Modulation of visceral pain and inflammation by protease-activated receptors

Abstract

The gastrointestinal (GI) tract is exposed to a large array of proteases, under both physiological and pathophysiological conditions. The discovery of G protein-coupled receptors activated by proteases, the protease-activated receptors (PARs), has highlighted new signaling functions for proteases in the GI tract, particularly in the domains of inflammation and pain mechanisms. Activation of PARs by selective peptidic agonists in the intestine or the pancreas leads to inflammatory events and changes in visceral nociception, suggesting that PARs could be involved in the modulation of visceral pain and inflammation. PARs are present in most of the cells that are potentially actors in the generation of irritable bowel syndrome (IBS) symptoms. Activation of PARs interferes with several pathophysiological factors that are involved in the generation of IBS symptoms, such as altered motility patterns, inflammatory mediator release, altered epithelial functions (immune, permeability and secretory) and altered visceral nociceptive functions. Although definitive studies using genetically modified animals, and, when available, pharmacological tools, in different IBS and inflammatory models have not yet confirmed a role for PARs in those pathologies, PARs appear as promising targets for therapeutic intervention in visceral pain and inflammation processes.

Introduction

Proteases represent 2% of the human genome and are present at particularly high levels in the gastrointestinal (GI) tract (Caughey, 1995). Thus, it is not surprising if, in addition to their role in protein degradation and/or digestion, certain proteases exert also a role as signaling molecules, regulating cell functions by cleaving receptors activated upon proteolysis. Those receptors constitute a family of G protein-coupled receptors (four members have been cloned thus far), which are called protease-activated receptors (PARs). Activation of those receptors has been shown to interfere with inflammatory and nociceptive pathways in different tissues (Vergnolle et al., 2001b). Understanding the functional role of PARs in a system as exposed to proteases as the GI tract represents an important and exciting challenge that could lead to define new pharmacological targets for GI diseases in general, and in particular for pathologies associating inflammatory and visceral pain disorders.

PAR structure and activation

PARs are activated by a unique mechanism that first involves recognition of an extracellular domain of the receptor situated on the N-terminus, by a protease (Rasmussen et al., 1991; Vu et al., 1991; Ishihara et al., 1997; Xu et al., 1998; Coughlin, 1999). Some proteases such as thrombin, bind the receptors. Others, such as trypsin or tryptase, do not need to bind the receptor in order to cleave it. Then, cleavage by proteolysis of this recognized site occurs, which exposes a new N-terminus domain that acts as a tethered ligand domain interacting with domains situated in the second extracellular loop of the receptor, to induce an intracellular signal (see Figure 1a). PARs respond to a variety of proteases, although thrombin for PAR₁, PAR₃ and PAR₄, and trypsin for PAR₂ and PAR₄ are usually regarded as the main activators of PARs. Proteases of the coagulation cascade such as factor Xa and VIIa can activate PAR₂ and PAR₄ (Dery et al., 1998; Coughlin, 1999; Hollenberg & Compton, 2002). However, co-factors such as tissue factor are needed to enhance the ability of factor VIIa to activate PAR₂. Cathepsin G released from neutrophils triggers PAR₄ activation and can activate PAR₁. Tryptase released from mast cells can activate PAR₂, while bacterial proteases such as gingipains can activate PAR₁, PAR₂ and PAR₄ (Dery et al., 1998; Coughlin, 1999; Hollenberg & Compton, 2002). Proteases released by dust mites are also able to activate PAR₂ (see Table 1) (Lourbakos et al., 2001; Asokananthan et al., 2002). Soluble forms of integral membrane proteases such as membrane-type serine protease 1 can also activate PARs in general, and PAR₂ in particular. Since PAR₂ and membrane-type serine protease 1 have similar tissue distribution, this further suggests a role for intra-membrane-type serine protease 1 as an endogenous activator of PAR₂. Although some proteases are common to the activation of different PARs, a single protease agonist activates distinct receptors with different potencies. For instance, thrombin can activate PAR₁, PAR₃ and PAR₄, with a highest potency for PAR₁, then a lower potency for PAR₃, and is finally weaker for the activation of PAR₄ (for reviews, see Dery et al., 1998; Coughlin, 1999; Hollenberg & Compton, 2002). In vivo activation of PARs requires the release of active proteases in the vicinity of the receptors. Depending on the physiological or pathophysiological situation, different proteases might activate the same receptor. Determining which proteases are endogenous activators of PARs in selected pathologies would be of great importance for the potential use of protease inhibitors as therapeutic options.

