Proteinases and signalling: pathophysiological and therapeutic implications via PARs and more

Authors

  • R Ramachandran,

    1. Inflammation Research Network, Department of Pharmacology & Therapeutics, Faculty of Medicine, University of Calgary, Calgary, Alberta, Canada
    Search for more papers by this author
  • M D Hollenberg

    Corresponding author
    1. Inflammation Research Network, Department of Pharmacology & Therapeutics, Faculty of Medicine, University of Calgary, Calgary, Alberta, Canada
    Search for more papers by this author

Department of Pharmacology & Therapeutics, Faculty of Medicine, University of Calgary, 3330 Hospital Drive NW, Calgary, Alberta, Canada T2N4N1. E-mail: mhollenb@ucalgary.ca

Abstract

Proteinases like thrombin, trypsin and tissue kallikreins are now known to regulate cell signaling by cleaving and activating a novel family of G-protein-coupled proteinase-activated receptors (PARs 1–4) via exposure of a tethered receptor-triggering ligand. On their own, short synthetic PAR-selective PAR-activating peptides (PAR-APs) mimicking the tethered ligand sequences can activate PARs 1, 2 and 4 and cause physiological responses both in vitro and in vivo. Using the PAR-APs as sentinel probes in vivo, it has been found that PAR activation can affect the vascular, renal, respiratory, gastrointestinal, musculoskeletal and nervous systems (both central and peripheral nervous system) and can promote cancer metastasis and invasion. In general, responses triggered by PARs 1, 2 and 4 are in keeping with an innate immune inflammatory response, ranging from vasodilatation to intestinal inflammation, increased cytokine production and increased or decreased nociception. Further, PARs have been implicated in a number of disease states, including cancer and inflammation of the cardiovascular, respiratory, musculoskeletal, gastrointestinal and nervous systems. In addition to activating PARs, proteinases can cause hormone-like effects by other signalling mechanisms, like growth factor receptor activation, that may be as important as the activation of PARs. We, therefore, propose that the PARs themselves, their activating serine proteinases and their associated signalling pathways can be considered as attractive targets for therapeutic drug development. Thus, proteinases in general must now be considered as ‘hormone-like’ messengers that can signal either via PARs or other mechanisms.

British Journal of Pharmacology (2008) 153, S263–S282; doi:10.1038/sj.bjp.0707507; published online 3 December 2007

Abbreviations:
APC

activated protein C

CNS

central nervous system

GPCR

G-protein coupled receptor

PAR

proteinase-activated receptor

PAR-AP

proteinase-activated-receptor-activating peptide

PGE2

prostaglandin E2

SERPIN

serine proteinase inhibitor

TE

tethered ligand

hPAR

human-PAR

mPAR

murine PAR

Introduction

Proteinases as signal generating enzymes

While traditionally regarded as digestive protein-degrading enzymes, proteinases are now gaining recognition as versatile and multifunctional hormone-like signalling molecules that are implicated in a number of physiological and pathophysiological events. Proteinases can regulate cellular signalling events through their interaction with a large variety of targets, including pro-hormones, kininogens, chemokines precursors and proteinase zymogens to name a few. In addition, proteinases can also potentially regulate integrin-extracellular matrix signalling and activate growth factor receptors like the one for insulin. Of key interest in terms of proteinase-mediated signalling is the ability of proteinases to regulate cell function by cleaving and activating the proteinase activated receptor (PAR) family of G-protein-coupled receptors (GPCRs).

Classes of proteinases that can trigger cell signalling

Proteinases may be divided into five different classes based on their mechanism of catalysis, namely the aspartate, mettalo, cysteine, serine and threonine proteinases. The aspartate and metalloproteinases use an activated water molecule as a nucleophile to attack peptide bonds, while a catalytic amino-acid residue in the active site of the proteinase serves as the nucleophile for the remaining classes of proteinases. The completion of the human genome sequencing project revealed approximately 550 genes encoding proteinase, with the metalloproteinases and serine proteinases making up the bulk of these (Puente et al., 2003). In principle, both of these classes of proteinases can signal via the mechanisms to be discussed in the following sections, which will focus in large part, but not entirely on PARs as targets for proteinase-mediated signal transduction. Thus, the proteinase-mediated signalling to be considered can proceed via both PAR and non-PAR mechanisms.

PARs and proteinase signalling

Many physiological responses mediated by serine proteinases can occur through the PAR family of GPCRs (for a comprehensive list, see review by Steinhoff et al., 2005). The PAR receptors, named PARs 1 to 4 in order of their discovery, have a unique mechanism of activation (Figure 1) that distinguishes them from other seven transmembrane GPCRs. Most GPCRs are activated reversibly by small hydrophilic molecules to elicit cellular responses (Wettschureck et al., 2005). PARs on the other hand effectively carry their own activating molecule in a masked state, with receptor activation achieved through proteolytic cleavage of a specific site within the receptor N-terminus to reveal a cryptic tethered ligand (TL) that binds to and activates the receptor (Hollenberg and Compton, 2002; Coughlin, 2005; Steinhoff et al., 2005). In addition, PARs, with the exception of PAR3, are also activated by short synthetic peptide sequences derived from the sequences of the proteolytically revealed TL (Vu et al., 1991; Scarborough et al., 1992). As outlined briefly in the following section and more extensively elsewhere (Hollenberg and Compton, 2002), these PAR-activating peptides (PAR-APs) have been very valuable tools in teasing out specific receptor function in systems that express more than one member of the PAR family. In addition to the cleavage/activation of PARs, proteinases can also negatively regulate functioning through the PARs by ‘disarming’ the receptor by cleavage at a non-receptor activating site to remove the TL (Figure 2). These truncated receptors nonetheless remain responsive to PAR-APs but would be unable to signal in a physiological setting.

Figure 1.

Mechanisms of PAR activation. (a) Activation of PAR signalling by proteinase-mediated cleavage of receptor N-terminus to reveal tethered ligand (TL). (b) Activation of PAR signalling by exogenous application of synthetic PAR agonist peptide (AP) without the need for proteolytic revealing of the TL. PAR, proteinase-activated receptor.

Figure 2.

Proteinases targeting PAR1 and PAR2. Proteinases can activate PARs by cleaving the receptor to reveal the tethered ligand (TL) (↓) or disarm them by cleaving at other sites (↑), making the TL unavailable to the receptor. Disarmed receptors remain responsive to AP-stimulated activation. hk, human kallekrien; TF, tissue factor; PAR, proteinase-activated receptor; PR3, proteinase 3.

Discovering pathophysiological roles for PARs: a pharmacological approach with receptor-selective agonists and antagonists

Receptor-selective PAR-activating peptides as probes for PAR function in bioassay systems: using structure–activity relationships to establish PAR-mediated responses

Studying PAR function with proteinase activation has been challenging especially in systems where more than one receptor is expressed. As alluded to above, synthetic peptides with sequences based on those of the proteolytically revealed PAR TLs can selectively activate the receptors without the need for proteolysis and have been key to teasing out responses mediated by these receptors. Importantly, it has also been possible to develop synthetic ‘scrambled’ TL peptide sequences that are not able to activate the PARs, so as to serve as appropriate ‘control peptides’ for studies done with cultured cells or tissues in vivo. Receptor-selective PAR-APs for PARs 1, 2 and 4 as well as the corresponding control peptides are outlined in Table 1. It has turned out surprisingly that PAR3 on its own does not appear to signal and does not respond either to thrombin or to the PAR-APs based on the thrombin-revealed PAR3 ‘tethered ligand sequence’. Rather, peptides with sequences derived from the thrombin-revealed PAR3 N-terminus are able to activate either PAR1 or PAR2 (Hansen et al., 2004). Studies have also provided evidence for a possible cofactor like role for PAR3 through its interaction with PAR1 and PAR4 (Nakanishi-Matsui et al., 2000; McLaughlin et al., 2007). The PAR-selective activating peptides, along with the appropriate PAR-inactive peptide sequences, have served as key reagents to explore the impact of activating a specific PAR in a tissue or cell, without the need for proteinase-stimulated activation of the receptor, which might result in effects other than simply activating a PAR.

Table 1.  PAR activating peptides and corresponding control peptides
ReceptorPAR-APsInactive control peptides
  1. Abbreviations: PAR, proteinase-activated receptor; PAR-AP, PAR-activating peptide.

