinolysis and fibrinogenolysis differ mainly due to covalent cross-linking of fibrin by Factor XIIIa/transglutaminase. Fibrinolysis in vivo is detected clinically by measuring the degradation product, D-dimer, which forms only when fibrin polymerizes and is cross-linked by Factor XIIIa, before becoming a substrate for plasmin (21).

uPA and tPA are similar to plasmin(ogen) in that these proteins exist in single-chain and two-chain forms (16, 22–24). The single-chain form of uPA demonstrates only minimal activity as a serine protease until it is converted into the two-chain form (24). By contrast, both single-chain and two-chain tPA are active as serine proteases and plasminogen activators (22, 23). Although a number of proteases may convert uPA and tPA into the two-chain forms, plasmin is paramount among these enzymes (23, 24). Thus, plasminogen activation may be viewed as a bi-directional reaction in which plasmin(ogen) and its activators function as enzymes and substrates.

When proteases in the fibrinolysis system circulate freely in the plasma, they are targeted by protease inhibitors in the Serpin gene family, including α₂-antiplasmin (α₂AP), which inhibits free plasmin with an initial association rate constant >1.0 × 10⁷ M⁻¹·s⁻¹ (25), and plasminogen activator inhibitor-1 (PAI-1), which inhibits both tPA and uPA, also with initial association rate constants exceeding 1.0 × 10⁷ M⁻¹·s⁻¹ (26, 27). The activity of α₂AP as a plasmin inhibitor is supported by α₂-macroglobulin (α₂M), which reacts slower with plasmin but is highly abundant in the plasma and thus, relevant as a regulator of fibrinolysis (28–30).

A number of mechanisms exist to limit fibrinolysis to mature clots and protect circulating fibrinogen so that the ability of the organism to respond to future injuries is not compromised. First, activation of plasminogen by tPA is greatly accelerated by fibrin, compared with the solution phase reaction (31, 32). Furthermore, once plasmin is associated with fibrin, it is relatively protected from its major inhibitors, α₂AP and α₂M (33, 34). These interactions are facilitated by multiple domains within the structure of plasmin and the plasminogen activators (Fig. 2). Binding of plasminogen to fibrin requires its tandem kringle domains, which contain binding sites for lysine and lysine-like compounds, including, most importantly, C-terminal lysine residues in fibrin and other proteins (16, 35, 36). tPA also binds to fibrin via its kringle-2 and fibronectin type-1/finger domains (37, 38). Thus, fibrin serves as a template, assembling enzyme (tPA), and substrate (plasminogen) so that the catalytic efficiency of plasminogen activation is increased. Importantly, plasmin is a lysine-specific serine protease (16). Thus, as fibrin is degraded, additional C-terminal lysine residues are exposed, allowing for additional plasminogen binding and acceleration of fibrinolysis (39).

Figure 2.

The multidomain structure of the fibrinolysis proteases, plasminogen (Plg), tissue-type plasminogen activator (tPA), and urokinase-type plasminogen activator (uPA) is shown. The structure of Plg is remarkable for five tandem kringle (K) domains. Plg is activated to form plasmin (Pm) by tPA and uPA (cleavage site shown). The 77 amino acid pre-activation peptide (PAP) is removed from Plg by plasmin cleavage, converting Glu-Plg into Lys-Plg. This reaction does not activate Plg but accelerates the kinetics of Plg activation by the plasminogen activators under certain circumstances. In addition to two kringle domains, the structure of tPA includes a fibronectin type-1/finger domain (FD) and an epidermal growth factor receptor (EGF)-like/growth factor domain (GFD). In uPA, the EGF-like/growth factor domain (GFD) and kringle domain are collectively referred to as the amino-terminal fragment, which is responsible for binding to urokinase receptor (uPAR). Both uPA and tPA may be cleaved by plasmin. These reactions convert the plasminogen activators into their two-chain forms.

Fibrin is less effective in promoting the activity of uPA as a plasminogen activator, compared with tPA (40, 41). In uPA-mediated plasminogen activation, the most important effect of fibrin may be to bind the full-length form of plasminogen (referred to as Glu-plasminogen) and convert it into a more open and readily activated conformation. Glu-plasminogen is cleaved by plasmin between Lys77 and Lys78, eliminating what was originally referred to as the “pre-activation peptide” and generating a form of the zymogen (Lys₇₈-plasminogen) with a more open conformation (16, 42, 43). Lys₇₈-plasminogen is a preferred substrate for the plasminogen activators in the absence of fibrin.