Figure 1.

Mechanism of activation of PARs. (a) Proteases cleave the extracellular N-terminus domain to release a new N-terminus domain that acts as a tethered ligand binding and activating the receptor to induce an intracellular signal. (b) Synthetic peptides corresponding to the tethered ligand domain can mimic the effects of the proteolytically cleaved N-terminal domain to specifically activate the receptor.

Table 1.

Structure and activation of protease-activated receptors

	Number of amino acids	Tethered ligand sequence	Activating protease	Peptidic selective agonists	Antagonists
PAR1	425 aa	Human: SFLLRMouse/rat: SFFLR	Thrombin, trypsin, cathepsin G, granzyme A, factor VIIa, factor Xa, gingipain, plasmin	TFLLR ApfFRChaCitY	RWJ-56110RWJ-58259SCH-79797SCH-73754FR-171113BMS-200261‘Merck isoxazole 1'Trans-cinnamoyl-(p-F-Phe)-(p-guanidino-Phe)-LRR
PAR2	397 aa	Human: SLIGKVMouse/rat: SLIGRL	Trypsin, tryptase, factor Xa, factor VIIa, tissue factor, acrosin, trypsin IV, membrane-type serine protease 1, dust mite proteases	SLIGRLTrans-cinnamoyl-LIGRLO 2-furoyl-LIGRLO
PAR3	374 aa	Human: TFRGAPMouse: SFNGGP	Thrombin, trypsin
PAR4	385 aa	Human: GYPGQVMouse: GYPGKF	Thrombin, trypsin, cathepsin G, factor VIIa, factor Xa, gingipain, trypsin IV	GYPGKFGYPGFKGYPGQVAYPGKF	Trans-cinnamoyl-YPGKF Palmitoyl-SGRRYGHALR

With the exception of PAR₃, synthetic peptides corresponding to the tethered ligand domain released upon proteolytic activation of PAR₁, PAR₂ and PAR₄ can directly activate the receptor (see Figure 1b) (Hollenberg & Compton, 2002). Those peptides constitute important pharmacological tools to understand the functions of PARs. Pharmacological studies have determined optimum peptidic sequences for selectivity and potency of PAR activation (see Table 1). For example, the tethered ligand peptide SFLLR, which corresponds to the human PAR₁, has been shown to also activate PAR₂, but a simple replacement of the serine residue by a threonine residue (TFLLR peptide) gives a more selective peptidic agonist for PAR₁ activation. Synthetic peptide corresponding to the mouse PAR₄ (GYPGKF) constitutes a selective agonist for mouse PAR₄, but substitution of the G residue by an A residue (AYPGKF peptide) creates a selective PAR₄ agonist, at least 10 times more potent than the tethered ligand peptide (Faruqi et al., 2000).

The understanding of the physiological and pathophysiological role of PARs has been hampered by difficult pharmacologic intervention to inhibit the activation of PARs. Different strategies and extensive chemical synthetic efforts have been employed to develop PAR antagonists (reviewed in Derian et al., 2003). Several PAR₁ antagonists are now available (see Table 1), and their use has demonstrated the prime role for PAR₁ in thrombotic and restenotic events. However, no studies have reported the use of PAR₂ or PAR₃ antagonists in bioassays or in vivo studies. Two peptidic PAR₄ antagonists have been tested in platelet assays (see Table 1), but their pharmacological properties in other tissues and bioassays still have to be established. Gene-deletion approach with the use of PAR-deficient mice appears as an alternate reliable approach to understand the role of PARs.

