Proteinase-Activated Receptors: Transducers of Proteinase-Mediated Signaling in Inflammation and Immune Response
http://www.100md.com
《内分泌研究杂志》
Abstract
Serine proteinases such as thrombin, mast cell tryptase, trypsin, or cathepsin G, for example, are highly active mediators with diverse biological activities. So far, proteinases have been considered to act primarily as degradative enzymes in the extracellular space. However, their biological actions in tissues and cells suggest important roles as a part of the body’s hormonal communication system during inflammation and immune response. These effects can be attributed to the activation of a new subfamily of G protein-coupled receptors, termed proteinase-activated receptors (PARs). Four members of the PAR family have been cloned so far. Thus, certain proteinases act as signaling molecules that specifically regulate cells by activating PARs. After stimulation, PARs couple to various G proteins and activate signal transduction pathways resulting in the rapid transcription of genes that are involved in inflammation. For example, PARs are widely expressed by cells involved in immune responses and inflammation, regulate endothelial-leukocyte interactions, and modulate the secretion of inflammatory mediators or neuropeptides. Together, the PAR family necessitates a paradigm shift in thinking about hormone action, to include proteinases as key modulators of biological function. Novel compounds that can modulate PAR function may be potent candidates for the treatment of inflammatory or immune diseases.
I. Introduction
II. PAR1 in Inflammation and Immune Response
A. Vasculature and heart: PAR1 antagonists—a novel potential approach for the treatment of cardiovascular diseases
B. Platelets
C. Immune cells
D. Airways
E. Gastrointestinal tract
F. Kidneys and urogenital tract
G. Brain and peripheral nervous system
H. Signaling by proteinases via PAR1
III. PAR2 in Inflammation and Immune Response
A. Vasculature
B. Immune cells
C. Arthritis
D. Skin
E. Airways
F. Brain and peripheral nervous system
G. Digestive tract and pancreas
H. Signaling by proteinases via PAR2
IV. PAR3 and PAR4
A. Biology and distribution of PAR3 and PAR4
B. Signaling by proteinases via PAR3 and PAR4
V. Conclusions
I. Introduction
Top
Abstract
I. Introduction
II. PAR1 in Inflammation...
III. PAR2 in Inflammation...
IV. PAR3 and PAR4
V. Conclusions
References
SERINE PROTEINASES CONSTITUTE a family of proteolytic enzymes characterized by a unique catalytic triad consisting of Ser, His, and Asp. These residues are able to hydrolyze peptide bonds (1). Serine proteinases are produced as inactive precursors or zymogens. Subsequent zymogen conversion into a mature physiologically active enzyme is mediated via a process called "limited proteolysis" or zymogen activation (2, 3, 4). In mammals, serine proteases, for example, regulate the hemostatic and fibrinolytic balance, degrade neuropeptides involved in neurogenic inflammation or serve as modulators of immune response during inflammation (5). Three different types of serine proteinase inhibitors can be distinguished based on their mechanism of action: canonical, noncanonical inhibitors, and serpins (6). An imbalance between these inhibitors and their targeted proteinases can affect immune/inflammatory responses and may result in disease. Moreover, the presence of proteinase inhibitors regulates and limits interactions between proteinases and their receptors (7, 8, 9, 10, 11).
Recent studies elucidated the ability of certain serine proteinases to regulate cell function via G protein-coupled receptors (GPCRs). At least two different types of proteinase receptors have been identified involving proteolytic cleavage in their activation mechanism: urokinase receptors and proteinase-activated receptors (PARs) (12, 13, 14).
PARs belong to a new subfamily of GPCRs with seven transmembrane domains activated via proteolytic cleavage by serine proteinases (13, 14, 15, 16). PAR1, PAR3, and PAR4 are targets for thrombin, trypsin, or cathepsin G (17, 18, 19, 20). In contrast, PAR2 is resistant to thrombin, but can be activated by trypsin, mast cell tryptase, factor Xa, acrosin, gingipain, and neuronal serine proteinases (21, 22, 23, 24, 25, 26, 27, 28) (Table 1). Interestingly, PARs are activated by a unique mechanism: proteinases activate PARs by proteolytic cleavage within the extracellular N terminus of their receptors, thereby exposing a novel "cryptic" receptor-activating N-terminal sequence that, remaining tethered, binds to and activates the receptor (Fig. 1) within the same receptor (17, 21, 22). Specific residues (about six amino acids) within this tethered ligand domain are believed to interact with extracellular loop 2 and other domains of the cleaved receptor (29), resulting in activation. This intramolecular activation process is followed by coupling to G proteins and the triggering of a variety of downstream signal transduction pathways (see Sections II.H, III.H, and IV.B; also see Table 2 and Fig. 2, and Refs.13 , 14 , and 16). Thus, PARs are not activated like "classical" receptors because the specific receptor-activating ligand is part of the receptor, whereas the circulating agonist is a relatively nonspecific serine proteinase that does not behave like a traditional hormonal regulator akin to insulin.
Several studies during the past few years have also demonstrated that several mechanisms exist to regulate stimulation and termination of PAR-initiated signaling (13, 14, 15, 16, 30). Importantly, the availability of PARs at the cell surface is governed by trafficking of the receptor from intracellular stores, and the signaling properties depend on the presence of G proteins and G protein-coupled receptor kinases (GRKs) that modify activity. For PAR1, PAR2, and PAR4, it is well established that short synthetic peptides [PAR-activating peptides (PAR-APs)] designed on their proteolytically revealed tethered ligand sequences can serve as selective receptor agonists (Table 1). Some PAR-APs activate more than one PAR, and they activate receptors at concentrations in the micromolar range as compared with nanomolar potencies of the proteinases themselves (31, 32). Although the PAR1-AP, SFLLRN-NH2 also activates PAR2, PAR2-APs, like SLIGRL-NH2 are not capable of activating other PARs. Unfortunately, the relatively low potency (10 to 100 μM EC50) and susceptibility to aminopeptidases (33) limit the utility of the PAR-APs in some bioassay systems. Recently, modified synthetic agonist peptides with higher potency, resistance to aminopeptidases, and greater receptor selectivity have been developed and characterized. These receptor-selective agonists are of use to study the consequences of activating PARs in vivo (Table 1) (13, 14, 16, 34). So far, antagonists for PAR1 (35, 36, 37) and PAR4 (38) have been synthesized, but are not yet available for PAR2. PAR3, as will be elaborated upon in Section IV, remains a puzzle, in that studies with the murine receptor indicate that it cannot be activated either by its cognate synthetic tethered ligand peptide or by thrombin (39, 40). Rather than acting as an independent cell regulator, PAR3 appears to function as a cofactor for the activation of PAR4 (39). Complementary data documenting PAR-mediated effects in various tissues have been obtained using PAR-deficient (PAR–/–) mice.
Taken together, data obtained using the enzyme activators themselves (trypsin, thrombin), the PAR-APs and using PAR gene-deficient mice provide compelling evidence that PARs play a critical role in the regulation of various physiological and pathophysiological functions in mammals, including humans. This review focuses on the biology and signaling properties of PARs in various mammalian tissues and highlights the current knowledge about the role of PARs during inflammation and immune response. To complement the information summarized in the sections that follow, the reader is encouraged to consult a number of other recent reviews concerning the activation mechanisms and cell biological aspects of PARs (13, 14, 16, 34, 41, 42, 43, 44, 45).
II. PAR1 in Inflammation and Immune Response
Thrombin is an important effector proteinase of the coagulation cascade that leads to formation of a hemostatic plug. Thrombin is thought to act near the site at which it is generated and it is activated when circulating coagulation factors in the blood plasma make contact with tissue factor. Tissue factor is a membrane protein that is usually produced by cells that are separated from blood (i.e., epithelial cells). However, it is also expressed at low levels on circulating monocytes and microparticles from leukocytes. Tissue factor is associated with the activation of zymogen factor X by factor VIIa. Factor Xa together with its cofactor Va subsequently converts prothrombin to the active enzyme. Thus, plasma coagulation can only take place when the vascular integrity is damaged (15). Thrombin causes shape change of endothelial cells and increased permeability of endothelial cell layers.
However, recent evidence revealed that thrombin is not only a clotting proteinase serving as both a pro- and anticoagulant molecule but also appears to play multifunctional roles related to inflammation, allergy, tumor growth, metastasis, tissue remodeling, thrombosis, and probably wound healing (13, 15, 42, 43). Subsequent to the cloning of PAR1 (17, 46), it was realized that many of the cellular actions of thrombin (e.g., platelet aggregation, angiogenesis, endothelial cell permeability, vasoregulation, gene regulation, leukocyte trafficking, immunomodulation) could be attributed to the activation of its GPCR (40, 47, 48, 49, 50, 51, 52, 53).
PAR1 has been detected in a variety of tissues, including platelets; endothelial cells; fibroblasts; monocytes; T cells positive for CD 8, CD 16, and either CD 56 or CD 57 (54, 55); natural killer (NK) cells (48); CD 34+ hematopoietic progenitor cells (56); dental pulp cells (57); smooth muscle cells (SMC); epithelial cells; neurons; glial cells; mast cells; and certain tumor cell lines (17, 54, 58, 59, 60, 61). A receptor with high affinity for thrombin has also been detected in rat peritoneal macrophages (62). It should be pointed out, however, that PAR1 very likely does not represent the only target for thrombin, in that other (non-PAR) high-affinity binding sites for thrombin [e.g., on platelets or macrophages (62)] and other conserved thrombin sequences apart from the catalytic domain (63) may also play a role in the cellular actions of thrombin in a variety of target tissues.
A. Vasculature and heart: PAR1 antagonists—a novel potential approach for the treatment of cardiovascular diseases
PAR1 can potentially regulate vascular function under both physiological as well as pathophysiological conditions (15, 42). A number of studies have revealed that thrombin and other agonists of PAR1 can affect the vascular tone. For example, before the discovery of PAR1, it was observed that thrombin can regulate vascular tone by an endothelial-dependent mechanism involving the release of nitric oxide (NO) (64). It is now recognized that this effect of thrombin is due to the activation of PAR1. Moreover, thrombin and PAR1 agonists can contract vascular SMC (VSMC) by a direct effect that requires extracellular Ca2+ (64). Thus, in isolated coronary artery and aorta preparations, PAR1 mediates relaxation (64, 65). In contrast, PAR1 agonists contract human placental and umbilical arteries (66). Furthermore, PAR1 mediates prostanoid generation and secretion, cellular contraction and barrier dysfunction, and enhanced expression of platelet-derived growth factor (PDGF) in human and bovine endothelial cells (67). Interestingly, progestins such as progesterone or gestodene up-regulate PAR1 expression and thereby stimulate thrombin-induced tissue factor-dependent surface procoagulant activity in the rat vascular system, suggesting a role of thrombin in hormone-induced thrombosis via PAR1 activation (68).
Recent observations support a role of thrombin and PAR1 in regulation of functions normal (69, 70, 71, 73) and atherosclerotic (75) endothelium. In normal human arteries, PAR1 is mostly confined to the endothelium, whereas during atherogenesis, its expression is enhanced in regions of inflammation associated with macrophage influx, smooth muscle cell proliferation, and an increase in mesenchymal-like intimal cells (75). In vivo, a neutralizing antibody to PAR1 has been observed to reduce expression of mRNA for the proliferating cell nuclear antigen, an index of intimal and neointimal smooth muscle cell accumulation in rat arteries during balloon angioplasty. These data suggest that PAR1 regulates proliferation and accumulation of neointimal SMC during tissue repair (76).
Thrombin and PAR1 agonists cause a rapid but transient contraction of endothelial cells in various tissues resulting in gap formation and increased permeability of plasma proteins and inflammatory cells. Several mediators are involved in this PAR1 modulated process such as cytokines, kinins, and biogenic amines (67, 72, 77, 78, 79, 80, 81, 82, 83, 84, 85, 86, 87, 88, 89, 90, 91, 92, 93, 94, 95).
It was also revealed that PAR1 plays a crucial role during vascular ontogenesis. Interestingly, about 50% of the PAR1-deficient mouse embryos die at midgestation with bleeding from multiple sites. However, a PAR1 transgene driven by an endothelial-specific promoter prevented death of PAR1-deficient embryos (96), indicating that PAR1 modulates endothelial cell function in developing blood vessels, thereby contributing to vascular development and homeostasis in mice (88, 96).
Vascular wall cells respond to the procoagulant factor Xa by an increase in intracellular Ca2+ ([Ca2+]i) and by assembly of this factor into prothrombinase complexes that even enhance this effect. Additionally, factor Xa stimulation of PAR1 leads to an increased production of tissue factor, a prothrombotic agent, underlining the important role of PAR1 for thrombosis (86). Together, these results point to a pivotal role of PAR1 in vascular homeostasis and thrombosis.
PAR1 agonists are also mitogenic, stimulating proliferation of endothelial cells (97, 98), mediating endothelium-dependent relaxation to thrombin and trypsin in human pulmonary arteries (99), and causing the release of IL-6 from human microvascular endothelial cells (HMEC) (72). Because PAR1 up-regulates 1(I)-procollagen synthesis in VSMC, one may speculate that PAR1 plays a role in vascular wound healing (100).
Finally, factors that regulate PAR1 function on endothelial cells have also been studied. For example, PAR1 is down-regulated by shear stress (101) and inflammatory mediators such as TNF- (102), and is directly modulated by thrombomodulin (103) in human endothelial cells. On the other hand, cyclic mechanical stress leads to an up-regulation of PAR1 in VSMC (104). In contrast to PAR2, little is known about the inflammatory mediators that regulate PAR1 expression in unstimulated or stimulated endothelial cells.
As of yet, only a few studies exist about the role of GRKs in regulating PAR1 in endothelial cells. Only recently, Tiruppathi et al. (105) found a crucial role of the isoform GRK-5 on thrombin-induced desensitization of PAR1.
It is also necessary to mention the recently demonstrated role of PAR1 in vascular associated pathological processes in which thrombin is involved. The role of PAR1 in thrombus formation was recently investigated in an animal model. Cook et al. (106) examined the role of PAR1 in intravascular thrombus formation in an experimental model of arterial thrombosis in the African green monkey. Using a blocking antibody to PAR1, this group demonstrated a dramatically diminished thrombin-stimulated aggregation and secretion of human platelets, whereas platelet activation induced by the PAR1 agonist SFLLR-NH2 was not affected. These results demonstrated that a specific blockade of PAR1 activation by thrombin can prevent arterial thrombosis in this animal model without significantly altering hemostatic parameters. The data suggest that PAR1 is crucially involved in this disease and is an attractive antithrombotic therapeutic target, although it probably has to be inhibited in concert with the low-affinity thrombin receptor PAR4 to fully prevent platelet activation (see Section II.A).
A demonstrated role of PAR1 in intravascular thrombus formation and knowledge about side effects of thrombin inhibitors made appealing an idea to investigate the potency of PAR1 antagonists as antithrombotic therapeutic drugs. Indirect thrombin inhibitors like heparin and warfarin, which prevent the circulation of thrombin, or the newly developed direct thrombin inhibitor ximelagatran (465) are currently used for the treatment of cardiovascular diseases, e.g., myocardial infarction, atherosclerosis, and thrombosis. However, they carry potential bleeding liabilities as an undesirable side effect when administered in chronic patients. Because it is now well known that human platelet responses to thrombin are mainly mediated via PAR1 (19), direct inhibition of this receptor instead of the proteinase represents an attractive way to circumvent possible side effects by thrombin inhibitors.
PAR1 possesses several extra- and intracellular sites that are crucial for its function and that might represent possible targets for antagonists (107). On the extracellular side, blocking of the cleavage site or the hirudin-like domain could inhibit N-terminal cleavage. Bradykinin or peptides derived thereof, also known as thrombostatins, have been shown to prevent thrombin-induced platelet aggregation (108). In addition, antibodies directed against both the cleavage site and the hirudin-binding site have been generated (106, 109). Binding of the antagonist to either the tethered ligand or extracellular loop 2 prevents interaction of the novel N terminus with the extracellular loop. Several small synthetic antagonists have been developed that block the binding site of the tethered ligand. However, peptide antagonists based on modifications of the tethered ligand were relatively unstable or showed only limited inhibitory activity (35). Thus, a second generation of indole or indazole-based peptide-mimetic antagonists was created that block binding of both the tethered ligand and the proteinase (110). One of these was recently shown to prevent thrombus formation in nonhuman primates (107). In addition, several nonpeptide inhibitors were developed. However, it was not reported whether these compounds were PAR1-specific (111). On the intracellular side, the G protein binding sites are possible targets for pharmacological intervention. One approach used the transfection of endothelial cells with a G minigene (112). However, this might suppress all signaling pathways that involve G. Another more specific approach used so-called "pepducins," peptides that were derived from the third intracellular loop of PAR1. These peptides were able to permeate the cells and also carried a membrane anchor (113, 114).
In summary, two preclinical studies with primates demonstrated that PAR1 antagonists have indeed a therapeutic potential for the treatment of cardiovascular diseases (106, 107). However, extensive research needs to be done for the development of orally administered PAR1 antagonists because only those drugs are thought to be suitable for chronic patients (111).
B. Platelets
Activation of platelets by thrombin or APs specific for PAR1 is characterized by calcium influx; cytoskeletal reorganization; platelet aggregation; degranulation (17, 115, 116); thromboxane production (117); mobilization of the adhesion molecule P-selectin and the CD40 ligand to the platelet surface (119); stimulation of serotonin and epinephrine release (120); enhanced expression of CD62, PDGF (AB) and PDGF(BB) (121); and exposure of anionic phospholipids (phosphatidylserine, phosphatidylethanolamine) that support blood clotting (122). Thrombin has also been demonstrated to stimulate vascular endothelial growth factor release in megakaryocytes and platelets (123). The growth factor in turn is able to induce proliferation of endothelial cells. In addition, tissue inhibitor of matrix metalloproteinase (MMP)-4 that colocalized with MMP-2 in resting platelets was released upon platelet aggregation induced by collagen and thrombin. However, a direct effect using specific PAR1 agonists has not been shown yet (124). The density of human PAR1 on the platelet cell surface was estimated to be about 1500 molecules per resting platelet (125). A recently reported intronic polymorphism in the PAR1 gene leads to decreased expression on the platelet surface and thus to a lower response to the AP (126, 127). PAR1 is a high-affinity thrombin receptor in humans: receptor blocking using an antagonist, domain specific antibodies, or simply by desensitization led to inhibition of responses at 1 nM thrombin but only to attenuated responses at 30 nM (128).
High-affinity thrombin binding of platelets is possibly associated via the membrane glycoprotein (GP)Ib-IX-V complex. This interaction is lost in patients suffering from the Bernard-Soulier syndrome, a disease where GPIb-IX-V is not expressed (129). The same condition can be mimicked using either monoclonal antibodies directed against GPIb or proteolytic removal of this protein from the platelet surface (130). Under these conditions, platelet aggregation is either delayed or requires higher thrombin concentrations.
Recently, Schmidt et al. (131) reported that the gene encoding IQ motif containing GTPase activating protein-2, a scaffolding protein for filopodal extension during platelet activation, was identified within the human PAR gene cluster at 5q13, flanked by the PAR1 gene and encompassing the PAR3 gene. Only thrombin- or PAR1-AP-activated platelets showed a rapid translocation of IQ motif containing GTPase activating protein-2 to the platelet cytoskeleton. This suggests that a functional genomic unit evolved to mediate thrombin signaling events in humans (132).
PAR1 on platelets also seems to play an important role during bacterial invasion: the cysteine proteinase gingipain from Porphyromonas gingivalis was reported to activate the receptor leading to uncontrolled activation of the host cells (28). The enzyme streptokinase derived from Streptococcus sp. is widely used in the treatment of coronary thrombosis. It functions as a plasminogen activator that forms an active complex with plasminogen of the host that is able to cleave PAR1 (133). In mouse platelets, PAR1 does not seem to play a role: PAR1 expression was hardly detectable, and a specific AP did not activate rodent platelets (23, 134, 135). Moreover, platelets derived from PAR1-deficient mice responded to thrombin-like wild-type platelets (135).
C. Immune cells
For some time, an important role for serine proteinases and PARs for the modulation of leukocyte effector functions has been proposed (136). As of yet, only limited data about the regulation of immune and inflammatory responses by PARs are available. It is well documented that PARs influence monocyte motility and chemotaxis, modulate pleiotropic cytokine responses, contribute to mononuclear cell proliferation, and induce apoptosis in various immune cells (84, 137). Recently, it was shown that PAR1 is capable of stimulating elastase secretion from macrophages (138). Moreover, functional thrombin receptors are expressed on human T lymphoblastoid cells (139). However, the receptor subtype(s) (PAR1, PAR3, and PAR4) have not been characterized as of yet. Human -thrombin stimulated five different T lymphoblastoid cell lines to increase intracellular free Ca2+ concentrations and to activate protein kinase C (PKC), whereas thrombin receptors were absent in B cell lines. Thus, PARs may regulate T cell function during inflammation and immune responses, but the precise mechanism is still unknown. Granzyme A from cytotoxic and helper T lymphocytes appears to interact with PAR1 in astrocytes to regulate thrombin function (140). Interestingly, granzyme A blocked the thrombin-induced platelet aggregation in a dose-dependent manner by cleaving PAR1, presumably downstream from its thrombin targeted activation site, thereby reducing the response to subsequent challenge with thrombin; but granzyme A itself did not induce a signal in thrombin-stimulated platelets. Thus, granzyme A may interact with PAR1 in a manner that is insufficient to cause aggregation, but sufficient to disarm the ability of the receptor to respond to thrombin. Finally, these observations demonstrate that granzyme A release occurring during immune responses within blood vessels would not directly cause platelet aggregation. Thus, the T cell-derived proteinase granzyme A seems to inhibit responses triggered by thrombin during inflammation or tissue injury.
PAR1 modulates chemotaxis in inflammatory cells. Besides IL-8 secretion (71), thrombin induces production of monocyte chemoattractant protein-1 (MCP-1) in monocytes, probably via PAR1 (58). However, other PAR1-specific agonists were not used in this study. Therefore, it cannot be excluded that other thrombin receptors are involved. IL-8 secretion is up-regulated by interferon- (IFN-) and diminished by prostaglandin (PG)E2 (51), suggesting a role in cytokine modulation. Furthermore, PAR1 is capable of inducing IL-1 as well as IL-6 production in monocytes (82). These cytokines are known to be proangiogenic, implicating PAR1 in angiogenesis and tissue repair. However, IFN--differentiated growth-arrested U937 cells also respond to PAR1 agonist administration by overcoming cell arrest and revert to a high proliferation rate via regulation of p21CIP1/WAF1 and cyclin D1. Together, these results may help to explain how thrombin promotes tissue repair and unrestricted proliferation in malignant tissues (137, 141).
Because thrombin, via PAR1, is chemotactic for large granular lymphocytes in humans (55), large granular lymphocytes can enhance the effects of PAR1 in patients with inflammatory disorders (142, 143, 144). Moreover, thrombin modulates activity of NK cells (48). For example, thrombin can enhance NK cell-mediated cytotoxicity and IL-2 production in rheumatoid arthritis, and PAR1 may therefore play an important role in this inflammatory process, as well as in NK cell responsiveness to IL-2.
D. Airways
Various cells in mammalian airways such as epithelial cells, SMC, and fibroblasts abundantly express PAR1 on the cell surface. Several recent studies also suggest that PAR1 may play an important role in inflammatory lung diseases such as neutrophilic alveolitis, pulmonary fibrosis, and asthma (50, 77, 91, 145, 146, 147, 148, 149).
In the airways of different species, PAR1 exerts a dual role leading to stimulation of contraction on the one hand and relaxation on the other hand. For example, thrombin stimulates contraction of human bronchial rings (150) and constriction in guinea pig bronchi (as does AP) (151), but activates constriction as well as relaxation in mouse tracheal SMC (152, 153). In these mouse studies, trypsin, thrombin, and peptide agonists of PAR1, PAR2, and PAR4 induced relaxant responses of isolated tracheal smooth muscle preparations, which were mediated by a prostanoid, probably PGE2. This effect was abolished by indomethacin, the cyclooxygenase (COX)-2 inhibitor, nimesulide, and a prostanoid receptor antagonist (AH6809). PAR1 and PAR4 synthetic peptides induced a rapid, transient, contractile response that preceded the relaxant response. Interestingly, the relaxations but not the contractions were inhibited by indomethacin, indicating that this response is mediated by cyclooxygenase products.