PAR1TFLLR-NH2FTLLR-NH2
PAR2SLIGRL-NH2; 2-furoyl- LIGRLO-NH2LSIGRL-NH2; 2-furoyl-OLRGIL-NH2
PAR3TFRGAP: does not activate PAR3 
PAR4AYPGKF-NH2YAPGKF-NH2

An example of the use of such receptor-selective PAR-APs can be seen in studies of PAR2-mediated calcium signalling in a PAR2-expressing KNRK cell line. In this test system, the relative potencies of the PAR2-selective agonist peptides, SLIGRL-NH2, trans-cinnamoyl-LIGRLO-NH2, 2-furoyl-LIGRLO-NH2 and of a potent PAR1-selective PAR1AP, AparafluoroFRChaChaCitY-NH2 would be 2-furoyl-LIGRLO-NH2trans-cinnamoyl-LIGRLO-NH2≅SLIGRL-NH2⋙AparafluoroFRChaChaCitY-NH2. A completely reversed structure–activity relationship (SAR) would be expected of a PAR1-mediated response, in which the three PAR2-activating peptides would be essentially inactive. The same order of peptide agonist potencies would be expected for the activation of PAR2 in tissue preparations. This principle has been used to establish in a variety of tissue and cell preparations, the PAR-mediated responses that would be anticipated upon activating the receptors by a proteinase. The use of these receptor-selective agonists, along with PAR-null animals, has led to the discovery of the ability of the PARs to regulate pain and inflammation and to affect a variety of tissues ranging from nerve cells to the vasculature and intestinal tract. In this manner, a number of physiological roles have been proposed for the PARs as summarized in Table 2 and outlined in more detail below.

Table 2.  Potential physiological roles for PARs
Potential roleCommentReviewed by
  1. Abbreviations: CNS, central nervous system; PAR, proteinase-activated receptor.

CNS neuronal and astrocyte functionUpregulation of PARs observed in the setting of CNS inflammationNoorbakhsh et al., 2003
Endothelial cell function: (PARs 1, 2 and 4)Regulate release of NO, von Willebrand factor; increase neutrophil adherence; promote cell migrationCoughlin, 2005
Intestinal function: (PARs 1, 2 and 4)Regulation of motility (GI smooth muscle) and secretion (GI epithelial cell)Vergnolle, 2005a
Myenteric neuron functionAlso affects GI motility and inflammatory responseVergnolle, 2003
Vascular smooth muscle functionActivation of contractility; angiogenesis?Coughlin, 2005
Renal vascular functionRegulation of flow and afferent arteriolar functionVesey et al., 2007
Skin pigmentationProteinase inhibition affects skin pigmentation. Involvement of PAR2 in ethnic skin color phenotypesSeiberg, 2001
Hyperalgesia and analgesiaPAR activation can increase (PAR2 mediated) or decrease (PAR1 and PAR4 mediated) pain sensationVergnolle, 2005b
Platelet activation, haemostasis: Thrombin-activated receptors (PARs 1, 3, 4)Regulate both secretion and aggregation; PARs 1 and 4 can play opposing roles possibly due to differential coupling to g-protiensCoughlin, 2005
Response to joint injuryUse of PAR2 null mice and PAR antagonists suggests key role for PAR2 in arthritisMcIntosh et al., 2007
Tumour cell growth and metastasisBoth PARs 1 and 2 may play roles, activated by tumor-derived serine proteinases and matrix metalloproteinases (e.g. MMP-1)Ruf et al., 2006; Soreide et al., 2006

Responses triggered by PAR-activating peptides that cannot be attributed to PAR activation

In the context of using the PAR-APs, it must be emphasized that the same SAR principle (Ahlquist, 1948) that can verify PAR-related responses to the peptides can also be used to establish that, in certain circumstances, the PAR-APs can stimulate a receptor other than one of the PARs. It thus came as a surprise that the SAR for these PAR2 agonists observed in a rat jejunal ion transport assay (SLIGRL-NH2>trans-cinnamoyl-LIGRLO-NH2>AparafluoroFRChaChaCitY-NH2)was different from the SAR expected of either PAR2 or PAR1 (Vergnolle et al., 1998). A plausible conclusion was that the short-circuit current response in the jejunal Ussing chamber due to the serosal application of the PAR-APs and trypsin was mediated by a receptor different from either PAR1 or PAR2. In a similar manner, some recent work with PAR4-derived agonists has been able to verify the presence of PAR4 in rat platelets, using a platelet aggregation assay, while pointing to a non-PAR4-mediated response in a rat gastric longitudinal muscle assay (Hollenberg et al., 2004). The work with the PAR4-derived peptides illustrates that a judicious choice of standard PAR-APs as well as a standard PAR-inactive peptide is required to establish whether or not a given response can be attributed to a given PAR.

Proteinase signalling by mechanisms other than PARs

Although the main focus of this review is on PARs as a target for proteinase signalling, it is important to recognize that quite a number of other mechanisms can account for the ability of proteinases to regulate cell function. Thus, at this point, having dealt with an overview of the PARs, it is of interest to consider briefly some alternate targets that can result in signal transduction.

Regulation of growth factor receptors

One of the first indications that proteinases can activate hormone-like cellular signals came from the observations in the early 1960s that trypsin and pepsin can exhibit an insulin-like action in rat diaphragm tissue (Rieser et al., 1964; Rieser, 1967). This hormone-like action of trypsin in striated muscle and adipocytes (Cuatrecasas, 1971; Kono et al., 1971) cannot be attributed to the activation of PARs, but is rather due to the effect of trypsin on the receptor for insulin. By cleaving at a dibasic residue of the insulin receptor alpha-subunit, trypsin generates a truncated receptor that has intrinsic signalling activity (Shoelson et al., 1988). In principle, this kind of action of proteinases, either activating or disarming growth factor receptors (for example at higher concentrations, trypsin can abolish the ability of the insulin receptor to bind to insulin) (Cuatrecasas, 1971) can in principle modulate cell function in a variety of settings, for example via the insulin-like growth factor-1 receptor. Another proteolytic mechanism that can lead to the activation of a growth factor receptor involves the proteolytic generation of a growth factor agonist in the cell environment. For instance, the trans-activation of the epidermal growth factor receptor can result from the metalloproteinase-mediated release from the cell surface of a receptor agonist (heparin-binding epidermal growth factor) (Prenzel et al., 1999). Thus, in principle, any of the receptors for growth factors or other comparable agonists like cytokines or interleukins (ILs) can be regulated either by activation or inactivation by proteinase action.

Non-receptor signalling targets

Apart from classical pharmacological receptors that exhibit the dual property of selective agonist recognition and signalling, other ‘non-receptor’ targets can also result in signalling by proteinases. For instance, disruption of extracellular matrix–integrin signalling by proteolysis of either the matrix or integrin molecules would in principle alter cell behaviour. In this regard, the ability of thrombin to activate metalloproteinases (Lafleur et al., 2001) could in principle lead to PAR-independent signalling via remodelling of the extracellular matrix. Another novel mechanism for proteinase-triggered signalling can be seen in the action of plasmin, which in addition to regulating PAR activity (below) can signal via an annexin A2 target mechanism, in which plasmin-mediated proteolysis of annexin A2 triggers chemotaxis in human monocytes (Laumonnier et al., 2006; Li et al., 2007). It can be presumed that serine proteinases other than plasmin will also be found to regulate cell behaviour via this novel annexin A2 proteolytic process. The signalling mechanism whereby annexin cleavage regulates chemotaxis or other cell responses remains to be determined. Whether thrombin at high concentrations might mimic the annexin A2 cleavage caused by plasmin, in the manner that plasmin mimics the action of thrombin on the PARs, is also an issue to be explored.

Non-catalytic mechanisms for proteinase-mediated signalling

Protein–protein interactions in addition to their catalytic function must also be considered in evaluating proteinase-mediated signalling. For instance, apart from its ability to signal catalytically via the PARs, thrombin can also yield from within its structure, chemotactic-mitogenic peptides released by proteolytic processing of its non-catalytic domain (Bar-Shavit et al., 1984, 1986). These thrombin-derived peptides cause their effects via receptors that are not PARs (Glenn et al., 1988). The ability of proteinases to affect signalling via their non-catalytic domains is an issue that can be often overlooked. The essence of the previous sections is that proteinases, apart from targeting the PARs, can affect signalling by a diverse set of mechanisms. This diversity of hormone-like signalling roles played by proteinases is exceeded only by the diversity of the proteinase families themselves. The next question, therefore, is which proteinases might be candidates for signalling via the PARs and other mechanisms?

Searching for the proteinases responsible for PAR activation in physiological settings

Although the approaches described in the preceding sections have led to an understanding of the potential physiological roles the PARs may play, identifying the proteinases that may activate the PARs in a variety of physiological settings has proved to be a challenge. The following sections deal with the proteinases that can potentially fulfill the roles of ‘physiological’ PAR regulators.