Proteases and PARs in the GI tract

PARs have been detected in several cell types throughout the entire GI tract and pancreas (see Figure 3 and Table 2). PAR₁, PAR₂ and PAR₄ have been detected in enterocytes (Kong et al., 1997; Buresi et al., 2001; Mule et al., 2004), where PAR₁ and PAR₂ have been shown to be functional both on the apical and baso-lateral membranes (Kong et al., 1997; Chin et al., 2003). PAR₁ and PAR₂ are also expressed in human colon cancer cell lines where their activation modulates proliferation and motility (Darmoul et al., 2001; 2003). PAR₁ and PAR₂ have been detected in enteric neurons (extrinsic and intrinsic submucosal and myenteric neurons), where they co-express with neuropeptides such as substance P and calcitonin gene-related peptide (CGRP) (reviewed in Vergnolle et al., 2003c). Although its expression on enteric neurons has not been clearly established, functional PAR₄ seems to be present on enteric neurons, which respond to PAR₄ agonists by evoking a depolarizing response in the guinea-pig small intestine (Gao et al., 2002). PAR₁ and PAR₂ have also been shown to be expressed in intestinal myofibroblasts (Seymour et al., 2001; 2003) and in mast cells (D'Andrea et al., 2000). The fact that PAR₁, PAR₂ and PAR₄ agonists cause contraction and/or relaxation in isolated GI segments suggest that those receptors are expressed in smooth muscle cells (Cocks et al., 1999; Kawabata et al., 2000a,2000b; Gao et al., 2002; Mule et al., 2002b; 2004; Zhao & Shea-Donohue, 2003). However, only the expression of PAR₂ in intestinal myocytes and PAR₁ in irradiated intestinal smooth muscle cells has been clearly established (Corvera et al., 1997; Wang et al., 2002). PAR₁, PAR₂ and PAR₄ expression both on endothelial surfaces and leukocytes has also been demonstrated (Hou et al., 1998; Vergnolle et al., 2002). Messengers of RNA for PAR₁ and PAR₂ have been found in parotid, sublingual and submaxillary glands, and PAR₂ protein has been detected in the pancreatic duct epithelium and pancreatic acinar cells (Nguyen et al., 1999; Kawabata et al., 2000c,2000d; 2002; Kawabata, 2003).

Figure 3.

Interactions of PARs with components of the pathogenesis of IBS. PAR₁ and PAR₂ are present on enterocytes, where their activation potentially induces an increase in permeability, which could lead to bacterial translocation, sepsis and inflammation, as well as hypersensitivity. Direct activation of PAR₁ and PAR₂ on enteric neurons interferes with nociceptive and motor functions. Activation of PARs on cells involved in inflammatory responses (mast cells, neutrophils, lymphocytes, macrophages, fibroblasts, endothelium or neuropeptides containing neurons) could lead to the generation of inflammatory mediators and provokes further signs of inflammation such as granulocyte recruitment and plasma extravasation. Ion secretion and potentially water flux disturbances (diarrhea) could be evoked by PAR activation on enterocytes, fibroblasts or enteric nerves. PAR activation in the central nervous system modulates nociceptive responses to peripheral stimulation, playing a potential role in hypersensitivity states and eventually (although never studied) in enteric motor dysfunctions. Activation of PARs on enterocytes, enteric neurons or lymphocytes by bacterial proteases from luminal or infiltrated pathogens could also participate in the generation of symptoms associated with IBS.

Table 2.

Localization and function of protease-activated receptors in the GI tract

	Localization in the GI tract	Known functions
PAR1	Enterocytes	Increase permeability, apoptosis, chloride secretion, prostaglandin release
	Human colon cancer epithelium	Proliferation and motility
	Myenteric neurons	Suppression of fast excitatory postsynaptic potential
	Submucosal neurons	Inhibition of chloride secretion,
	Fibroblasts	Prostaglandin release
	Mast cells
	Smooth muscle	Relaxation/contraction
	Endothelium	Gap formation
PAR2	Enterocytes	Chloride secretion, prostaglandin production, eicosanoid production
	Human colon cancer epithelium	Proliferation
	Myenteric neurons	Neuropeptide release, increased excitability, suppression of fast excitatory postsynaptic potential
	Submucosal neurons	Neuropeptide release, stimulate epithelial ion secretion, hyperexcitability
	Fibroblasts	Prostaglandin release, proliferation
	Mast cells
	Smooth muscle	Relaxation/contraction
	Pancreatic duct epithelium	Ion channel activation
	Pancreatic acinar cells	Amylase secretion
	Endothelium/leukocyte interface	Rolling, adhesion, transmigration, gap formation
PAR3	Detected by RT–PCR in whole GI tissues (stomach and small intestine), but unidentified cell type
PAR4	Enterocytes
	Submucosa	Contraction of longitudinal muscle
	Enteric neurons	Depolarization
	Endothelium/leukocyte interface	Rolling, adhesion, transmigration