E. Gastrointestinal tract
PAR1 was detected in the lamina propria, the submucosa, endothelial cells, and nerves of the gastrointestinal (GI) tract (154, 155, 156, 157). The GI tract expresses relatively high levels of PAR1 mRNA compared with other tissues, both in mice and humans. The functional role of PAR1 as a receptor mediating thrombin-induced effects in the GI tract is far from being resolved because thrombin can also interact with PAR3 and PAR4 (18, 20, 74). Several studies suggest a role for PAR1 in regulating GI motility. In guinea pigs, PAR1 mediates contraction of longitudinal smooth muscle tissue, dependent on extracellular Ca2+ (158, 159). In mouse intestine, PAR1 agonists modulate the function of L-type Ca-channels and also cause contraction (161). Similar results have been observed in rats (162, 163). Irradiation by fractionated X-radiation leads to up-regulation of PAR1 at the protein level, indicating a regulatory role of external trigger factors such as irradiation on PAR1 expression (164). Stimulation of PAR1 upon adhesion of pancreatic carcinoma cells to extracellular matrix proteins such as laminin, collagen IV, and fibronectin has been observed in a pancreatic carcinoma cell line (MIA PaCa-2) (165, 166). Other studies have pointed out a role for PAR1 in intestinal secretory pathways. Buresi et al. (156) have shown that selective PAR1 agonists stimulated Ca2+-dependent chloride secretion in intestinal epithelial cells. This PAR1-induced intestinal chloride secretion involves protein tyrosin kinase Src (Src), epidermal growth factor (EGF) receptor (EGFR) transactivation, activation of a MAPK pathway, phosphorylation of cytosolic phospholipase A2 (cPLA2), and cyclooxygenase activity. The presence of functional PAR1 on intestinal epithelium and the fact that PAR1 activation leads to ion secretion suggest that PAR1 might have important implications for intestinal barrier functions. PAR1 activation on intestinal surfaces could lead to a secretory response, thus contributing to diarrhea, a symptom of intestinal inflammation. More recent observations further suggest a role of PAR1 in intestinal inflammation. We have shown that intracolonic administration of PAR1 agonists caused inflammation and disruption of intestinal barrier integrity (53). Taken together, these results suggest a proinflammatory role of PAR1 in the gut (reviewed in Ref.167). However, additional studies using PAR1 antagonists and/or PAR1-deficient mice will have to verify the role of PAR1 in the pathogenesis of inflammatory bowel diseases.
F. Kidneys and urogenital tract
Recently, a crucial role of PAR1 in the cell-mediated renal inflammation of crescentic glomerulonephritis has been demonstrated in vivo (49). In wild-type mice treated with hirudin (a direct thrombin inhibitor, characterized by a bifunctional mechanism of inhibition, exclusive specificity and strong ability to bind the enzyme) and in PAR1-deficient animals, the proinflammatory effect of thrombin was significantly reduced. Moreover, treatment of wild-type mice with the PAR1 peptide agonist (SFLLRN-NH2) augmented the inflammatory response, suggesting that PAR1 plays an important role in renal inflammation in vivo. Unfortunately, as mentioned above, the peptide SFLLRN-NH2 can also activate PAR2 with a comparable potency, and therefore a role of PAR2 in renal inflammation cannot be ruled out by this study. Surprisingly, although highly expressed in human kidneys, data clarifying a role for PAR1 in inflammation of human kidneys are still lacking.
Recent data revealed that PAR1 is involved in cellular invasion of a transfected canine kidney cell line. PAR1 agonists abrogated G(olf)-mediated invasion of MDCKts.src cells in collagen gels, indicating an important role for PAR1 in tumor metastasis (168, 169). Using the human urogenital cell line RT4, studies done in vitro have revealed that PAR1 and PAR2 agonists are capable of activating iPLA2 accompanied by release of PGE2, which may provide cytoprotection during an acute inflammatory reaction (170).
G. Brain and peripheral nervous system
Recent studies are in favor of an important role of thrombin and PARs in the brain under normal and pathophysiological conditions such as trauma, inflammation, or tumorigenesis (171, 172). Under pathophysiological conditions, i.e., during breakdown of the blood-brain barrier, circulating thrombin in the bloodstream may enter the central nervous system (CNS), leading to activation of PAR1 or PAR4. Additionally, neurons and glia cells are capable of generating prothrombin (173). Most data so far have been achieved by investigating PAR1 (13, 14, 174). In rat brain, PAR1 is expressed by neurons of the neocortex, cingulate cortex, subsets of thalamic and hypothalamic nuclei, discrete layers of the hippocampus, cerebellum, and olfactory bulb, as well as by astroglia (60). Striggow et al. (175) showed that all four PAR subtypes are expressed in rat brain. PAR1 expression was most abundant in the hippocampus, amygdala, and cortex. Interestingly, the expression of PAR1, PAR2, and PAR3 was up-regulated during experimentally induced ischemia.
In human brain, PAR1 is expressed in neurons and astrocytes (175). Cultured rat glia cells (C6) express both functional PAR1 and PAR2 (176, 177). Moreover, PAR1 agonists induce up-regulation of inducible NO synthase (iNOS) in these cells (85) and promote neuronal survival after ischemia (178) or brain trauma (179). PAR1 also protects astrocytes and neurons from apoptosis induced by hypoglycemia and oxidative stress during inflammation (180). In contrast, thrombin may also exert cytotoxic effects on neurons and induce neurite retraction (181, 182), which may be at least in part due to PAR1. Functional studies further revealed that PAR1 agonists cause retraction of neurites by neuroblastoma cells (181) and induce Ca2+ mobilization by hippocampal neurons (178).
In astrocytes, thrombin stimulates aggregation, morphological changes, and proliferation via PAR1 and induces intracellular caspase pathways leading to apoptosis in a cultured motor neuron cell line (NSC19) (183, 184). Both PAR1 and PAR2 stimulate enhanced proliferation of astrocytes (185). Friedmann et al. (186) reported a key role of thrombin in PAR1-mediated postinjury neuron survival. They demonstrated that the toxicity of thrombin on neurons can be controlled by down-regulation of PAR1 and/or release of antithrombin III by T cells. Interestingly, prothrombin expression was enhanced 24 h after injury, whereas PAR1, PAR3, and nexin-1 mRNA expression was unchanged (187). Nexin-1 is a thrombin inhibitor that also appears to control thrombin-PAR1 interactions in a rat trauma model (171, 187, 188). Thus, the whole factory needed to regulate and modulate thrombin function, including receptor as well as activating and inhibiting proteinases, can be generated in the brain. However, PAR1 also induces the reversal of astrocyte stellation in mice and rat (189, 190, 191). A comparable process is caused by exogenous or endogenous injuries of the CNS, triggering astrogliosis.
Taken together, these data clearly indicate a functional role of PAR1 during inflammation and injury in the CNS. However, in some studies using thrombin as an agonist, activation of other PARs and a role of other molecules like thrombomodulin cannot be excluded. For example, thrombin itself up-regulates thrombomodulin in astrocytes in a dose-dependent manner (192). Thus, future studies taking into account all CNS-derived PARs, their proteinases, and proteinase inhibitors are necessary to fully explore the role of PARs in the brain. Finally, some authors suggest a role of thrombin in Alzheimer’s disease, amyotrophic lateral sclerosis, or HIV encephalitis, probably via activation of PARs (182, 193, 194).
Accumulating data also indicate the role of PAR1 in the peripheral nervous system. In rat peripheral nerves, PAR1 is expressed by primary afferent neurons (195). From these observations, one may speculate that thrombin may also regulate inflammatory responses in the peripheral nervous system via PAR1. Indeed, very recently it has been demonstrated that PAR1 is expressed by a large proportion of primary spinal afferent neurons (155). Thus, thrombin directly signals to sensory neurons by cleaving PAR1. Moreover, administration of PAR1 agonists can induce plasma extravasation and edema, which were blocked by ablation of sensory nerves and administration of antagonists to the neurokinin-1 receptor, supporting the idea that thrombin cleaves PAR1 on sensory nerves to stimulate release of SP, which in turn interacts with the NK1 receptor to induce neurogenic inflammation. Thus, thrombin triggers neurogenic inflammation via PAR1 (155). However, because PAR2 and PAR4 mRNA are also expressed in the CNS, it cannot be excluded that proinflammatory effects of thrombin in the nervous system may, in addition, be mediated by other PARs. Functional studies further revealed that PAR1 agonists induce Ca2+ mobilization in enteric neurons, and regulate both excitatory and inhibitory neurons of the myenteric plexus in guinea pig small intestine, e.g., by releasing neuropeptides (196).
Recently, Vergnolle and co-workers (197) as well as other groups (198) examined the effects of PAR1 agonists on nociceptive responses to mechanical and thermal noxious stimuli. Interestingly, intraplantar injection of selective PAR1 agonists induced an enhanced nociceptive threshold and withdrawal latency resulting in mechanical and thermal analgesia. However, thrombin was analgesic in response to mechanical, but not to thermal, stimuli. Moreover, application of PAR1-AP with carrageenan significantly reduced the hyperalgesia. Thus, thrombin may play a dual nociceptive-analgesic role, and future studies will be required to determine whether PAR1 agonists might be of therapeutic use for the treatment of pain.
H. Signaling by proteinases via PAR1
Investigations during the last few years revealed that the above-mentioned PAR1-mediated effects are transduced by various signaling pathways leading to diverse functions of PAR1 under physiological and pathophysiological conditions. These findings were recently reviewed (13). However, in the current work we focus on the signaling events mediated via PAR1 in different tissues and cell types (Table 2).
1. Platelets.
After its generation from prothrombin, thrombin plays multiple roles in the blood coagulation cascade that are mediated by interaction with a number of physiological substrates, effectors, and inhibitors. The accumulation of thrombin at sites of vascular injury provides the recruitment of platelets into a growing hemostatic plug. When added to human platelets in vitro, thrombin causes platelets to change shape, stick to each other, and secrete the contents of their storage granules. How this is accomplished is still not fully understood, but a major step forward occurred since the identification of PAR1 in the 1990s. This finding shed light on the way by which an extracellular protease may initiate intracellular events.
Mammalian GTP-binding proteins (G proteins) fall into four families that are typically referred to by the designation of the -subunit: Gs, Gi, Gq/11, and G12/13. Human platelets express at least one member of the Gs family and four members of the Gi family (Gi1, Gi2, Gi3, and Gz), which, among other functions, stimulate or inhibit cAMP formation by adenylyl cyclase (199). In addition, platelets express one or more members of the Gq/11 family and members of the G12/13 family (G12 and G13) (199).
A number of different G protein -subunits have been shown to bind to PAR1, including members of the Gi, Gq/11 and G12/13 families (200). However, the knowledge of these interactions in humans under physiological and pathophysiological conditions is still far from completion.
Kim et al. (201) have demonstrated that thrombin and PAR1-AP, as well as PAR4-AP, induce Gi pathway stimulation in human platelets. Additionally, the authors demonstrated that thrombin, PAR1-AP, and PAR4-AP cause platelet aggregation independently of Gi stimulation (201). Offermanns et al. (202) also demonstrated a direct interaction between G12, G13, and PAR1. Additionally, it was reported that thrombin-mediated cleavage of the PAR1 (and PAR4; see Section IV.B) receptor leads to calcium-dependent and calcium-independent shape changes of human platelets in consequence of direct activation of both Gq and G12/13 pathways, respectively (115, 202). Therefore, current evidence suggests that PAR1 interacts with Gq, G12/13, and possibly Gi protein family members in human platelets. In turn, these data suggest an activation of downstream signaling cascade members after PAR1 stimulation on human platelets. Among such members are phospholipase C (PLC)-?, phosphoinositide 3 (PI3)-kinase, and monomeric G proteins.
Indeed, PI3 kinase was demonstrated to play an important role in some PAR1 or thrombin-mediated cellular effects such as cytoskeletal reorganization, alterations in cell motility, cell survival, and mitogenesis. For example, thrombin is able to activate multiple PI3-kinase isoforms, including the recently discovered 110-kDa isoforms that can be directly activated by G protein ?-subunits (145, 150, 203). Stimulation of platelets by PAR1 leads to the activation of PI3 kinase, which is dependent on the small G protein Rho (204). Recently, Trumel et al. (205) have demonstrated that PI3 kinase plays an important role in the PAR1-dependent reorganization of the platelet cytoskeleton via myosin heavy chain translocation and stable association of signaling complexes with the actin cytoskeleton. Interestingly, the activity of small G proteins such as Rac and cdc42 in platelets may be regulated through PAR1 stimulation, but the role of PI3 kinase in these events remains to be determined (206). In their recent study, Vaidyula and Rao (207) provided evidence that in human platelets PLC-?2 plays a major role in responses to PAR1 and PAR4 activation, and that PLC-?2 is required for the sustained rise in [Ca2+]i concentration upon thrombin activation.
Putting all of this together, current data suggest that thrombin activates human platelets by cleaving and activating PAR1 and PAR4. In turn, PAR1 activates the members of Gq/11, G12/13, and perhaps Gi families, leading to the activation of PI3 kinase, PLC-?, and monomeric G proteins (Rho, Rac, and possibly RapI), and also causes an increase of cytosolic Ca2+ concentration and inhibition of cAMP formation.
Additionally, it is important to note that despite the fact that human platelets express both PAR1 and PAR4, these receptors appear to be activated by different thrombin concentrations. Cleavage of human PAR4 requires a higher concentration of thrombin than does cleavage of PAR1, and it is likely that PAR1 is the predominant signaling receptor at low thrombin concentrations (208, 209).
Mouse platelets provide an interesting contrast to human platelets: whereas human platelets express functional PAR1 and PAR4, mouse platelets express PAR3 and PAR4, although in the latter case signaling appears to be mediated entirely by PAR4, with PAR3 serving to facilitate PAR4 cleavage at low thrombin concentrations (for details, see Section IV.A) (39, 40).
2. Cells in the nervous system.
PAR1, as mentioned above, appears to affect various processes in both the central and peripheral nervous systems (45, 210). Among such processes are: neuroinflammation and neurodegeneration, neuritogenesis, astrocyte proliferation, and synaptic plasticity. However, here we would like to focus on data concerning the involvement of PAR1 in intracellular signaling cascades in neuronal cells.
PAR1 may have an effect on various intracellular signaling cascades within neuronal cells (171, 181, 189). By coupling to different G proteins, PAR1 affects a wide range of neuronal cell functions. For example, the necessity of PAR1 coupling to G12 was demonstrated for thrombin-stimulated DNA synthesis in 1321N1 astrocytoma cells (211). Interestingly, injection of antibodies directed against G12 abolished the thrombin-stimulated DNA synthesis (212). Additional studies performed on 1321N1 astrocytoma cells revealed that thrombin treatment causes a concentration-dependent rounding. One may speculate that such an effect of thrombin could be Rho-dependent. The Rho family of small GTPases is known to be involved in the control of cytoskeletal changes via modulation of actin polymerization. Indeed, this thrombin-induced rounding of 1321N1 astrocytoma cells was Rho-dependent and mediated via G12 (213).
Moreover, LaMorte et al. (214) demonstrated binding of PAR1 to Gq by using the same cell line. The authors also showed that thrombin stimulation induces Ras-GTP complex formation and that Ras is required for PAR1-mediated activation of PLC in 1321N1 astrocytoma cells (214). This is especially interesting because thrombin is known to be a factor regulating the proliferation of astrocytes. This effect of thrombin appears to be associated with Go/i- and Gq-mediated pathways. Wang and coworkers recently shed some light on the downstream cascade events of PAR1 signaling (174, 210). Gq-Mediated PLC activity results in Ca2+ mobilization and activation of PKC, which phosphorylates a nonreceptor tyrosine kinase, proline-rich tyrosine kinase (Pyk2). Interestingly, the activation of other nonreceptor tyrosine kinases like Src and focal adhesion kinase (FAK) was also demonstrated after thrombin stimulation. Pyk2 is a factor that is able to connect GPCRs to ERK1/2 activation. Therefore, Pyk2, acting together with Src-tyrosine kinase, causes the activation of the ERK/MAPK pathway, which mediates the proliferative effect of thrombin. Additionally, the same group of authors demonstrated that the Go/i-mediated PI3 kinase pathway is also involved in thrombin-induced astrocyte proliferation (174).
Thrombin is also known to exhibit both beneficial and unfavorable effects in hippocampal neurons and astrocytes (178, 180, 215). Donovan and colleagues (184, 216) found that tyrosine kinases, serine/threonine kinases, and the actin cytoskeleton are involved in both pro- and antiapoptotic effects of thrombin in neuronal cells. However, there was no involvement of Go/i and the PI3 kinase pathway observed in these thrombin effects.
Additionally, Zieger et al. (217) demonstrated the existence of a novel PAR1-associated signaling pathway in the nervous system. According to these data, PAR1 participates in cAMP-independent PKA activation in SNB-19 glioblastoma cells. Moreover, PAR1 stimulation causes the activation of transcriptional factor nuclear factor B (NFB) in this cell type (217).
In summary, besides morphological and proliferative effects described above that are mediated via G12/13, Gq, and Go/i, respectively, thrombin also affects activation of transcriptional factors such as activator protein-1 (AP-1) in neuronal cells (214).
3. SMC.
Thrombin exerts direct effects on vascular cells such as SMC and endothelial cells via interactions with PARs (218). As mentioned in Sections II.H.1 and II.H.2, PAR1 interacts with G12/13, Gq, and Gi to elicit diverse downstream signaling events. For example, thrombin-stimulated DNA synthesis and cell migration are associated with activation of the G13 signaling pathway in SMC. The G13 signaling cascade includes the activation of Rho and thus induction of cytoskeletal changes affecting cell migration (219). The Gq-dependent signaling pathway includes the activation of PLC, which in turn leads to MAPK phosphorylation and receptor tyrosine kinase transactivation, both necessary events in thrombin-mediated proliferation. It is interesting that the thrombin-mediated activation of MAPKs such as ERK1/2 was recently demonstrated in SMC. As noticed, PAR1 caused rapid phosphorylation, whereas PAR4 induced prolonged phosphorylation of these kinases in SMC (220). Additionally, Ghosh et al. (221) have shown that thrombin activates p38 MAPK in a time-dependent manner in VSMC. Furthermore, Sabri et al. (222) demonstrated that PAR1 agonists induce activation of Jun N-terminal kinase (JNK) and Akt/PKB in rat ventricular cardiomyocytes.
After dissociation of the G protein heterodimer, G? interactions activate phosphoinositide 3-kinase, which promotes [Ca2+]i release that is required for SMC growth in response to thrombin stimulation (223).
It is also interesting that PAR1 signaling can modulate gene transcription induced by cytokines in SMC. Thrombin, acting via PAR1, can block IL-6-induced signal transducer and activator of transcription 3/Sis-inducible factor-A (Stat3/SIF-A) activation (224).
In summary, since the discovery of PARs, considerable progress in our understanding of thrombin signaling in SMC was achieved. This allowed some light to be shed on signaling events associated with SMC proliferation, migration, and synthesis of extracellular matrix proteins (e.g., collagen) after thrombin stimulation. However, additional studies are necessary to reveal the role of thrombin in SMC apoptosis, vascular lesion formation, and wound-healing associated pathways.
4. Endothelial cells.
Thrombin is known to affect various functions of endothelial cells. Among these are cell rounding, changes of cell-cell junctions, proliferation, barrier function, and permeability. These thrombin-induced effects appear to be mediated via PARs, particularly PAR1 (225, 226). A major step forward was reached in recent studies, which revealed important intracellular signaling events underlying PAR1-mediated effects in endothelial cells (227).
PAR1 was demonstrated to interact with Go/i, Gq, and probably G12/13 in endothelial cells (228, 229). As well as it was demonstrated for neuronal cells, PAR1-mediated cytoskeletal changes in endothelial cells (cell rounding) are associated with the activation of RhoA (225). Additionally, in the same work it was demonstrated that PAR1-AP stimulation rapidly enhanced vascular permeability in a mouse skin assay (225). Moreover, it was found that binding of PAR1 to pertussis toxin (PTX)-sensitive G proteins (however only to Go, but not to Gi) is also necessary for thrombin-induced changes of endothelial barrier permeability (229). The activation of the MAPK signaling pathway by thrombin in endothelial cells was also demonstrated. This activation plays a crucial role in thrombin-induced effects of endothelial cell functions such as chemokine and cytokine production as well as the expression of cell adhesion molecules (89, 230, 231).
The role of thrombin and PAR1 in the activation of transcriptional factor networks was intensively investigated in recent publications. Malik and coauthors (228, 231, 232) studied the involvement of PAR1 in the activation of NFB in endothelial cells. Their studies showed that Gq and the G? dimer are responsible for NFB activation and intercellular adhesion molecule-1 (ICAM-1) transcription in endothelial cells induced by the PAR1 agonist thrombin and the PAR1-specific AP TFLLRNPNDK. Furthermore, transfection experiments strongly supported simultaneous activation of G?/PKC-/p38 and G?/PI3-kinase pathways that converge into the Akt pathway, leading to subsequent NFB activation and ICAM-1 expression (228, 231, 232). Moreover, thrombin-induced stimulation of vascular cell adhesion molecule-1 (VCAM-1) production involves the inducible binding of p65 NFB to a tandem NFB motif in the 5' flanking region (233). Taken together, these findings suggest that in the case of ICAM-1 and VCAM-1, p65 NFB is necessary for transducing the thrombin response in endothelial cells.
Additionally, thrombin-mediated induction of VCAM-1 was shown to involve the inducible binding of GATA-2 to a tandem GATA motif in the upstream promoter region (234). Interestingly, the effect of thrombin on GATA-2 DNA binding and transcriptional activity was found to be mediated by a PI3K, PKC--dependent signaling pathway (233).
It is important to note that thrombin effects on transcriptional factor networks have been more deeply investigated in endothelial cells than in other cell types. These data explain, at least in part, thrombin-mediated effects on the production of cell adhesion molecules and some chemokines in endothelial cells. Subsequently, this accounts for thrombin effects at leukocyte migration via endothelial barrier (ICAM-1 serves as a ligand for leukocyte ?2 integrins and promotes leukocyte adhesion) and leukocyte recruitment.
5. Immune cells.
The participation of PARs in T cell signaling pathways has also been demonstrated (235). Recently, Bar-Shavit et al. (236) have examined a possible involvement of Vav 1 in PAR-mediated signaling in human Jurkat T cells. The Vav family has three known members in mammalian cells (Vav, Vav2, and Vav3) and one in nematodes (CelVav) (237). Tyrosine phosphorylation of Vav1 regulates its activity as a guanine-nucleotide exchange factor for the Rho-like small GTPases RhoA, Rac1, and cdc 42, which affect cytoskeletal reorganization and activation of stress-activated protein kinases/JNKs. Bar-Shavit et al. (236) clearly showed that activation of PARs induces tyrosine phosphorylation of Vav1 in Jurkat T-leukemic cells. Because -chain-associated protein kinase of 70 kDa (ZAP-70) and SH2-domain containing leukocyte-specific phosphoprotein of 76 kDa (SLP-76) have been shown to associate physically with Vav1 and because this association is critical for normal functioning of T cells, it was tempting to investigate whether ZAP-70 and SLP-76 are involved in PAR-induced signaling cascades. Indeed, an increase of tyrosine phosphorylation of ZAP-70 and SLP-76 was observed after activation of Jurkat cells with PAR-APs. Moreover, phosphorylation of Vav1 in response to PAR stimulation was dependent on p56lck (236). Unfortunately, because nonselective PAR-APs were used in the studies done with the Jurkat cells (236), it is difficult to distinguish between the effects of PAR1 and PAR2 in this T cell model system. Furthermore, because of the lack of the ability of PAR3-derived peptides to activate PAR3 (39), the role of PAR3 in T cell signaling remains unknown.
6. Factor Xa signaling mediated via PAR1.
Most of the PAR1-associated signaling events have been observed subsequent to cell stimulation by thrombin. However, recently a possible role of PAR1 in coagulation factor Xa-mediated signaling has been demonstrated (238, 239, 240). The coagulation proteinase factor Xa is generated at sites of vascular injury and inflammation after formation of the tissue factor/VIIa (TF/VIIa) complex. Coagulation factor Xa has been shown to be mitogenic for SMC (238) and elicits inflammatory responses in endothelial cells. Riewald and Ruf (241) have presented new data indicating the involvement of PAR1 in Xa-mediated signaling. They used HeLa cells expressing only PAR1 and demonstrated that factor Xa induces NFB activation and MAPK phosphorylation in these cells. Inhibition studies with specific antibodies revealed that factor Xa responses were mediated via PAR1 (241). Thus, factor Xa (or potentially other serine proteinases) may substitute for thrombin in proteolytic signaling via PAR activation. In endothelial cells, which also express PAR2, Camerer et al. (240) have shown that factor Xa signaling was mediated not only by PAR1, but also to a large extent via PAR2. Together, PAR-1 and PAR-2 appear to account for more than 90% of factor Xa signaling in endothelial cells (240).
In summary, the multiple biological as well as inflammatory and immune responses of thrombin that are mediated by PAR1 [including 1) vasoregulation; 2) increased vascular permeability; 3) cellular adhesion and infiltration of leukocytes; 4) angiogenesis; 5) stimulation of the production of inflammatory mediators such as cytokines, neuropeptides, NO and prostanoids, for example; 6) regulation of extracellular matrix proteins; and 7) induction of signal transduction pathways which are involved in immunomodulation] suggest an important role of PAR1 during inflammation and immune response.