An exosite interaction involving the lysine binding sites in plasmin is very important for the rapid and irreversible inhibition of free plasmin by α₂AP (25, 33). The fact that these lysine binding sites are occupied when plasmin associates with fibrin explains the substantial decrease in the rate of plasmin inhibition by α₂AP. A similar principle also applies when plasmin binds other proteins, including specific ECM proteins and cell surface receptors. For example, when plasmin associates with receptors on cell surfaces, the response is mediated by its kringle domains and the bound plasmin is protected from inhibition by α₂AP and α₂M (44, 45).

PAI-2 has been identified as a rapid inhibitor of tPA and uPA (46). Because of an inefficient signal sequence, PAI-2 is usually retained within cells, where it expresses unique biological activities (47). In specific conditions, such as pregnancy, PAI-2 becomes available in extracellular spaces and plays a role in controlling fibrinolysis (46). Other protease inhibitors that may regulate fibrinolysis under specific conditions and in specific tissue microenvironments include: PAI-3/protein C inhibitor (48, 49); protease nexin-1 (50); and neuroserpin, which functions mainly in the central nervous system (CNS) (51). Thrombin-activatable fibrinolysis inhibitor (TAFI) is the precursor for plasma carboxypeptidase B, which inhibits fibrinolysis by cleaving and removing C-terminal Lys residues from fibrin and other substrates (52).

PLASMINOGEN ACTIVATION AT THE CELL SURFACE

The macromolecules that maintain an anticoagulant state on the surfaces of uninjured cells (10–14) are supported by cell-surface proteins with plasminogen-binding activity (35, 53, 54). Plasminogen receptors that have been identified using discovery-based methodologies are listed in Table 1, together with representative references (55–64). Plasminogen receptors are structurally diverse and include both integral plasma membrane proteins and extrinsic membrane-associated proteins. In many cells, the total plasminogen binding capacity exceeds 10⁶ sites/cell. The dissociation constant (K_D) for plasminogen binding to most receptors is ∼1.0 µM (53, 54). Because the concentration of plasminogen in plasma is ∼1.7 µM, for cells such as endothelium and inflammatory cells that are in direct contact with plasma, ∼60% of the plasminogen receptors should be occupied, as determined by the formula: B/B_max = [plasminogen, Plg]/([Plg] + K_D). Some plasminogen receptors have been described as moonlighting proteins because they have, in addition to their role as plasminogen-binding proteins, alternative activities.

Table 1.

Examples of known cell-surface receptors for plasminogen

Receptor	References
α-Enolase	Miles et al. (55)
Annexin-A2/S100 complex	Hajjar et al. (56)
	Miller et al. (57)
Cytokeratin-8	Hembrough et al. (58)
Histone H2B	Das et al. (59)
Actin	Dudani et al. (60)
Plg-RKT	Andronicos et al. (61)
Integrin αMβ2/Mac-1	Pluskota et al. (62)
Amphoterin	Parkkinen et al. (63)
Megalin (LRP-2/gp330)	Kanalas et al. (64)

Plasminogen receptors share common properties. First, plasminogen receptors typically are synthesized with C-terminal lysines or posttranslationally modified so that they have C-terminal lysines; this is critical for plasmin(ogen) binding (35, 53, 54). Second, many identified plasminogen receptors increase the catalytic efficiency (k_cat/K_M) of plasminogen activation, either in isolation or in concert with other cellular proteins that bind plasminogen activators. Finally, once plasmin is generated in association with receptors, it is protected from inhibitors (44, 45). These properties allow healthy cells that are directly exposed to blood plasma to coat themselves with sufficient quantities of plasmin so that any nidus of inappropriate fibrin formation is reversed.

Plasmin that is generated on the surfaces of cells targets important substrates beyond fibrin and fibrinogen. Thus, many biological activities, which have been attributed to plasminogen activators, reflect the activity of plasmin as a protease. For example, plasmin cleaves the amino-terminal component of latent transforming growth factor-β (TGF-β), resulting in release of active TGF-β (65). Plasmin also targets the chondroitin sulfate/heparan sulfate proteoglycan and TGF-β receptor, betaglycan, allowing for redistribution of active TGF-β between its diverse receptors and between neighboring cells (66).