The GI tract and pancreas are particularly exposed to a large array of proteases (see Figure 2). Trypsin is released in the upper GI tract lumen, and in pancreatic duct under its inactive (proactive) form trypsinogen, for physiological digestive purposes. On mucosal surfaces, a balance between proteolytic activity and the presence of protease inhibitors such as pancreatic secretory trypsin inhibitor (PSTI) is constantly present. PSTI, which is released by mucus-secreting cells throughout the GI tract, prevents premature activation of pancreatic proteases, and protects mucosal surfaces from exposure to active proteolytic enzymes (Marchbank et al., 1996; 1998). Thus, the balance between proteolytic activity in the center of the lumen and the presence of protease inhibitors at mucosal surfaces warrants in the upper intestine efficient digestive processes and mucosal protection. In the lower GI tract, trypsin is not released into the lumen for digestive purposes. However, inflammatory bowel disease (IBD) patients showed an increased trypsin activity in their colonic luminal content, suggesting that the balance between tryptic proteolytic activity and protease inhibitors is broken in this particular pathophysiological situation, and that trypsin is present in the lumen of the lower GI tract, associated with inflammatory conditions. Trypsinogen can also be synthesized by several different extrapancreatic cell types including endothelium (Koshikawa et al., 1997; 1998). Trypsinogen IV and trypsin IV, which have been shown to activate PAR₂ and PAR₄ (see Table 1), have been found in epithelial cell lines from prostate, colon and airways (Cottrell et al., 2003; 2004). Thus, it is conceivable that endothelium- or epithelium-derived trypsin can also be present on the basolateral side of the intestinal barrier. Tryptase, which is expressed by almost all subsets of human mast cells (Caughey, 1995), is released upon mast cell degranulation. Although tryptase is considered poorly diffusible, in pathologies such as intestinal inflammation, allergy, or even stress, large amounts of tryptase have been found in the vasculature and the gut lumen of patients or animals (Plebani et al., 1992; Miller, 1996; Santos et al., 1998; 1999; Gelbmann et al., 1999). Proteases of the coagulation cascade such as thrombin, factor VIIa and Xa are also potentially present in the GI tract during inflammation or tissue trauma. These events have been shown to lead to the presence of pro-thrombin and active thrombin into the intestinal lumen and deeper within the intestinal tissues (McHugh et al., 1996), where it could activate PARs. Cathepsin G, which is released upon neutrophil activation, can also be massively present in GI tissues associated with inflammatory conditions. The transepithelial migration of neutrophils towards mucosal surfaces upon inflammation (Edens et al., 2002) also renders plausible the presence of cathepsin G into the gut lumen. Finally, mucosal surfaces are constantly exposed to bacterial products, and particularly to bacterial proteases. Although no studies have reported the effects of intestinal bacteria (pathogens or nonpathogens) on PAR activation, studies have reported the possible activation of PAR₁, PAR₂ and PAR₄ by proteases from pathogens such as Porphyromonas gingivalis (Lourbakos et al., 2001) or dust mites (Sun et al., 2001; Asokananthan et al., 2002). Thus, a variety of proteases may act on PARs and influence GI functions at several levels depending on the PAR-expressing target cells. Whether PARs are activated in physiological or pathophysiological settings might depend on the proximity of the receptors to digestive enzymes, inflammation-associated protease or proteases released by pathogen or nonpathogen organisms. However, several studies using different PAR agonists and antagonists point to a role for PARs in inflammatory and nociceptive mechanisms of the GI tract.

Figure 2.

Presence of proteases in the GI tract. Trypsin is released in the lumen of the GI tract for digestive purposes, but can also be present deeper into the tissues released from endothelial or epithelial cells. Proteases of the coagulation cascade such as thrombin, factor Xa and VIIa can be released into the lumen or the GI tissues upon tissue damage. When they degranulate, mast cells release massive amounts of tryptase both in the gut lumen and vasculature. Neutrophils present either in the tissues or translocated in the lumen release cathepsin G. Epithelial cells are exposed to bacterial proteases from the luminal side.

Proteases and PARs: role in inflammation of the GI tract

Direct injection of thrombin, trypsin, tryptase or selective agonists for PAR₁ and PAR₂ into the paw of rodents produces edema and granulocyte infiltration, two of the main features of inflammation (Cirino et al., 1996; Vergnolle et al., 1999a,1999b). Since those receptors and proteases are also highly present in the GI tract, they could be involved in GI inflammatory processes.