III. PAR2 in Inflammation and Immune Response
Top
Abstract
I. Introduction
II. PAR1 in Inflammation...
III. PAR2 in Inflammation...
IV. PAR3 and PAR4
V. Conclusions
References
PAR2 is expressed in brain, lia (DRG), eye, airway, heart, GI tract, pancreas, kidney, liver, prostate, ovary, testes, and skin (21, 22, 24, 154, 242, 243, 244, 245) and is found in various cell types such as epithelial cells, endothelial cells, SMC, osteoblasts, as well as immune cells such as T cells, neutrophils, mast cells, or eosinophils (97, 154, 246, 247, 248, 249, 250, 251, 252, 253, 254). On the other hand, platelets do not express PAR2 (248). Recent findings point to an important role for PAR2 under physiological and pathophysiological conditions in many tissues (Table 2). However, the endogenous enzymes responsible for activating PAR2 in many tissues remain to be determined. Many endogenous or exogenous trypsin-like enzymes may cleave and activate PAR2. Expression of trypsinogen-2 mRNA and its translation product has been demonstrated in endothelial cells (255). Interestingly, various types of human cancer cells secrete enzymes with trypsin-like specificity (255) that may activate PAR2. In human skin, trypsinogen-4 generated by keratinocytes and trypsinogen-2 from human dermal microvascular endothelial cells can activate PAR2 in vitro (M. Steinhoff, unpublished observation). Another candidate is mast cell tryptase, a major secretory protein of human mast cells. Mast cells from mice and rats are more heterogeneous regarding their protease content, although they also produce proteases with tryptic specificity. Tryptase has been shown to activate PAR2 on epithelial as well as endothelial cells and neurons (23, 196, 243, 251, 256). The observation that tryptase can activate PAR2 suggests a role of this receptor in humans under circumstances when mast cells are involved, e.g., during inflammation, hypersensitivity reactions, and wound repair (24, 257). However, the ability of human tryptase to activate PAR2 appears to be restricted by receptor glycosylation at an N-terminal residue just proximal to the receptor’s cleavage activation site (258, 259). Notwithstanding, the effects of tryptase on cells in vitro often mimic those of PAR2 activation: tryptase up-regulates IL-1? and IL-8 secretion, enhances the presence of intracellular adhesion molecules/selectins on endothelial cells, mediates accumulation of neutrophils and eosinophils, produces vascular leakage, and is mitogenic for epithelial cells, fibroblasts, and SMC (260, 261, 262, 263). That said, direct proof that PAR2 mediates the effects of tryptase during inflammation in vivo is still lacking.
A. Vasculature
In the vasculature, PAR2 has many effects that are proinflammatory. Agonists of PAR2 induce relaxation in the rings of rat aorta or porcine coronary artery dependent on endothelial NO synthase activity (246, 248, 250). This effect is abolished in the absence of endothelium (246, 250). In contrast, trypsin stimulates contraction of the rabbit aorta in the absence of endothelium (264). In the intact rat, iv injection of SLIGRL-NH2 produces a marked fall in blood pressure, consistent with release of NO from endothelial cells (248). Furthermore, PAR2 agonists increase IL-6 production (72), induce von Willebrand factor release, and serve as a mitogen for human umbilical vein endothelial cells (HUVEC) (97, 265, 266).
Moreover, it was demonstrated that some inflammatory mediators are able to affect the expression of PAR2 in endothelial cells. In HUVEC, PAR2 mRNA is up-regulated by TNF- and IL-1, cytokines that act together to orchestrate the acute inflammatory response (25, 265).
In summary, PAR2 mediates vasodilation, plasma protein extravasation, as well as endothelial cell proliferation in the cardiovascular system. Thus, this receptor can be regarded as an important factor in neovascularization. Moreover, PAR2 can be considered as a vascular sensor for trypsin-like proteinases of the coagulation cascade, which play an important role in cardiovascular medicine.
B. Immune cells
Although information has been acquired about the role of PAR2 in epithelial and endothelial function, relatively little is known so far about the role of this receptor in the immune system. As was recently demonstrated, PAR2 is expressed by various cells involved in immune response, such as T cell lines, eosinophils, neutrophils, and mast cells (147, 160, 249, 252, 267). Accumulating evidence points to a role of PAR2 in the regulation of leukocyte function. Some of the PAR2- mediated effects on leukocyte behavior have been observed in vivo in rodents (262). In this model, it has been reported that the PAR2-AP (SLIGRL-NH2) caused a significant increase in leukocyte migration into the peritoneal cavity after ip injection. Furthermore, PAR2-APs induced a significant increase in leukocyte rolling and adherence by a mechanism depending on the release of platelet-activating factor (262). Lindner et al. (268) have also demonstrated the importance of PAR2 for leukocyte adhesion and rolling. In a model of acute inflammation induced by tissue trauma, they showed that PAR2-deficient mice have a decreased leukocyte rolling capacity compared with wild-type mice. Conversely, activation of PAR2 by a specific PAR2-AP in wild-type mice induced a significant reduction of leukocyte rolling velocity, an increased leukocyte rolling flux as well as increased leukocyte adhesion (268).
Howells et al. (252) reported that a specific PAR2-AP as well as trypsin are able to activate PAR2 on human neutrophils. PAR2 activation caused shape changes and up-regulation of CD11b/CD18 in these cells. In a coculture of human neutrophils with endothelial cells, PAR2 agonists induced L-selectin shedding and CD11b/CD18 up-regulation in neutrophils (268). We demonstrated that PAR2 agonists induce [Ca2+]i release in human neutrophils and enhance neutrophil motility in 3-D collagen gel lattices (269). Moreover, Lourbakos et al. (27) demonstrated that a bacterial proteinase, gingipain-R from P. gingivalis, activates PAR2 on human neutrophils. Because many bacteria with pathogenicity in humans produce serine proteinases with trypsin-like activity, it can be assumed that certain pathogenic effects of bacteria in many tissues can be mediated through PAR2. For example, Lourbakos et al. (28) reported that activation of PAR2 by bacterial gingipain R leads to IL-6 secretion in human oral epithelial cells. In mouse airways, PAR2 agonists were capable of inhibiting the recruitment of neutrophils induced by bacterial proteinase (270), suggesting a direct role of PAR2 on neutrophil function in vivo. From these data, one may speculate that PAR2 may serve as a receptor for serine proteinases derived from bacteria, viruses, or even fungi and may thus directly trigger inflammatory or immune responses in vivo induced by microorganisms. However, this intriguing hypothesis remains to be verified.
Recent work has shed some light on the role of PAR2 on eosinophil function. Miike et al. (267) clearly demonstrated that human eosinophils express functional PAR2. Furthermore, trypsin as well as a specific PAR2-AP was able to induce degranulation and superoxide production in human eosinophils. More recently, Temkin et al. (79) reported that human mast cell tryptase induced up-regulation of IL-8 mRNA and caused IL-8 release from human eosinophils. A comparable effect of tryptase on eosinophil IL-6 expression and release was also observed. This effect of tryptase appears to be activated via the protein kinase/AP-1 signaling pathway (79). Schmidlin et al. (271) have also demonstrated that PAR2 mediates infiltration of eosinophils and hyperreactivity in allergic inflammation of the airway. Furthermore, PAR2 activation releases survival factors such as granulocyte-macrophage colony-stimulating factor (GM-CSF) from eosinophils (272). Together, these results are clearly in favor of an important role of PAR2 in eosinophil regulation and in inflammatory/allergic diseases in which eosinophils are involved.
Several T cell lines also express PAR2 (244). Trypsin as well as a synthetic PAR2-AP induce [Ca2+]i mobilization in the Jurkat and HPB.ALL T cell lines (249). In Jurkat cells, the involvement of PARs, possibly PAR2, in Vav1-mediated signaling was clearly demonstrated (236).
The localization of PAR1 and PAR2 on human mast cells was first reported by D’Andrea et al. (147). As shown by immunohistochemistry, PAR2 was localized not only on the cell membrane of mast cells, but also on the membrane of the intracellular tryptase-positive granules (147). Moreover, rat peritoneal mast cells also express PAR2 mRNA (273). However, the precise role of PAR2 in mast cells especially with regard to a potential autocrine regulatory mechanism via tryptase activation requires additional investigation. In summary, these results provide evidence for a role of serine proteinases in directly regulating immune cells and immune responses via PARs. An understanding of the molecular mechanisms by which PAR2 is involved in immune regulation will aid the design of new antiinflammatory drugs that can target proteinase-activated receptors.
C. Arthritis
Arthritis is a chronic inflammatory disease that is characterized by joint swelling, vasodilatation, edema, hyperemia, pain, and recruitment of inflammatory cells, especially neutrophils and macrophages, but also lymphocytes. Very recently, it was demonstrated that PAR2 participates in the pathophysiology of arthritis (274). Using PAR2-deficient mice, the authors showed that PAR2 deficiency results in a marked reduction of swelling responses of experimentally induced monoarthritis. However, no differences were observed concerning joint damage when compared with wild-type mice. Using an enzymatic approach (?-galactosidase), they also showed positive ?-gal staining in endothelial cells of blood vessels. Two weeks after induction of monoarthritis, however, a significant increase of ?-gal staining of extravascular inflammatory cells was observed in PAR2+/+ mice compared with PAR2–/– mice. Examination of the knee joint showed dramatic arthritic changes in PAR2+/+ mice characterized by synovial hyperplasia, an enhanced inflammatory infiltrate, and cartilage damage. Thirty days after the induction of monoarthritis, the cartilage was completely replaced by pannus in PAR2+/+ mice. In contrast, PAR2–/– mice showed an intact appearance of the cartilage similar to normal and control joints. This is in favor of an important role of PAR2 as a mediator of proinflammatory responses in the joints. Moreover, a novel synthetic PAR2 peptide agonist (ASKH95, phenylacetyl-LIGKV-OH) revealed proinflammatory effects after intraarticular injection like synovial hyperemia and joint swelling. These results clearly demonstrate that ASKH95 induces signs of chronic inflammation such as long-lasting swelling responses, vasodilatation, and tissue destruction. Interestingly, ASKH95 showed a longer-lasting proinflammatory effect in the tissue compared with the classical SLIGRL PAR2 agonist peptide. This may be due to different rates of degradation, longer receptor activation by ASKH95, or most likely different lipophilicity. The underlying mechanisms of the PAR2-mediated inflammatory effect in this chronic disease, however, are not fully explored yet. Potential mediators may be cytokines or prostanoids, for example. Thus, PAR2 may play an essential role in the pathophysiology of arthritis, and PAR2 antagonists may be beneficial for the treatment of this chronic inflammatory disease.
D. Skin
PAR2 may play an important role in cutaneous inflammation. Keratinocytes and dermal endothelial cells express functional PAR2 and are important targets of inflammatory mediators (247, 275, 276). In humans, PAR2 is also expressed by dermal immunocompetent cells, which so far have not been characterized (24), and by dermal sensory nerves (277). Potential endogenous activators of PAR2 in human skin are mast cell tryptase, proteinases of the trypsin-family produced by keratinocytes (trypsinogen-4), and proteinases of the fibrinolysis cascade, such as factor VII/Xa, which could be released during inflammation and wound healing. Exogenous activators of PAR2 may be serine proteinases generated by bacteria, fungi, and house dust mites, although direct evidence on keratinocytes is still lacking. Indirect evidence for bacterial-induced activation of PAR2 is coming from the observation that bacterial gingipain activates the receptor on buccal keratinocytes (28). At certain stages of atopic dermatitis or psoriasis, mast cells are in close contact to basal keratinocytes or are recruited into the epidermal layer (278). Numerous mast cells are found in the dermis in close proximity to blood vessels and at the dermal-epidermal border in atopic dermatitis and psoriasis. Thus, tryptase or other skin-derived proteinases may activate PAR2 to induce inflammatory changes and thereby contribute to the pathophysiology of atopic dermatitis and psoriasis. It is well known that intradermal injection of tryptase into the skin results in vasodilatation and erythema, followed by leukocyte infiltration and local induration (279), indicating a role of tryptase in cutaneous inflammation. Moreover, tryptase is an important long-term marker in body fluids for systemic anaphylaxis and other immediate hypersensitivity allergic reactions (280). Thus, it is possible that PAR2 mediates the systemic effects of tryptase in certain skin diseases.
Serine proteinases other than tryptase may also activate PAR2. Cytotoxic T cells, which express PAR2 (249), also produce the trypsin-like enzyme granzyme A (281), suggesting a potential role for PAR2 in regulating T cells in inflammatory dermatoses. In summary, proteinases in the skin are signaling molecules that may directly regulate the function of target cells by activating PAR2. PAR2 as well as PAR1 agonists regulate cytokine production such as IL-8 (282), IL-6, and GM-CSF (283) in cultured human keratinocytes. Furthermore, PAR2 may also regulate proliferation and differentiation in the skin. Similar to TGF-?, PAR2 agonists inhibit proliferation or differentiation of human neonatal keratinocytes, whereas PAR1 agonists stimulate proliferation (275). In contrast, agonists of both PAR2 and PAR1 are mitogenic for endothelial cells (248). Thus, our observations in normal and inflamed human skin suggest a functional role of PAR2 in different cell types of human skin during inflammation.
In human skin, PAR2 is only weakly expressed by microvascular endothelial cells. Studies done in vitro revealed that PAR2 is functionally expressed by human dermal microvascular endothelial cells and that PAR2 agonists induce up-regulation and release of IL-6 and IL-8 in a concentration-dependent fashion (263). Moreover, PAR2 regulates expression of cell adhesion molecules in primary cultures of human endothelial cells (284) and cell lines (268). These findings are supported by in vivo findings demonstrating a role of PAR2 in the adhesion, rolling, and extravasation of leukocytes in rat venules. Finally, PAR2 appears to play a role in the regulation of leukocyte/endothelial interactions in humans and mice in vivo, as shown for atopic dermatitis or experimentally induced contact dermatitis (285). Together, these results strongly support the idea that PAR2 plays an important role in leukocyte/endothelial interactions during inflammation and immune response.
The knowledge about the signaling cascades involved in PAR2-induced inflammatory and immune responses by keratinocytes is still incomplete. Recently, Kanke et al. (286) have shown that PAR2 agonists stimulate activation of inhibitory B kinases (IKK) and IKK?, and also stimulate NFB-DNA binding. Our own data using primary human keratinocytes also clearly demonstrate that PAR2 activates NFB (287). Moreover, PAR2 is directly involved in the NFB-mediated regulation of keratinocyte ICAM-1, which is an important molecule for the regulation of keratinocyte immune responses such as host defense, allergy, and recruitment of inflammatory cells into the epidermis (287).
Finally, recent data suggest a role of PAR2 during cutaneous inflammation in vivo. In a murine model of experimentally induced allergic contact dermatitis, a proinflammatory role of PAR2 agonists was demonstrated (285, 288). Stimulation of PAR2 resulted in increased ear swelling responses with a maximum after 48 h. Independently, our group confirmed the proinflammatory effects of PAR2 agonists in a model of experimentally induced allergic and irritant contact dermatitis using PAR2-deficient mice. Moreover, we examined the underlying mechanisms responsible for PAR2-induced effects during contact dermatitis (285). These results clearly indicate that PAR2 is involved in edema formation, plasma extravasation, up-regulation of cytokines (IL-6), cell adhesion molecules (ICAM-1), and selectins (E-selectin) in mice. Intravital microscopy studies further showed that velocity and adhesion of leukocytes is impaired in PAR2-deficient mice compared with controls. Moreover, in vivo studies by microdialysis confirmed a role of PAR2 in mediating vascular responses such as edema formation and plasma extravasation also in human skin. Finally, ex vivo studies in skin biopsies of human volunteers are in favor of a role for PAR2 in selectin regulation of dermal blood vessels because PAR2 agonists induced a marked increase of E- selectin immunoreactivity in dermal endothelial cells in comparison to control tissues. Together, these results support a regulatory role of PAR2 during cutaneous inflammation and immune response.
E. Airways
Several observations are in favor of an important role of PAR2 in airway inflammation. Firstly, tryptase (251, 289) seems to play an important role in airway inflammation and hyperresponsiveness (253). Secondly, PAR2 is markedly up-regulated after exposure to proinflammatory stimuli or cytokines (265) that have been shown to play a role in chronic airway diseases. Morphological studies have demonstrated PAR2 immunoreactivity in endothelium and smooth muscle of bronchial vessels. These observations are consistent with a similar distribution in other tissues and with the known ability of PAR2 agonists to cause arterial relaxation as well as contraction in certain arteries (290). Moreover, PAR2 is expressed by bronchial, tracheal, epithelial and SMC which may result in direct or indirect bronchomotor effects leading to pathophysiological conditions in the airways such as asthma or chronic obstructive pulmonary disease. Interestingly, PAR2 agonists are capable of causing bronchoconstriction (291). Whether this response is dependent on secondary effects such as the generation of bronchoactive peptides or other mediators remains to be determined. However, studies done in vitro suggest further that PAR2 agonists also produce a relaxant effect in isolated main bronchi (292). Thus, PAR2 agonists may cause dose-dependent bronchoconstrictor or bronchodilatator effects in the airways, with high agonist concentrations favoring bronchoconstriction. NO and prostanoids like PGE2 may be involved in this process, because potentiation of PAR2-induced bronchoconstriction has been observed upon inhibiting NO synthase or after prostanoid generation and COX-2 activation (153). In conclusion, there is strong evidence that PAR2 stimulation activates contractile and dilator mechanisms in the airways. The reason why the predominant effect in guinea pig airways in vivo is bronchoconstriction, however, is not known at present. In addition to the potentially bronchoprotective effects observed by Cocks et al. (292) in rat bronchial preparations, others have shown that PAR2 may also mediate bronchoprotection in guinea pig airways (293). However, other observations show that PAR2 can act as an enhancer of histamine-mediated contraction (294). The study of Cocks et al. (292) showed that trypsin colocalized with PAR2 in airway epithelial cells, where PAR2 activation resulted in relaxation of airway preparations by the release of cytoprotective PGE2. Prostanoids may be important mediators of PAR2 activity in the airways not only in the lung, but also in the GI tract (152, 153, 253, 295). It is tempting to speculate that tryptase and/or other proteinases with trypsin-like activity, released in close proximity to the cells expressing PAR2, may stimulate this receptor during airway inflammation. Finally, intensive staining of PAR2 in the apical region of the airway epithelium might suggest an activating role of proteinases, for instance bacterial proteinases, present in the airway lumen (296). Very recently, PAR2 was shown to stimulate the release of MMP-9 in a human epithelial cell line (297), providing a strong hint that PAR2 is involved in the orchestration and reorganization of lung extracellular matrix proteins.
It was also revealed that PAR2 stimulation of airway epithelial cells leads to the release of eosinophil survival-promoting factors such as GM-CSF (272). Moreover, in human respiratory epithelial cells, PAR2, like PAR1, stimulates IL-6, IL-8, and PGE2 release (77).
Profound evidence for an involvement of PAR2 in allergic inflammation of the airways is provided by the recent work of Schmidlin et al. (271). To study the involvement of PAR2 in airway inflammation, they used PAR2-deficient mice and mice overexpressing human PAR2 (PAR2tg). PAR2 is known to be up-regulated by inflammatory agents. Sensitization of wild-type mice by injection of ovalbumin led to infiltration of immune cells into the lumen of the airways and induced hyperreactivity, whereas both infiltration of eosinophils into the lumen and airway hyperreactivity was exacerbated in PAR2tg mice. In contrast, infiltration of immune cells and airway hyperreactivity was markedly diminished in PAR2-deficient mice. Additionally, in PAR2–/– mice the IgE response was strongly reduced, implicating a role of PAR2 in immune response. Remarkably, intranasal administration of PAR2-AP in wild-type mice stimulated increased recruitment of macrophages, confirming the proinflammatory role of PAR2 in the airways, whereas altered PAR2 expression mainly influenced eosinophil infiltration.
In summary, evidence exists for a proinflammatory as well as antiinflammatory role of PAR2 in airway inflammation. This dual role may be explained by the various tissues, cells, methodological approaches, and inflammatory model systems (acute vs. chronic, mouse strains) that have been used to evaluate this role of PAR2 in the airways. So far, it appears that PAR2 is a sensor receptor that releases proinflammatory mediators in the early phase and antiinflammatory molecules in the late phase of "regulated inflammation." Under "dysregulated" inflammatory situations such as chronic disease states, PAR2 may exert antiinflammatory effects, as shown in a colitis model (298), or proinflammatory effects, as demonstrated in an asthma model (271). Thus, additional studies using human and mouse (knockout) models of pulmonary dysfunction are necessary to fully explain the role of PAR2 during inflammation and immune response.
F. Brain and peripheral nervous system
By immunohistochemistry, PAR2 has been localized in various compartments of the nervous system such as brain, spinal cord, DRG, and peripheral nerves, with different receptor densities found during various stages of mouse development (299). At d 14, PAR2-staining was observed throughout the mouse brain. Among the hippocampal formation, PAR2 was localized in the subiculum, in pyramidal cells throughout the CA1, CA2, CA3, and hilus region, and in granular cells of the dentate area. Additionally, staining for PAR2 was observed in all cortical layers, amygdaloid nuclei, striatum, thalamus, and hypothalamus (175). Peripheral nerves were intensively stained after d 14. Similarly, PAR2 has also been detected in human brain (276). In rat brain, PAR2 is also expressed by neurons of the hippocampus (175, 300), meningeal cells (301), as well as astrocytes (176, 185). PAR2 is abundantly expressed in various subareas and is transiently up-regulated during oxygen and glucose deprivation (175). Moreover, PAR2 activation is cytotoxic for isolated rat hippocampal neurons in a concentration-dependent manner (300).
Domotor et al. (302) recently reported that agonists of PAR2 induce Ca2+ responses in brain microvascular endothelial cells. This effect may be modulated by elastase and plasmin, which regulate PAR2 signaling. Studies of the actions of PAR2 agonists on CNS targets suggest an important role of PAR2 during injury, growth, apoptosis, and probably memory.
Very recently, PAR2 has been localized on rat sensory neurons (243). Moreover, functional data strongly support the idea that the peripheral nervous system is directly regulated by PAR2 during neurogenic inflammation. Neuropeptides such as calcitonin gene-related peptide and substance P (SP) from primary spinal afferent neurons are known as important mediators of neurogenic inflammation in many organs. Because of the close proximity of tryptase-containing mast cells to spinal afferent fibers, and because agonists of PAR2 cause effects similar to those of tryptase in many tissues, comprising many of the characteristics of neurogenic inflammation, it was speculated that serine proteinases or other PAR2 agonists may activate PAR2 on sensory neurons to stimulate neuropeptide release and may thus regulate inflammation. Indeed, it was shown that agonists of PAR2 can induce inflammation by a neurogenic mechanism that depends on the release of calcitonin-gene related peptide (CGRP) and SP from primary spinal afferent neurons. However, there also appears to be a nonneurogenic component of PAR2-induced inflammation because granulocyte infiltration triggered by PAR2 in a paw edema inflammation model was unaffected by sensory denervation or by neuropeptide receptor antagonists. Thus, PAR2 agonists may also act directly on endothelial cells or on neutrophils themselves to stimulate inflammation-related granulocyte adhesion and infiltration (14, 250). It is possible that serine proteinases may activate PAR2 on spinal afferent neurons under pathophysiological inflammatory circumstances. In a similar manner, serine proteinases may regulate enteric neuronal function by cleaving PAR2 in a Ca2+-dependent manner (196). Trypsin, which is released by airway epithelial cells, but not the PAR2 agonist peptide sequence SLIGRL, induced a marked contraction in guinea pig bronchi via SP release. Thus, non-PAR mediated effects of trypsin may also regulate neuropeptide function in the lung (303).
Several reports strongly suggest a role of PAR2 in the central transmission of pain and nociception in rats and mice (304, 305, 306). Intraplantar administration of low doses of PAR2 agonists, which do not induce inflammation, resulted in a long-lasting hyperalgesia. Interestingly, these effects of PAR1 agonists were more efficacious in comparison with other inducers of hyperalgesia such as PGE2 (0.3 μg). In the spinal cord, PAR2 agonists induced an up-regulation of c-fos, a marker of activated nociceptive neurons (305, 307, 308). Another study also reported a role of PAR2 in a rat visceral pain model (308). PAR2 agonists administered in vivo clearly increased abdominal colonic contractions. This effect was inhibited by an NK1 receptor antagonist, but not by the PG inhibitor indomethacin (308). However, additional studies are necessary to fully explain the underlying direct or indirect effects of PAR2-induced nociception. For example, neuropeptides released from neurons upon PAR2 stimulation may activate the release of nociceptive mast cell mediators such as kinins or prostanoids (309).