Plasmin activates the “pro-form” of matrix metalloprotease-2 (MMP2) by a mechanism that requires the plasma membrane-anchored protein, membrane-type 1 matrix metalloprotease/MT1-MMP (67). Plasmin also activates the pro-form of MMP-3/stromelysin-1 in a reaction that leads to activation of MMP9 (68). Multiple glycoproteins in the ECM are plasmin substrates (69–71). Cleavage of these proteins may result in remodeling of the ECM, altered interaction with cellular integrins, or release of growth factors from the ECM. Many complex interactions involving plasminogen activators, plasmin(ogen), MMPs, ECM proteins, growth factors, and cell surface protease receptors were elegantly reviewed a number of years ago by Mignatti and Rifkin (72). How these interactions may regulate innate immunity, inflammation, and the response to pathogens was more recently reviewed by Medcalf and Keragala (73).

THE uPA RECEPTOR ASSEMBLY: STRUCTURE AND CELL-SIGNALING ACTIVITY

The principal high-affinity receptor for uPA is urokinase receptor (uPAR), a three-domain glycosylphosphatidylinositol (GPI)-anchored membrane protein (74). The crystal structure of uPAR revealed a central cavity formed by its three domains (D1–D3) into which the amino terminal fragment of uPA, which includes the growth factor and kringle domains, inserts (75). When early studies were published indicating that uPA binding to uPAR activates cell signaling, these results were unanticipated because uPAR lacks transmembrane and intracellular domains (76–81). The solution to this problem emerged as uPAR coreceptors, with known cell-signaling activity, were discovered.

One way to conceptualize this system is to consider uPAR a membrane-anchored ligand as opposed to a receptor. The function of uPAR as a ligand is activated by uPA binding, which triggers cell signaling through uPAR coreceptors. In support of this hypothesis, soluble forms of uPAR activate cell signaling similarly to uPA, apparently by direct binding to uPAR coreceptors (82–86). This interaction is facilitated by cleavage of soluble uPAR between its N-terminal and central domains, which yields a D2–D3 fragment that is active in cell signaling (82, 84, 85, 87, 88). When uPAR is membrane-anchored, the principal protease responsible for uPAR cleavage between D1 and D2 is uPA (88). Full-length soluble uPAR is less readily cleaved by uPA but may be converted into D1 and the bioactive D2D3 fragment by chymotrypsin (85, 87–89). Full-length soluble uPAR, the D1 fragment, and the D2D3 fragment have been identified as useful biomarkers for neoplastic and inflammatory diseases (88, 89).

Diverse integrins have been identified as uPAR coreceptors, which directly associate with uPAR and mediate cell signaling in response to uPA-uPAR complex and cleaved soluble uPAR, including but not limited to: α₅β₁, α₄β₁, α_vβ₅, and α_Mβ₂/Mac-1 (83, 90–94). The pertussis toxin-sensitive G protein-coupled receptor, N-formyl peptide receptor-2 (FPR2; formerly FPRL1/LXA4R), also was identified as an essential uPAR coreceptor (84). FPR2 responds to membrane-anchored uPA-uPAR complex and to cleaved soluble uPAR, suggesting collaboration between FPR2 with integrins within a single cell-signaling receptor assembly. Finally, receptor tyrosine kinases (RTKs) and specifically the epidermal growth factor receptor (EGF-R) have been implicated in uPA-initiated cell signaling (95, 96). The EGF-R may be transactivated by Src family kinases (SFKs), downstream of uPA-uPAR-integrin complexes. Ligand-independent EGF-R transactivation does not require physical association of the EGF-R with the uPA cell-signaling receptor assembly; however, physical association has been demonstrated (95). The uPA cell-signaling receptor assembly is shown in Fig. 3. Activation of cell signaling by uPA has diverse effects on cell physiology. Prominent among these is promotion of cell migration and cell survival (74, 78, 79, 92, 97, 98).

Figure 3.

The urokinase-type plasminogen activator (uPA) cell-signaling receptor assembly is shown, including urokinase receptor (uPAR), N-formyl peptide receptor-2 (FPR2), integrins, and the epidermal growth factor receptor (EGF-R). Other plasma membrane proteins also may participate in this receptor assembly. The uPA cell-signaling receptor assembly is activated to trigger cell-signaling when uPA binds to glycosylphosphatidyl inositol (GPI)-anchored uPAR. Soluble uPAR also may trigger cell-signaling through the uPA cell-signaling receptor assembly, especially when soluble uPAR is cleaved between its first and second domains.

The collection of uPAR coreceptors shown in Fig. 3 may not be comprehensive. For example, at least in some cells, glycoprotein 130 (gp130) may associate with uPAR to mediate uPA-initiated cell signaling (99). gp130 is best known as a transmembrane protein that associates with cytokine receptors to mediate signal transduction in response to cytokines (100). Interestingly, gp130 also may directly bind uPA (101). These gp130 interactions appear to be essential for activation of JAK/STAT signaling by uPA (99).