Intestine

Recently, we have shown that luminal administration of selective peptidic agonists for PAR₁ (TFFLR), PAR₂ (SLIGRL) and PAR₄ (AYPGKF) provoked a colonic inflammation within a few hours (from 4 to 24 h). This inflammation was characterized by an increased wall thickness, the presence of erythema and significant infiltration of granulocytes (Cenac et al., 2002; Ferazzini et al., 2004; Vergnolle et al., 2004). PAR₂-induced colitis was dependent on sensory neuron activation, substance P and CGRP release (Cenac et al., 2003; Nguyen et al., 2003). Intracolonic administration of PAR₂ also resulted in increased paracellular permeability, as observed by the passage of ⁵¹Cromium-EDTA from the lumen to the blood and by the presence of translocated bacteria in different intraperitoneal organs (Cenac et al., 2002). Tight junction blockers and inhibitors of myosin-light-chain kinase blocked this increased permeability without affecting the level of granulocyte recruitment (Cenac et al., 2003). This suggests that although increased permeability and granulocyte infiltration might both participate in the generation of the inflammatory response, they constitute separated events that can occur independently. Activation of PAR₁ also caused increased intestinal permeability both in vivo (intracolonic administration of TFLLR) and in vitro (on enterocytes monolayers), by a mechanism involving the induction of apoptosis of the intestinal epithelium, and the activation of tyrosine and myosin-light-chain kinases (Chin et al., 2003). The fact that PAR₁ activation can compromise the epithelial barrier function suggests that PAR₁ could be implicated in the pathogenesis of a number of disorders affecting mucosal surfaces, including IBD and infectious diseases. Using PAR₁-deficient mice, we have been able to show that PAR₁ activation is implicated in the pathogenesis of trinitrobenzene sulfonic acid (TNBS)-induced colitis. PAR₁^−/− mice showed significantly less inflammatory damage than wild-type controls after the induction of TNBS colitis (Vergnolle et al., 2003a). Another recent study showed that PAR₁ agonist-induced epithelial cell ion transport was altered after nematode infection, also suggesting a role for PAR₁ in infectious intestinal diseases (Fernandez et al., 2003). The fact that PAR₂-induced colonic inflammation is regulated by a neurogenic mechanism is in favor of a role for PAR₂ in infectious intestinal diseases, if we consider the fact that enteric infections are largely mediated by neurogenic mechanisms (Spiller, 2002; Vergnolle et al., 2003c) and the fact that PAR₂ can be activated by bacterial proteases (Lourbakos et al., 2001; Sun et al., 2001; Asokananthan et al., 2002). However, no study has reported yet such role for PAR₂. In IBD models, however, activation of the enteric nervous system (ENS) and further release of neuropeptides protects against inflammatory damage (Collins, 2000). Thus, in the setting of chronic inflammation, PAR₂-induced ENS activation might exert protective effects rather than pro-inflammatory effects. This hypothesis is further supported by the findings of Fiorucci et al. (2001), who have observed that daily systemic treatments with PAR₂ agonists diminished inflammatory damage caused by the intracolonic injection of TNBS. These results suggest that PAR₂ agonists might be beneficial in the setting of chronic inflammation such as IBD, where they might exert protective effects though ENS activation. However, the use of such agonists might also enhance visceral hypersensitivity (see next chapter), which then would be detrimental, potentially causing more pain to patients. The use of PAR₂-deficient mice in different models of infectious colitis or IBD should help, in the very near future, to clarify the role of PAR₂ in those pathologies. However, the overexpression of PAR₂ observed in biopsies from ulcerative colitis patients strongly suggests a role for PAR₂ in IBD (Kim et al., 2003). We have reported that treatment of mice with the PAR₄ antagonist palmitoyl-SGRRYGHALR reduced inflammation induced by dextran sodium sulfate (Ferazzini et al., 2003). However, the specificity of such antagonist could still be questioned and only experiments performed with PAR₄-deficient mice would completely clarify the role of PAR₄ in IBD models.

Stomach

PAR₂ can potentially modulate a variety of gastric functions, through its ability to induce the secretion of mucus and pepsinogen, to suppress acid output, to increase mucosal blood flow and to induce gastric strip contraction/relaxation (reviewed in Kawabata, 2003; Nishikawa & Kawabata, 2003). Except for pepsinogen secretion, all these effects of PAR₂ in the stomach favor a protective role for PAR₂ in gastric mucosa. As a matter of fact, treatment of rats with the selective PAR₂ agonist SLIGRL has been shown to be protective in models of gastritis induced by indomethacin or ethanol and HCl (Kawabata et al., 2001a). Here again, this protective effect of PAR₂ agonist has been shown to be mediated, at least in part, by the activation of sensory neurons (Kawabata et al., 2001a). The selective agonist for PAR₁ also reveals protective effects for gastric mucosa, inhibiting acid secretion and increasing gastric mucosal blood flow (Kawabata, 2003). However, this protective effect of PAR₁ activation does not seem to be mediated by a neurogenic mechanism, but rather by an enhancement of endogenous protective prostaglandin production (Kawabata, 2003).