The observation that PAR2 agonists are involved in the transmission of neurogenic inflammation and pain subsequently draws the question whether PAR2 may be involved in the central transmission of pruritus. Itching is one of the most frequent symptoms in dermatological diseases and accompanies inflammatory and immune responses of many diseases such as hypersensitivity reactions or urticaria, for example. Indeed, neuronal PAR2 appears to be involved in the induction of pruritus in human skin (277). Moreover, the endogenous PAR2 agonist tryptase was increased up to 4-fold in atopic dermatitis patients, and PAR2 expression was markedly enhanced on primary afferent nerve fibers in skin biopsies of atopic dermatitis patients. Intracutaneous injection of specific PAR2 agonists provoked enhanced and prolonged itch when applied intralesionally. Thus, PAR2 activation on cutaneous sensory nerves may be a novel pathway for the transmission of itch and inflammatory responses during atopic dermatitis and probably other skin diseases. PAR2 antagonists may be promising therapeutic targets for the treatment of cutaneous neurogenic inflammation and pruritus (277).
In summary, PAR2 may play an important regulatory role in the central and peripheral nervous system in normal and disease states, including neurogenic inflammation.
G. Digestive tract and pancreas
In rats, Kawabata et al. (167, 310, 311) detected PAR2 mRNA in the sublingual, submaxillary, and parotid salivary glands. Of these three distinct salivary glands, the sublingual exhibited the strongest responses to PAR2 activation resulting in the secretion of mucin. This response appeared to be mediated in part via a tyrosine kinase signal pathway, because genistein, an inhibitor of several tyrosine and other protein kinases, attenuated the effect. Pretreatment with the alkaloid capsaicin, which activates the transient receptor potential of vanilloid type 1 and which, in turn, is known to play a major role in inflammatory thermal nociception, failed to abrogate the secretion of mucin from salivary glands but abolished the PAR2-triggered cytoprotective secretion of mucus in the stomach. These results provide both a sensory neuron-independent and -dependent mechanism for PAR2-regulated exocrine secretion (312). Additional support for a role of PAR2 in salivary gland secretion was provided by data showing that a PAR2 agonist can stimulate amylase secretion from rat parotid gland slices in vitro (311).
In the GI tract, PAR2 is highly expressed by enterocytes, where it is localized on the apical and basolateral membranes (253, 313). Moreover, myocytes of the muscularis mucosae and muscularis externa as well as neuronal elements are immunoreactive for PAR2. Recent observations indicate that trypsin may regulate enterocytes by cleaving and triggering PAR2 at the apical membrane (253) to induce the generation of PGE2 and PGF1a, suggesting that PAR2 may have a cytoprotective as well as an inflammatory effect on the GI tract. Moreover, mucosal mast cell proteinases may possibly activate PAR2 on enterocytes and colonic myocytes. PAR2-mediated inhibition of intestinal motility may contribute to inflammatory conditions in which mast cells are involved.
Trypsin may activate PAR2 in the pancreas itself under physiological and pathophysiological conditions because trypsin can be prematurely activated in the inflamed pancreas (314). PAR2 seems to play an important role also in pancreatic nociception because injection of an AP specific for PAR2 can activate and sensitize pancreas-specific afferent neurons in vivo (314). Moreover, Hoogerwerf et al. (314) found that stimulating DRG with a PAR2-AP resulted in enhanced capsaicin- and KCl-stimulated release of calcitonin gene-related peptide, which is a marker for nociceptive signaling.
PAR2 is expressed by both acinar cells, which release digestive enzymes, and duct cells, which produce fluid and bicarbonate. In isolated pancreatic acini, trypsin and PAR2 agonists stimulate amylase release (154) (N. W. Bunnett, personal communication). In vivo studies revealed the secretion of pancreatic juice after PAR2 activation (167). Interestingly, PAR2 agonists applied to the basolateral but not apical membrane of monolayers of pancreatic duct cells increase short circuit currents due to activation of Ca2+-sensitive Cl– and K+ channels (315). These effects may be of relevance in pancreatitis when trypsin is released across the basolateral membrane.
H. Signaling by proteinases via PAR2
There are relatively few studies examining the involvement of PAR2 in signaling cascades compared with the relatively large number of studies concerning PAR1-mediated intracellular signaling (13, 14). So far, there are only indirect data indicating that PAR2 interacts with Gq/G11 (activation of calcium signaling) and possibly with G0/Gi (316). Whether or not PAR2 binds to other G proteins, such as G12 or G13, remains unclear.
Interaction of PAR2 with Gi and Gq suggests a subsequent activation of PLC, PKC, and MAPK pathways. These signaling pathways can affect various cell activities including cell proliferation, morphological changes, motility and survival, and gene transcription regulation (Table 2 and Fig. 3). Indeed, as demonstrated for neuronal cells and SMC, stimulation by tryptase as well as PAR2-AP leads to subsequent activation of PLC and PKC (317, 318). Kanke et al. (286) demonstrated that PAR2 agonists (trypsin as well as the AP SLIGKV-NH2) stimulate JNK and p38 MAPK activation in a human keratinocyte cell line (NCTC2544). Moreover, Vouret-Craviari et al. (225) described PAR2-induced activation of RhoA in HUVEC revealing a possible mechanism underlying PAR2-mediated effects at cytoskeleton in macrovascular endothelial cells. Together, these data shed some light on the potential signaling mechanisms underlying PAR2-associated cellular effects such as depolarization response in neuronal cells, proliferation and mitogenic response in SMC, proliferation and differentiation of keratinocytes, and cytoskeletal changes in endothelial cells (225, 275, 319, 320).
View larger version (47K):
[in this window]
[in a new window]
FIG. 3. Diagram showing the major G protein-mediated signaling pathways coupled to PAR2. Dashed lines or circles represent signaling pathways or intermediates that are not fully revealed to be activated by PAR2, but are in favor of other G protein-coupled receptors or are typical intermediate-accompanying molecules. Drawing represents a composition of signaling events of different tissues, cell types, and species (also see Table 2).
Additionally, the effects of trypsin as well as PAR2-AP on the activation of nuclear transcriptional factors have recently been demonstrated. It was shown that PAR2 agonists stimulate NFB-DNA binding activity and activation of upstream kinases IKK and IKK? (286). The effects of PAR2 stimulation on NFB or IKK also have been described in other studies performed on different cell types (13, 263, 287, 297, 321).
It is also important to pay attention to factor Xa signaling mediated via PAR2. It was mentioned in Section II.H.6 that this factor induces signaling events via PAR1; the potential role of PAR2 in such signaling remained unclear for a long time. Recently, however, the involvement of PAR2 activation in coagulation factor Xa signaling has been reported. In HUVEC, factor Xa interacts with a protein called effector cell proteinase receptor-1 (EPR-1). This interaction is associated with signal transduction, generation of intracellular second messengers, and modulation of cytokine gene expression. Inhibitors of factor Xa blocked these responses (322). Additionally, direct cleavage of PAR2 by factor Xa has been demonstrated, reflecting the complexity of PAR2-induced signaling. These data allowed the authors to suggest that factor Xa induces endothelial cell activation via a novel cascade of receptor activation involving docking to EPR-1 and proteolytic cleavage of PAR2 (322).
As already mentioned extensively in Section II.H, PAR1 and PAR2 both account for around 90% of endothelial factor Xa-mediated signaling in mice. In contrast, PAR1 virtually accounts for all factor Xa-induced activation in fibroblasts, indicating potential cell-specific synergistic effects of PAR receptors on target cells (25).
Moreover, Koo et al. (323) demonstrated an involvement of PAR2 in factor Xa signaling by analyzing factor Xa-induced stimulation of coronary artery SMC using cell proliferation and ERK1/2 activation as indices of response. Furthermore, factor Xa-induced ERK1/2 activation was not desensitized by preincubation of the cells with thrombin. However, ERK1/2 activation was markedly attenuated by prior exposure of the cells to PAR2-AP (SLIGKV-NH2). The mitogenic effect of factor Xa was significantly reduced in the presence of an anti-PAR2 monoclonal antibody that attenuates receptor activation, demonstrating the specificity of these effects (323). Together, these observations suggest that various signaling cascades are involved in PAR2-mediated signaling. Clearly, we are just beginning to understand the variety of PAR2-induced cell signaling pathways regulated under physiological conditions and during disease.
IV. PAR3 and PAR4
Top
Abstract
I. Introduction
II. PAR1 in Inflammation...
III. PAR2 in Inflammation...
IV. PAR3 and PAR4
V. Conclusions
References
A. Biology and distribution of PAR3 and PAR4
As already mentioned, murine platelets do not express PAR1. The observation that thrombin was still capable of inducing Ca responses in these cells led to the identification of PAR3 (18). In humans, PAR3 is expressed in bone marrow, heart, brain, placenta, liver, pancreas, thymus, small intestine, stomach, lymph nodes, and trachea, although the cell types remain to be identified. In mouse, PAR3 is expressed by megakaryocytes and platelets, among other cell types. The receptor is necessary for normal thrombin signaling in mouse platelets because blocking of the hirudin-like domain of PAR3 using a specific antibody prevented mouse platelet activation by low but not by high concentrations of thrombin. The same result was observed with PAR3-deficient platelets (19, 324). This also shows that PAR3 is a high-affinity thrombin receptor in mouse that is activated by proteolytic cleavage. Interestingly, murine PAR3 (mPAR3) itself does not lead to thrombin signaling even when overexpressed, indicating that the receptor has lost its ability to function autonomously during evolution (39, 40). Knocking out the PAR3 gene in mice leads to protection of the animals against thrombosis but has a relatively mild effect on hemostasis (325) (Table 2). However, analysis of PAR3 expression in human platelets showed that the receptor is not produced or is hardly produced by these cells (128). This suggests that in humans PAR3 does not play a major role for platelet activation in contrast to the mouse system.
Antibodies specific for PAR1 inhibited human platelet activation by low but not by high concentrations of thrombin (125, 326). In mice, PAR3 is necessary for normal thrombin signaling in platelets. This indicates the presence of more than one thrombin receptor on the surface of these cells. Indeed, a fourth receptor was cloned, named PAR4 (19, 20, 128). So far, PAR4 has been cloned from human, mouse, and rat tissues (19, 20, 327). In humans, PAR4 is widely expressed in brain (175), testes, placenta, lung, liver, pancreas, thyroid, skeletal muscle, and small intestine (20). In rats, PAR4 is expressed in esophagus, stomach, duodenum, jejunum, distal colon, spleen, and brain (327, 328).
Both in mice and humans, platelets utilize two thrombin receptors. PAR4 is a low-affinity receptor in platelets both in humans and mice. However, mouse platelets use PAR3 and PAR4 instead of PAR1 and PAR4 to respond to thrombin (19). In humans, platelet effects of thrombin appear to be mediated predominantly by PAR1. Only at high concentrations and when PAR1 activation has been inhibited is platelet activation by thrombin dependent on PAR4. Signals mediated by PAR4 result in calcium influx (19, 122, 329), thromboxane production (117), endostatin secretion in rat platelets (118), and platelet aggregation (122) (Table 2). However, selective activation of human PAR4 (hPAR4) resulted in a weaker response compared with hPAR1-mediated signaling because anionic phospholipids were not exposed on the surface of hPAR4-activated platelets. Interestingly, hPAR4 can also be activated by cathepsin G, a neutrophil granule proteinase (330). Cathepsin G mediates neutrophil-platelet interactions at sites of vascular injury or inflammation. Inhibition of hPAR1 had no effect on platelet responses to cathepsin G, indicating a specific activation of PAR4.
Patients with Hermansky-Pudlak syndrome, an autosomal recessive disorder, lack platelet dense granules and have no ADP autocrine response. However, these patients show only a mild bleeding phenotype (331, 332). It was hypothesized that the defect of ADP autocrine response was compensated by signaling through PAR4 because the activation of this receptor occurs well after ADP release in normal individuals (333). Therefore, the ADP-autocrine response does not seem to be necessary for platelet aggregation as long as PAR4 is strongly activated. However, it was observed by another laboratory that PAR4-induced, but not PAR1-induced, aggregation was entirely ADP-dependent using a specific AP for PAR4 (334). Moreover, the authors found that subthreshold concentrations of an AP-activating PAR1 potentiated the effects of a PAR4-AP to stimulate maximal aggregation. In addition, both prostacyclin (PGI2) and S-nitroso-glutathione, an NO-releasing agent, reduced AP-stimulated aggregation and fibrinogen-receptor up-regulation.
PAR4–/– mice had markedly increased bleeding times (40). In addition, the platelets of these mice failed to change shape, mobilize calcium, or aggregate in response to thrombin. Ma et al. (118) observed that a specific AP for PAR4 could induce endostatin release in rat platelets. A selective PAR4 antagonist prevented endostatin release. In human platelets, specific agonists for PAR1 or PAR4 stimulated thromboxane production (117). Thromboxane produced by the combined stimulation of PAR1 and PAR4 was additive, suggesting the presence of two pathways for thrombin-induced thromboxane production in platelets. Blocking PAR1 function with a domain-specific antibody resulted in substantial inhibition of thrombin signaling. However, PAR4 can mediate platelet activation in response to high concentrations of thrombin (128). hPAR4 itself showed no pharmacological effect at either 1 nM or 30 nM thrombin. However, blocking of both hPAR1 and hPAR4 resulted in a profound inhibitory response even at high concentrations of the proteinase (128). This shows that hPAR4 activation is not necessary for robust responses in platelets when hPAR1 function is intact. Thus, inhibition of hPAR1 alone is probably not sufficient when seeking new antithrombotic therapies; it might be necessary to block both receptors simultaneously (106, 128). hPAR4 seems to play a role as a backup signaling device that might mediate thrombin signaling to distinct effectors or with different kinetics compared with PAR1. It might also allow platelets to respond to proteinases other than thrombin. Moreover, both receptors might be able to directly interact with each other.
The fact that mPAR3 alone does not result in thrombin signaling on mouse platelets showed that mPAR3 functions as a cofactor that promotes cleavage and activation of PAR4 at low concentrations of thrombin (39, 40). Interestingly, knocking out the PAR3 or the PAR4 gene leads to a similar degree of protection against thrombosis in mice (40, 325). In humans, both hPAR1 and hPAR4 may independently mediate thrombin signaling (20, 128, 177). Thus, the mouse system does not have a direct analog in the human platelet. However, mPAR3 promoted cleavage and activation of hPAR4 as effectively as mPAR4. Thus, hPAR4 can be "cofactored" by PAR3 in vitro, but it remains unclear whether or not cofactors such as PAR1 play a direct role in hPAR4 activation in vivo (39).
This reservoir of multiple thrombin receptors including PAR1, PAR3, and PAR4 may allow for a precise regulation of thrombin-induced inflammatory stimuli under different pathophysiological conditions and the fine-tuned induction of different signal transduction pathways.
In endothelial cells, thrombin induces a rapid but transient activation of endothelial cells by stimulating the secretion of PGI2 or platelet-activating factor (PAF) and by inducing the production of cell adhesion molecules (P-selectin, E-selectin) (47, 69, 70). Thrombin and a nonselective AP also induce synthesis and release of cytokines from endothelial cells such as IL-1, IL-6, and IL-8 (71, 72). Moreover, they regulate the cytokine-independent expression of ICAM-1 and VCAM-1 on human endothelial cells and cause an increased adhesion of monocytes to endothelial cells. This effect can be diminished by blocking antibodies (anti-CD 18, anti-CD 49d) (73). In the past, the effects observed above had been attributed to the activation of PAR1. However, a recent in vivo study has shown that thrombin-induced leukocyte rolling and adhesion to vascular walls is mediated via PAR4 (74). Although thrombin can trigger the recruitment of leukocytes to sites of inflammation, the PAR1 antagonist RWJ-56110 does not block this effect. In contrast, a PAR4-AP is able to reproduce the effects of thrombin on leukocyte rolling and adherence, indicating that this proinflammatory effect of thrombin is due to the activation of PAR4 and not PAR1 (74). Thus, PAR4 appears to function in early events of inflammatory reaction, in terms of recruiting leukocytes to the site of injury.
In contrast to a potential participation of PAR1 in stimulating the angiogenic process (335), PAR4 may play an opposing role. Apart from its antiangiogenic role via the activation of platelets, PAR4 can also affect the arterial system by stimulating vascular smooth muscle mitogenesis (220).
PAR4, like PAR1 and PAR2, also appears to be linked to signal transduction processes that modulate airway smooth muscle tone, because a PAR4-AP caused a rapid transient contractile response in a murine tracheal preparation. This contraction was followed by a transient relaxant phase. The underlying molecular mechanism is still unclear, but the relaxation appears to be dependent on PAR4-stimulated and COX-2-generated PGE2 production, which activates the E-type prostanoid 2 receptor (152, 153).
In conclusion, preliminary data strongly suggest an important pathophysiological role of PAR4 in vascular homeostasis and platelet function. However, the role of PAR3 as a signaling molecule and the precise function of PAR4 during inflammation and immune response remain to be clarified.
B. Signaling by proteinases via PAR3 and PAR4
PAR3 and PAR4 are the most recent members of the proteinase-activated receptor family. They are activated by thrombin. Subsequent to the cloning of PAR3 (18) and PAR4 (19, 20), relatively few publications have appeared examining the involvement of these receptors in signaling cascades. One puzzle that is yet to be resolved, as already mentioned, is the inability of PAR3 to signal in response to its tethered ligand-derived peptide or to thrombin (39, 40). Until this issue is resolved unequivocally, a meaningful discussion about PAR3 signaling is not possible. In contrast, PAR4 appears to be capable of activating both G12/G13 and Gq pathways (201, 336, 337).
As mentioned above, PAR4 appears to be involved in the same signaling cascades as PAR1 in human platelets. In contrast, mouse platelets express PAR3 and PAR4, but only PAR4 appears to serve as a real signaling receptor, and PAR3 serves solely to facilitate cleavage of PAR4 by thrombin (39).
Studying downstream signaling events in which PAR1 and PAR4 could be involved in VSMC, Bretschneider et al. (220) revealed that these receptors have distinct downstream signaling kinetics. Later, activation of MAPKs after stimulation of PAR4 was demonstrated. In their recent work, Sabri et al. (338) investigated PAR4-mediated signaling in cardiomyocytes derived from PAR1–/– mice. Using AYPGKF-NH2, a modified PAR4 agonist with an increased binding potency to PAR4, they were able to demonstrate p38 phosphorylation as well as slight activation of PLC and ERK1/2. Additionally, thrombin and PAR4-AP, but not PAR1-AP, were able to activate Src in these cells, clearly indicating that the action of thrombin on Src activation is mediated by PAR4, and not by PAR1 in these cells. Further studies implicated the involvement of Src and EGFR kinase activity in the PAR4-dependent p38 signaling pathway (338).
Recently, the involvement of PAR4 in factor Xa-mediated signaling has been demonstrated. Camerer et al. (25) revealed that expression of PAR1, PAR2, and PAR4 in Xenopus oocytes confers calcium signaling in response to factor Xa. They further showed that PAR4-AP (AYPGKF-NH2) stimulated increased phosphoinositide hydrolysis in endothelial cells and that responses to AYPGKF-NH2 are absent in PAR4–/– endothelial cells (240). Accordingly, PAR4-mediated phosphoinositide hydrolysis in response to factor Xa has been clearly demonstrated (25).
V. Conclusions
Top
Abstract
I. Introduction
II. PAR1 in Inflammation...
III. PAR2 in Inflammation...
IV. PAR3 and PAR4
V. Conclusions
References
There is no doubt that PARs play an important regulatory role during inflammation and immune response. Recent findings consolidate the concept that serine proteinases act as autocrine, paracrine, or endocrine mediators that talk directly to cells (13, 14, 15, 30). Some of these proteinase-stimulated effects are mediated through activation of proteinase-activated receptors, resulting in signal transduction pathways that are involved in inflammation and immune response. In many cases, PARs appear to play a proinflammatory role due to activation of proinflammatory mediators and cytokines (28, 59, 67, 72, 77, 78, 79, 80, 81, 82, 83, 84, 85, 86, 87, 88, 89, 90, 94, 99, 263, 272, 274, 282, 283, 295, 339, 340, 341, 342, 343, 344, 345). In other instances, a protective and antiinflammatory role of PARs has been observed (52, 292, 298, 312, 346, 347, 348). Various serine proteinases, which serve as PAR activators, are relevant mediators during inflammation (Table 2). However, the precise role of PARs and serine proteinases in different tissues, cell types, and states of inflammation remains to be determined.
Open questions include: 1) For the several proteinases that are released in various tissues from different cell types during inflammation, what is their mechanism of regulation and what is their functional relevance? 2) What is the significance of the existence of multiple receptor subtypes for one proteinase, and how might these common proteinase targets be differentially regulated? 3) Which immune cells express functional PARs in humans in vivo, and what is their biological relevance? 4) Which are the endogenous proteinases that activate PAR2 in human inflammation in vivo? 5) Which factors influence the regulation of PARs during inflammation or host defense? 6) Which effects of PAR-activating proteinases are non-PAR-mediated effects? 7) Which are the cell-specific signaling pathways and molecular mechanisms (transcription factors) after selective PAR activation in different inflammatory states and tissues? 8) What role do PARs play in neuronal transmission in humans? 9) Can PAR agonists or antagonists be used as therapeutic agents during inflammation or immune response in human diseases?
Thus, an integrative understanding of a regulatory role of serine proteinases as extracellular degradative enzymes as well as hormone-like signaling messengers in part via PARs and their counterregulation by extracellular proteinase inhibitors and intracellular molecules should lead to effective therapeutic approaches for various inflammatory/immune diseases such as thrombosis, sepsis, bacterial infections, gingivitis, asthma, hyperreactivity reaction, lung fibrosis, renal inflammation, rheumatoid arthritis, colitis ulcerosa, Crohn’s disease, pancreatitis, Alzheimer’s disease, amyotrophic lateral sclerosis, HIV encephalitis, atopic dermatitis, contact dermatitis, rosacea, wound repair, and infertility, for example, as well as pathophysiological symptoms such as pain and pruritus, in the future.
Footnotes
This work was supported by grants from the Federal Ministry of Education and Research (Interdisciplinary Center for Clinical Research, Münster F?. II-703-04; German Research Association, STE 1014; Collaborative Research Center (SFB) 293; SFB 492 to M.S.), Schering Foundation (to M.S., V.S.), C.E.R.I.E.S., Paris, France, and Boltzmann-Institute, Münster, Germany (to T.A.L., M.S.); Novartis Research Foundation (Vienna, Austria); Rosacea Foundation (to M.S.), and Boehringer-Ingelheim Fonds (to J.B.). The work of M.D.H. and N.V. is supported by funds from the Canadian Institutes for Health Research, the Kidney Foundation of Canada (to M.D.H.), the Heart and Stroke Foundation of Canada (to M.D.H.), by a Johnson & Johnson Focused Giving Grant, and by the Ileitis Colitis Foundation (to N.V.).
First Published Online October 12, 2004
Abbreviations: AP, Activating peptide; AP-1, activator protein-1; [Ca2+]i, intracellular calcium; CGRP, calcitonin-gene related peptide; CNS, central nervous system; COX, cyclooxygenase; cPLA2, cytosolic phospholipase A2; DRG, dorsal root ganglion; EGF, epidermal growth factor; EGFR, EGF receptor; EPR-1, effector cell proteinase receptor-1; FAK, focal adhesion kinase; GI, gastrointestinal; GM-CSF, granulocyte-macrophage colony-stimulating factor; GP, glycoprotein; GPCR, G protein-coupled receptor; GRK, G protein-coupled receptor kinase; HMEC, human microvascular endothelial cells; hPAR, human PAR; HUVEC, human umbilical vein endothelial cells; ICAM-1, intercellular adhesion molecule-1; IFN-, interferon ; IKK, inhibitor of NFB kinase; iNOS, inducible NO synthase; JNK, Jun N-terminal kinase; MCP-1, monocyte chemoattractant protein-1; MDCK cells, Madin-Darby canine kidney cells; MMP, matrix metalloproteinase; mPAR, murine PAR; NFB, nuclear factor B; NK cells, natural killer cells; NK receptor, neurokinin receptor; NO, nitric oxide; PAF, platelet-activating factor; PAR, proteinase-activated receptor; PAR-AP, proteinase-activated receptor activating peptide; PDGF, platelet-derived growth factor; PG, prostaglandin; PI3, phosphoinositide 3; PKA, protein kinase A; PKB, protein kinase B; PKC, protein kinase C; PLC, phospholipase C; PTX, pertussis toxin; Pyk2, proline-rich tyrosine kinase 2; SLP-76, SH2-domain containing leukocyte-specific phosphoprotein of 76 kDa; SMC, smooth muscle cells; SP, substance P; Src, protein tyrosin kinase Src; TK, tyrosine kinase; VCAM-1 , vascular cell adhesion molecule-1; VSMC, vascular SMC; ZAP-70, -chain-associated protein kinase of 70 kDa.