Vitronectin is a second, distinct ligand for uPAR, which activates uPAR-dependent signal transduction (96, 102–104). Vitronectin-triggered uPAR signaling does not require uPA and does not require the EGF-R as a coreceptor (96, 103). A major downstream mediator of vitronectin-initiated uPAR signaling is Rac1, as opposed to ERK1/2, which plays an important role in uPA-initiated cell signaling. Rac1 and ERK1/2 are complementary promoters of cell migration by distinct mechanisms (105).

THE tPA RECEPTOR ASSEMBLY: STRUCTURE AND CELL-SIGNALING ACTIVITY

For over twenty years, it has been apparent that tPA expresses biological activities unrelated to fibrinolysis, including, for example, regulation of endothelial cell proliferation (106), smooth muscle cell proliferation (107), and activation of microglia (108). These activities were described as either dependent on the protease activity of tPA or not. tPA activities that required its protease activity probably reflected plasminogen activation and modification of one or more plasmin substrates, including growth factors, ECM proteins, and MMP pro-forms. Plasmin that is generated on cell surfaces also may activate cell-signaling receptors in the protease-activated receptor (PAR) family, affecting diverse cell physiologic processes (109–112).

Low-density lipoprotein receptor-related protein-1 (LRP1) is a 600-kDa transmembrane receptor in the LDL receptor gene family that undergoes endocytosis in clathrin-coated pits and rapid recycling to the cell surface (113, 114). LRP1-associated ligands are typically delivered to lysosomes. LRP1 was first identified as a tPA receptor in studies demonstrating the ability of hepatic LRP1 to clear tPA from the circulation (115, 116). As the LRP1 field advanced, it became apparent that LRP1 is capable of coupling ligand endocytosis with activation of cell signaling (113, 114). This occurs by at least two mechanisms. First, the cytoplasmic tail of LRP1 serves as a scaffold for assembling cell-signaling enzymes and adaptor proteins, facilitating activation of cell-signaling pathways (113, 117–119). Second, based on its broad ligand-binding activity, LRP1 may sequester ligands and deliver these ligands to cell-signaling coreceptors.

Activation of cell-signaling by binding of tPA to LRP1 induces expression of genes encoding MMPs (120, 121); opens the blood-brain barrier (122); promotes hippocampal long-term potentiation (123); promotes cell survival (124–126); increases cell migration (127); and promotes neurite outgrowth (128). To confirm that activities attributed to tPA reflect receptor-binding and not plasminogen activation, studies have been performed with enzymatically inactive tPA (EI-tPA), in which the active-site serine is mutated. To implicate LRP1, Lrp1 gene-silencing approaches have been applied. LRP receptor-associated protein (RAP) is a useful reagent because it blocks the interaction of LRP1 with other LRP1 ligands (129). In many cell culture model systems, RAP does not independently regulate cell signaling and thus, serves as a competitive LRP1 cell-signaling inhibitor; however, this is not uniformly true (130). A pitfall that should be recognized when studying RAP is its ability to bind to receptors in addition to LRP1 in the LDL receptor gene family (113). A variety of Lrp1 conditional gene knockout mice have been generated and have been highly useful in studying the diverse activities of this receptor (131).

A second receptor that has been implicated in tPA-initiated cell signaling is the N-methyl-d-aspartate receptor (NMDA-R) (132). Although the NMDA-R is best characterized as a neuronal ionotropic glutamate receptor (133), the NMDA-R has been identified in many cell types, including macrophages and Schwann cells in which tPA-initiated cell signaling is dependent on the NMDA-R (134, 135). Direct binding of tPA to the N-terminal region of the essential GluN1 NMDA-R subunit has been demonstrated (136, 137). This event may result in GluN1 cleavage (132, 138), although, at least in some model systems, NMDA-R subunit cleavage does not appear to be required for NMDA-R-mediated tPA-signaling because EI-tPA is active (127, 134, 135). Other NMDA-R subunits also may be cleaved when enzymatically active tPA interacts with the NMDA-R (139).

In the nervous system, the tPA-NMDA-R interaction has been reported to support neuronal survival and paradoxically, to be neurotoxic (126, 140). Experimental conditions may determine which result is observed. The effects of tPA-activated cell signaling on neuronal survival and other CNS responses also may depend on which member of the NMDA-R family mediates the tPA response. For example, blocking the activity of GluN2D-containing NMDA-Rs specifically inhibits tPA-mediated neurotoxicity (141). Interaction of tPA with GluN2B-containing NMDA-Rs in the hippocampus regulates contextual fear responses (142).