Pancreas

Inflammation of the pancreas leads to the premature activation of trypsin, which can then signal to pancreatic acini and duct cells through the activation of PAR₂ (Cottrell et al., 2003). Trypsin stimulates fluid and electrolyte secretion in the pancreatic ducts through a mechanism involving the activation of PAR₂ (Nguyen et al., 1999). PAR₂ selective agonist causes a prompt increase, followed by transient decrease of pancreatic juice secretion, and also facilitates amylase secretion (Kawabata et al., 2000d; 2002; Kawabata, 2003). These PAR₂-induced effects in the pancreas may also be protective, although in vivo studies using PAR₂-deficient mice and/or PAR₂ antagonists still have to be performed to demonstrate such hypothesis.

Proteases and PARs: mediators of visceral perception

Afferent sensory fibers of the ENS convey sensory data to the central nervous system, and the presence of PAR₁ and PAR₂ on those fibers has suggested a role for those receptors in visceral nociception mechanisms.

PAR₂

It has been shown that peripheral (intraplantar) administration of sub-inflammatory doses of the PAR₂ agonist SLIGRL, and also trypsin and tryptase, provoked nociceptor activation at a spinal level, together with a severe and prolonged (>24 h) thermal and mechanical hyperalgesia (Vergnolle et al., 2001a). Transient receptor potential vanilloid-like 1 (TRPV1) mediated PAR₂-induced thermal, but not mechanical hyperalgesia (N. Vergnolle, unpublished work). When injected into the colon lumen, sub-inflammatory doses of PAR₂-activating peptide and trypsin caused visceral hyperalgesia, as observed by an increased number of abdominal contractions in response to colorectal distension (Coelho et al., 2002). The increased Fos expression observed in the superficial laminae of the dorsal horn in response to intracolonic administration of PAR₂ agonist also suggests that visceral activation of PAR₂ induces the activation of second-order neurons at a spinal level (Coelho et al., 2002). This was confirmed by a similar study, which showed that pancreatic activation of PAR₂ also provoked an increased Fos expression in the superficial laminae of the dorsal horn (Hoogerwerf et al., 2001). Since in both cases (colonic lumen or pancreatic duct exposure to PAR₂ agonists) increased Fos expression was observed primarily in laminae I and II of the dorsal horn, which contain nociceptive nerve terminals, this suggests that a central nociceptive signal is triggered by visceral activation of PAR₂. Although these experiments showed that peripheral activation of PAR₂ caused nociceptor activation, they did not unequivocally show that this effect was due to a direct activation of PAR₂ on sensory neurons. The presence of functional PAR₂ on dorsal root ganglia neurons (Steinhoff et al., 2000) and the fact that, in those cells, PAR₂ agonists enhanced KCl- and capsaicin (TRPV1 agonist)-evoked release of CGRP (a spinal mediator of nociception) (Hoogerwerf et al., 2001) strongly suggest that PAR₂ agonists directly activate primary afferents to induce a nociceptive signal. Other evidence suggesting that PAR₂ agonists signal directly to neurons to induce hyperalgesia comes from a study by Reed et al. (2003), which showed that trypsin, tryptase and PAR₂-activating peptide induced prolonged hyperexcitability of submucosal neurons isolated from the guinea-pig ileum. In a recent study, we have shown that not only peripheral but also spinal activation of PAR₂ can participate in inflammatory visceral hyperalgesia (Vergnolle et al., 2003b). Intrathecal injection of the selective PAR₂-activating peptide SLIGRL, but not the control peptide or their vehicle, increased (in a dose-dependent manner) the number of writhing behaviors in response to intraperitoneal injection of acetic acid. Intrathecal injections of tachykinin-1 receptor and CGRP receptor antagonists were able to block this PAR₂-induced enhancement of visceral hyperalgesia, suggesting that spinal release of substance P and CGRP are involved in PAR₂-induced spinal effect (Vergnolle et al., 2003b). These results suggest that PAR₂ activation from the periphery, but also at a spinal level, might play an important role in states of hypersensitivity.