References
Scott G, Leopardi S, Parker L, Babiarz L, Seiberg M, Han R 2003 The proteinase-activated receptor-2 mediates phagocytosis in a rho-dependent manner in human keratinocytes. J Invest Dermatol 121:529–541(Martin Steinhoff, J?rg Bu)
Serine proteinases such as thrombin, mast cell tryptase, trypsin, or cathepsin G, for example, are highly active mediators with diverse biological activities. So far, proteinases have been considered to act primarily as degradative enzymes in the extracellular space. However, their biological actions in tissues and cells suggest important roles as a part of the body’s hormonal communication system during inflammation and immune response. These effects can be attributed to the activation of a new subfamily of G protein-coupled receptors, termed proteinase-activated receptors (PARs). Four members of the PAR family have been cloned so far. Thus, certain proteinases act as signaling molecules that specifically regulate cells by activating PARs. After stimulation, PARs couple to various G proteins and activate signal transduction pathways resulting in the rapid transcription of genes that are involved in inflammation. For example, PARs are widely expressed by cells involved in immune responses and inflammation, regulate endothelial-leukocyte interactions, and modulate the secretion of inflammatory mediators or neuropeptides. Together, the PAR family necessitates a paradigm shift in thinking about hormone action, to include proteinases as key modulators of biological function. Novel compounds that can modulate PAR function may be potent candidates for the treatment of inflammatory or immune diseases.
I. Introduction
II. PAR1 in Inflammation and Immune Response
A. Vasculature and heart: PAR1 antagonists—a novel potential approach for the treatment of cardiovascular diseases
B. Platelets
C. Immune cells
D. Airways
E. Gastrointestinal tract
F. Kidneys and urogenital tract
G. Brain and peripheral nervous system
H. Signaling by proteinases via PAR1
III. PAR2 in Inflammation and Immune Response
A. Vasculature
B. Immune cells
C. Arthritis
D. Skin
E. Airways
F. Brain and peripheral nervous system
G. Digestive tract and pancreas
H. Signaling by proteinases via PAR2
IV. PAR3 and PAR4
A. Biology and distribution of PAR3 and PAR4
B. Signaling by proteinases via PAR3 and PAR4
V. Conclusions
I. Introduction
Top
Abstract
I. Introduction
II. PAR1 in Inflammation...
III. PAR2 in Inflammation...
IV. PAR3 and PAR4
V. Conclusions
References
SERINE PROTEINASES CONSTITUTE a family of proteolytic enzymes characterized by a unique catalytic triad consisting of Ser, His, and Asp. These residues are able to hydrolyze peptide bonds (1). Serine proteinases are produced as inactive precursors or zymogens. Subsequent zymogen conversion into a mature physiologically active enzyme is mediated via a process called "limited proteolysis" or zymogen activation (2, 3, 4). In mammals, serine proteases, for example, regulate the hemostatic and fibrinolytic balance, degrade neuropeptides involved in neurogenic inflammation or serve as modulators of immune response during inflammation (5). Three different types of serine proteinase inhibitors can be distinguished based on their mechanism of action: canonical, noncanonical inhibitors, and serpins (6). An imbalance between these inhibitors and their targeted proteinases can affect immune/inflammatory responses and may result in disease. Moreover, the presence of proteinase inhibitors regulates and limits interactions between proteinases and their receptors (7, 8, 9, 10, 11).
Recent studies elucidated the ability of certain serine proteinases to regulate cell function via G protein-coupled receptors (GPCRs). At least two different types of proteinase receptors have been identified involving proteolytic cleavage in their activation mechanism: urokinase receptors and proteinase-activated receptors (PARs) (12, 13, 14).
PARs belong to a new subfamily of GPCRs with seven transmembrane domains activated via proteolytic cleavage by serine proteinases (13, 14, 15, 16). PAR1, PAR3, and PAR4 are targets for thrombin, trypsin, or cathepsin G (17, 18, 19, 20). In contrast, PAR2 is resistant to thrombin, but can be activated by trypsin, mast cell tryptase, factor Xa, acrosin, gingipain, and neuronal serine proteinases (21, 22, 23, 24, 25, 26, 27, 28) (Table 1). Interestingly, PARs are activated by a unique mechanism: proteinases activate PARs by proteolytic cleavage within the extracellular N terminus of their receptors, thereby exposing a novel "cryptic" receptor-activating N-terminal sequence that, remaining tethered, binds to and activates the receptor (Fig. 1) within the same receptor (17, 21, 22). Specific residues (about six amino acids) within this tethered ligand domain are believed to interact with extracellular loop 2 and other domains of the cleaved receptor (29), resulting in activation. This intramolecular activation process is followed by coupling to G proteins and the triggering of a variety of downstream signal transduction pathways (see Sections II.H, III.H, and IV.B; also see Table 2 and Fig. 2, and Refs.13 , 14 , and 16). Thus, PARs are not activated like "classical" receptors because the specific receptor-activating ligand is part of the receptor, whereas the circulating agonist is a relatively nonspecific serine proteinase that does not behave like a traditional hormonal regulator akin to insulin.
Several studies during the past few years have also demonstrated that several mechanisms exist to regulate stimulation and termination of PAR-initiated signaling (13, 14, 15, 16, 30). Importantly, the availability of PARs at the cell surface is governed by trafficking of the receptor from intracellular stores, and the signaling properties depend on the presence of G proteins and G protein-coupled receptor kinases (GRKs) that modify activity. For PAR1, PAR2, and PAR4, it is well established that short synthetic peptides [PAR-activating peptides (PAR-APs)] designed on their proteolytically revealed tethered ligand sequences can serve as selective receptor agonists (Table 1). Some PAR-APs activate more than one PAR, and they activate receptors at concentrations in the micromolar range as compared with nanomolar potencies of the proteinases themselves (31, 32). Although the PAR1-AP, SFLLRN-NH2 also activates PAR2, PAR2-APs, like SLIGRL-NH2 are not capable of activating other PARs. Unfortunately, the relatively low potency (10 to 100 μM EC50) and susceptibility to aminopeptidases (33) limit the utility of the PAR-APs in some bioassay systems. Recently, modified synthetic agonist peptides with higher potency, resistance to aminopeptidases, and greater receptor selectivity have been developed and characterized. These receptor-selective agonists are of use to study the consequences of activating PARs in vivo (Table 1) (13, 14, 16, 34). So far, antagonists for PAR1 (35, 36, 37) and PAR4 (38) have been synthesized, but are not yet available for PAR2. PAR3, as will be elaborated upon in Section IV, remains a puzzle, in that studies with the murine receptor indicate that it cannot be activated either by its cognate synthetic tethered ligand peptide or by thrombin (39, 40). Rather than acting as an independent cell regulator, PAR3 appears to function as a cofactor for the activation of PAR4 (39). Complementary data documenting PAR-mediated effects in various tissues have been obtained using PAR-deficient (PAR–/–) mice.
Taken together, data obtained using the enzyme activators themselves (trypsin, thrombin), the PAR-APs and using PAR gene-deficient mice provide compelling evidence that PARs play a critical role in the regulation of various physiological and pathophysiological functions in mammals, including humans. This review focuses on the biology and signaling properties of PARs in various mammalian tissues and highlights the current knowledge about the role of PARs during inflammation and immune response. To complement the information summarized in the sections that follow, the reader is encouraged to consult a number of other recent reviews concerning the activation mechanisms and cell biological aspects of PARs (13, 14, 16, 34, 41, 42, 43, 44, 45).
II. PAR1 in Inflammation and Immune Response
Thrombin is an important effector proteinase of the coagulation cascade that leads to formation of a hemostatic plug. Thrombin is thought to act near the site at which it is generated and it is activated when circulating coagulation factors in the blood plasma make contact with tissue factor. Tissue factor is a membrane protein that is usually produced by cells that are separated from blood (i.e., epithelial cells). However, it is also expressed at low levels on circulating monocytes and microparticles from leukocytes. Tissue factor is associated with the activation of zymogen factor X by factor VIIa. Factor Xa together with its cofactor Va subsequently converts prothrombin to the active enzyme. Thus, plasma coagulation can only take place when the vascular integrity is damaged (15). Thrombin causes shape change of endothelial cells and increased permeability of endothelial cell layers.
However, recent evidence revealed that thrombin is not only a clotting proteinase serving as both a pro- and anticoagulant molecule but also appears to play multifunctional roles related to inflammation, allergy, tumor growth, metastasis, tissue remodeling, thrombosis, and probably wound healing (13, 15, 42, 43). Subsequent to the cloning of PAR1 (17, 46), it was realized that many of the cellular actions of thrombin (e.g., platelet aggregation, angiogenesis, endothelial cell permeability, vasoregulation, gene regulation, leukocyte trafficking, immunomodulation) could be attributed to the activation of its GPCR (40, 47, 48, 49, 50, 51, 52, 53).
PAR1 has been detected in a variety of tissues, including platelets; endothelial cells; fibroblasts; monocytes; T cells positive for CD 8, CD 16, and either CD 56 or CD 57 (54, 55); natural killer (NK) cells (48); CD 34+ hematopoietic progenitor cells (56); dental pulp cells (57); smooth muscle cells (SMC); epithelial cells; neurons; glial cells; mast cells; and certain tumor cell lines (17, 54, 58, 59, 60, 61). A receptor with high affinity for thrombin has also been detected in rat peritoneal macrophages (62). It should be pointed out, however, that PAR1 very likely does not represent the only target for thrombin, in that other (non-PAR) high-affinity binding sites for thrombin [e.g., on platelets or macrophages (62)] and other conserved thrombin sequences apart from the catalytic domain (63) may also play a role in the cellular actions of thrombin in a variety of target tissues.
A. Vasculature and heart: PAR1 antagonists—a novel potential approach for the treatment of cardiovascular diseases
PAR1 can potentially regulate vascular function under both physiological as well as pathophysiological conditions (15, 42). A number of studies have revealed that thrombin and other agonists of PAR1 can affect the vascular tone. For example, before the discovery of PAR1, it was observed that thrombin can regulate vascular tone by an endothelial-dependent mechanism involving the release of nitric oxide (NO) (64). It is now recognized that this effect of thrombin is due to the activation of PAR1. Moreover, thrombin and PAR1 agonists can contract vascular SMC (VSMC) by a direct effect that requires extracellular Ca2+ (64). Thus, in isolated coronary artery and aorta preparations, PAR1 mediates relaxation (64, 65). In contrast, PAR1 agonists contract human placental and umbilical arteries (66). Furthermore, PAR1 mediates prostanoid generation and secretion, cellular contraction and barrier dysfunction, and enhanced expression of platelet-derived growth factor (PDGF) in human and bovine endothelial cells (67). Interestingly, progestins such as progesterone or gestodene up-regulate PAR1 expression and thereby stimulate thrombin-induced tissue factor-dependent surface procoagulant activity in the rat vascular system, suggesting a role of thrombin in hormone-induced thrombosis via PAR1 activation (68).
Recent observations support a role of thrombin and PAR1 in regulation of functions normal (69, 70, 71, 73) and atherosclerotic (75) endothelium. In normal human arteries, PAR1 is mostly confined to the endothelium, whereas during atherogenesis, its expression is enhanced in regions of inflammation associated with macrophage influx, smooth muscle cell proliferation, and an increase in mesenchymal-like intimal cells (75). In vivo, a neutralizing antibody to PAR1 has been observed to reduce expression of mRNA for the proliferating cell nuclear antigen, an index of intimal and neointimal smooth muscle cell accumulation in rat arteries during balloon angioplasty. These data suggest that PAR1 regulates proliferation and accumulation of neointimal SMC during tissue repair (76).
Thrombin and PAR1 agonists cause a rapid but transient contraction of endothelial cells in various tissues resulting in gap formation and increased permeability of plasma proteins and inflammatory cells. Several mediators are involved in this PAR1 modulated process such as cytokines, kinins, and biogenic amines (67, 72, 77, 78, 79, 80, 81, 82, 83, 84, 85, 86, 87, 88, 89, 90, 91, 92, 93, 94, 95).
It was also revealed that PAR1 plays a crucial role during vascular ontogenesis. Interestingly, about 50% of the PAR1-deficient mouse embryos die at midgestation with bleeding from multiple sites. However, a PAR1 transgene driven by an endothelial-specific promoter prevented death of PAR1-deficient embryos (96), indicating that PAR1 modulates endothelial cell function in developing blood vessels, thereby contributing to vascular development and homeostasis in mice (88, 96).
Vascular wall cells respond to the procoagulant factor Xa by an increase in intracellular Ca2+ ([Ca2+]i) and by assembly of this factor into prothrombinase complexes that even enhance this effect. Additionally, factor Xa stimulation of PAR1 leads to an increased production of tissue factor, a prothrombotic agent, underlining the important role of PAR1 for thrombosis (86). Together, these results point to a pivotal role of PAR1 in vascular homeostasis and thrombosis.
PAR1 agonists are also mitogenic, stimulating proliferation of endothelial cells (97, 98), mediating endothelium-dependent relaxation to thrombin and trypsin in human pulmonary arteries (99), and causing the release of IL-6 from human microvascular endothelial cells (HMEC) (72). Because PAR1 up-regulates 1(I)-procollagen synthesis in VSMC, one may speculate that PAR1 plays a role in vascular wound healing (100).
Finally, factors that regulate PAR1 function on endothelial cells have also been studied. For example, PAR1 is down-regulated by shear stress (101) and inflammatory mediators such as TNF- (102), and is directly modulated by thrombomodulin (103) in human endothelial cells. On the other hand, cyclic mechanical stress leads to an up-regulation of PAR1 in VSMC (104). In contrast to PAR2, little is known about the inflammatory mediators that regulate PAR1 expression in unstimulated or stimulated endothelial cells.
As of yet, only a few studies exist about the role of GRKs in regulating PAR1 in endothelial cells. Only recently, Tiruppathi et al. (105) found a crucial role of the isoform GRK-5 on thrombin-induced desensitization of PAR1.
It is also necessary to mention the recently demonstrated role of PAR1 in vascular associated pathological processes in which thrombin is involved. The role of PAR1 in thrombus formation was recently investigated in an animal model. Cook et al. (106) examined the role of PAR1 in intravascular thrombus formation in an experimental model of arterial thrombosis in the African green monkey. Using a blocking antibody to PAR1, this group demonstrated a dramatically diminished thrombin-stimulated aggregation and secretion of human platelets, whereas platelet activation induced by the PAR1 agonist SFLLR-NH2 was not affected. These results demonstrated that a specific blockade of PAR1 activation by thrombin can prevent arterial thrombosis in this animal model without significantly altering hemostatic parameters. The data suggest that PAR1 is crucially involved in this disease and is an attractive antithrombotic therapeutic target, although it probably has to be inhibited in concert with the low-affinity thrombin receptor PAR4 to fully prevent platelet activation (see Section II.A).
A demonstrated role of PAR1 in intravascular thrombus formation and knowledge about side effects of thrombin inhibitors made appealing an idea to investigate the potency of PAR1 antagonists as antithrombotic therapeutic drugs. Indirect thrombin inhibitors like heparin and warfarin, which prevent the circulation of thrombin, or the newly developed direct thrombin inhibitor ximelagatran (465) are currently used for the treatment of cardiovascular diseases, e.g., myocardial infarction, atherosclerosis, and thrombosis. However, they carry potential bleeding liabilities as an undesirable side effect when administered in chronic patients. Because it is now well known that human platelet responses to thrombin are mainly mediated via PAR1 (19), direct inhibition of this receptor instead of the proteinase represents an attractive way to circumvent possible side effects by thrombin inhibitors.
PAR1 possesses several extra- and intracellular sites that are crucial for its function and that might represent possible targets for antagonists (107). On the extracellular side, blocking of the cleavage site or the hirudin-like domain could inhibit N-terminal cleavage. Bradykinin or peptides derived thereof, also known as thrombostatins, have been shown to prevent thrombin-induced platelet aggregation (108). In addition, antibodies directed against both the cleavage site and the hirudin-binding site have been generated (106, 109). Binding of the antagonist to either the tethered ligand or extracellular loop 2 prevents interaction of the novel N terminus with the extracellular loop. Several small synthetic antagonists have been developed that block the binding site of the tethered ligand. However, peptide antagonists based on modifications of the tethered ligand were relatively unstable or showed only limited inhibitory activity (35). Thus, a second generation of indole or indazole-based peptide-mimetic antagonists was created that block binding of both the tethered ligand and the proteinase (110). One of these was recently shown to prevent thrombus formation in nonhuman primates (107). In addition, several nonpeptide inhibitors were developed. However, it was not reported whether these compounds were PAR1-specific (111). On the intracellular side, the G protein binding sites are possible targets for pharmacological intervention. One approach used the transfection of endothelial cells with a G minigene (112). However, this might suppress all signaling pathways that involve G. Another more specific approach used so-called "pepducins," peptides that were derived from the third intracellular loop of PAR1. These peptides were able to permeate the cells and also carried a membrane anchor (113, 114).
In summary, two preclinical studies with primates demonstrated that PAR1 antagonists have indeed a therapeutic potential for the treatment of cardiovascular diseases (106, 107). However, extensive research needs to be done for the development of orally administered PAR1 antagonists because only those drugs are thought to be suitable for chronic patients (111).
B. Platelets
Activation of platelets by thrombin or APs specific for PAR1 is characterized by calcium influx; cytoskeletal reorganization; platelet aggregation; degranulation (17, 115, 116); thromboxane production (117); mobilization of the adhesion molecule P-selectin and the CD40 ligand to the platelet surface (119); stimulation of serotonin and epinephrine release (120); enhanced expression of CD62, PDGF (AB) and PDGF(BB) (121); and exposure of anionic phospholipids (phosphatidylserine, phosphatidylethanolamine) that support blood clotting (122). Thrombin has also been demonstrated to stimulate vascular endothelial growth factor release in megakaryocytes and platelets (123). The growth factor in turn is able to induce proliferation of endothelial cells. In addition, tissue inhibitor of matrix metalloproteinase (MMP)-4 that colocalized with MMP-2 in resting platelets was released upon platelet aggregation induced by collagen and thrombin. However, a direct effect using specific PAR1 agonists has not been shown yet (124). The density of human PAR1 on the platelet cell surface was estimated to be about 1500 molecules per resting platelet (125). A recently reported intronic polymorphism in the PAR1 gene leads to decreased expression on the platelet surface and thus to a lower response to the AP (126, 127). PAR1 is a high-affinity thrombin receptor in humans: receptor blocking using an antagonist, domain specific antibodies, or simply by desensitization led to inhibition of responses at 1 nM thrombin but only to attenuated responses at 30 nM (128).
High-affinity thrombin binding of platelets is possibly associated via the membrane glycoprotein (GP)Ib-IX-V complex. This interaction is lost in patients suffering from the Bernard-Soulier syndrome, a disease where GPIb-IX-V is not expressed (129). The same condition can be mimicked using either monoclonal antibodies directed against GPIb or proteolytic removal of this protein from the platelet surface (130). Under these conditions, platelet aggregation is either delayed or requires higher thrombin concentrations.
Recently, Schmidt et al. (131) reported that the gene encoding IQ motif containing GTPase activating protein-2, a scaffolding protein for filopodal extension during platelet activation, was identified within the human PAR gene cluster at 5q13, flanked by the PAR1 gene and encompassing the PAR3 gene. Only thrombin- or PAR1-AP-activated platelets showed a rapid translocation of IQ motif containing GTPase activating protein-2 to the platelet cytoskeleton. This suggests that a functional genomic unit evolved to mediate thrombin signaling events in humans (132).
PAR1 on platelets also seems to play an important role during bacterial invasion: the cysteine proteinase gingipain from Porphyromonas gingivalis was reported to activate the receptor leading to uncontrolled activation of the host cells (28). The enzyme streptokinase derived from Streptococcus sp. is widely used in the treatment of coronary thrombosis. It functions as a plasminogen activator that forms an active complex with plasminogen of the host that is able to cleave PAR1 (133). In mouse platelets, PAR1 does not seem to play a role: PAR1 expression was hardly detectable, and a specific AP did not activate rodent platelets (23, 134, 135). Moreover, platelets derived from PAR1-deficient mice responded to thrombin-like wild-type platelets (135).
C. Immune cells
For some time, an important role for serine proteinases and PARs for the modulation of leukocyte effector functions has been proposed (136). As of yet, only limited data about the regulation of immune and inflammatory responses by PARs are available. It is well documented that PARs influence monocyte motility and chemotaxis, modulate pleiotropic cytokine responses, contribute to mononuclear cell proliferation, and induce apoptosis in various immune cells (84, 137). Recently, it was shown that PAR1 is capable of stimulating elastase secretion from macrophages (138). Moreover, functional thrombin receptors are expressed on human T lymphoblastoid cells (139). However, the receptor subtype(s) (PAR1, PAR3, and PAR4) have not been characterized as of yet. Human -thrombin stimulated five different T lymphoblastoid cell lines to increase intracellular free Ca2+ concentrations and to activate protein kinase C (PKC), whereas thrombin receptors were absent in B cell lines. Thus, PARs may regulate T cell function during inflammation and immune responses, but the precise mechanism is still unknown. Granzyme A from cytotoxic and helper T lymphocytes appears to interact with PAR1 in astrocytes to regulate thrombin function (140). Interestingly, granzyme A blocked the thrombin-induced platelet aggregation in a dose-dependent manner by cleaving PAR1, presumably downstream from its thrombin targeted activation site, thereby reducing the response to subsequent challenge with thrombin; but granzyme A itself did not induce a signal in thrombin-stimulated platelets. Thus, granzyme A may interact with PAR1 in a manner that is insufficient to cause aggregation, but sufficient to disarm the ability of the receptor to respond to thrombin. Finally, these observations demonstrate that granzyme A release occurring during immune responses within blood vessels would not directly cause platelet aggregation. Thus, the T cell-derived proteinase granzyme A seems to inhibit responses triggered by thrombin during inflammation or tissue injury.
PAR1 modulates chemotaxis in inflammatory cells. Besides IL-8 secretion (71), thrombin induces production of monocyte chemoattractant protein-1 (MCP-1) in monocytes, probably via PAR1 (58). However, other PAR1-specific agonists were not used in this study. Therefore, it cannot be excluded that other thrombin receptors are involved. IL-8 secretion is up-regulated by interferon- (IFN-) and diminished by prostaglandin (PG)E2 (51), suggesting a role in cytokine modulation. Furthermore, PAR1 is capable of inducing IL-1 as well as IL-6 production in monocytes (82). These cytokines are known to be proangiogenic, implicating PAR1 in angiogenesis and tissue repair. However, IFN--differentiated growth-arrested U937 cells also respond to PAR1 agonist administration by overcoming cell arrest and revert to a high proliferation rate via regulation of p21CIP1/WAF1 and cyclin D1. Together, these results may help to explain how thrombin promotes tissue repair and unrestricted proliferation in malignant tissues (137, 141).
Because thrombin, via PAR1, is chemotactic for large granular lymphocytes in humans (55), large granular lymphocytes can enhance the effects of PAR1 in patients with inflammatory disorders (142, 143, 144). Moreover, thrombin modulates activity of NK cells (48). For example, thrombin can enhance NK cell-mediated cytotoxicity and IL-2 production in rheumatoid arthritis, and PAR1 may therefore play an important role in this inflammatory process, as well as in NK cell responsiveness to IL-2.
D. Airways
Various cells in mammalian airways such as epithelial cells, SMC, and fibroblasts abundantly express PAR1 on the cell surface. Several recent studies also suggest that PAR1 may play an important role in inflammatory lung diseases such as neutrophilic alveolitis, pulmonary fibrosis, and asthma (50, 77, 91, 145, 146, 147, 148, 149).
In the airways of different species, PAR1 exerts a dual role leading to stimulation of contraction on the one hand and relaxation on the other hand. For example, thrombin stimulates contraction of human bronchial rings (150) and constriction in guinea pig bronchi (as does AP) (151), but activates constriction as well as relaxation in mouse tracheal SMC (152, 153). In these mouse studies, trypsin, thrombin, and peptide agonists of PAR1, PAR2, and PAR4 induced relaxant responses of isolated tracheal smooth muscle preparations, which were mediated by a prostanoid, probably PGE2. This effect was abolished by indomethacin, the cyclooxygenase (COX)-2 inhibitor, nimesulide, and a prostanoid receptor antagonist (AH6809). PAR1 and PAR4 synthetic peptides induced a rapid, transient, contractile response that preceded the relaxant response. Interestingly, the relaxations but not the contractions were inhibited by indomethacin, indicating that this response is mediated by cyclooxygenase products.