The possibility that LRP1 and the NMDA-R function together, as members of a single tPA receptor assembly, was first proposed over ten years ago (143, 144) and supported by more recent studies (127, 134, 135). A deeper dive into this synergy has shown that in PC12 cells, N2a cells, and Schwann cells, the requirement for the NMDA-R to mediate tPA-activated cell signaling is absolute (127, 135). By contrast, neutralizing LRP1 does not completely block tPA-activated cell signaling but instead, increases the tPA concentration required to elicit cell-signaling events. These results are consistent with a “sequester and deliver” model for the function of LRP1 in tPA-activated cell signaling. According to this model, most tPA signaling events are triggered by tPA, which associates first with LRP1 and is then delivered to the NMDA-R. In the absence of LRP1, signaling is limited to events in which tPA binds directly to the GluN1 NMDA-R subunit without first being sequestered at the cell surface. In support of this model, the NMDA-R and LRP1 have been reported to form a multiprotein complex, in tPA-treated neurons and neuron-like cells, bridged by the bifunctional intracellular adapter protein, postsynaptic density protein-95 (PSD-95) (127, 143).

The tPA cell-signaling receptor assembly is shown in Fig. 4. Compared with uPA, less is known about the various cell-signaling cascades activated by tPA. When tPA engages the NMDA-R/LRP1 complex, calcium influx occurs, which results in activation of SFKs (127, 128). SFKs are capable of activating diverse downstream substrates and controlling diverse cellular processes (145). RTKs are transactivated by SFKs, which expands the continuum of cell-signaling proteins that may be regulated (146). The receptor assembly shown in Fig. 4 may not be complete. For example, it has been shown that the nonpathogenic form of cellular prion protein/PrP^C associates with LRP1 to mediate cell signaling in response to tPA (147). Furthermore, it has been reported that Annexin A2 mediates tPA-activated cell signaling (148). Whether Annexin A2 functions in concert with the NMDA-R/LRP1 receptor system or independently is an important question.

Figure 4.

The tissue-type plasminogen activator (tPA) cell-signaling receptor assembly is shown. In the “sequester and deliver” model, LRP1 functions to initially bind tPA and then transfer tPA to the N-methyl-d-aspartate receptor (NMDA-R). In the absence of LRP1, tPA may bind directly to the NMDA-R; however, under these conditions, higher tPA concentrations are required to elicit cell-signaling responses. At least in some cell types, tPA binding to the tPA cell-signaling receptor assembly causes postsynaptic density protein-95 (PSD-95) to associate with both LRP1 and the NMDA-R, providing an intracytoplasmic bridge between these two receptors.

uPA RECEPTOR ASSEMBLIES IN INNATE IMMUNITY AND INFLAMMATION

In innate immunity, pattern recognition receptors (PRRs) recognize conserved products released by infectious agents and rapidly reprogram cells, generating proinflammatory mediators that support organism defense (149, 150). In mammalian systems, innate immunity provides an early response to infection. Unfortunately, dysregulated innate immunity contributes to many illnesses that involve chronic inflammation (151). Elucidating novel pathways that regulate innate immunity is very important in the anti-inflammatory therapeutics development field.

To understand the function of uPAR in innate immunity and inflammation, first, its expression pattern should be considered. In healthy adult humans, uPAR expression is fairly restricted, which has contributed to interest in developing uPAR-targeting drugs to treat cancer (152, 153). CD34-positive hematopoietic precursor cells in the bone marrow are uPAR-negative (154–156). Monocytes and granulocytes acquire uPAR as they mature and differentiate. Resting lymphocytes also are uPAR negative; however, activated lymphocytes and NK cells may express uPAR (155, 157). The effects of cellular differentiation and/or activation on uPAR expression are probably related to the ability of uPAR to promote cell migration and cell survival. The effects of uPAR on migration of neutrophils and monocytes/macrophages have been convincingly demonstrated using both in vitro and in vivo model systems (158, 159). When the gene encoding uPAR (Plaur) is deleted in mice, neutrophil migration into lungs infected with Pseudomonas aeruginosa is impaired and, as a result, clearance of the pathogen also is impaired (159).