Although a clear role for PAR₂ in inflammation-induced somatic hyperalgesia and particularly in mast cell degranulation-induced hyperalgesia has been demonstrated using PAR₂-deficient mice (Steinhoff et al., 2000), such role in visceral hyperalgesia has yet to be demonstrated.

PAR₁

In contrast to what has been shown for PAR₂, sub-inflammatory doses of PAR₁ agonist did not provoke hyperalgesia when injected into the rat paw, but increased nociceptive threshold and significantly inhibited inflammatory hyperalgesia induced by intraplantar injection of carrageenan (Asfaha et al., 2002). Intraperitoneal injection of the selective PAR₁ agonist TFLLR also inhibited visceral pain behaviors induced by the intraperitoneal injection of acetic acid, by a mechanism involving the activation of opioid receptors (Vergnolle et al., 2003a). Intracolonic administration of thrombin and TFLLR has also been shown to produce visceral analgesia, reducing the number of abdominal contractions in response to colorectal distension (Coelho & Bunnett, 2003). Here again, there are no evidences that such analgesic effect of PAR₁ activation is due to a direct activation on primary afferents. Further in vitro studies would be necessary to determine whether or not PAR₁ agonists could cause hyperpolarization of the sensory neuron membrane.

Proteases and PARs in the pathogenesis of irritable bowel syndrome (IBS)

The IBS is characterized by abdominal pain or discomfort associated with disturbed defecation and often bloating. Several pathophysiological factors are involved in the generation of symptoms of IBS: psychological factors, altered motility patterns, inflammatory mediator release, altered epithelial functions (permeability, ion exchanges and immune) and altered visceral nociceptive functions (hypersensitivity) (for a review, see Mayer & Collins, 2002). PARs are present in most of the cells that are potentially actors in the generation of IBS symptoms, and PAR activation interferes with several components of the pathogenesis of IBS (see Figure 3).

First, as discussed in the precedent paragraph, PARs are present on sensory neurons, where they can potentially interfere in a direct manner with the transmission of nociceptive signal. Activation of PARs on enteric neurons (PAR₁ and PAR₂) can also provoke the release of neuropeptides such as substance P and CGRP (Steinhoff et al., 2000; de Garavilla et al., 2001), which in turn activate their receptors present on endothelium to induce plasma extravasation. Such PAR-induced micro-inflammation might participate in the generation of IBS symptoms, as low levels of inflammation have been proposed to be involved in the pathogenesis of hypersensitivity (Collins, 2001; Collins et al., 2001; Bueno & Fioramonti, 2002). Activation of PARs on cells that are involved in inflammatory or immune responses, such as neutrophils, mast cells, lymphocytes, macrophages, or fibroblasts, might also be involved in causing hypersensitivity, by indirect stimulation of primary afferents (through the release of prostaglandins, neuropeptides or cytokines). Such indirect pro-algesic effect is in accordance with the hyperalgesia and nociceptor activation that have been demonstrated in response to visceral exposure to PAR₂ agonists (Hoogerwerf et al., 2001; Coelho et al., 2002). However, PAR₁ agonists did not cause hyperalgesia, but decreased nociceptive threshold and inhibited painful behaviors (Asfaha et al., 2002; Coelho & Bunnett, 2003; Vergnolle et al., 2003a). A possible explanation is that sub-inflammatory doses were used for PAR₁ agonists to cause analgesia, while larger doses of TFLLR were necessary to cause inflammation.

Compromised intestinal barrier function has been associated with IBS in child and adult patients (Barau & Dupont, 1990; Spiller et al., 2000), suggesting that increased intestinal permeability plays a role in the generation of the symptoms associated with IBS. Since both PAR₁ and PAR₂ activation disrupted the integrity of the intestinal barrier, it can be hypothesized that PAR-induced increased permeability plays a role in visceral hypersensitivity states. As a matter of fact, a recent study has shown that a tight junction blocker (2,4,6 triaminopyrimidine – TAP) was able to inhibit PAR₂-induced permeability, but also at the same time inhibited PAR₂-induced rectal hypersensitivity (Moriez et al., 2003). The fact that PAR₁ agonists induce increased permeability, and also analgesia, might also depend on the doses of PAR₁ agonists used. In order to cause bacterial translocation and significant increase of ⁵¹CrEDTA passage from the lumen to the blood, doses of 200 μg per mouse were used (Chin et al., 2003), while doses of only 1–100 μg per animal were used in rats to observe analgesia (Coelho & Bunnett, 2003; Vergnolle et al., 2003a). This suggests that at low doses PAR₁ activation induces analgesia independently of permeability dysfunctions, while high doses of PAR₁ agonists might cause hyperalgesia through a mechanism dependent on the integrity of intestinal barrier.