E. Gastrointestinal tract
PAR1 was detected in the lamina propria, the submucosa, endothelial cells, and nerves of the gastrointestinal (GI) tract (154, 155, 156, 157). The GI tract expresses relatively high levels of PAR1 mRNA compared with other tissues, both in mice and humans. The functional role of PAR1 as a receptor mediating thrombin-induced effects in the GI tract is far from being resolved because thrombin can also interact with PAR3 and PAR4 (18, 20, 74). Several studies suggest a role for PAR1 in regulating GI motility. In guinea pigs, PAR1 mediates contraction of longitudinal smooth muscle tissue, dependent on extracellular Ca2+ (158, 159). In mouse intestine, PAR1 agonists modulate the function of L-type Ca-channels and also cause contraction (161). Similar results have been observed in rats (162, 163). Irradiation by fractionated X-radiation leads to up-regulation of PAR1 at the protein level, indicating a regulatory role of external trigger factors such as irradiation on PAR1 expression (164). Stimulation of PAR1 upon adhesion of pancreatic carcinoma cells to extracellular matrix proteins such as laminin, collagen IV, and fibronectin has been observed in a pancreatic carcinoma cell line (MIA PaCa-2) (165, 166). Other studies have pointed out a role for PAR1 in intestinal secretory pathways. Buresi et al. (156) have shown that selective PAR1 agonists stimulated Ca2+-dependent chloride secretion in intestinal epithelial cells. This PAR1-induced intestinal chloride secretion involves protein tyrosin kinase Src (Src), epidermal growth factor (EGF) receptor (EGFR) transactivation, activation of a MAPK pathway, phosphorylation of cytosolic phospholipase A2 (cPLA2), and cyclooxygenase activity. The presence of functional PAR1 on intestinal epithelium and the fact that PAR1 activation leads to ion secretion suggest that PAR1 might have important implications for intestinal barrier functions. PAR1 activation on intestinal surfaces could lead to a secretory response, thus contributing to diarrhea, a symptom of intestinal inflammation. More recent observations further suggest a role of PAR1 in intestinal inflammation. We have shown that intracolonic administration of PAR1 agonists caused inflammation and disruption of intestinal barrier integrity (53). Taken together, these results suggest a proinflammatory role of PAR1 in the gut (reviewed in Ref.167). However, additional studies using PAR1 antagonists and/or PAR1-deficient mice will have to verify the role of PAR1 in the pathogenesis of inflammatory bowel diseases.
F. Kidneys and urogenital tract
Recently, a crucial role of PAR1 in the cell-mediated renal inflammation of crescentic glomerulonephritis has been demonstrated in vivo (49). In wild-type mice treated with hirudin (a direct thrombin inhibitor, characterized by a bifunctional mechanism of inhibition, exclusive specificity and strong ability to bind the enzyme) and in PAR1-deficient animals, the proinflammatory effect of thrombin was significantly reduced. Moreover, treatment of wild-type mice with the PAR1 peptide agonist (SFLLRN-NH2) augmented the inflammatory response, suggesting that PAR1 plays an important role in renal inflammation in vivo. Unfortunately, as mentioned above, the peptide SFLLRN-NH2 can also activate PAR2 with a comparable potency, and therefore a role of PAR2 in renal inflammation cannot be ruled out by this study. Surprisingly, although highly expressed in human kidneys, data clarifying a role for PAR1 in inflammation of human kidneys are still lacking.
Recent data revealed that PAR1 is involved in cellular invasion of a transfected canine kidney cell line. PAR1 agonists abrogated G(olf)-mediated invasion of MDCKts.src cells in collagen gels, indicating an important role for PAR1 in tumor metastasis (168, 169). Using the human urogenital cell line RT4, studies done in vitro have revealed that PAR1 and PAR2 agonists are capable of activating iPLA2 accompanied by release of PGE2, which may provide cytoprotection during an acute inflammatory reaction (170).
G. Brain and peripheral nervous system
Recent studies are in favor of an important role of thrombin and PARs in the brain under normal and pathophysiological conditions such as trauma, inflammation, or tumorigenesis (171, 172). Under pathophysiological conditions, i.e., during breakdown of the blood-brain barrier, circulating thrombin in the bloodstream may enter the central nervous system (CNS), leading to activation of PAR1 or PAR4. Additionally, neurons and glia cells are capable of generating prothrombin (173). Most data so far have been achieved by investigating PAR1 (13, 14, 174). In rat brain, PAR1 is expressed by neurons of the neocortex, cingulate cortex, subsets of thalamic and hypothalamic nuclei, discrete layers of the hippocampus, cerebellum, and olfactory bulb, as well as by astroglia (60). Striggow et al. (175) showed that all four PAR subtypes are expressed in rat brain. PAR1 expression was most abundant in the hippocampus, amygdala, and cortex. Interestingly, the expression of PAR1, PAR2, and PAR3 was up-regulated during experimentally induced ischemia.
In human brain, PAR1 is expressed in neurons and astrocytes (175). Cultured rat glia cells (C6) express both functional PAR1 and PAR2 (176, 177). Moreover, PAR1 agonists induce up-regulation of inducible NO synthase (iNOS) in these cells (85) and promote neuronal survival after ischemia (178) or brain trauma (179). PAR1 also protects astrocytes and neurons from apoptosis induced by hypoglycemia and oxidative stress during inflammation (180). In contrast, thrombin may also exert cytotoxic effects on neurons and induce neurite retraction (181, 182), which may be at least in part due to PAR1. Functional studies further revealed that PAR1 agonists cause retraction of neurites by neuroblastoma cells (181) and induce Ca2+ mobilization by hippocampal neurons (178).
In astrocytes, thrombin stimulates aggregation, morphological changes, and proliferation via PAR1 and induces intracellular caspase pathways leading to apoptosis in a cultured motor neuron cell line (NSC19) (183, 184). Both PAR1 and PAR2 stimulate enhanced proliferation of astrocytes (185). Friedmann et al. (186) reported a key role of thrombin in PAR1-mediated postinjury neuron survival. They demonstrated that the toxicity of thrombin on neurons can be controlled by down-regulation of PAR1 and/or release of antithrombin III by T cells. Interestingly, prothrombin expression was enhanced 24 h after injury, whereas PAR1, PAR3, and nexin-1 mRNA expression was unchanged (187). Nexin-1 is a thrombin inhibitor that also appears to control thrombin-PAR1 interactions in a rat trauma model (171, 187, 188). Thus, the whole factory needed to regulate and modulate thrombin function, including receptor as well as activating and inhibiting proteinases, can be generated in the brain. However, PAR1 also induces the reversal of astrocyte stellation in mice and rat (189, 190, 191). A comparable process is caused by exogenous or endogenous injuries of the CNS, triggering astrogliosis.
Taken together, these data clearly indicate a functional role of PAR1 during inflammation and injury in the CNS. However, in some studies using thrombin as an agonist, activation of other PARs and a role of other molecules like thrombomodulin cannot be excluded. For example, thrombin itself up-regulates thrombomodulin in astrocytes in a dose-dependent manner (192). Thus, future studies taking into account all CNS-derived PARs, their proteinases, and proteinase inhibitors are necessary to fully explore the role of PARs in the brain. Finally, some authors suggest a role of thrombin in Alzheimer’s disease, amyotrophic lateral sclerosis, or HIV encephalitis, probably via activation of PARs (182, 193, 194).
Accumulating data also indicate the role of PAR1 in the peripheral nervous system. In rat peripheral nerves, PAR1 is expressed by primary afferent neurons (195). From these observations, one may speculate that thrombin may also regulate inflammatory responses in the peripheral nervous system via PAR1. Indeed, very recently it has been demonstrated that PAR1 is expressed by a large proportion of primary spinal afferent neurons (155). Thus, thrombin directly signals to sensory neurons by cleaving PAR1. Moreover, administration of PAR1 agonists can induce plasma extravasation and edema, which were blocked by ablation of sensory nerves and administration of antagonists to the neurokinin-1 receptor, supporting the idea that thrombin cleaves PAR1 on sensory nerves to stimulate release of SP, which in turn interacts with the NK1 receptor to induce neurogenic inflammation. Thus, thrombin triggers neurogenic inflammation via PAR1 (155). However, because PAR2 and PAR4 mRNA are also expressed in the CNS, it cannot be excluded that proinflammatory effects of thrombin in the nervous system may, in addition, be mediated by other PARs. Functional studies further revealed that PAR1 agonists induce Ca2+ mobilization in enteric neurons, and regulate both excitatory and inhibitory neurons of the myenteric plexus in guinea pig small intestine, e.g., by releasing neuropeptides (196).
Recently, Vergnolle and co-workers (197) as well as other groups (198) examined the effects of PAR1 agonists on nociceptive responses to mechanical and thermal noxious stimuli. Interestingly, intraplantar injection of selective PAR1 agonists induced an enhanced nociceptive threshold and withdrawal latency resulting in mechanical and thermal analgesia. However, thrombin was analgesic in response to mechanical, but not to thermal, stimuli. Moreover, application of PAR1-AP with carrageenan significantly reduced the hyperalgesia. Thus, thrombin may play a dual nociceptive-analgesic role, and future studies will be required to determine whether PAR1 agonists might be of therapeutic use for the treatment of pain.
H. Signaling by proteinases via PAR1
Investigations during the last few years revealed that the above-mentioned PAR1-mediated effects are transduced by various signaling pathways leading to diverse functions of PAR1 under physiological and pathophysiological conditions. These findings were recently reviewed (13). However, in the current work we focus on the signaling events mediated via PAR1 in different tissues and cell types (Table 2).
1. Platelets.
After its generation from prothrombin, thrombin plays multiple roles in the blood coagulation cascade that are mediated by interaction with a number of physiological substrates, effectors, and inhibitors. The accumulation of thrombin at sites of vascular injury provides the recruitment of platelets into a growing hemostatic plug. When added to human platelets in vitro, thrombin causes platelets to change shape, stick to each other, and secrete the contents of their storage granules. How this is accomplished is still not fully understood, but a major step forward occurred since the identification of PAR1 in the 1990s. This finding shed light on the way by which an extracellular protease may initiate intracellular events.
Mammalian GTP-binding proteins (G proteins) fall into four families that are typically referred to by the designation of the -subunit: Gs, Gi, Gq/11, and G12/13. Human platelets express at least one member of the Gs family and four members of the Gi family (Gi1, Gi2, Gi3, and Gz), which, among other functions, stimulate or inhibit cAMP formation by adenylyl cyclase (199). In addition, platelets express one or more members of the Gq/11 family and members of the G12/13 family (G12 and G13) (199).
A number of different G protein -subunits have been shown to bind to PAR1, including members of the Gi, Gq/11 and G12/13 families (200). However, the knowledge of these interactions in humans under physiological and pathophysiological conditions is still far from completion.
Kim et al. (201) have demonstrated that thrombin and PAR1-AP, as well as PAR4-AP, induce Gi pathway stimulation in human platelets. Additionally, the authors demonstrated that thrombin, PAR1-AP, and PAR4-AP cause platelet aggregation independently of Gi stimulation (201). Offermanns et al. (202) also demonstrated a direct interaction between G12, G13, and PAR1. Additionally, it was reported that thrombin-mediated cleavage of the PAR1 (and PAR4; see Section IV.B) receptor leads to calcium-dependent and calcium-independent shape changes of human platelets in consequence of direct activation of both Gq and G12/13 pathways, respectively (115, 202). Therefore, current evidence suggests that PAR1 interacts with Gq, G12/13, and possibly Gi protein family members in human platelets. In turn, these data suggest an activation of downstream signaling cascade members after PAR1 stimulation on human platelets. Among such members are phospholipase C (PLC)-?, phosphoinositide 3 (PI3)-kinase, and monomeric G proteins.
Indeed, PI3 kinase was demonstrated to play an important role in some PAR1 or thrombin-mediated cellular effects such as cytoskeletal reorganization, alterations in cell motility, cell survival, and mitogenesis. For example, thrombin is able to activate multiple PI3-kinase isoforms, including the recently discovered 110-kDa isoforms that can be directly activated by G protein ?-subunits (145, 150, 203). Stimulation of platelets by PAR1 leads to the activation of PI3 kinase, which is dependent on the small G protein Rho (204). Recently, Trumel et al. (205) have demonstrated that PI3 kinase plays an important role in the PAR1-dependent reorganization of the platelet cytoskeleton via myosin heavy chain translocation and stable association of signaling complexes with the actin cytoskeleton. Interestingly, the activity of small G proteins such as Rac and cdc42 in platelets may be regulated through PAR1 stimulation, but the role of PI3 kinase in these events remains to be determined (206). In their recent study, Vaidyula and Rao (207) provided evidence that in human platelets PLC-?2 plays a major role in responses to PAR1 and PAR4 activation, and that PLC-?2 is required for the sustained rise in [Ca2+]i concentration upon thrombin activation.
Putting all of this together, current data suggest that thrombin activates human platelets by cleaving and activating PAR1 and PAR4. In turn, PAR1 activates the members of Gq/11, G12/13, and perhaps Gi families, leading to the activation of PI3 kinase, PLC-?, and monomeric G proteins (Rho, Rac, and possibly RapI), and also causes an increase of cytosolic Ca2+ concentration and inhibition of cAMP formation.
Additionally, it is important to note that despite the fact that human platelets express both PAR1 and PAR4, these receptors appear to be activated by different thrombin concentrations. Cleavage of human PAR4 requires a higher concentration of thrombin than does cleavage of PAR1, and it is likely that PAR1 is the predominant signaling receptor at low thrombin concentrations (208, 209).
Mouse platelets provide an interesting contrast to human platelets: whereas human platelets express functional PAR1 and PAR4, mouse platelets express PAR3 and PAR4, although in the latter case signaling appears to be mediated entirely by PAR4, with PAR3 serving to facilitate PAR4 cleavage at low thrombin concentrations (for details, see Section IV.A) (39, 40).
2. Cells in the nervous system.
PAR1, as mentioned above, appears to affect various processes in both the central and peripheral nervous systems (45, 210). Among such processes are: neuroinflammation and neurodegeneration, neuritogenesis, astrocyte proliferation, and synaptic plasticity. However, here we would like to focus on data concerning the involvement of PAR1 in intracellular signaling cascades in neuronal cells.
PAR1 may have an effect on various intracellular signaling cascades within neuronal cells (171, 181, 189). By coupling to different G proteins, PAR1 affects a wide range of neuronal cell functions. For example, the necessity of PAR1 coupling to G12 was demonstrated for thrombin-stimulated DNA synthesis in 1321N1 astrocytoma cells (211). Interestingly, injection of antibodies directed against G12 abolished the thrombin-stimulated DNA synthesis (212). Additional studies performed on 1321N1 astrocytoma cells revealed that thrombin treatment causes a concentration-dependent rounding. One may speculate that such an effect of thrombin could be Rho-dependent. The Rho family of small GTPases is known to be involved in the control of cytoskeletal changes via modulation of actin polymerization. Indeed, this thrombin-induced rounding of 1321N1 astrocytoma cells was Rho-dependent and mediated via G12 (213).
Moreover, LaMorte et al. (214) demonstrated binding of PAR1 to Gq by using the same cell line. The authors also showed that thrombin stimulation induces Ras-GTP complex formation and that Ras is required for PAR1-mediated activation of PLC in 1321N1 astrocytoma cells (214). This is especially interesting because thrombin is known to be a factor regulating the proliferation of astrocytes. This effect of thrombin appears to be associated with Go/i- and Gq-mediated pathways. Wang and coworkers recently shed some light on the downstream cascade events of PAR1 signaling (174, 210). Gq-Mediated PLC activity results in Ca2+ mobilization and activation of PKC, which phosphorylates a nonreceptor tyrosine kinase, proline-rich tyrosine kinase (Pyk2). Interestingly, the activation of other nonreceptor tyrosine kinases like Src and focal adhesion kinase (FAK) was also demonstrated after thrombin stimulation. Pyk2 is a factor that is able to connect GPCRs to ERK1/2 activation. Therefore, Pyk2, acting together with Src-tyrosine kinase, causes the activation of the ERK/MAPK pathway, which mediates the proliferative effect of thrombin. Additionally, the same group of authors demonstrated that the Go/i-mediated PI3 kinase pathway is also involved in thrombin-induced astrocyte proliferation (174).
Thrombin is also known to exhibit both beneficial and unfavorable effects in hippocampal neurons and astrocytes (178, 180, 215). Donovan and colleagues (184, 216) found that tyrosine kinases, serine/threonine kinases, and the actin cytoskeleton are involved in both pro- and antiapoptotic effects of thrombin in neuronal cells. However, there was no involvement of Go/i and the PI3 kinase pathway observed in these thrombin effects.
Additionally, Zieger et al. (217) demonstrated the existence of a novel PAR1-associated signaling pathway in the nervous system. According to these data, PAR1 participates in cAMP-independent PKA activation in SNB-19 glioblastoma cells. Moreover, PAR1 stimulation causes the activation of transcriptional factor nuclear factor B (NFB) in this cell type (217).
In summary, besides morphological and proliferative effects described above that are mediated via G12/13, Gq, and Go/i, respectively, thrombin also affects activation of transcriptional factors such as activator protein-1 (AP-1) in neuronal cells (214).
3. SMC.
Thrombin exerts direct effects on vascular cells such as SMC and endothelial cells via interactions with PARs (218). As mentioned in Sections II.H.1 and II.H.2, PAR1 interacts with G12/13, Gq, and Gi to elicit diverse downstream signaling events. For example, thrombin-stimulated DNA synthesis and cell migration are associated with activation of the G13 signaling pathway in SMC. The G13 signaling cascade includes the activation of Rho and thus induction of cytoskeletal changes affecting cell migration (219). The Gq-dependent signaling pathway includes the activation of PLC, which in turn leads to MAPK phosphorylation and receptor tyrosine kinase transactivation, both necessary events in thrombin-mediated proliferation. It is interesting that the thrombin-mediated activation of MAPKs such as ERK1/2 was recently demonstrated in SMC. As noticed, PAR1 caused rapid phosphorylation, whereas PAR4 induced prolonged phosphorylation of these kinases in SMC (220). Additionally, Ghosh et al. (221) have shown that thrombin activates p38 MAPK in a time-dependent manner in VSMC. Furthermore, Sabri et al. (222) demonstrated that PAR1 agonists induce activation of Jun N-terminal kinase (JNK) and Akt/PKB in rat ventricular cardiomyocytes.
After dissociation of the G protein heterodimer, G? interactions activate phosphoinositide 3-kinase, which promotes [Ca2+]i release that is required for SMC growth in response to thrombin stimulation (223).
It is also interesting that PAR1 signaling can modulate gene transcription induced by cytokines in SMC. Thrombin, acting via PAR1, can block IL-6-induced signal transducer and activator of transcription 3/Sis-inducible factor-A (Stat3/SIF-A) activation (224).
In summary, since the discovery of PARs, considerable progress in our understanding of thrombin signaling in SMC was achieved. This allowed some light to be shed on signaling events associated with SMC proliferation, migration, and synthesis of extracellular matrix proteins (e.g., collagen) after thrombin stimulation. However, additional studies are necessary to reveal the role of thrombin in SMC apoptosis, vascular lesion formation, and wound-healing associated pathways.
4. Endothelial cells.
Thrombin is known to affect various functions of endothelial cells. Among these are cell rounding, changes of cell-cell junctions, proliferation, barrier function, and permeability. These thrombin-induced effects appear to be mediated via PARs, particularly PAR1 (225, 226). A major step forward was reached in recent studies, which revealed important intracellular signaling events underlying PAR1-mediated effects in endothelial cells (227).
PAR1 was demonstrated to interact with Go/i, Gq, and probably G12/13 in endothelial cells (228, 229). As well as it was demonstrated for neuronal cells, PAR1-mediated cytoskeletal changes in endothelial cells (cell rounding) are associated with the activation of RhoA (225). Additionally, in the same work it was demonstrated that PAR1-AP stimulation rapidly enhanced vascular permeability in a mouse skin assay (225). Moreover, it was found that binding of PAR1 to pertussis toxin (PTX)-sensitive G proteins (however only to Go, but not to Gi) is also necessary for thrombin-induced changes of endothelial barrier permeability (229). The activation of the MAPK signaling pathway by thrombin in endothelial cells was also demonstrated. This activation plays a crucial role in thrombin-induced effects of endothelial cell functions such as chemokine and cytokine production as well as the expression of cell adhesion molecules (89, 230, 231).
The role of thrombin and PAR1 in the activation of transcriptional factor networks was intensively investigated in recent publications. Malik and coauthors (228, 231, 232) studied the involvement of PAR1 in the activation of NFB in endothelial cells. Their studies showed that Gq and the G? dimer are responsible for NFB activation and intercellular adhesion molecule-1 (ICAM-1) transcription in endothelial cells induced by the PAR1 agonist thrombin and the PAR1-specific AP TFLLRNPNDK. Furthermore, transfection experiments strongly supported simultaneous activation of G?/PKC-/p38 and G?/PI3-kinase pathways that converge into the Akt pathway, leading to subsequent NFB activation and ICAM-1 expression (228, 231, 232). Moreover, thrombin-induced stimulation of vascular cell adhesion molecule-1 (VCAM-1) production involves the inducible binding of p65 NFB to a tandem NFB motif in the 5' flanking region (233). Taken together, these findings suggest that in the case of ICAM-1 and VCAM-1, p65 NFB is necessary for transducing the thrombin response in endothelial cells.
Additionally, thrombin-mediated induction of VCAM-1 was shown to involve the inducible binding of GATA-2 to a tandem GATA motif in the upstream promoter region (234). Interestingly, the effect of thrombin on GATA-2 DNA binding and transcriptional activity was found to be mediated by a PI3K, PKC--dependent signaling pathway (233).
It is important to note that thrombin effects on transcriptional factor networks have been more deeply investigated in endothelial cells than in other cell types. These data explain, at least in part, thrombin-mediated effects on the production of cell adhesion molecules and some chemokines in endothelial cells. Subsequently, this accounts for thrombin effects at leukocyte migration via endothelial barrier (ICAM-1 serves as a ligand for leukocyte ?2 integrins and promotes leukocyte adhesion) and leukocyte recruitment.
5. Immune cells.
The participation of PARs in T cell signaling pathways has also been demonstrated (235). Recently, Bar-Shavit et al. (236) have examined a possible involvement of Vav 1 in PAR-mediated signaling in human Jurkat T cells. The Vav family has three known members in mammalian cells (Vav, Vav2, and Vav3) and one in nematodes (CelVav) (237). Tyrosine phosphorylation of Vav1 regulates its activity as a guanine-nucleotide exchange factor for the Rho-like small GTPases RhoA, Rac1, and cdc 42, which affect cytoskeletal reorganization and activation of stress-activated protein kinases/JNKs. Bar-Shavit et al. (236) clearly showed that activation of PARs induces tyrosine phosphorylation of Vav1 in Jurkat T-leukemic cells. Because -chain-associated protein kinase of 70 kDa (ZAP-70) and SH2-domain containing leukocyte-specific phosphoprotein of 76 kDa (SLP-76) have been shown to associate physically with Vav1 and because this association is critical for normal functioning of T cells, it was tempting to investigate whether ZAP-70 and SLP-76 are involved in PAR-induced signaling cascades. Indeed, an increase of tyrosine phosphorylation of ZAP-70 and SLP-76 was observed after activation of Jurkat cells with PAR-APs. Moreover, phosphorylation of Vav1 in response to PAR stimulation was dependent on p56lck (236). Unfortunately, because nonselective PAR-APs were used in the studies done with the Jurkat cells (236), it is difficult to distinguish between the effects of PAR1 and PAR2 in this T cell model system. Furthermore, because of the lack of the ability of PAR3-derived peptides to activate PAR3 (39), the role of PAR3 in T cell signaling remains unknown.
6. Factor Xa signaling mediated via PAR1.
Most of the PAR1-associated signaling events have been observed subsequent to cell stimulation by thrombin. However, recently a possible role of PAR1 in coagulation factor Xa-mediated signaling has been demonstrated (238, 239, 240). The coagulation proteinase factor Xa is generated at sites of vascular injury and inflammation after formation of the tissue factor/VIIa (TF/VIIa) complex. Coagulation factor Xa has been shown to be mitogenic for SMC (238) and elicits inflammatory responses in endothelial cells. Riewald and Ruf (241) have presented new data indicating the involvement of PAR1 in Xa-mediated signaling. They used HeLa cells expressing only PAR1 and demonstrated that factor Xa induces NFB activation and MAPK phosphorylation in these cells. Inhibition studies with specific antibodies revealed that factor Xa responses were mediated via PAR1 (241). Thus, factor Xa (or potentially other serine proteinases) may substitute for thrombin in proteolytic signaling via PAR activation. In endothelial cells, which also express PAR2, Camerer et al. (240) have shown that factor Xa signaling was mediated not only by PAR1, but also to a large extent via PAR2. Together, PAR-1 and PAR-2 appear to account for more than 90% of factor Xa signaling in endothelial cells (240).
In summary, the multiple biological as well as inflammatory and immune responses of thrombin that are mediated by PAR1 [including 1) vasoregulation; 2) increased vascular permeability; 3) cellular adhesion and infiltration of leukocytes; 4) angiogenesis; 5) stimulation of the production of inflammatory mediators such as cytokines, neuropeptides, NO and prostanoids, for example; 6) regulation of extracellular matrix proteins; and 7) induction of signal transduction pathways which are involved in immunomodulation] suggest an important role of PAR1 during inflammation and immune response.
