Molecular Mechanisms of PAR Activation

The four cloned PARs share the typical structural features of GPCRs, with seven transmembrane domains, three intracellular and extracellular loops, an extracellular amino terminus, and an intracellular carboxyl terminus. PAR₁ was initially identified by expressing libraries of RNA from thrombin-responsive cells in Xenopus oocytes, and identifying clones that exhibited thrombin responsiveness [8,9]. The discovery of PAR₂ was accidental and occurred while screening a mouse genomic library using degenerate primers to the bovine neurokinin 2 receptor [10]. The existence of at least one further PAR, PAR₃, was inferred from the observation that platelets from PAR₁ knockout mice still respond to thrombin [11]. Finally, PAR₄ was identified by searching expressed sequence tag libraries for PAR-like sequences [12, 13]. All four PARs share a general common mechanism of activation by proteases, although there are some subtle differences in activation that depend on the receptor and protease in question (these differences will be addressed below). In general, activating proteases cleave receptors at specific sites within the extracellular amino terminus to reveal a new amino terminus that acts as a tethered ligand, binding to conserved regions of the second extracellular loop, thereby activating the receptor (Table 31.1, Figure 31.1). Support for this mechanism of activation comes from the observation that synthetic activating peptides (APs), as short as six amino acids, that correspond to the tethered ligand domain can directly bind to and activate PAR₁, PAR₂, and PAR₄. Although APs generally have a very low potency compared to proteases, they are valuable reagents for investigating the physiological functions of PARs, avoiding the use of proteases that can have many other mechanisms of action. Moreover, APs have been used as the starting materials for developing specific and potent antagonists and agonists of certain PARs [14, 15]. However, the peptide corresponding to the tethered ligand for PAR₃ is unable to activate this receptor, for unknown reasons. The activating proteases and peptides, intracellular signals, and in vivo locations of the four PARs are summarized in Table 31.1.

TABLE 31.1

A Summary of the Major Known PAR-Activating Proteases, Cleavage Sites, Sequences of Activating Peptides, Signaling Mechanisms, and the Sites of PAR Expression

FIGURE 31.1

Protease cleavage and activation of PAR₂ on primary spinal afferent neurons to cause neurogenic inflammation and pain. (1) Injury and inflammation trigger the generation and release of proteases from mast cells (tryptase), epithelial cells, and neurons (more...)

PAR Activating Proteases

The great diversity of proteases found within living organisms means that there are many different sources for potential activators of PARs. However, the presence of a particular protease within a system does not mean that it will necessarily be able to initiate a signal. Indeed, the capacity of proteases to activate PARs may depend on the presence of binding domains that tether proteases to receptors, of accessory proteins that tether proteases to the cell surface, and of protease inhibitors that modulate enzyme activity.

Certain proteases potently activate PARs only when they are concentrated at the cell surface, either through direct interaction with the receptor or by binding to other membrane proteins. As an example of receptor interaction, the anion-binding site of the coagulation factor thrombin binds to charged domains of PAR₁ and PAR₃, enabling the potent activation of these receptors [16, 17]. Higher concentrations of thrombin are required to activate PAR₄, which lacks a thrombin-binding site [12, 13]. As an example of interaction with other membrane proteins, the coagulation factors VIIa and Xa interact with tissue factor, an integral membrane protein that is normally expressed by extravascular cells and expressed by endothelial cells and monocytes during inflammation, and thereby can potently activate PAR₁ and PAR₂ [18, 19]. Activated protein C, an anticoagulant protease that degrades coagulation factors Va and VIIa when released from the cell surface, can activate PAR₁ when concentrated at the cell surface by the endothelial protein C receptor [20]. Interactions between PARs can also influence the capacity of proteases to activate receptors. The binding of an enzyme to one receptor can activate a different PAR, as in the case of PAR₃ and PAR₄, the thrombin receptors on murine platelets, which lack PAR₁ [21]. Thrombin interacts with the charged domains of PAR₃, which concentrates thrombin at the cell surface of murine platelets to promote thrombin cleavage and activation of PAR₄. In addition, there is evidence for intermolecular signaling of certain PARs. The activating peptide sequence for PAR₁ (SFLLRN) can activate PAR₂, with reduced potency. When coexpressed in the same cell, the tethered ligand domain of PAR₁ can interact with the binding site of PAR₂, activating an uncleaved receptor [22].