Enterocytes that secrete electrolytes, such as chloride, promote the movement of water in the intestinal lumen, and can thereby regulate IBS symptoms such as diarrhea. Recent studies have demonstrated that PAR₁ activation in human non-transformed intestinal epithelia cell line (SCBN) results in calcium-dependent chloride secretion through a pathway that involves MAP-kinase and cyclooxygenase (Buresi et al., 2001). In contrast, activation of PAR₁ in isolated segments of mouse colon resulted in a decrease of neurally evoked ion secretion (Buresi et al., 2003). Thus, depending on the target cells where PAR₁ is activated, this activation can lead to different diarrheal or anti-diarrheal symptoms. PAR₂ activation in isolated intestinal segments also provoked chloride secretion (Vergnolle et al., 1998). Several cells could constitute the primary target for PAR₂-induced chloride secretion: enterocytes which express functional PAR₂ (Kong et al., 1997), fibroblasts which stimulate PGE₂ release in response to PAR₂ agonists (Seymour et al., 2003), or enteric neurons which have been shown to be involved at least in part in PAR₂-induced ion transport in porcine ileum (Green et al., 2000). It is not yet known whether or not PAR₁ and PAR₂ activation participate in the generation of diarrhea symptoms associated with IBS, but considering the effects of PAR agonists on epithelial secretory functions PARs appear as potential therapeutic targets in the treatment of secretory dysfunctions.

Thrombin, trypsin and mast cell tryptase, as well as selective agonists for PAR₁, PAR₂ and PAR₄, enhanced the excitability and firing of enteric neurons (Gao et al., 2002). In that study, the authors suggest that these PARs can be activated on inhibitory motor enteric neurons, which would lead to inhibited contractile and motor activity. Several studies have reported the effects of PAR agonists on motor functions of the GI tract (Kawabata et al., 2001b; Mule et al., 2002a,2002b; 2004; Zhao & Shea-Donohue, 2003), suggesting that PAR activation could participate in motor dysfunctions associated with IBS. Whether direct activation of PARs on smooth muscle cells or on enteric neurons is involved in PAR-induced GI motor changes still has to be determined.

Finally, bi-directional interactions between the central nervous system and the gut-directed pathogenetic mechanisms are playing a major role in the development of IBS symptoms. Whether or not proteases and PARs are implicated in such interactions still has to be investigated. However, a recent study reports that central activation of PAR₁ by thrombin or selective PAR₁ agonist inhibits NMDA-mediated nociception (both somatic and visceral), by a pathway involving endothelin type A receptors (Fang et al., 2003). We also reported that spinal activation of PAR₂ exacerbated visceral pain behaviors (Vergnolle et al., 2003b). These results suggest an important role for central activation of PARs in states of hypersensitivity, in accordance with a pro-algesic role for PAR₂ activation and an analgesic role for PAR₁ activation.

Conclusions

The development of drugs for IBS remains a therapeutic challenge. Scientists are constantly in search of potential new therapeutic targets for the treatment of IBS, and PARs appear as interesting receptors that clearly interfere with visceral nociceptive pathways. PARs are expressed in diverse cell types of the GI tract and their activation is associated with several pathophysiological factors that are involved in the generation of IBS symptoms, such as altered motility patterns, inflammatory mediator release, altered epithelial functions and altered visceral nociceptive functions. Further studies using genetically modified animals, and, when available, pharmacological tools, in different IBS and IBD models would definitively contribute to our understanding of the role of these receptors and their activating proteases in visceral pain and inflammation processes.

Abbreviations

CGRP

calcitonin gene-related peptide

ENS

enteric nervous system

GI

gastrointestinal

IBD

inflammatory bowel disease

IBS

irritable bowel syndrome

NK-1 R

tachykinin-1 receptor

PARs

protease-activated receptors

PAR₁

protease-activated receptor-1

PAR₂

protease-activated receptor-2

PAR₃

protease-activated receptor-3

PAR₄

protease-activated receptor-4

PSTI

pancreatic secretory trypsin inhibitor