Coagulation cascade proteinases

The coagulation cascade mediated by a variety of serine proteinases is one of the best characterized systems where such enzymes play key roles. Haemostasis is characterized essentially by a fine balance between the formation and the lysis of clots. The level of coagulation factors circulating in the blood, if totally activated, would significantly shift this fine balance towards the uncontrolled clotting of blood (Esmon, 2004). The coagulation cascade allows the body to control this process and consists of a number of serine proteinase zymogens and their glycoprotein cofactors, which are activated in sequence to propagate the cascade (Davie and Ratnoff, 1964; Macfarlane, 1964). Quite apart from their critical involvement in the clotting of blood, a number of coagulation cascade proteinases are reported to target the PAR family of GPCRs, and there is now little doubt that a number of the coagulation cascade proteinases represent physiological regulators of PARs 1, 2 and 4 (Coughlin, 2005). Indeed, the initial impetus for the search for PARs came from the inability of known mechanisms to explain the numerous functions that thrombin performed in various cellular environments (Weksler et al., 1978; Bar-Shavit et al., 1983a, 1983b; Babich et al., 1990). As it turns out, thrombin is one of the most potent agonists for PAR1 and to a lesser extent PAR4 (Vu et al., 1991; Kahn et al., 1998; Xu et al., 1998). While thrombin is also known to cleave PAR3, the ambiguity surrounding the function of this receptor makes it difficult to assign physiological relevance to this event. PAR2, the remaining member of the PAR family, is not activated by thrombin and was thought not to be a target for the coagulation proteinases. However, PAR2 is now known to be activated by the Tissue Factor/VIIa binary complex and by Factor Xa, which can cleave PAR2 either independently, or with greater efficiency as part of a ternary complex with TF and VIIa (Camerer et al., 2000; Riewald et al., 2001). PAR2 activation by coagulation proteases is now known to initiate a number of important signalling events, including signalling and migration of cancer cells (Morris et al., 2006), induction of angiogenesis (Belting et al., 2004; Uusitalo-Jarvinen et al., 2007) and signalling in Osteosarcoma cells (Daubie et al., 2007).

In addition to the proteinases involved in clot formation, trypsin-related serine proteinases of the fibrinolytic system such as plasmin can also regulate signalling in part via the PARs. However, the role of plasmin is complex since this enzyme can both activate and disarm a PAR like PAR1 (Kimura et al., 1996; Kuliopulos et al., 1999) and can also activate PAR4 (Quinton et al., 2004). Further, as mentioned above, plasmin can activate cell signalling by a mechanism that, apart from involving the PARs, signals via the cleavage of the heterotetramer composed of annexin A2 and S100A10. Yet another clotting cascade-associated serine proteinase with anticoagulant and anti-inflammatory activities, activated protein C (APC), can exert its cytoprotective/anti-inflammatory effects by activating PAR1 employing a mechanism that involves both binding to the endothelial cell surface via an APC-targeted endothelial adsorption site (endothelial cell protein C receptor) and a specific interaction with PAR1 via a specific APC exosite domain (Riewald et al., 2002; Yang et al., 2007). The impact that APC has on the disarming of PAR2 or the potential activation of PAR4 has not yet been evaluated. Apart from these coagulation-associated proteinases, for which a physiological PAR-regulatory role can be seen, the tissue-localized enzymes that may regulate PARs in vivo, including PARs 2 and 4, which are both potently activated by trypsin (and presumably other serine proteinases), have yet to be identified.

Proteinases in the immune system

Neutrophils are immune cells that are well known to carry a variety of proteinases to aid in their essential roles in host defence against invading pathogens. Neutrophil proteinases consist mainly of proteinase-3, cathepsin-G and elastase (Pham, 2006). Neutrophils can be seen to play a ‘signalling role’ in terms of their processing and enhancing the chemotactic activity of chemokines such CXCL8, CXCL5, and CCL15 as well as proteolytically activating cytokines like IL-1 and tumour necrosis factor. Neutrophil proteinases can also negatively regulate cytokine function with elastase and cathepsin-G reported to degrade tumour necrosis factor, while proteinase-3, cathepsin-G and elastase are reported to inactivate IL-6 (Pham, 2006).

Proteinase-3 is reported to activate PAR2 on epithelial cells (Uehara et al., 2002, 2003). Of relevance to PAR signalling as well is the recent report that proteinase-3 can cleave and inactivate the endothelial cell protein C receptor, consequently inhibiting the ability of APC to regulate PAR1 activation (Villegas-Mendez et al., 2007). Cathepsin-G preferentially activates PAR4 and has been reported to activate this receptor in a number of cell types (Sambrano et al., 2000; Ramachandran et al., 2007). In addition, cathepsin-G is known to disarm PAR1 (Molino et al., 1995) and in murine platelets to inactivate PAR3 and abolish its cofactor like function (Cumashi et al., 2001). The elastase-mediated regulation of PARs occurs through disarming the receptors, with cleavage of the PAR2 receptor reported at Ser37-Leu38 and Gly52-Val53 sites (Dulon et al., 2003, 2005), while PAR1 is susceptible to elastase cleavage at the Val72-Ser73 and Ile74-Asn75 sites downstream of the activation site (Renesto et al., 1997).

Proteinases in the nervous system

The expression of a number of PARs and their possible activating proteinases has been described in both the central (CNS) and the peripheral nervous system. The potential activating proteinases in the peripheral nervous system are essentially as found in the organ system being innervated, while the CNS is also seen to express a number of known serine proteinases that may regulate PAR function, including thrombin, plasmin, urokinase plasminogen activator, tissue plasminogen activator and tissue kallikreins. In the CNS, such proteinases are believed to play an important role in neural plasticity (Yoshida et al., 1999; Shiosaka et al., 2000). Thrombin expression in the brain is well established (Yoshida et al., 1999; Rohatgi et al., 2004) and would constitute a key regulator of the PARs in the CNS. As outlined above, the plasmin–tissue plasminogen activator system is also reported to activate PAR1 in the brain (Nagai et al., 2006). In a number of neuropathological conditions, levels of thrombin are seen to be upregulated (Turgeon et al., 1997), whereas levels of plasmin are reduced in Alzheimer's disease (Dotti et al., 2004). The differential expression and levels of proteinases in the brain could thus potentially result in selective activation of the various PARs expressed in CNS tissue and may work in concert with the upregulation of PAR expression as seen in disease settings (Boven et al., 2003; Noorbakhsh et al., 2006). The regulators of PAR2 in the brain are harder to describe; however, recent evidence has pointed to the involvement of the kallikrein family of serine proteinases as key regulators of PAR2 function (Oikonomopoulou et al., 2006). A number of kallikreins have been described in the brain with an important role indicated for neuropsin (KLK8) and neurosin (KLK6) (Yousef et al., 2003). Neurosin levels are downregulated in Alzheimer's disease (Ogawa et al., 2000), while elevated levels of neuropsin and neurosin are reported in experimental autoimmune encephalomyelitis (Terayama et al., 2005) and spinal chord injury (Terayama et al., 2004).

In addition to the kallikreins, a number of trypsin species are reported to be expressed in the brain and to target PARs (Minn et al., 1998; Wang et al., 2006). Further, following an inflammatory event, as is seen in a number of neurological disorders, immune cells are detectable in the CNS and therefore the potential exists for immune cell proteinases described above to regulate PAR function in the brain.

Proteinases in the airways

The airways are exposed to numerous proteinases both from resident cells and from infiltrating immune cells. In addition, the airways are constantly exposed to airborne microorganisms, a number of which can also present proteinases to the airways. Proteinases produced by mast cells, tryptase, chymase and cathepsin-G, are some of the most abundant serine proteinases detectable in the airways. Mast cell proteinases are also implicated in a variety of cellular responses, including inflammation, matrix destruction and remodelling, hydrolysing chemokines and cytokines as well as inactivating allergens and neuropeptides to name but a few. Importantly, mast cell proteinases are also reported to regulate the PARs, both through activating and disarming the receptors (Reed et al., 2004).