Agents that activate PRRs in the Toll-like receptor (TLR) family upregulate uPAR expression (160, 161). Moreover, uPA and the uPA cell-signaling receptor assembly regulate TLR activity. Kiyan et al. (162) demonstrated that uPAR is a member of the TLR4 interactome. uPAR-deficient leukocytes demonstrate attenuated cell-signaling responses and decreased proinflammatory cytokine expression when treated with the TLR4 agonist, lipopolysaccharide (LPS). The toxicity of systemically administered LPS is attenuated in uPAR-deficient mice (162). uPAR deficiency also protects mice from the effects of LPS on kidney barrier function (163). uPA potentiates activation of neutrophils by LPS, by a mechanism that depends on the integrin, α_vβ₃ (164). Similarly, uPAR may be necessary for optimal neutrophil activation in response to TLR2 agonists (165). Collectively, these studies point to a role for the uPA cell-signaling receptor assembly in supporting TLR activity. However, in some complex models of human disease, the role of uPAR and its coreceptors may be less straightforward. For example, in mouse models of colitis, uPAR deficiency is associated with increased inflammation, an effect that was explained by the ability of uPAR to support integrin α_Mβ₂/Mac-1-dependent phagocytosis and bacterial product tolerance (166).

tPA RECEPTOR ASSEMBLIES IN INNATE IMMUNITY AND INFLAMMATION

Our current understanding of the activity of tPA in innate immunity and inflammation may be traced in part to earlier studies that focused on LRP1. Overton et al. (167) deleted Lrp1 in monocytes, macrophages, and neutrophils, under the control of the LysM promoter, in mice. When bone marrow from these mice was transplanted into irradiated LDL receptor-deficient mice, atherosclerosis in response to a Western diet was significantly increased. This result was explained mechanistically by studies demonstrating that LRP1 deficiency increases expression of proinflammatory mediators, such as monocyte chemoattractant protein-1/MCP1 and MMP9. LRP1-deficient macrophages demonstrated increased migration in vitro and LRP1-deficient monocytes migrated in greater numbers into atherosclerotic lesions in vivo (167). Similarly, we demonstrated that LRP1-deficient monocytes migrate in increased numbers into pancreatic tumor isografts in mice (168). Mechanistically, we attributed this result to increased expression of the promigratory chemokine, macrophage inflammatory protein-1α/CCL3, and its receptor CCR5.

May et al. (169) demonstrated that in macrophages, LRP1 deficiency is associated with increased expression of genes that define the proinflammatory M1 phenotype and decreased expression of gene products that define the M2 phenotype (169). The ability of LRP1 to regulate the expression of proinflammatory mediators is not restricted to macrophages. In a number of cell lines, deletion or silencing of the gene encoding LRP1 (Lrp1) enhanced the basal level of activity of the IκBα kinase-NFκB pathway, derepressing expression of proinflammatory mediators, including MCP-1, iNOS, and complement proteases (170).

The activities of LRP1 and tPA in inflammation were linked conceptually when we demonstrated that the effects of LRP1 on proinflammatory mediator expression are evident not only when Lrp1 is deleted but also in LRP1-expressing macrophages, when these cells are treated with LRP1 ligands (130). Importantly, the effects of LRP1 ligands on gene expression in macrophages are ligand-specific (130). In the absence of TLR agonists, the LRP1 ligands, RAP and lactoferrin, independently activate NFκB in macrophages, inducing expression of proinflammatory cytokines, promoting macrophage migration, and increasing the expression of the microRNA, miR-155. These changes mimic those observed when Lrp1 is deleted. By contrast, the LRP1 ligands, EI-tPA and activated α₂M, block NFκB activation and cytokine expression in LPS-treated macrophages (130, 134).

LRP1 ligands number over 100, including proteins released by injured and dying cells (113, 171, 172). Although the number of ligands we have examined in macrophage gene expression studies is still limited, it is intriguing to consider a model in which LRP1 ligands are divided into “cell-signaling agonists,” which oppose production of proinflammatory mediators, and “cell-signaling antagonists,” which stimulate production of proinflammatory mediators. Under basal conditions, macrophages in cell culture probably produce sufficient levels of LRP1 cell-signaling agonists endogenously to suppress activation of the NFκB pathway. Addition of exogenous LRP1 agonists, such as EI-tPA, amplifies anti-inflammatory LRP1-signaling sufficiently so that even robust LPS responses are attenuated (130, 134). Similarly, therapeutically administered EI-tPA blocks the toxicity of LPS in mice (134).

When Lrp1 is deleted, autocrine LRP1 signaling is eliminated and with it, the effects of this system on NFκB activation. In LRP1-expressing cells, antagonists such as RAP competitively block binding of endogenously produced ligands to LRP1, creating a situation similar to that observed with Lrp1 deletion. Our model, which reconciles the effects of LRP1 ligands and Lrp1 deletion, is presented in Fig. 5. Further studies will be required to determine whether this model is correct. One interesting observation is that the intensity of the proinflammatory effect observed when Lrp1 is deleted depends on the approach applied to delete the gene (130, 168). This raises the question of whether other members of the LDL receptor gene family may compensate for LRP1 when Lrp1 is deleted during development.