Other proteases can potently activate PARs without being concentrated at the cell surface by interactions with receptors or other membrane proteins. For example, trypsins potently activate PAR₂ and PAR₄ without interacting with receptors or accessory membrane proteins. There are three trypsinogen genes in humans: protease serine (PRSS) I encodes trypsinogen I; PRSS2 encodes trypsinogen II; and PRSS3 encodes mesotrypsinogen [23, 24]. Trypsinogen IV is a splice variant of mesotrypsinogen [25]. Trypsinogen I and II are highly expressed in the pancreas and are secreted into the intestinal lumen, where they can cleave PAR₂ on enterocytes [26]. These trypsins may also cleave PAR₂ in the pancreas during inflammation, when trypsins are prematurely activated [27, 28]. Trypsinogen IV is expressed by brain neurons and epithelial cells of the intestine, airway, and prostate [25, 29]. Although the physiological functions of trypsinogen IV and mesotrypsinogen are unknown, they can activate PAR₁, PAR₂, and PAR₄ [29] (Bunnett, unpublished). The observations that trypsinogen IV is resistant to polypeptide inhibitors, such as the pancreatic secretory trypsin inhibitor and soybean trypsin inhibitor, and also degrades such inhibitors [24, 29–31] implies that trypsinogen IV could show sustained activity in tissues, in contrast to trypsinogen I and II, which are susceptible to endogenous inhibitors and unable to degrade them.

Posttranslational modifications of PARs may also affect the ability of certain proteases to activate receptors. This appears to be the case for tryptase, an enzyme that may play an important role in activating PAR₂ in the nervous system [32]. Tryptase is found in almost all subsets of human mast cells, where it forms up to 25 percent of total cellular protein [33, 34]. Tryptase is released from mast cells when they degranulate, and because mast cells are often in close proximity to sensory nerve fibers, both in normal and inflamed tissues [35], the activation of neuronal PARs by tryptase may be one mechanism by which mast cells regulate neuronal functions. Compared to trypsinogen, tryptase is a weak activator of PAR₂. However, deglycosylation of PAR₂ at a residue near the cleave site markedly increases the potency with which tryptase activates this receptor [36, 37]. The reason for this effect is unknown, although glycosylation may impede the interaction of the tryptase tetramer, with its active sites in the center of a doughnut shape, with the PAR₂ cleavage site. Different forms of tryptase could also activate PAR₂ with different potencies. In humans there are five different tryptase genes, and further splice variants are present in human mast cells. Recombinant forms of these tryptases have demonstrated mixed results in PAR₂ activation assays, with α and βII tryptase activating PAR₂ in a lymphoid cell line [38], while βI tryptase showed no PAR activation [39]. In addition to tryptase, cells of the immune system release several other PAR-activating proteases. Leukocytes, particularly neutrophils, release cathepsin G, elastase, and proteinase C, all of which can activate PARs.

PAR Coupling to G Proteins and Downstream Signaling Events

The unique nature of PAR activation, where a single protease molecule can activate multiple receptors, suggests that differences must exist between their signal transduction pathways and those for other GPCRs. Typical GPCR responses involve producing secondary messenger molecules in proportion to the number of receptors occupied by ligands at any one time. A single protease molecule could eventually cleave and activate every PAR on the surface of a cell, so the intensity of PAR responses is determined by the rate at which the receptors are cleaved, in proportion to the number of protease molecules present [40].