Tryptase is expressed in the airways as both soluble (α, β and γ tryptase) and membrane-anchored (δ tryptase) forms. Only the β form of human mast cell tryptase is believed to play important roles in the extracellular environment (Caughey, 2007) and is known to promote allergic inflammation, airway hyperresponsiveness and tissue remodelling (Cairns, 1998; Berger et al., 1999). Tryptase is also reported to cleave and activate PAR2 (Molino et al., 1997) and a number of the tryptase-stimulated cellular responses have been attributed to PAR2 activation (Compton et al., 1999; Akers et al., 2000; Frungieri et al., 2002; Berger et al., 2003). Mast cell chymase is predominantly involved in matrix destruction by cleaving fibronectin and collagens and is also reported to activate PAR1 (Schechter et al., 1998). Cathepsin-G on the other hand has been reported to activate PAR4 (Sambrano et al., 2000) as well as disarming PAR1 and PAR2 (Ramachandran et al., 2007).

In addition to mast cell proteinases, the expression of PAR-activating extrapancreatic trypsins, such as trypsin, trypsin IV and human airway trypsin, have also been reported in the airways (Cocks et al., 1999a; Cottrell et al., 2004; Matsushima et al., 2005).

PARs are also targeted by proteinases produced by the house dust mite. Der-p1 is reported to cleave the PAR2 activation site and disarm PAR1 (Asokananthan et al., 2002b). Der-p3 and Der-p9 are also shown to have the potential to cleave the PAR2 activation site (Sun et al., 2001), although there are alternative interpretations to those data, since Der-p1 responses have been reported to result clearly from mechanisms that do not involve PAR2 (Adam et al., 2006). Under inflammatory conditions, the airways are also infiltrated with large numbers of polymorphonuclear neutrophils. The proteinases produced by these cells also contribute to regulating responses via the PARs as discussed above.

Proteinases in the gastrointestinal system

The gastrointestinal tract is exposed to very high levels of proteinases, including digestive gland proteinases involved in digestion of food as well as exogenous proteinases produced by the commensal gut microflora or by invading pathogenic microorganisms (Antalis et al., 2007). It is believed that in the gastrointestinal tract, members of the trypsin family represent the major physiological regulators of the PARs (Kong et al., 1997). In a murine model of infectious colitis, colon-derived trypsin family members triggered by the infectious process have been identified and have been shown to be capable of activating PAR2 (Hansen et al., 2005). In a study on colonic tissue from individuals with inflammatory bowel syndrome, it was shown further that levels of trypsin, tryptase and other unidentified serine proteinase that favoured cleavage at an arginine site are significantly upregulated (Cenac et al., 2007). These proteinases have been shown to target PAR2. High concentrations of trypsin like proteinases are also potentially activators of PAR1 (Knecht et al., 2007) and PAR4 (Xu et al., 1998) and increased levels of trypsin are reported in patients with conditions such as inflammatory bowel disease (Playford et al., 1995).

Pathogen-derived proteinases, such as gingipains, arginine-specific proteinases produced by the oral pathogen Poryphyromonas gingivalis, are also known to activate PARs 1, 2 and 4 (Lourbakos et al., 1998, 2001; Holzhausen et al., 2006).

Inflammation is a key event in a number of disorders of the gut, including conditions such as inflammatory bowel disease. The resultant infiltration of mast cells and immune cells would allow the presence of proteinases secreted by these cells to regulate PARs in the gastrointestinal tract as described in the sections above. The precise relationship between PARs, their regulating proteinases present in the enteric tract and gastrointestinal pathophysiology is currently a topic of intensive investigation.

PAR-mediated signal transduction pathways

PAR1 is known to couple with multiple G-proteins, including Gi, Gq/11 and G12/13 families (Macfarlane et al., 2001; Steinhoff et al., 2005). However, our knowledge of PAR interaction with G proteins is still far from complete. Evidence for PAR1 interaction with Gi has come from multiple cell types, such as HEL cells (Brass et al., 1991), osteosarcoma cells (Babich et al., 1990), vascular smooth muscle cells (Kanthou et al., 1996), fibroblasts (Hung et al., 1992a), platelets (Kim et al., 2002), astrocytes (Wang et al., 2002) and endothelial cells (Vanhauwe et al., 2002). Coupling of PAR1 to Gq/11 was revealed through the use of G-protein-specific antibodies, which inhibited PAR1-mediated calcium signalling in CCL-39 fibroblast cells (Baffy et al., 1994) and through co-immunoprecipitation of Gq/11 with PAR1 (Ogino et al., 1996). Interaction between PAR1 and Gq/11 was also reported in astrocytes (Wang et al., 2002) and indicated in astrocytoma cells (LaMorte et al., 1993). Evidence for PAR1 coupling to G12/13 comes from studies on thrombin-stimulated platelets (Offermanns et al., 1994) and for coupling to G12 in astocytoma cells (Aragay et al., 1995; Post et al., 1996). PAR1 is also known to activate multiple signalling pathways downstream of coupling to G-proteins, including the activation of PI3 kinase (Malarkey et al., 1995), Src family tyrosine kinases (Sabri et al., 2000) and the extracellular signal-regulated kinase (ERK)/mitogen-activated protein kinase (MAPK) pathway (Vouret-Craviari et al., 1993). PAR1 signalling to nuclear factor (NF)-κB (Rahman et al., 2002) has also been reported, but the precise signal pathways involved are not known.

PAR2 signalling has been studied in less detail than for PAR1. The activation of PAR2 leads to an elevation of intracellular calcium subsequent to the release of inositol tris-phosphate along with diacyl-glycerol (Nystedt et al., 1995). This result suggests that PAR2 signals via Gq/11 and possibly Gi (Macfarlane et al., 2001). Coupling of PAR2 with G12/13 has not been reported. Another very important aspect of cell activation via PAR2 relates to its ability to signal via an arrestin-mediated process that can be independent of Gq–protein interactions (Zoudilova et al., 2007). This arrestin-dependent mechanism may explain the ‘dual’ actions that PAR2 can have in certain settings, triggering either inflammatory or anti-inflammatory responses. The ability of arrestin-mediated, G-protein-independent signalling that might involve PAR heterodimers (for example PAR2/PAR1) is a topic that merits an in-depth evaluation.

PAR2 activation on smooth muscle and neuronal cells has been observed to activate phospholipase C (Berger et al., 2001b) and protein kinase C, respectively (Okamoto et al., 2001). A recent study has revealed further that regions within the PAR2 C-terminal tail are critical for PLC-mediated IP3 formation and calcium signalling as opposed to the activation of MAPkinase that is triggered by other intracellular PAR2 sequences (Seatter et al., 2004). PAR2 was also reported to activate the MAPK cascade in rat aortic smooth muscle cells (Belham et al., 1996) and in human keratinocytes PAR2-mediated activation of JNK and p38 MAPK was shown to mediate NF-κB DNA binding (Kanke et al., 2001; Buddenkotte et al., 2005). Further, in keratinocytes, intracellular calcium mobilization is an important determinant of PAR2 NF-κB signalling (Macfarlane et al., 2005). PAR2-mediated activation of NF-κB has also been demonstrated in epithelial cells (Vliagoftis et al., 2000) and coronary artery smooth muscle cells (Bretschneider et al., 1999). PAR2 activation of ERK1/2 is now established in a number of cell including cardiomyocytes (Sabri et al., 2000) and smooth muscle cells (Belham et al., 1996). Factor Xa-mediated activation of ERK1/2 in coronary artery smooth muscle cells was also shown to be PAR2 mediated (Koo et al., 2002). Endothelial PAR2 stimulation can trigger the activation of the ‘small G-proteins’ such as Rho-A, Rac leading to the regulation of p21-activated kinase. These responses are believed to be involved in mediating cytoskeletal effects (Vouret-Craviari et al., 2003), possibly involving arrestin-mediated signalling (Zoudilova et al., 2007).

The signalling role of PAR3 is not yet clear, except for its ability to act as a cofactor for PAR1 activation. On its own, PAR3 does not appear to signal, and its TL sequence can activate other PARs, like PARs 1 and 2 (Hansen et al., 2004; Kaufmann et al., 2005). However, the ability of PAR3 to dimerize with other PARs (McLaughlin et al., 2007) points to another mechanism whereby PAR3 can regulate signalling, apart from employing its TL.

PAR4 is known to activate calcium signalling, presumably via Gq (Kahn et al., 1998; Xu et al., 1998; Camerer et al., 2002) and PAR4 activation of MAPK signalling has been observed in vascular smooth muscle cells (Bretschneider et al., 2001). PAR4-mediated Src-dependent p38 phosphorylation and activation of ERK and PLC has been described in cardiomyocytes derived from PAR1-knockout mice (Sabri et al., 2003). A recent study has also reported PAR4 activation of p38 MAPK in endothelial cells (Fujiwara et al., 2005). Whether or not this PAR4-mediated activation of MAPkinase involves Gi or a separate arrestin-mediated process remains to be determined.