III. PAR2 in Inflammation and Immune Response
Top
Abstract
I. Introduction
II. PAR1 in Inflammation...
III. PAR2 in Inflammation...
IV. PAR3 and PAR4
V. Conclusions
References
PAR2 is expressed in brain, lia (DRG), eye, airway, heart, GI tract, pancreas, kidney, liver, prostate, ovary, testes, and skin (21, 22, 24, 154, 242, 243, 244, 245) and is found in various cell types such as epithelial cells, endothelial cells, SMC, osteoblasts, as well as immune cells such as T cells, neutrophils, mast cells, or eosinophils (97, 154, 246, 247, 248, 249, 250, 251, 252, 253, 254). On the other hand, platelets do not express PAR2 (248). Recent findings point to an important role for PAR2 under physiological and pathophysiological conditions in many tissues (Table 2). However, the endogenous enzymes responsible for activating PAR2 in many tissues remain to be determined. Many endogenous or exogenous trypsin-like enzymes may cleave and activate PAR2. Expression of trypsinogen-2 mRNA and its translation product has been demonstrated in endothelial cells (255). Interestingly, various types of human cancer cells secrete enzymes with trypsin-like specificity (255) that may activate PAR2. In human skin, trypsinogen-4 generated by keratinocytes and trypsinogen-2 from human dermal microvascular endothelial cells can activate PAR2 in vitro (M. Steinhoff, unpublished observation). Another candidate is mast cell tryptase, a major secretory protein of human mast cells. Mast cells from mice and rats are more heterogeneous regarding their protease content, although they also produce proteases with tryptic specificity. Tryptase has been shown to activate PAR2 on epithelial as well as endothelial cells and neurons (23, 196, 243, 251, 256). The observation that tryptase can activate PAR2 suggests a role of this receptor in humans under circumstances when mast cells are involved, e.g., during inflammation, hypersensitivity reactions, and wound repair (24, 257). However, the ability of human tryptase to activate PAR2 appears to be restricted by receptor glycosylation at an N-terminal residue just proximal to the receptor’s cleavage activation site (258, 259). Notwithstanding, the effects of tryptase on cells in vitro often mimic those of PAR2 activation: tryptase up-regulates IL-1? and IL-8 secretion, enhances the presence of intracellular adhesion molecules/selectins on endothelial cells, mediates accumulation of neutrophils and eosinophils, produces vascular leakage, and is mitogenic for epithelial cells, fibroblasts, and SMC (260, 261, 262, 263). That said, direct proof that PAR2 mediates the effects of tryptase during inflammation in vivo is still lacking.
A. Vasculature
In the vasculature, PAR2 has many effects that are proinflammatory. Agonists of PAR2 induce relaxation in the rings of rat aorta or porcine coronary artery dependent on endothelial NO synthase activity (246, 248, 250). This effect is abolished in the absence of endothelium (246, 250). In contrast, trypsin stimulates contraction of the rabbit aorta in the absence of endothelium (264). In the intact rat, iv injection of SLIGRL-NH2 produces a marked fall in blood pressure, consistent with release of NO from endothelial cells (248). Furthermore, PAR2 agonists increase IL-6 production (72), induce von Willebrand factor release, and serve as a mitogen for human umbilical vein endothelial cells (HUVEC) (97, 265, 266).
Moreover, it was demonstrated that some inflammatory mediators are able to affect the expression of PAR2 in endothelial cells. In HUVEC, PAR2 mRNA is up-regulated by TNF- and IL-1, cytokines that act together to orchestrate the acute inflammatory response (25, 265).
In summary, PAR2 mediates vasodilation, plasma protein extravasation, as well as endothelial cell proliferation in the cardiovascular system. Thus, this receptor can be regarded as an important factor in neovascularization. Moreover, PAR2 can be considered as a vascular sensor for trypsin-like proteinases of the coagulation cascade, which play an important role in cardiovascular medicine.
B. Immune cells
Although information has been acquired about the role of PAR2 in epithelial and endothelial function, relatively little is known so far about the role of this receptor in the immune system. As was recently demonstrated, PAR2 is expressed by various cells involved in immune response, such as T cell lines, eosinophils, neutrophils, and mast cells (147, 160, 249, 252, 267). Accumulating evidence points to a role of PAR2 in the regulation of leukocyte function. Some of the PAR2- mediated effects on leukocyte behavior have been observed in vivo in rodents (262). In this model, it has been reported that the PAR2-AP (SLIGRL-NH2) caused a significant increase in leukocyte migration into the peritoneal cavity after ip injection. Furthermore, PAR2-APs induced a significant increase in leukocyte rolling and adherence by a mechanism depending on the release of platelet-activating factor (262). Lindner et al. (268) have also demonstrated the importance of PAR2 for leukocyte adhesion and rolling. In a model of acute inflammation induced by tissue trauma, they showed that PAR2-deficient mice have a decreased leukocyte rolling capacity compared with wild-type mice. Conversely, activation of PAR2 by a specific PAR2-AP in wild-type mice induced a significant reduction of leukocyte rolling velocity, an increased leukocyte rolling flux as well as increased leukocyte adhesion (268).
Howells et al. (252) reported that a specific PAR2-AP as well as trypsin are able to activate PAR2 on human neutrophils. PAR2 activation caused shape changes and up-regulation of CD11b/CD18 in these cells. In a coculture of human neutrophils with endothelial cells, PAR2 agonists induced L-selectin shedding and CD11b/CD18 up-regulation in neutrophils (268). We demonstrated that PAR2 agonists induce [Ca2+]i release in human neutrophils and enhance neutrophil motility in 3-D collagen gel lattices (269). Moreover, Lourbakos et al. (27) demonstrated that a bacterial proteinase, gingipain-R from P. gingivalis, activates PAR2 on human neutrophils. Because many bacteria with pathogenicity in humans produce serine proteinases with trypsin-like activity, it can be assumed that certain pathogenic effects of bacteria in many tissues can be mediated through PAR2. For example, Lourbakos et al. (28) reported that activation of PAR2 by bacterial gingipain R leads to IL-6 secretion in human oral epithelial cells. In mouse airways, PAR2 agonists were capable of inhibiting the recruitment of neutrophils induced by bacterial proteinase (270), suggesting a direct role of PAR2 on neutrophil function in vivo. From these data, one may speculate that PAR2 may serve as a receptor for serine proteinases derived from bacteria, viruses, or even fungi and may thus directly trigger inflammatory or immune responses in vivo induced by microorganisms. However, this intriguing hypothesis remains to be verified.
Recent work has shed some light on the role of PAR2 on eosinophil function. Miike et al. (267) clearly demonstrated that human eosinophils express functional PAR2. Furthermore, trypsin as well as a specific PAR2-AP was able to induce degranulation and superoxide production in human eosinophils. More recently, Temkin et al. (79) reported that human mast cell tryptase induced up-regulation of IL-8 mRNA and caused IL-8 release from human eosinophils. A comparable effect of tryptase on eosinophil IL-6 expression and release was also observed. This effect of tryptase appears to be activated via the protein kinase/AP-1 signaling pathway (79). Schmidlin et al. (271) have also demonstrated that PAR2 mediates infiltration of eosinophils and hyperreactivity in allergic inflammation of the airway. Furthermore, PAR2 activation releases survival factors such as granulocyte-macrophage colony-stimulating factor (GM-CSF) from eosinophils (272). Together, these results are clearly in favor of an important role of PAR2 in eosinophil regulation and in inflammatory/allergic diseases in which eosinophils are involved.
Several T cell lines also express PAR2 (244). Trypsin as well as a synthetic PAR2-AP induce [Ca2+]i mobilization in the Jurkat and HPB.ALL T cell lines (249). In Jurkat cells, the involvement of PARs, possibly PAR2, in Vav1-mediated signaling was clearly demonstrated (236).
The localization of PAR1 and PAR2 on human mast cells was first reported by D’Andrea et al. (147). As shown by immunohistochemistry, PAR2 was localized not only on the cell membrane of mast cells, but also on the membrane of the intracellular tryptase-positive granules (147). Moreover, rat peritoneal mast cells also express PAR2 mRNA (273). However, the precise role of PAR2 in mast cells especially with regard to a potential autocrine regulatory mechanism via tryptase activation requires additional investigation. In summary, these results provide evidence for a role of serine proteinases in directly regulating immune cells and immune responses via PARs. An understanding of the molecular mechanisms by which PAR2 is involved in immune regulation will aid the design of new antiinflammatory drugs that can target proteinase-activated receptors.
C. Arthritis
Arthritis is a chronic inflammatory disease that is characterized by joint swelling, vasodilatation, edema, hyperemia, pain, and recruitment of inflammatory cells, especially neutrophils and macrophages, but also lymphocytes. Very recently, it was demonstrated that PAR2 participates in the pathophysiology of arthritis (274). Using PAR2-deficient mice, the authors showed that PAR2 deficiency results in a marked reduction of swelling responses of experimentally induced monoarthritis. However, no differences were observed concerning joint damage when compared with wild-type mice. Using an enzymatic approach (?-galactosidase), they also showed positive ?-gal staining in endothelial cells of blood vessels. Two weeks after induction of monoarthritis, however, a significant increase of ?-gal staining of extravascular inflammatory cells was observed in PAR2+/+ mice compared with PAR2–/– mice. Examination of the knee joint showed dramatic arthritic changes in PAR2+/+ mice characterized by synovial hyperplasia, an enhanced inflammatory infiltrate, and cartilage damage. Thirty days after the induction of monoarthritis, the cartilage was completely replaced by pannus in PAR2+/+ mice. In contrast, PAR2–/– mice showed an intact appearance of the cartilage similar to normal and control joints. This is in favor of an important role of PAR2 as a mediator of proinflammatory responses in the joints. Moreover, a novel synthetic PAR2 peptide agonist (ASKH95, phenylacetyl-LIGKV-OH) revealed proinflammatory effects after intraarticular injection like synovial hyperemia and joint swelling. These results clearly demonstrate that ASKH95 induces signs of chronic inflammation such as long-lasting swelling responses, vasodilatation, and tissue destruction. Interestingly, ASKH95 showed a longer-lasting proinflammatory effect in the tissue compared with the classical SLIGRL PAR2 agonist peptide. This may be due to different rates of degradation, longer receptor activation by ASKH95, or most likely different lipophilicity. The underlying mechanisms of the PAR2-mediated inflammatory effect in this chronic disease, however, are not fully explored yet. Potential mediators may be cytokines or prostanoids, for example. Thus, PAR2 may play an essential role in the pathophysiology of arthritis, and PAR2 antagonists may be beneficial for the treatment of this chronic inflammatory disease.
D. Skin
PAR2 may play an important role in cutaneous inflammation. Keratinocytes and dermal endothelial cells express functional PAR2 and are important targets of inflammatory mediators (247, 275, 276). In humans, PAR2 is also expressed by dermal immunocompetent cells, which so far have not been characterized (24), and by dermal sensory nerves (277). Potential endogenous activators of PAR2 in human skin are mast cell tryptase, proteinases of the trypsin-family produced by keratinocytes (trypsinogen-4), and proteinases of the fibrinolysis cascade, such as factor VII/Xa, which could be released during inflammation and wound healing. Exogenous activators of PAR2 may be serine proteinases generated by bacteria, fungi, and house dust mites, although direct evidence on keratinocytes is still lacking. Indirect evidence for bacterial-induced activation of PAR2 is coming from the observation that bacterial gingipain activates the receptor on buccal keratinocytes (28). At certain stages of atopic dermatitis or psoriasis, mast cells are in close contact to basal keratinocytes or are recruited into the epidermal layer (278). Numerous mast cells are found in the dermis in close proximity to blood vessels and at the dermal-epidermal border in atopic dermatitis and psoriasis. Thus, tryptase or other skin-derived proteinases may activate PAR2 to induce inflammatory changes and thereby contribute to the pathophysiology of atopic dermatitis and psoriasis. It is well known that intradermal injection of tryptase into the skin results in vasodilatation and erythema, followed by leukocyte infiltration and local induration (279), indicating a role of tryptase in cutaneous inflammation. Moreover, tryptase is an important long-term marker in body fluids for systemic anaphylaxis and other immediate hypersensitivity allergic reactions (280). Thus, it is possible that PAR2 mediates the systemic effects of tryptase in certain skin diseases.
Serine proteinases other than tryptase may also activate PAR2. Cytotoxic T cells, which express PAR2 (249), also produce the trypsin-like enzyme granzyme A (281), suggesting a potential role for PAR2 in regulating T cells in inflammatory dermatoses. In summary, proteinases in the skin are signaling molecules that may directly regulate the function of target cells by activating PAR2. PAR2 as well as PAR1 agonists regulate cytokine production such as IL-8 (282), IL-6, and GM-CSF (283) in cultured human keratinocytes. Furthermore, PAR2 may also regulate proliferation and differentiation in the skin. Similar to TGF-?, PAR2 agonists inhibit proliferation or differentiation of human neonatal keratinocytes, whereas PAR1 agonists stimulate proliferation (275). In contrast, agonists of both PAR2 and PAR1 are mitogenic for endothelial cells (248). Thus, our observations in normal and inflamed human skin suggest a functional role of PAR2 in different cell types of human skin during inflammation.
In human skin, PAR2 is only weakly expressed by microvascular endothelial cells. Studies done in vitro revealed that PAR2 is functionally expressed by human dermal microvascular endothelial cells and that PAR2 agonists induce up-regulation and release of IL-6 and IL-8 in a concentration-dependent fashion (263). Moreover, PAR2 regulates expression of cell adhesion molecules in primary cultures of human endothelial cells (284) and cell lines (268). These findings are supported by in vivo findings demonstrating a role of PAR2 in the adhesion, rolling, and extravasation of leukocytes in rat venules. Finally, PAR2 appears to play a role in the regulation of leukocyte/endothelial interactions in humans and mice in vivo, as shown for atopic dermatitis or experimentally induced contact dermatitis (285). Together, these results strongly support the idea that PAR2 plays an important role in leukocyte/endothelial interactions during inflammation and immune response.
The knowledge about the signaling cascades involved in PAR2-induced inflammatory and immune responses by keratinocytes is still incomplete. Recently, Kanke et al. (286) have shown that PAR2 agonists stimulate activation of inhibitory B kinases (IKK) and IKK?, and also stimulate NFB-DNA binding. Our own data using primary human keratinocytes also clearly demonstrate that PAR2 activates NFB (287). Moreover, PAR2 is directly involved in the NFB-mediated regulation of keratinocyte ICAM-1, which is an important molecule for the regulation of keratinocyte immune responses such as host defense, allergy, and recruitment of inflammatory cells into the epidermis (287).
Finally, recent data suggest a role of PAR2 during cutaneous inflammation in vivo. In a murine model of experimentally induced allergic contact dermatitis, a proinflammatory role of PAR2 agonists was demonstrated (285, 288). Stimulation of PAR2 resulted in increased ear swelling responses with a maximum after 48 h. Independently, our group confirmed the proinflammatory effects of PAR2 agonists in a model of experimentally induced allergic and irritant contact dermatitis using PAR2-deficient mice. Moreover, we examined the underlying mechanisms responsible for PAR2-induced effects during contact dermatitis (285). These results clearly indicate that PAR2 is involved in edema formation, plasma extravasation, up-regulation of cytokines (IL-6), cell adhesion molecules (ICAM-1), and selectins (E-selectin) in mice. Intravital microscopy studies further showed that velocity and adhesion of leukocytes is impaired in PAR2-deficient mice compared with controls. Moreover, in vivo studies by microdialysis confirmed a role of PAR2 in mediating vascular responses such as edema formation and plasma extravasation also in human skin. Finally, ex vivo studies in skin biopsies of human volunteers are in favor of a role for PAR2 in selectin regulation of dermal blood vessels because PAR2 agonists induced a marked increase of E- selectin immunoreactivity in dermal endothelial cells in comparison to control tissues. Together, these results support a regulatory role of PAR2 during cutaneous inflammation and immune response.
E. Airways
Several observations are in favor of an important role of PAR2 in airway inflammation. Firstly, tryptase (251, 289) seems to play an important role in airway inflammation and hyperresponsiveness (253). Secondly, PAR2 is markedly up-regulated after exposure to proinflammatory stimuli or cytokines (265) that have been shown to play a role in chronic airway diseases. Morphological studies have demonstrated PAR2 immunoreactivity in endothelium and smooth muscle of bronchial vessels. These observations are consistent with a similar distribution in other tissues and with the known ability of PAR2 agonists to cause arterial relaxation as well as contraction in certain arteries (290). Moreover, PAR2 is expressed by bronchial, tracheal, epithelial and SMC which may result in direct or indirect bronchomotor effects leading to pathophysiological conditions in the airways such as asthma or chronic obstructive pulmonary disease. Interestingly, PAR2 agonists are capable of causing bronchoconstriction (291). Whether this response is dependent on secondary effects such as the generation of bronchoactive peptides or other mediators remains to be determined. However, studies done in vitro suggest further that PAR2 agonists also produce a relaxant effect in isolated main bronchi (292). Thus, PAR2 agonists may cause dose-dependent bronchoconstrictor or bronchodilatator effects in the airways, with high agonist concentrations favoring bronchoconstriction. NO and prostanoids like PGE2 may be involved in this process, because potentiation of PAR2-induced bronchoconstriction has been observed upon inhibiting NO synthase or after prostanoid generation and COX-2 activation (153). In conclusion, there is strong evidence that PAR2 stimulation activates contractile and dilator mechanisms in the airways. The reason why the predominant effect in guinea pig airways in vivo is bronchoconstriction, however, is not known at present. In addition to the potentially bronchoprotective effects observed by Cocks et al. (292) in rat bronchial preparations, others have shown that PAR2 may also mediate bronchoprotection in guinea pig airways (293). However, other observations show that PAR2 can act as an enhancer of histamine-mediated contraction (294). The study of Cocks et al. (292) showed that trypsin colocalized with PAR2 in airway epithelial cells, where PAR2 activation resulted in relaxation of airway preparations by the release of cytoprotective PGE2. Prostanoids may be important mediators of PAR2 activity in the airways not only in the lung, but also in the GI tract (152, 153, 253, 295). It is tempting to speculate that tryptase and/or other proteinases with trypsin-like activity, released in close proximity to the cells expressing PAR2, may stimulate this receptor during airway inflammation. Finally, intensive staining of PAR2 in the apical region of the airway epithelium might suggest an activating role of proteinases, for instance bacterial proteinases, present in the airway lumen (296). Very recently, PAR2 was shown to stimulate the release of MMP-9 in a human epithelial cell line (297), providing a strong hint that PAR2 is involved in the orchestration and reorganization of lung extracellular matrix proteins.
It was also revealed that PAR2 stimulation of airway epithelial cells leads to the release of eosinophil survival-promoting factors such as GM-CSF (272). Moreover, in human respiratory epithelial cells, PAR2, like PAR1, stimulates IL-6, IL-8, and PGE2 release (77).
Profound evidence for an involvement of PAR2 in allergic inflammation of the airways is provided by the recent work of Schmidlin et al. (271). To study the involvement of PAR2 in airway inflammation, they used PAR2-deficient mice and mice overexpressing human PAR2 (PAR2tg). PAR2 is known to be up-regulated by inflammatory agents. Sensitization of wild-type mice by injection of ovalbumin led to infiltration of immune cells into the lumen of the airways and induced hyperreactivity, whereas both infiltration of eosinophils into the lumen and airway hyperreactivity was exacerbated in PAR2tg mice. In contrast, infiltration of immune cells and airway hyperreactivity was markedly diminished in PAR2-deficient mice. Additionally, in PAR2–/– mice the IgE response was strongly reduced, implicating a role of PAR2 in immune response. Remarkably, intranasal administration of PAR2-AP in wild-type mice stimulated increased recruitment of macrophages, confirming the proinflammatory role of PAR2 in the airways, whereas altered PAR2 expression mainly influenced eosinophil infiltration.
In summary, evidence exists for a proinflammatory as well as antiinflammatory role of PAR2 in airway inflammation. This dual role may be explained by the various tissues, cells, methodological approaches, and inflammatory model systems (acute vs. chronic, mouse strains) that have been used to evaluate this role of PAR2 in the airways. So far, it appears that PAR2 is a sensor receptor that releases proinflammatory mediators in the early phase and antiinflammatory molecules in the late phase of "regulated inflammation." Under "dysregulated" inflammatory situations such as chronic disease states, PAR2 may exert antiinflammatory effects, as shown in a colitis model (298), or proinflammatory effects, as demonstrated in an asthma model (271). Thus, additional studies using human and mouse (knockout) models of pulmonary dysfunction are necessary to fully explain the role of PAR2 during inflammation and immune response.
F. Brain and peripheral nervous system
By immunohistochemistry, PAR2 has been localized in various compartments of the nervous system such as brain, spinal cord, DRG, and peripheral nerves, with different receptor densities found during various stages of mouse development (299). At d 14, PAR2-staining was observed throughout the mouse brain. Among the hippocampal formation, PAR2 was localized in the subiculum, in pyramidal cells throughout the CA1, CA2, CA3, and hilus region, and in granular cells of the dentate area. Additionally, staining for PAR2 was observed in all cortical layers, amygdaloid nuclei, striatum, thalamus, and hypothalamus (175). Peripheral nerves were intensively stained after d 14. Similarly, PAR2 has also been detected in human brain (276). In rat brain, PAR2 is also expressed by neurons of the hippocampus (175, 300), meningeal cells (301), as well as astrocytes (176, 185). PAR2 is abundantly expressed in various subareas and is transiently up-regulated during oxygen and glucose deprivation (175). Moreover, PAR2 activation is cytotoxic for isolated rat hippocampal neurons in a concentration-dependent manner (300).
Domotor et al. (302) recently reported that agonists of PAR2 induce Ca2+ responses in brain microvascular endothelial cells. This effect may be modulated by elastase and plasmin, which regulate PAR2 signaling. Studies of the actions of PAR2 agonists on CNS targets suggest an important role of PAR2 during injury, growth, apoptosis, and probably memory.
Very recently, PAR2 has been localized on rat sensory neurons (243). Moreover, functional data strongly support the idea that the peripheral nervous system is directly regulated by PAR2 during neurogenic inflammation. Neuropeptides such as calcitonin gene-related peptide and substance P (SP) from primary spinal afferent neurons are known as important mediators of neurogenic inflammation in many organs. Because of the close proximity of tryptase-containing mast cells to spinal afferent fibers, and because agonists of PAR2 cause effects similar to those of tryptase in many tissues, comprising many of the characteristics of neurogenic inflammation, it was speculated that serine proteinases or other PAR2 agonists may activate PAR2 on sensory neurons to stimulate neuropeptide release and may thus regulate inflammation. Indeed, it was shown that agonists of PAR2 can induce inflammation by a neurogenic mechanism that depends on the release of calcitonin-gene related peptide (CGRP) and SP from primary spinal afferent neurons. However, there also appears to be a nonneurogenic component of PAR2-induced inflammation because granulocyte infiltration triggered by PAR2 in a paw edema inflammation model was unaffected by sensory denervation or by neuropeptide receptor antagonists. Thus, PAR2 agonists may also act directly on endothelial cells or on neutrophils themselves to stimulate inflammation-related granulocyte adhesion and infiltration (14, 250). It is possible that serine proteinases may activate PAR2 on spinal afferent neurons under pathophysiological inflammatory circumstances. In a similar manner, serine proteinases may regulate enteric neuronal function by cleaving PAR2 in a Ca2+-dependent manner (196). Trypsin, which is released by airway epithelial cells, but not the PAR2 agonist peptide sequence SLIGRL, induced a marked contraction in guinea pig bronchi via SP release. Thus, non-PAR mediated effects of trypsin may also regulate neuropeptide function in the lung (303).
Several reports strongly suggest a role of PAR2 in the central transmission of pain and nociception in rats and mice (304, 305, 306). Intraplantar administration of low doses of PAR2 agonists, which do not induce inflammation, resulted in a long-lasting hyperalgesia. Interestingly, these effects of PAR1 agonists were more efficacious in comparison with other inducers of hyperalgesia such as PGE2 (0.3 μg). In the spinal cord, PAR2 agonists induced an up-regulation of c-fos, a marker of activated nociceptive neurons (305, 307, 308). Another study also reported a role of PAR2 in a rat visceral pain model (308). PAR2 agonists administered in vivo clearly increased abdominal colonic contractions. This effect was inhibited by an NK1 receptor antagonist, but not by the PG inhibitor indomethacin (308). However, additional studies are necessary to fully explain the underlying direct or indirect effects of PAR2-induced nociception. For example, neuropeptides released from neurons upon PAR2 stimulation may activate the release of nociceptive mast cell mediators such as kinins or prostanoids (309).