Figure 5.

A model showing how tissue-type plasminogen activator (tPA) may regulate NFκB activation in macrophages. It is assumed that LRP1 ligands include proteins secreted by macrophages, a fraction of which activates cell signaling via the tPA cell-signaling receptor assembly, similar to tPA. This autocrine pathway, involving endogenously produced LRP1 ligands, suppresses NFκB activation resulting from low levels of activation of pattern recognition receptors (PRRs) and other proinflammatory receptors under basal conditions (top left). When an LRP1 antagonist such as LRP receptor-associated protein (RAP) is added, or when Lrp1 is silenced or deleted, cell-signaling via the tPA receptor assembly is aborted, eliminating its activity in maintaining immune quiescence and increasing NFκB activity (top right). In the presence of LRP1, cell-signaling via the tPA receptor assembly, triggered by endogenously produced ligands, is insufficient to neutralize cell-signaling activated by high levels of PRR agonists such as lipopolysaccharide (LPS) (bottom left). Under these conditions, NFκB activation is observed. However, if tPA is administered exogenously, to supplement endogenously produced ligands, even responses triggered by high levels of PRR agonists may be neutralized (bottom right).

If the model shown in Fig. 5 is proven correct, it will be important to determine how LRP1 ligands that are cell-signaling agonists and antagonists differ. One possible explanation focuses on the NMDA-R, which is expressed by macrophages and essential for the anti-inflammatory activity of tPA (134). LRP1 agonists may be ligands that bind to LRP1 and also engage the NMDA-R. Expression of the NMDA-R in macrophages is differentiation state-dependent (173). Macrophages that are harvested from the mouse peritoneum without eliciting or activating agents express very low levels of the NMDA-R and are not responsive to tPA. However, when these cells are treated with colony-stimulating factor-1, the abundance of cell-surface NMDA-R increases and responsiveness to tPA is acquired. The importance of the NMDA-R as an essential LRP1 coreceptor in cell signaling has been demonstrated for activated α₂M as well, in early studies with cultured neurons (174).

In addition to TLR4, tPA counteracts proinflammatory responses mediated by ligands that activate other TLRs, including TLR2 and TLR9 (173). In unpublished studies, we have shown that tPA also is effective at inhibiting TLR7; however, not all PRRs are antagonized by tPA. Responses to the PRRs, nucleotide binding oligomerization domain-containing 1 (NOD1) and NOD2, are not antagonized by tPA and instead, may be amplified (173). Furthermore, plasmin that is activated on macrophage cell surfaces stimulates proinflammatory responses by activating PARs, which also cannot be blocked by tPA (112). Thus, it may be most correct to refer to the tPA receptor assembly as an “anti-TLR system” as opposed to a generalized anti-inflammatory system.

Given the PRR-selective nature of tPA, we undertook studies to assess the efficacy of EI-tPA as a candidate anti-inflammatory agent in the dextran sulfate sodium (DSS) colitis model. DSS colitis is a complex model of a human inflammatory bowel disease (IBD) in which TLRs and diverse other PRRs contribute to the symptoms and pathology observed (175–178). DSS is administered in the drinking water and causes a chemical colitis focused in the large intestines. Although this IBD model has little in common with the development of IBD in humans, once pathology develops in DSS colitis, it is characterized by epithelial denudation, crypt disruption, and extensive inflammatory infiltrates in the mucosa and submucosa.

We treated C57BL/6 mice with DSS for 5 or 7 days. A single dose of EI-tPA (2.5 mg⋅kg⁻¹) was administered by intravenous injection immediately after DSS treatment was terminated. The effects of the EI-tPA on the rate of recovery were monitored. In studies examining mouse weight, abdominal tenderness, stool character and blood content, and in posteuthanasia studies examining colon length (an index of irreversible tissue damage and scarring) and colon histology, EI-tPA significantly and robustly accelerated recovery (179).

Our DSS colitis study justifies additional studies to examine EI-tPA and other ligands that engage the tPA cell-signaling receptor assembly as candidate therapeutics for complex human diseases in which inflammation plays a pivotal role. However, many important questions still remain. For example, although most of our research has focused on macrophages, the cell type or types that are targeted by EI-tPA in the DSS colitis model remains to be determined. Given the broad range of effects on cell biology attributed to the NMDA-R/LRP1 receptor complex, we cannot be certain whether the beneficial effects of EI-tPA in the DSS model system reflects its anti-TLR activity or, for example, effects on cell survival and proliferation. Considerable opportunities for future work exist.