PARs in physiology and pathophysiology

Since their identification within the last 16 years or so, a myriad of physiological and pathophysiological events have been attributed to PAR-mediated signalling. The specifics of their involvement in various processes, however, remain unknown. The complex interplay within PARs, multiple proteinases, proteinase inhibitors and other molecules that may be present in different organ systems, and under different pathological conditions, makes the understanding of PAR-mediated events a complex exercise. The lack of potent, readily available antagonists for all of the PARs and the challenge of extrapolating data obtained from animal models to human disease are some of the other issues to deal with in determining the precise role that PARs may play in humans. Notwithstanding, the results of work with PAR-null animals and the documented expression and activation of PARs in various organs in humans point to key roles for the PARs in many pathophysiological settings.

PARs in the cardiovascular and circulatory system

PARs are expressed by multiple cells in the cardiovascular and circulatory system, including the circulating cells as well as the vascular endothelial and smooth muscle cells. Human platelets express PAR1 and PAR4 (Kahn et al., 1999), while murine platelets express PAR3 and PAR4 (Kahn et al., 1998). PAR1 (Vu et al., 1991), PAR2 (Hwa et al., 1996; Mirza et al., 1996), PAR3 (Schmidt et al., 1998) and PAR4 (Fujiwara et al., 2005) are expressed on endothelial cells, while vascular smooth muscle cells express PAR1 (McNamara et al., 1993; D'Andrea et al., 1998) and PAR2 (D'Andrea et al., 1998; Molino et al., 1998). Both PAR3 (Bretschneider et al., 2003) and PAR4 (Bretschneider et al., 2001) are also found on vascular smooth muscle cells.

Activation of PAR1 or PAR4 on human platelets is sufficient to trigger calcium signalling and aggregation (Vu et al., 1991; Xu et al., 1998; Kahn et al., 1999). At low thrombin concentrations, PAR1 appears to be responsible for mediating thrombin-mediated responses in platelets, while PAR4-mediated responses are limited to systems where PAR1 is not expressed or where high thrombin concentrations are employed (Hung et al., 1992b; Kahn et al., 1999). It has also been claimed that PAR4 may be involved in responding to proteinases other than thrombin, and hPAR4 activation by the neutrophil proteinase cathepsin-G has been revealed (Sambrano et al., 2000). It has also been suggested that hPAR4 might mediate slow and prolonged responses to thrombin, while hPAR1-mediated signalling events may be more rapid. Support for this hypothesis comes from different kinetics observed in hPAR1 and hPAR4 desensitization (Shapiro et al., 2000). A differential function of PARs 1 and 4 for regulating human platelets can be seen in the ability of PAR1 activation to trigger the release of vascular endothelial growth factor, whereas PAR4 activation causes the release of endostatin, which would be counter-regulatory for vascular endothelial growth factor action (Ma et al., 2005). Recent studies in human platelets have indicated that activation of PAR1 and PAR4 triggers distinct pathways following thrombin cleavage (Bilodeau et al., 2007; Holinstat et al., 2007; Voss et al., 2007) and may form the basis for the differential functions observed. Murine platelets express mPAR3 and mPAR4, yet mPAR4 is the sole functional thrombin receptor, with mPAR3 serving as a cofactor to enhance thrombin action (Nakanishi-Matsui et al., 2000). The presence of both receptors is, however, essential for haemostasis, since genetically deleting either receptor in mice results in an increased tail bleeding time and compromised thrombosis (Sambrano et al., 2001). Moreover, the administration of the peptide pepducin P4pal-10 PAR4 antagonist also prolongs tail bleeding times in mice, providing further evidence for a key role for PAR4 in thrombin signalling in rodents (Covic et al., 2002b). In other mammals, however, PAR1 appears to be more important than PAR4 for regulating platelet aggregation. Thus, by inhibiting PAR1 function selectively with specific antagonists, it is possible to prevent thrombus formation in monkeys and presumably also in humans (Derian et al., 2003).

Activation of either rat or human endothelial PAR1 and PAR2 in organ bath preparations causes vascular relaxation, which is mediated, at least in part by the release of endothelial nitric oxide (Al-Ani et al., 1995; Hamilton et al., 2001, 2002). Comparable results have been obtained for observations done in vivo in rats and mice (Cicala et al., 2001). A study done in vivo in humans has revealed further that PAR2-AP causes vasodilation in an NO- and prostaglandin-dependent manner (Robin et al., 2003). Moreover, activation of PAR1 and PAR2 can cause proliferation of cultured vascular smooth muscle (McNamara et al., 1993; Chaikof et al., 1995; Bretschneider et al., 1999; Koo et al., 2002) and PAR2 activation can also stimulate endothelial cell proliferation (Mirza et al., 1996). Another report has shown that PAR2 can stimulate angiogenesis in vivo in a murine model of hindlimb ischaemia (Milia et al., 2002). PAR4 activation has also been implicated in vascular smooth muscle proliferation (Bretschneider et al., 2001).

PAR1, PAR2 and PAR4 have also been implicated in mediating vascular responses in the setting of models of inflammation (for example increased leukocyte endothelial adherence and rolling; migration of leukocytes from the vasculature into inflamed tissues). The increased expression of proteinases from inflammatory cells and from activation of the coagulation cascade very likely serve as triggers for the PARs in this type of setting. Activation of rat paw PAR1 in response to the intraplantar injection of a PAR1-activating peptide causes oedema that is, in part, due to the extravasation of plasma into the peripheral tissue via a breach in endothelial integrity (Vergnolle et al., 1999b). The involvement of PAR4 in inflammatory responses has also been implicated by observations in vivo demonstrating that thrombin and a PAR4-AP (but not a PAR1-AP) induce leukocyte endothelial adhesion and rolling in rat mesenteric venules (Vergnolle et al., 2002). Like PAR1, PAR2 activation resulting from the intraplantar administration of a PAR2-AP also causes oedema and granulocyte infiltration from the vasculature (Vergnolle et al., 1999a). The oedema caused by PAR2 activation has been found to depend on the release of neuropeptides from sensory nerves, whereas the extravasation of granulocytes is believed to be due to a direct effect of activating PAR2 on endothelial cells.

PARs in the immune system

PAR expression has been revealed on many immune cells. PAR1 is expressed on macrophages (Colognato et al., 2003), monocytes (Colognato et al., 2003) and mast cells (D'Andrea et al., 2000), while other immune cells such as T cells (Mari et al., 1996) and Jurkat-T leukaemic cells (Bar-Shavit et al., 2002) are seen to be thrombin responsive presumably via PAR1. PAR2 expression has been detected on mast cells (D'Andrea et al., 2000), eosinophils (Miike et al., 2001), monocytes (Colognato et al., 2003; Roche et al., 2003), macrophages (Colognato et al., 2003) and neutrophils (Howells et al., 1997; Lourbakos et al., 1998; Shpacovitch et al., 2004), but the function of the PARs in these cells is not always clear. PAR3 expression has been detected on eosinophils (Miike et al., 2001) and Jurkat-T leukaemic cells (Bar-Shavit et al., 2002), while PAR4 expression is reported to be functional on circulating leukocytes (Vergnolle et al., 2002).

The primary response mediated by PARs in the immune system appears to be related to chemotaxis and cytokine release from inflammatory cells. Thrombin is known to stimulate the release of IL-8, IL-6 (Naldini et al., 2000) and monocyte chemoattractant protein 1 (Colotta et al., 1994). Thrombin also activates natural killer cells to stimulate IL-2 release (Naldini et al., 1996). Eosinophil PAR2 activation results in degranulation and superoxide production (Miike et al., 2001), while exposure to mast cell tryptase also induces eosinophil IL-8 release implicating an involvement for PAR2 in mediating this response (Temkin et al., 2002). Neutrophil PAR2 activation causes shape changes and enhanced CD11b/CD18 expression on these cells (Howells et al., 1997). PAR2 agonists are also known to cause calcium signalling in neutrophils and to enhance neutrophil motility in 3D collagen gel lattices through upregulation of cell surface integrins (Shpacovitch et al., 2002).

PARs in the nervous system

The expression of all four PARs has been detected in CNS of rodents (Striggow et al., 2001), as revealed in rat brain tissue and in cultured rat glioma cells and astrocytes (Ubl et al., 1998). PAR1 expression has also been documented in human astrocytoma cells (Grishina et al., 2005) and PAR2 expression was been found on normal human astrocytes and neurons (D'Andrea et al., 1998). PAR4 expression has also been reported in human astrocytes (Kaufmann et al., 2000). Further, PAR1 and PAR2 are also found in the spinal cord and in isolated rat dorsal root ganglia (Noorbakhsh et al., 2003) as well as in guinea pig myenteric (Corvera et al., 1999) and submucosal neurons (Reed et al., 2003).