The observation that PAR2 agonists are involved in the transmission of neurogenic inflammation and pain subsequently draws the question whether PAR2 may be involved in the central transmission of pruritus. Itching is one of the most frequent symptoms in dermatological diseases and accompanies inflammatory and immune responses of many diseases such as hypersensitivity reactions or urticaria, for example. Indeed, neuronal PAR2 appears to be involved in the induction of pruritus in human skin (277). Moreover, the endogenous PAR2 agonist tryptase was increased up to 4-fold in atopic dermatitis patients, and PAR2 expression was markedly enhanced on primary afferent nerve fibers in skin biopsies of atopic dermatitis patients. Intracutaneous injection of specific PAR2 agonists provoked enhanced and prolonged itch when applied intralesionally. Thus, PAR2 activation on cutaneous sensory nerves may be a novel pathway for the transmission of itch and inflammatory responses during atopic dermatitis and probably other skin diseases. PAR2 antagonists may be promising therapeutic targets for the treatment of cutaneous neurogenic inflammation and pruritus (277).
In summary, PAR2 may play an important regulatory role in the central and peripheral nervous system in normal and disease states, including neurogenic inflammation.
G. Digestive tract and pancreas
In rats, Kawabata et al. (167, 310, 311) detected PAR2 mRNA in the sublingual, submaxillary, and parotid salivary glands. Of these three distinct salivary glands, the sublingual exhibited the strongest responses to PAR2 activation resulting in the secretion of mucin. This response appeared to be mediated in part via a tyrosine kinase signal pathway, because genistein, an inhibitor of several tyrosine and other protein kinases, attenuated the effect. Pretreatment with the alkaloid capsaicin, which activates the transient receptor potential of vanilloid type 1 and which, in turn, is known to play a major role in inflammatory thermal nociception, failed to abrogate the secretion of mucin from salivary glands but abolished the PAR2-triggered cytoprotective secretion of mucus in the stomach. These results provide both a sensory neuron-independent and -dependent mechanism for PAR2-regulated exocrine secretion (312). Additional support for a role of PAR2 in salivary gland secretion was provided by data showing that a PAR2 agonist can stimulate amylase secretion from rat parotid gland slices in vitro (311).
In the GI tract, PAR2 is highly expressed by enterocytes, where it is localized on the apical and basolateral membranes (253, 313). Moreover, myocytes of the muscularis mucosae and muscularis externa as well as neuronal elements are immunoreactive for PAR2. Recent observations indicate that trypsin may regulate enterocytes by cleaving and triggering PAR2 at the apical membrane (253) to induce the generation of PGE2 and PGF1a, suggesting that PAR2 may have a cytoprotective as well as an inflammatory effect on the GI tract. Moreover, mucosal mast cell proteinases may possibly activate PAR2 on enterocytes and colonic myocytes. PAR2-mediated inhibition of intestinal motility may contribute to inflammatory conditions in which mast cells are involved.
Trypsin may activate PAR2 in the pancreas itself under physiological and pathophysiological conditions because trypsin can be prematurely activated in the inflamed pancreas (314). PAR2 seems to play an important role also in pancreatic nociception because injection of an AP specific for PAR2 can activate and sensitize pancreas-specific afferent neurons in vivo (314). Moreover, Hoogerwerf et al. (314) found that stimulating DRG with a PAR2-AP resulted in enhanced capsaicin- and KCl-stimulated release of calcitonin gene-related peptide, which is a marker for nociceptive signaling.
PAR2 is expressed by both acinar cells, which release digestive enzymes, and duct cells, which produce fluid and bicarbonate. In isolated pancreatic acini, trypsin and PAR2 agonists stimulate amylase release (154) (N. W. Bunnett, personal communication). In vivo studies revealed the secretion of pancreatic juice after PAR2 activation (167). Interestingly, PAR2 agonists applied to the basolateral but not apical membrane of monolayers of pancreatic duct cells increase short circuit currents due to activation of Ca2+-sensitive Cl– and K+ channels (315). These effects may be of relevance in pancreatitis when trypsin is released across the basolateral membrane.
H. Signaling by proteinases via PAR2
There are relatively few studies examining the involvement of PAR2 in signaling cascades compared with the relatively large number of studies concerning PAR1-mediated intracellular signaling (13, 14). So far, there are only indirect data indicating that PAR2 interacts with Gq/G11 (activation of calcium signaling) and possibly with G0/Gi (316). Whether or not PAR2 binds to other G proteins, such as G12 or G13, remains unclear.
Interaction of PAR2 with Gi and Gq suggests a subsequent activation of PLC, PKC, and MAPK pathways. These signaling pathways can affect various cell activities including cell proliferation, morphological changes, motility and survival, and gene transcription regulation (Table 2 and Fig. 3). Indeed, as demonstrated for neuronal cells and SMC, stimulation by tryptase as well as PAR2-AP leads to subsequent activation of PLC and PKC (317, 318). Kanke et al. (286) demonstrated that PAR2 agonists (trypsin as well as the AP SLIGKV-NH2) stimulate JNK and p38 MAPK activation in a human keratinocyte cell line (NCTC2544). Moreover, Vouret-Craviari et al. (225) described PAR2-induced activation of RhoA in HUVEC revealing a possible mechanism underlying PAR2-mediated effects at cytoskeleton in macrovascular endothelial cells. Together, these data shed some light on the potential signaling mechanisms underlying PAR2-associated cellular effects such as depolarization response in neuronal cells, proliferation and mitogenic response in SMC, proliferation and differentiation of keratinocytes, and cytoskeletal changes in endothelial cells (225, 275, 319, 320).
View larger version (47K):
[in this window]
[in a new window]
FIG. 3. Diagram showing the major G protein-mediated signaling pathways coupled to PAR2. Dashed lines or circles represent signaling pathways or intermediates that are not fully revealed to be activated by PAR2, but are in favor of other G protein-coupled receptors or are typical intermediate-accompanying molecules. Drawing represents a composition of signaling events of different tissues, cell types, and species (also see Table 2).
Additionally, the effects of trypsin as well as PAR2-AP on the activation of nuclear transcriptional factors have recently been demonstrated. It was shown that PAR2 agonists stimulate NFB-DNA binding activity and activation of upstream kinases IKK and IKK? (286). The effects of PAR2 stimulation on NFB or IKK also have been described in other studies performed on different cell types (13, 263, 287, 297, 321).
It is also important to pay attention to factor Xa signaling mediated via PAR2. It was mentioned in Section II.H.6 that this factor induces signaling events via PAR1; the potential role of PAR2 in such signaling remained unclear for a long time. Recently, however, the involvement of PAR2 activation in coagulation factor Xa signaling has been reported. In HUVEC, factor Xa interacts with a protein called effector cell proteinase receptor-1 (EPR-1). This interaction is associated with signal transduction, generation of intracellular second messengers, and modulation of cytokine gene expression. Inhibitors of factor Xa blocked these responses (322). Additionally, direct cleavage of PAR2 by factor Xa has been demonstrated, reflecting the complexity of PAR2-induced signaling. These data allowed the authors to suggest that factor Xa induces endothelial cell activation via a novel cascade of receptor activation involving docking to EPR-1 and proteolytic cleavage of PAR2 (322).
As already mentioned extensively in Section II.H, PAR1 and PAR2 both account for around 90% of endothelial factor Xa-mediated signaling in mice. In contrast, PAR1 virtually accounts for all factor Xa-induced activation in fibroblasts, indicating potential cell-specific synergistic effects of PAR receptors on target cells (25).
Moreover, Koo et al. (323) demonstrated an involvement of PAR2 in factor Xa signaling by analyzing factor Xa-induced stimulation of coronary artery SMC using cell proliferation and ERK1/2 activation as indices of response. Furthermore, factor Xa-induced ERK1/2 activation was not desensitized by preincubation of the cells with thrombin. However, ERK1/2 activation was markedly attenuated by prior exposure of the cells to PAR2-AP (SLIGKV-NH2). The mitogenic effect of factor Xa was significantly reduced in the presence of an anti-PAR2 monoclonal antibody that attenuates receptor activation, demonstrating the specificity of these effects (323). Together, these observations suggest that various signaling cascades are involved in PAR2-mediated signaling. Clearly, we are just beginning to understand the variety of PAR2-induced cell signaling pathways regulated under physiological conditions and during disease.
IV. PAR3 and PAR4
Top
Abstract
I. Introduction
II. PAR1 in Inflammation...
III. PAR2 in Inflammation...
IV. PAR3 and PAR4
V. Conclusions
References
A. Biology and distribution of PAR3 and PAR4
As already mentioned, murine platelets do not express PAR1. The observation that thrombin was still capable of inducing Ca responses in these cells led to the identification of PAR3 (18). In humans, PAR3 is expressed in bone marrow, heart, brain, placenta, liver, pancreas, thymus, small intestine, stomach, lymph nodes, and trachea, although the cell types remain to be identified. In mouse, PAR3 is expressed by megakaryocytes and platelets, among other cell types. The receptor is necessary for normal thrombin signaling in mouse platelets because blocking of the hirudin-like domain of PAR3 using a specific antibody prevented mouse platelet activation by low but not by high concentrations of thrombin. The same result was observed with PAR3-deficient platelets (19, 324). This also shows that PAR3 is a high-affinity thrombin receptor in mouse that is activated by proteolytic cleavage. Interestingly, murine PAR3 (mPAR3) itself does not lead to thrombin signaling even when overexpressed, indicating that the receptor has lost its ability to function autonomously during evolution (39, 40). Knocking out the PAR3 gene in mice leads to protection of the animals against thrombosis but has a relatively mild effect on hemostasis (325) (Table 2). However, analysis of PAR3 expression in human platelets showed that the receptor is not produced or is hardly produced by these cells (128). This suggests that in humans PAR3 does not play a major role for platelet activation in contrast to the mouse system.
Antibodies specific for PAR1 inhibited human platelet activation by low but not by high concentrations of thrombin (125, 326). In mice, PAR3 is necessary for normal thrombin signaling in platelets. This indicates the presence of more than one thrombin receptor on the surface of these cells. Indeed, a fourth receptor was cloned, named PAR4 (19, 20, 128). So far, PAR4 has been cloned from human, mouse, and rat tissues (19, 20, 327). In humans, PAR4 is widely expressed in brain (175), testes, placenta, lung, liver, pancreas, thyroid, skeletal muscle, and small intestine (20). In rats, PAR4 is expressed in esophagus, stomach, duodenum, jejunum, distal colon, spleen, and brain (327, 328).
Both in mice and humans, platelets utilize two thrombin receptors. PAR4 is a low-affinity receptor in platelets both in humans and mice. However, mouse platelets use PAR3 and PAR4 instead of PAR1 and PAR4 to respond to thrombin (19). In humans, platelet effects of thrombin appear to be mediated predominantly by PAR1. Only at high concentrations and when PAR1 activation has been inhibited is platelet activation by thrombin dependent on PAR4. Signals mediated by PAR4 result in calcium influx (19, 122, 329), thromboxane production (117), endostatin secretion in rat platelets (118), and platelet aggregation (122) (Table 2). However, selective activation of human PAR4 (hPAR4) resulted in a weaker response compared with hPAR1-mediated signaling because anionic phospholipids were not exposed on the surface of hPAR4-activated platelets. Interestingly, hPAR4 can also be activated by cathepsin G, a neutrophil granule proteinase (330). Cathepsin G mediates neutrophil-platelet interactions at sites of vascular injury or inflammation. Inhibition of hPAR1 had no effect on platelet responses to cathepsin G, indicating a specific activation of PAR4.
Patients with Hermansky-Pudlak syndrome, an autosomal recessive disorder, lack platelet dense granules and have no ADP autocrine response. However, these patients show only a mild bleeding phenotype (331, 332). It was hypothesized that the defect of ADP autocrine response was compensated by signaling through PAR4 because the activation of this receptor occurs well after ADP release in normal individuals (333). Therefore, the ADP-autocrine response does not seem to be necessary for platelet aggregation as long as PAR4 is strongly activated. However, it was observed by another laboratory that PAR4-induced, but not PAR1-induced, aggregation was entirely ADP-dependent using a specific AP for PAR4 (334). Moreover, the authors found that subthreshold concentrations of an AP-activating PAR1 potentiated the effects of a PAR4-AP to stimulate maximal aggregation. In addition, both prostacyclin (PGI2) and S-nitroso-glutathione, an NO-releasing agent, reduced AP-stimulated aggregation and fibrinogen-receptor up-regulation.
PAR4–/– mice had markedly increased bleeding times (40). In addition, the platelets of these mice failed to change shape, mobilize calcium, or aggregate in response to thrombin. Ma et al. (118) observed that a specific AP for PAR4 could induce endostatin release in rat platelets. A selective PAR4 antagonist prevented endostatin release. In human platelets, specific agonists for PAR1 or PAR4 stimulated thromboxane production (117). Thromboxane produced by the combined stimulation of PAR1 and PAR4 was additive, suggesting the presence of two pathways for thrombin-induced thromboxane production in platelets. Blocking PAR1 function with a domain-specific antibody resulted in substantial inhibition of thrombin signaling. However, PAR4 can mediate platelet activation in response to high concentrations of thrombin (128). hPAR4 itself showed no pharmacological effect at either 1 nM or 30 nM thrombin. However, blocking of both hPAR1 and hPAR4 resulted in a profound inhibitory response even at high concentrations of the proteinase (128). This shows that hPAR4 activation is not necessary for robust responses in platelets when hPAR1 function is intact. Thus, inhibition of hPAR1 alone is probably not sufficient when seeking new antithrombotic therapies; it might be necessary to block both receptors simultaneously (106, 128). hPAR4 seems to play a role as a backup signaling device that might mediate thrombin signaling to distinct effectors or with different kinetics compared with PAR1. It might also allow platelets to respond to proteinases other than thrombin. Moreover, both receptors might be able to directly interact with each other.
The fact that mPAR3 alone does not result in thrombin signaling on mouse platelets showed that mPAR3 functions as a cofactor that promotes cleavage and activation of PAR4 at low concentrations of thrombin (39, 40). Interestingly, knocking out the PAR3 or the PAR4 gene leads to a similar degree of protection against thrombosis in mice (40, 325). In humans, both hPAR1 and hPAR4 may independently mediate thrombin signaling (20, 128, 177). Thus, the mouse system does not have a direct analog in the human platelet. However, mPAR3 promoted cleavage and activation of hPAR4 as effectively as mPAR4. Thus, hPAR4 can be "cofactored" by PAR3 in vitro, but it remains unclear whether or not cofactors such as PAR1 play a direct role in hPAR4 activation in vivo (39).
This reservoir of multiple thrombin receptors including PAR1, PAR3, and PAR4 may allow for a precise regulation of thrombin-induced inflammatory stimuli under different pathophysiological conditions and the fine-tuned induction of different signal transduction pathways.
In endothelial cells, thrombin induces a rapid but transient activation of endothelial cells by stimulating the secretion of PGI2 or platelet-activating factor (PAF) and by inducing the production of cell adhesion molecules (P-selectin, E-selectin) (47, 69, 70). Thrombin and a nonselective AP also induce synthesis and release of cytokines from endothelial cells such as IL-1, IL-6, and IL-8 (71, 72). Moreover, they regulate the cytokine-independent expression of ICAM-1 and VCAM-1 on human endothelial cells and cause an increased adhesion of monocytes to endothelial cells. This effect can be diminished by blocking antibodies (anti-CD 18, anti-CD 49d) (73). In the past, the effects observed above had been attributed to the activation of PAR1. However, a recent in vivo study has shown that thrombin-induced leukocyte rolling and adhesion to vascular walls is mediated via PAR4 (74). Although thrombin can trigger the recruitment of leukocytes to sites of inflammation, the PAR1 antagonist RWJ-56110 does not block this effect. In contrast, a PAR4-AP is able to reproduce the effects of thrombin on leukocyte rolling and adherence, indicating that this proinflammatory effect of thrombin is due to the activation of PAR4 and not PAR1 (74). Thus, PAR4 appears to function in early events of inflammatory reaction, in terms of recruiting leukocytes to the site of injury.
In contrast to a potential participation of PAR1 in stimulating the angiogenic process (335), PAR4 may play an opposing role. Apart from its antiangiogenic role via the activation of platelets, PAR4 can also affect the arterial system by stimulating vascular smooth muscle mitogenesis (220).
PAR4, like PAR1 and PAR2, also appears to be linked to signal transduction processes that modulate airway smooth muscle tone, because a PAR4-AP caused a rapid transient contractile response in a murine tracheal preparation. This contraction was followed by a transient relaxant phase. The underlying molecular mechanism is still unclear, but the relaxation appears to be dependent on PAR4-stimulated and COX-2-generated PGE2 production, which activates the E-type prostanoid 2 receptor (152, 153).
In conclusion, preliminary data strongly suggest an important pathophysiological role of PAR4 in vascular homeostasis and platelet function. However, the role of PAR3 as a signaling molecule and the precise function of PAR4 during inflammation and immune response remain to be clarified.
B. Signaling by proteinases via PAR3 and PAR4
PAR3 and PAR4 are the most recent members of the proteinase-activated receptor family. They are activated by thrombin. Subsequent to the cloning of PAR3 (18) and PAR4 (19, 20), relatively few publications have appeared examining the involvement of these receptors in signaling cascades. One puzzle that is yet to be resolved, as already mentioned, is the inability of PAR3 to signal in response to its tethered ligand-derived peptide or to thrombin (39, 40). Until this issue is resolved unequivocally, a meaningful discussion about PAR3 signaling is not possible. In contrast, PAR4 appears to be capable of activating both G12/G13 and Gq pathways (201, 336, 337).
As mentioned above, PAR4 appears to be involved in the same signaling cascades as PAR1 in human platelets. In contrast, mouse platelets express PAR3 and PAR4, but only PAR4 appears to serve as a real signaling receptor, and PAR3 serves solely to facilitate cleavage of PAR4 by thrombin (39).
Studying downstream signaling events in which PAR1 and PAR4 could be involved in VSMC, Bretschneider et al. (220) revealed that these receptors have distinct downstream signaling kinetics. Later, activation of MAPKs after stimulation of PAR4 was demonstrated. In their recent work, Sabri et al. (338) investigated PAR4-mediated signaling in cardiomyocytes derived from PAR1–/– mice. Using AYPGKF-NH2, a modified PAR4 agonist with an increased binding potency to PAR4, they were able to demonstrate p38 phosphorylation as well as slight activation of PLC and ERK1/2. Additionally, thrombin and PAR4-AP, but not PAR1-AP, were able to activate Src in these cells, clearly indicating that the action of thrombin on Src activation is mediated by PAR4, and not by PAR1 in these cells. Further studies implicated the involvement of Src and EGFR kinase activity in the PAR4-dependent p38 signaling pathway (338).
Recently, the involvement of PAR4 in factor Xa-mediated signaling has been demonstrated. Camerer et al. (25) revealed that expression of PAR1, PAR2, and PAR4 in Xenopus oocytes confers calcium signaling in response to factor Xa. They further showed that PAR4-AP (AYPGKF-NH2) stimulated increased phosphoinositide hydrolysis in endothelial cells and that responses to AYPGKF-NH2 are absent in PAR4–/– endothelial cells (240). Accordingly, PAR4-mediated phosphoinositide hydrolysis in response to factor Xa has been clearly demonstrated (25).
V. Conclusions
Top
Abstract
I. Introduction
II. PAR1 in Inflammation...
III. PAR2 in Inflammation...
IV. PAR3 and PAR4
V. Conclusions
References
There is no doubt that PARs play an important regulatory role during inflammation and immune response. Recent findings consolidate the concept that serine proteinases act as autocrine, paracrine, or endocrine mediators that talk directly to cells (13, 14, 15, 30). Some of these proteinase-stimulated effects are mediated through activation of proteinase-activated receptors, resulting in signal transduction pathways that are involved in inflammation and immune response. In many cases, PARs appear to play a proinflammatory role due to activation of proinflammatory mediators and cytokines (28, 59, 67, 72, 77, 78, 79, 80, 81, 82, 83, 84, 85, 86, 87, 88, 89, 90, 94, 99, 263, 272, 274, 282, 283, 295, 339, 340, 341, 342, 343, 344, 345). In other instances, a protective and antiinflammatory role of PARs has been observed (52, 292, 298, 312, 346, 347, 348). Various serine proteinases, which serve as PAR activators, are relevant mediators during inflammation (Table 2). However, the precise role of PARs and serine proteinases in different tissues, cell types, and states of inflammation remains to be determined.
Open questions include: 1) For the several proteinases that are released in various tissues from different cell types during inflammation, what is their mechanism of regulation and what is their functional relevance? 2) What is the significance of the existence of multiple receptor subtypes for one proteinase, and how might these common proteinase targets be differentially regulated? 3) Which immune cells express functional PARs in humans in vivo, and what is their biological relevance? 4) Which are the endogenous proteinases that activate PAR2 in human inflammation in vivo? 5) Which factors influence the regulation of PARs during inflammation or host defense? 6) Which effects of PAR-activating proteinases are non-PAR-mediated effects? 7) Which are the cell-specific signaling pathways and molecular mechanisms (transcription factors) after selective PAR activation in different inflammatory states and tissues? 8) What role do PARs play in neuronal transmission in humans? 9) Can PAR agonists or antagonists be used as therapeutic agents during inflammation or immune response in human diseases?
Thus, an integrative understanding of a regulatory role of serine proteinases as extracellular degradative enzymes as well as hormone-like signaling messengers in part via PARs and their counterregulation by extracellular proteinase inhibitors and intracellular molecules should lead to effective therapeutic approaches for various inflammatory/immune diseases such as thrombosis, sepsis, bacterial infections, gingivitis, asthma, hyperreactivity reaction, lung fibrosis, renal inflammation, rheumatoid arthritis, colitis ulcerosa, Crohn’s disease, pancreatitis, Alzheimer’s disease, amyotrophic lateral sclerosis, HIV encephalitis, atopic dermatitis, contact dermatitis, rosacea, wound repair, and infertility, for example, as well as pathophysiological symptoms such as pain and pruritus, in the future.
Footnotes
This work was supported by grants from the Federal Ministry of Education and Research (Interdisciplinary Center for Clinical Research, Münster F?. II-703-04; German Research Association, STE 1014; Collaborative Research Center (SFB) 293; SFB 492 to M.S.), Schering Foundation (to M.S., V.S.), C.E.R.I.E.S., Paris, France, and Boltzmann-Institute, Münster, Germany (to T.A.L., M.S.); Novartis Research Foundation (Vienna, Austria); Rosacea Foundation (to M.S.), and Boehringer-Ingelheim Fonds (to J.B.). The work of M.D.H. and N.V. is supported by funds from the Canadian Institutes for Health Research, the Kidney Foundation of Canada (to M.D.H.), the Heart and Stroke Foundation of Canada (to M.D.H.), by a Johnson & Johnson Focused Giving Grant, and by the Ileitis Colitis Foundation (to N.V.).
First Published Online October 12, 2004
Abbreviations: AP, Activating peptide; AP-1, activator protein-1; [Ca2+]i, intracellular calcium; CGRP, calcitonin-gene related peptide; CNS, central nervous system; COX, cyclooxygenase; cPLA2, cytosolic phospholipase A2; DRG, dorsal root ganglion; EGF, epidermal growth factor; EGFR, EGF receptor; EPR-1, effector cell proteinase receptor-1; FAK, focal adhesion kinase; GI, gastrointestinal; GM-CSF, granulocyte-macrophage colony-stimulating factor; GP, glycoprotein; GPCR, G protein-coupled receptor; GRK, G protein-coupled receptor kinase; HMEC, human microvascular endothelial cells; hPAR, human PAR; HUVEC, human umbilical vein endothelial cells; ICAM-1, intercellular adhesion molecule-1; IFN-, interferon ; IKK, inhibitor of NFB kinase; iNOS, inducible NO synthase; JNK, Jun N-terminal kinase; MCP-1, monocyte chemoattractant protein-1; MDCK cells, Madin-Darby canine kidney cells; MMP, matrix metalloproteinase; mPAR, murine PAR; NFB, nuclear factor B; NK cells, natural killer cells; NK receptor, neurokinin receptor; NO, nitric oxide; PAF, platelet-activating factor; PAR, proteinase-activated receptor; PAR-AP, proteinase-activated receptor activating peptide; PDGF, platelet-derived growth factor; PG, prostaglandin; PI3, phosphoinositide 3; PKA, protein kinase A; PKB, protein kinase B; PKC, protein kinase C; PLC, phospholipase C; PTX, pertussis toxin; Pyk2, proline-rich tyrosine kinase 2; SLP-76, SH2-domain containing leukocyte-specific phosphoprotein of 76 kDa; SMC, smooth muscle cells; SP, substance P; Src, protein tyrosin kinase Src; TK, tyrosine kinase; VCAM-1 , vascular cell adhesion molecule-1; VSMC, vascular SMC; ZAP-70, -chain-associated protein kinase of 70 kDa.
References
Scott G, Leopardi S, Parker L, Babiarz L, Seiberg M, Han R 2003 The proteinase-activated receptor-2 mediates phagocytosis in a rho-dependent manner in human keratinocytes. J Invest Dermatol 121:529–541(Martin Steinhoff, J?rg Bu)