Another area of importance is the cell-signaling cascade that couples activation of the tPA receptor assembly to regulation of NFκB activation. This cascade remains incompletely defined. When macrophages are treated with a TLR agonist and EI-tPA simultaneously, EI-tPA does not prevent IκBα phosphorylation, which is required for activation of NFκB as a transcription factor (134, 173). Instead, EI-tPA rapidly reverses IκBα phosphorylation. This process is sufficient to prevent proinflammatory cytokine expression. An important mediator may be phosphatidylinositol 3-kinase-γ, an inhibitor of TLR-induced NFκB activation, known to be activated downstream of the NMDA-R (180, 181). In response to TLR stimulation, the cytoplasmic tail of LRP1 is phosphorylated and recruits Rab8a, which leads to activation of phosphatidylinositol 3-kinase-γ (119). The activity of this pathway in cells treated with tPA will constitute an important area of investigation.

Finally, the effects of the principal tPA inhibitor, PAI-1, on tPA cell signaling deserves consideration. PAI-1 is present in the plasma at increased levels in conditions associated with inflammation, including metabolic syndrome, cardiovascular disease, and diabetes (182–184). Whether PAI-1 is simply a biomarker of these serious illnesses or contributes mechanistically is not clear. We have shown that complex formation between catalytically active tPA and PAI-1 completely blocks the ability of tPA to inhibit NFκB activation and cytokine expression in macrophages treated with LPS (185). Similarly, PAI-1 blocks the cell-signaling activity of tPA in experiments with PC12 cells and Schwann cells. The effects of PAI-1 on tPA-activated cell signaling were unanticipated because the affinity of tPA-PAI-1 complex for LRP1 is actually higher than the affinity of tPA (186). Inhibiting the anti-TLR activity of tPA represents a plausible mechanism by which PAI-1 may be proinflammatory in diverse diseases.

CONCLUDING REMARKS/FUTURE DIRECTIONS

Although major components of the hemostasis system are, for the most part, well described, hemostasis factors have important activities in cell biology that remain only partially understood. The effect of fibrinolysis factors on innate immunity and inflammation is an important example. Considerably more work will be required to fully understand how fibrinolysis proteases and their inhibitors regulate these processes. It also is important to better understand how inflammatory mediators regulate hemostasis. This topic is not covered in this review.

Mining the hemostasis system to develop novel therapeutics to treat chronic inflammation constitutes an important opportunity. In this review, we focused on EI-tPA as an example of a candidate therapeutic derived from the fibrinolysis system. The activity of EI-tPA in a mouse model of IBD was discussed (179). In a second study, using a mouse model of spinal cord injury, EI-tPA again demonstrated efficacy, improving the ability of induced pluripotent stem cells to restore motor function (187). Although the activity of EI-tPA in this study was attributed to direct effects of EI-tPA on the stem cells, EI-tPA was administered into the spinal cord injury site with the stem cells. Inflammation is a prominent contributor to the pathology observed in spinal cord injury (188) and may have been modified by the EI-tPA.

The activity of EI-tPA, observed in the LPS challenge and DSS colitis models (134, 179), was somewhat surprising because the circulating half-life of inactive tPA in mice is less than 5 min (189). We assume that in these model systems, EI-tPA rapidly activated cell-signaling pathways in target cells, which induced sustained effects on the physiology of those cells. Because LRP1 may have opposing effects on the activity of EI-tPA in vivo, decreasing the EI-tPA concentration required to elicit NMDA-R-dependent cell signaling (114, 127, 134, 135), while, at the same time, clearing EI-tPA from the circulation (115, 116), there may be an opportunity to genetically engineer tPA and/or other ligands that engage the tPA cell-signaling receptor assembly to optimize the activity of candidate therapeutics.

Our studies in the DSS model system justify additional studies of EI-tPA in other models of diseases with chronic inflammation, including but not limited to rheumatoid arthritis, multiple sclerosis, and psoriasis. Finally, TLRs have been identified as important therapeutic targets in SARS-CoV-2 (190). Given the anti-TLR activity of the tPA cell-signaling receptor assembly, it may be legitimate to ask whether EI-tPA should be studied as a candidate therapeutic for this illness.

GRANTS

Dr. S. L. Gonias was partially supported by NIH Grants R01 HL136395 and R01 NS097590.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the author.

AUTHOR CONTRIBUTIONS

S.L.G. prepared figures; drafted manuscript; edited and revised manuscript; and approved final version of manuscript.