Both thrombin and PAR1-AP have been revealed to have effects on astrocyte morphology and proliferation. Also, for some time, thrombin has been known to affect the morphology and outgrowth of neurites in nerve cells, very likely by activating PAR1 (Gurwitz and Cunningham, 1990; Suidan et al., 1992; Olianas et al., 2007). At low concentrations, thrombin is observed to reverse astrocyte process extension perhaps leading to compromised integrity of the blood–brain barrier (Noorbakhsh et al., 2003). Higher concentrations of thrombin on the other hand, can cause cell proliferation (Noorbakhsh et al., 2003), and indeed an astrocyte proliferation-dependant disorder, astrogliosis, is revealed to be triggered by PAR1 activation (Nicole et al., 2005). Thrombin-mediated PAR1 activation has also been shown to protect neuronal cells and astrocytes from hypoglycaemia or oxidative stress-induced cell death (Vaughan et al., 1995). Thrombin and PAR1-APs were also reported to exert cytotoxic effects, affecting the morphology of neuronal cells, particularly with respect to shortening the length of the secondary and tertiary neurites (Suidan et al., 1992; Debeir et al., 1998). Indirect PAR1 involvement in synaptic function has also been indicated through potentiating the effects of N-Methyl-D-aspartate receptors by activation of protein kinase C (Gingrich et al., 2000). Both PAR1 and PAR4 have also been implicated in activation of microglial cells (Suo et al., 2002, 2003a).

An important role for thrombin in CNS injury is indicated by a number of studies. Increased thrombin levels have been observed at sites of cerebrovascular injury (Smirnova et al., 1996) and high levels of thrombin have been detected in plaques of Alzheimer's disease patient brains (Akiyama et al., 2000). Further, thrombin, very likely acting via PAR1 in concert with metalloproteinase 9 activity, has been observed to exacerbate neurotoxicity in a murine model of intracerebral haemorrhage (Xue et al., 2006). In this regard, it is of interest that thrombin can either protect neurons and astrocytes from cell death produced by environmental insults or can promote neuronal and astrocytes apoptosis, presumably acting via PAR1 (Vaughan et al., 1995; Donovan et al., 1997). The importance of thrombin signalling in Alzheimer's disease pathology is further underlined by the observed decrease in levels of protease nexin-1, a thrombin inhibitor, in the Alzheimer's disease affected brain (Vaughan et al., 1994; Choi et al., 1995). In vitro studies on cultured neurons show that low concentrations of thrombin and PAR1-AP can reduce β-amyloid neurotoxicity and astrocytes stellation while also increasing secretion of basic fibroblast growth factor, an index of neurotoxicity (Pike et al., 1996), while PAR4 activation by thrombin is thought to partly mediate rapid tau aggregation and delayed hippocampal neuronal death (Suo et al., 2003b), thus suggesting that thrombin at least in part through PAR1 and PAR4 is involved in Alzheimer's disease pathophysiology.

Recent studies have also revealed a role for PAR1 in HIV-associated neurodegenerative disorders (Boven et al., 2003; Noorbakhsh et al., 2005). Increased levels of PAR1 were observed in astrocytes through staining of brain sections from patients with HIV encephalitis. Cultured human astrocytes stimulated with thrombin or PAR1-AP revealed increased levels of IL-1β and nitric oxide synthase (NOS), with levels of PAR1 mRNA also being upregulated in these cells (Boven et al., 2003). PAR2 expression is increased by tumour necrosis factor α and IL-1β in the HIV-associated dementia brains, and PAR2 activation exerts a neuroprotective effect, preventing HIV-encoded protein Tat-mediated neuronal death (Noorbakhsh et al., 2005). High levels of mast cells are also seen in brains and cerebrospinal fluid of multiple sclerosis patients (Rozniecki et al., 1995; Noorbakhsh et al., 2003). Thus, PAR2 activation could potentially play a role in this disease, but it remains to be seen whether these effects are protective or degenerative. PAR2-mediated coronary vasodilation appears to involve a neurogenic mechanism (McLean et al., 2002), but it is not clear if PAR2 activation occurs directly on the neuronal elements or on other cells in the nerve cell environment (or both).

Studies on PAR function in the peripheral nervous system have pointed to a role for these receptors in neurogenic inflammation (Steinhoff et al., 2000), hyperalgesia, analgesia (Vergnolle et al., 2001) and itching (Steinhoff et al., 2003), as well as playing roles in nerve regeneration (Vergnolle et al., 2003), gastric epithelial ion secretion (Green et al., 2000) and mucous secretion function (Kawabata et al., 2001a, 2002). Recent evidence has also implicated a role for PAR2 in potentiating transient receptor potential vanilloid-1 (TRPV-1) and TRPV-4 ion channel function to mediate hyperalgesia and inflammation (Amadesi et al., 2004; Dai et al., 2004; Grant et al., 2007).

PARs in the respiratory system

PAR expression is detectable in a wide range of airway cells, and evidence so far suggests that the PARs play critical roles in pulmonary physiology and pathology. Signalling via PARs may be regulated by a large number of potential PAR proteinase agonists in the airways, including, but not limited to trypsin (Cocks et al., 1999a), mast cell tryptase (Molino et al., 1997), neutrophil proteinases (Uehara et al., 2003), allergens (Asokananthan et al., 2002b) and human airway trypsin (Matsushima et al., 2005). PAR1 is found on pulmonary fibroblasts (Trejo et al., 1996; Chambers et al., 1998; Ramachandran et al., 2006, 2007), epithelial cells (Asokananthan et al., 2002a), endothelial cells (D'Andrea et al., 1998; Kataoka et al., 2003) and smooth muscle cells (Lan et al., 2000; Walker et al., 2005). PAR2 expression is also widely distributed in the respiratory system and has been detected on epithelial cells (Knight et al., 2001; Vliagoftis et al., 2001; Ubl et al., 2002; Asokananthan et al., 2002a), smooth muscle cells (Schmidlin et al., 2001; Berger et al., 2001a), endothelial cells (D'Andrea et al., 1998; Schmidlin et al., 2001) and fibroblasts (Akers et al., 2000; Matsushima et al., 2005; Ramachandran et al., 2006, 2007). PAR3 expression in the airways has not been clearly demonstrated. However, PAR3 mRNA expression is reported in airway smooth muscle cells (Hauck et al., 1999), epithelial cells (Shimizu et al., 2000) and fibroblasts (Ramachandran et al., 2006). PAR4 expression appears to be limited to respiratory epithelial (Shimizu et al., 2000; Asokananthan et al., 2002a), endothelial (Kataoka et al., 2003) and smooth muscle cells (Lan et al., 2000).

An important function of PARs in the respiratory system involves the regulation of airway tone by causing either a contraction or relaxation of smooth muscle cells (Lan et al., 2002). Their involvement has also been demonstrated in remodelling of the lung, through promoting secretion of pro-inflammatory and profibrotic mediator release (Mercer et al., 2007), the production of extracellular matrix components and through stimulating cell mitogenesis (Moffatt et al., 2004). An important role for PARs is also indicated in regulating the inflammatory responses in the airways through recruitment of inflammatory cells (Cocks et al., 2001; Lan et al., 2002).

PAR1 involvement in the airways appears to be important in mediating cell mitogenesis (Blanc-Brude et al., 2005; Walker et al., 2005; Ramachandran et al., 2007), although in some reports thrombin but not PAR1-AP-mediated mitogenesis has been reported, suggesting a non-PAR mechanism for this action of thrombin (Tran et al., 2003; Walker et al., 2005). Thrombin acting through PAR1 is also shown to stimulate fibroblast procollagen production (Chambers et al., 1998) and upregulates connective tissue growth factor (Chambers et al., 2000), an autocrine agent which promotes fibroblast mitogenesis and extracellular matrix production. The importance of PAR1 in mediating fibrotic responses described above is supported further by findings that PAR1 expression is significantly increased in a bleomycin induced model of pulmonary fibrosis, and that thrombin inhibitors can reduce connective tissue growth factor gene expression and collagen accumulation (Howell et al., 2001, 2002). Interestingly, intratracheal administration of APC, a coagulation cascade inhibitory protein that can activate PAR1, is reported to be protective in bleomycin induced pulmonary fibrosis (Yasui et al., 2001). A further study has confirmed that when activating PAR1, APC can stimulate signalling pathways distinct from those activated by thrombin (Riewald et al., 2005). Thus, depending on the activating proteinase, thrombin or APC, PAR1 activation could potentially cause fibrosis or protect from fibrotic lung disease. PAR1-mediated thrombin signalling is also shown to upregulate IL-8 secretion from lung fibroblasts as well as IL-6, IL-8 and prostaglandin E2 release from lung epithelial cells (Asokananthan et al., 2002a). Thrombin, acting via PAR1 in lung epithelial cells is also known to stimulate expression of platelet-derived growth factor (Shimizu et al., 2000). Studies of the ability of PAR1 to regulate airway tone have revealed a dual role that may be species-dependent. In isolated human bronchial tissue, thrombin and a PAR1-AP causes contraction (Hauck et al., 1999). A contractile response to PAR1 activation is also observed in vivo for guinea pig bronchi (Cicala et al., 1999). In contrast, in mouse trachea, PAR1 activation causes an epithelial-dependent relaxation through prostaglandin E2 release (Lan et al., 2001).

PAR2 plays an important role in regulating airway tone, with studies showing both a protective bronchodilatory response as well as detrimental bronchoconstriction. Pulmonary epithelial PAR2 activation in isolated pre contracted bronchi from mice (Cocks et al., 1999a; Lan et al., 2001), rats (Cocks et al., 1999a; Chow et al., 2000), guinea pigs (Cocks et al., 1999a) and humans (Cocks et al., 1999a) results in relaxation. Similar results were demonstrated through in vivo studies where PAR2-AP administration was found to attenuate drug-induced bronchoconstriction in mice (Cocks et al., 1999a). On the other hand, activation of PAR2 on isolated human bronchi has been reported to cause contraction (Schmidlin et al., 2001) and studies done in vivo showed that intratracheal, intravenous or intranasal administration of PAR2 agonists causes bronchoconstriction (Ricciardolo et al., 2000; Schmidlin et al., 2002). A possible explanation for these contradictory results was provided by another study which found that PAR2 activation results in relaxation in the main bronchi and the trachea, but causes a contractile response in tissues isolated from smaller intrapulmonary bronchi (Ricciardolo et al., 2000). Thus, PAR2 may be protective in the larger airways, but activation in the smaller bronchioles might increase airway resistance and may thus be detrimental. PAR2 involvement in mediating eosinophil infiltration into the airways has also been reported (Schmidlin et al., 2002).

One of the major features of airway disease is increased structural cell proliferation and an associated narrowing of airways with reduced airflow. PAR2 is implicated in mediating some of these responses through inducing proliferation of smooth muscle cells (Berger et al., 2001a), epithelial cells (Cairns et al., 1996) and fibroblasts (Akers et al., 2000; Matsushima et al., 2005). Mast cell tryptase, thought to be acting via PAR2, is also known to cause hyperresponsiveness and increased mast cell infiltration into subepithelial tissue in isolated human bronchi (Berger et al., 1999). Similar results are reported in in vivo studies where airway hyperresponsiveness and inflammatory cell infiltration have been observed in response to PAR2 agonist administration (Schmidlin et al., 2002; Ebeling et al., 2005). Moreover, studies using airway epithelial cells grown in vitro have shown PAR2-mediated release of IL-6, IL-8 and prostaglandin E2 (Asokananthan et al., 2002a) and allergen stimulated (possibly PAR2 mediated), GM-CSF and eotaxin release from pulmonary epithelial cells (Sun et al., 2001). Recent reports in cultured airway fibroblasts have also suggested a role for PAR2 in stimulating cytokine production and adhesion molecule expression (Ramachandran et al., 2006).

PARs in the gastrointestinal system

Of all the body systems, the gastrointestinal tract is most exposed to proteinases, both from physiologically relevant processes such as digestion and from exposure to the bacterial flora of the gut. Under pathological conditions, the gut is also exposed to proteinases released by infectious agents and immune cells. Thus, expression and function of PARs may be especially relevant in the gastrointestinal system.

All of the PARs have been found to be expressed in the gastrointestinal tract, with PAR1 present on endothelial cells (Vergnolle et al., 2004), epithelial cells (Buresi et al., 2001, 2002), smooth muscle cells (Kawabata et al., 2004), myofibroblasts (Seymour et al., 2003) and on enteric neurons (Corvera et al., 1997). PAR2 expression has been observed on colonic myocytes (Cenac et al., 2002), enterocytes (Cenac et al., 2002), epithelial cells (Bohm et al., 1996), smooth muscle cells (Al-Ani et al., 1995), endothelial cells (Vergnolle, 2005a), neurons (Reed et al., 2003) and myofibroblasts (Seymour et al., 2005). PAR3 mRNA has also been detected in a variety of tissues including the stomach and small intestine (Ishihara et al., 1997), but the cellular expression patterns have not been studied. Mechanical responses to PAR4-activating peptides indicate that functional PAR4 is present in rat gastric and colonic tissue (Hollenberg et al., 1999; Mule et al., 2004).

PAR activation in the gastrointestinal system is involved in ion transport, permeability, motility and inflammation. Activation of both PAR1 (Buresi et al., 2001, 2002) and PAR2 (Vergnolle et al., 1998; Nguyen et al., 1999; Green et al., 2000; Cuffe et al., 2002) can induce intestinal epithelial chloride secretion, thus enhancing the accumulation of intralumenal fluid, possibly contributing to diarrhoea as well as potentially serving a protective function by flushing away pathogens and toxins. Both PAR1 and PAR2 are reported to be involved in mediating intestinal barrier permeability (Chin et al., 2003; Cenac et al., 2004) and thus can regulate the passage of fluids and microorganisms across the gut mucosa. Organ bath studies with isolated tissue have shown that gastric smooth muscle PAR1 and PAR2 activation stimulate contraction (Saifeddine et al., 1996, 2001), whereas other studies have shown that elsewhere in the GI tract the activation of PAR1 and PAR2 can cause relaxation (Corvera et al., 1997; Cocks et al., 1999b; Mule et al., 2002). Studies done in vivo have demonstrated that PAR1 and PAR2 can enhance gastrointestinal transit, presumably due to a coordinated contraction–relaxation process (Kawabata et al., 2001b). PAR4 activation, due to the administration of PAR4-activating peptides in vivo has also been reported to cause contractile responses in colonic longitudinal muscle (Mule et al., 2004). Thus, PARs may mediate both contractile and relaxation responses in the gastrointestinal tract with the net effect of their actions being to enhance gastrointestinal transit. PARs also play an important role in mediating intestinal inflammation and indeed, important roles in diseases such as inflammatory bowel disease and colitis have been identified (Vergnolle et al., 2004; Cenac et al., 2005; Hansen et al., 2005). Investigation of PAR1 and PAR2 involvement in vivo with the use of models of intestinal inflammation has found that the intraluminal administration of PAR-APs results in a marked inflammatory response characterized by oedema and granulocyte infiltration (Cenac et al., 2002; Vergnolle et al., 2004). The PAR2-mediated inflammatory response was shown further to be mediated by the activation of enteric nerves and the release of neuropeptides whereas the PAR1-mediated response was independent of neuropeptide release.

PARs in the renal system

The initial studies cloning both PAR1 (Rasmussen et al., 1991) and PAR2 (Nystedt et al., 1995) reported abundant expression of mRNA for these receptors in the kidney. Further work has documented the presence of PARs 1, 2 and 4, widely distributed throughout the kidney (Kahn et al., 1998; Gui et al., 2003). In normal human kidney tissue, PAR1 is expressed on endothelial, mesangial and epithelial cells (Grandaliano et al., 2000) and both PAR1 and PAR3 expression has been reported in primary human renal carcinoma cells (Kaufmann et al., 2002). Another study done in vivo has identified an important role for PAR1 in a murine model of crescentic glomerular nephritis (Cunningham et al., 2000). PAR1 activation enhanced glomerular crescent formation, T cell and macrophage infiltration, fibrin deposition and elevated serum creatinine, all of which are prominent features of glomerular nephritis. Exposure to hirudin, an inhibitor of thrombin, and studies in PAR1-knockout mice showed attenuated responses with respect to all of the above indices of glomerular nephritis.

PAR2 activation has been found to cause vasodilation in a perfused rat kidney model (Trottier et al., 2002; Gui et al., 2003). In this model, both trypsin and a PAR2-AP increase renal perfusion flow, whereas the PAR1-AP, TFLLR-NH2 and thrombin cause a marked reduction in renal flow (Gui et al., 2003). These data suggest that in the renal vasculature, PAR1 and PAR2 mediate opposing effects in terms of regulating blood flow and glomerular filtration. These differences between the effects of PAR1 and PAR2 activation may be significant in the settings of disease where proteinases may activate one or the other receptor. In addition, a recent report has implicated PAR2 activation by factor Xa in mediating mesangial cell proliferation (Tanaka et al., 2005).

Therapeutic implications: antagonizing proteinase-mediated signalling

Inhibiting PARs and their activating proteinases

Proteinases represent an important class of therapeutic targets, perhaps best typified by the development of the angiotensin converting enzyme inhibitors (Cushman et al., 1999). The success of the angiotensin converting enzyme inhibitors as a therapy for hypertension can be greatly tempered by the failure of metalloproteinase inhibitors to prove of value for cancer chemotherapy in humans, despite promising data obtained with these inhibitors in animal model systems (Coussens et al., 2002). Even the success of anticoagulant therapy with inhibitors of thrombin is qualified by the bleeding risk that accompanies the use of this approach (Hirsh, 2003; Weitz, 2003). Thus, in thinking of therapeutic approaches that target one of the PARs, efforts to develop receptor inhibitors rather than targeting the activating enzymes have taken priority. As outlined below, substantial success has been achieved to develop PAR1 antagonists, and there has been some progress in developing inhibitors of PAR4. PAR2, however, has presented the greatest challenge, with very limited success in the development of potent receptor antagonists. Largely because of (1) the difficulty in synthesizing serine proteinase inhibitors that are enzyme-selective and (2) the ability of several serine proteinases to activate PARs 2 and 4, the receptor-activating proteinases themselves do not appear to be attractive therapeutic targets for blocking the activation of these two PARs.

PAR antagonists

Targeting PAR1

A number of peptide and peptidomimetic PAR1 antagonists, for both experimental studies and pharmaceutical use in humans are currently available. Work focused on the development of selective and potent non-peptide PAR1 antagonists has succeeded with the compounds described by the Johnson & Johnson team (RWJ-56110 and RWJ-58259) (Andrade-Gordon et al., 1999; Zhang et al., 2001; Maryanoff et al., 2003) and by the Schering group (Chackalamannil et al., 2005). The RWJ compounds, while effective in non-human models of thrombosis (Derian et al., 2003) may not be clinically suitable as systemic antithrombotics, but may prove successful as local inhibitors of thrombosis for example for arterial stent adducts. The orally active PAR1 antagonist developed by Schering (PAR1 antagonist 55 or SCH-205831), which inhibits PAR1 by competitively inhibiting TL-binding site (Chackalamannil et al., 2005) has been found to be effective as an antithrombotic agent in humans. A recent report describes the successful completion of a phase-II clinical trial of a related PAR1 antagonist, SCH5303048, providing hope for the availability of an effective orally available PAR1-targeted antithrombotic compound (Camerer, 2007), (http://www.sch-plough.com/schering_plough/news/release.jspreleaseID=977603). A number of other PAR1 antagonists have also been previously described, but their use is limited by relatively low efficacies (Pakala et al., 2000; Nantermet et al., 2002; Kato et al., 2003; Seiler and Bernatowicz, 2003).

Targeting PAR2

PAR2 has represented a substantial challenge in terms of developing a high potency receptor-selective antagonist. Some success has been achieved with antagonists based on modified PAR2 TL peptides, but the potencies of these agents (for example FSLLRY-NH2 and LSIGRL-NH2) are too low to be considered for general systemic use (Al-Ani et al., 2002). These peptides are able to inhibit trypsin activation of PAR2 through possible interactions with the TL docking site, but are of limited use to block activation of PAR2 by PAR-APs. The peptide, LIGK-NH2, has also been found to block trypsin and PAR2-AP activation of the receptor and a peptidomimetic based on this peptide sequence, ENMD-1068 (N1-3-methlybutyryl-N4-6-aminohexanoyl-piperazine), has proved of use in studies done in vivo, despite its very low potency (Kelso et al., 2006). Although ENMD-1068 can serve as a ‘lead’ compound, the synthesis of high potency orally available PAR2 antagonists is a goal that has yet to be reached. Notwithstanding, the use of PAR2-targeted antibodies that block proteinase activation of the receptor may ultimately prove of therapeutic value. Proof of principle for the use of such antisera (for example a ‘B5’ polyclonal or SAM11 monoclonal antibody) has been demonstrated in a murine arthritis model in which both antisera blocked the development of joint swelling (Kelso et al., 2006). That same study demonstrated the possibility of using PAR2-targeted short interfering RNA as a mode of therapy, but neither the antibody nor the short interfering RNA approach is sufficiently developed to be considered for therapeutic use as of yet.

Targeting PAR4

PAR4 antagonism has been demonstrated using the peptide trans-cinnamoyl-YPGKF-NH2 in rat and human platelets (Hollenberg and Saifeddine, 2001), but this peptide can have non-PAR actions that limit its use in vivo (Hollenberg et al., 2004). A novel approach to receptor inhibition, through targeting the receptor intracellular loops with palmitoylated membrane penetrating peptides termed pepducins, has succeeded in developing a relatively high potency (micromolar range) PAR4 antagonist (Covic et al., 2002a, 2002b; Kuliopulos et al., 2003). Although not completely selective for PAR4 and although the pepducins can have unusual non-PAR pharmacological actions (Hollenberg et al., 2004), the pepducin, P4pal-10 has proved of utility to block PAR4 activation both in vivo and in vitro (Covic et al., 2002a, 2002b).

Targeting the proteinases

As pointed out above, the PAR-activating proteinases are challenging therapeutic targets. That said, proteinase inhibition in a restricted environment may have value. For instance, the serine proteinase-dependent, PAR2-mediated inflammation caused by Citrobacter rodentium in a murine model of colitis can be attenuated by the oral administration of the SERPIN, soya trypsin inhibitor. Further, the systemic administration of proteinase inhibitors in a murine model of joint inflammation is able to diminish joint swelling significantly (Kelso et al., 2006). Thus, selective targeting of specific PAR-activating proteinases may have value in selected therapeutic settings.

Conclusions

Proteinases must now be considered as important hormone-like mediators that can signal to cells and tissues by a number of mechanisms including, but by no means restricted to the regulation of PARs, as outlined in Figure 3. Proteinases mediate a number of physiological and pathophysiological settings; and their target PARs are also now recognized for the many pathophysiological roles they can play. The complexity of this system, comprising so many proteinase agonists and their PAR and non-PAR mechanisms of signalling makes it a challenge to study because of the number of variables to be considered. The expression of proteinases and their inhibitory SERPINs, the expression of the PARs and the regulation of their expression by factors such as inflammatory mediators, as well as the specific proteinase-PAR combination present in a local environment, resulting in either receptor activation or silencing, can all contribute to protective or harmful responses in vivo. Further, while there is now a growing recognition of the signalling role of proteinases, the next frontier to face is developing an understanding of the biology and physiology of the proteinase inhibitors (for example SERPINs, or tissue inhibitors of metalloproteinases). Clearly, the proteinase–PAR combination that plays a number of roles in a variety of disease states ranging from inflammatory to neurodegenerative disorders, represents an attractive target for the design of new therapeutic modalities. It is suggested that the reader ‘stay tuned’ for exciting developments in this area in the future.

Figure 3.

Signalling by proteinases: PARs and more. The scheme illustrates the main mechanisms that proteinases can use to regulate cell signalling in addition to targeting the PARs. The potential proteolytic targets for proteinase-mediated signalling include growth factor receptors, such as the one for insulin that can be activated by trypsin (Cuatrecasas, 1971; Shoelson et al., 1988), membrane-tethered agonists like heparin-binding EGF that can be released by metalloproteinase action to activate the EGF receptor (Prenzel et al., 1999), and either processing of peptide agonist precursors, like angiotensinogen, or inactivation of peptide agonists like neurokinins by membrane-tethered proteinases. Extracellular matrix–integrin interactions that regulate intracellular signalling are also shown as potential targets for enzymes, like the matrix metalloproteinases. Not shown is the ability of a proteinase like thrombin to regulate cell function via its non-catalytic domains (Bar-Shavit et al., 1984, 1986; Glenn et al., 1988). EGF, epidermal growth factor; PAR, proteinase-activated receptor.

Acknowledgments

The studies from the author's laboratory cited in this article are supported in part by grants from the Canadian Institutes of Health Research (CIHR) and the NIH, USA. RR is the recipient of a Canadian Association of Gastroenterology/CIHR/Ortho-Jensen post-doctoral fellowship.

Conflict of interest

The authors state no conflict of interest.

Ancillary