The identity of the G proteins through which PARs couple to mediate their cellular effects varies depending on the receptor in question and the cell type, allowing PARs to produce a wide range of physiological responses. The majority of studies of the mechanism of signal transduction by PARs have been carried out with PAR₁, which can couple through Gα_q/11, Gα_I, and Gα_12/13. In platelets, thrombin activation of PAR₁ activates Gα_q/11, which in turn activates phospholipase C-β_1. [41] This enzyme catalyzes the hydrolysis of the membrane phospholipid PIP₂, releasing inositol 1,4,5-trisphosphate (IP₃) and diacylglycerol (DAG), thus releasing Ca²⁺ from intracellular stores and activating protein kinase C (PKC). Activation of PAR₁ on platelets also activates Gα_12/13, coupling the receptor to the Rho family of guanine-nucleotide exchange factors that regulate Rho-kinase and myosin light chain kinase, mediating thrombin-induced changes in platelet shape [42]. In contrast, IP₃ and DAG mediate the platelet aggregation and degranulation seen in response to thrombin [43]. The potential of PAR₁ to also activate Gα_I, and the ability of the βγ subunits of the G proteins to themselves couple to other signaling pathways (such as phosphatidylinositol 3-kinase activation), further expands the range of intracellular processes mediated by PAR₁.

Less is known about the signal transduction mechanisms of the other PARs. Activation of PAR₂ induces generation of IP₃ and a subsequent release of Ca²⁺ into the cytosol, in transfected cell lines [44], and in cells that endogenously express PAR₂, such as neurons and enterocytes [45, 46], suggesting that PAR₂ couples to Gα_q/11. PAR₂-mediated release of arachidonic acid and a subsequent generation of prostaglandins occur in enterocytes and transfected epithelial cells, indicating activation of phospholipase A₂, although the coupling mechanism by which this occurs is not identified [26]. In addition, there is evidence of an interaction between PAR₂ and TRPV channels requiring activation of protein kinase A (PKA), although it is not clear whether this is a result of Gα_s activation and the resultant cAMP generation by adenylyl cyclase or whether it is a downstream component of a separate pathway (Bunnett, in press).

PAR agonists also couple to the activation of MAP kinases by a variety of mechanisms. One mechanism by which PAR₂ activates ERK1/2 involves the β-arrestins, cytosolic proteins that interact with activated GPCRs at the plasma membrane. In addition to their role in receptor desensitization and endocytosis, β-arrestins are molecular scaffolds that form an ERK1/2 “signaling module” at the plasmamembrane or in endosomes. Thus, PAR₂ agonists stimulate the assembly of a large molecular weight complex containing PAR₂, β-arrestins, raf-1, and activated ERK1/2, which forms at the plasma membrane or in early endosomes [47]. The function of this module is to retain activated ERK1/2 in the cytosol, where they may regulate cytosolic or membrane targets, including ion channels.

Mechanisms of PAR Signal Arrest

Proteases activate PARs in an irreversible manner; once exposed, the tethered ligand is always available to interact with the cleaved receptor, which could lead to sustained signaling without the existence of efficient mechanisms to attenuate the response. Transient and permanent mechanisms dampen PAR signaling. The transient mechanism involves phosphorylation of PARs by members of the family of G-protein receptor kinases (GRKs), followed by interaction with β-arrestins, which uncouple PARs from heterotrimeric G proteins to desensitize G-protein signaling. GRK3 mediates agonist-induced phosphorylation of PAR₁, because GRK3 overexpression enhances thrombin-stimulated PAR₁ phosphorylation and suppresses thrombin signaling [48]. The overexpression of GRK3 in transgenic mice also attenuates thrombin-induced signaling in vivo [49]. In endothelial cells GRK5 mediates PAR₁ desensitization; GRK5 expression suppresses thrombin signaling, and a dominant-negative GRK5 mutant prolongs thrombin signals [50]. The purpose of GRK-induced phosphorylation of a GPCR is to increase its affinity for β-arrestins to interact with GRK-phosphorylated GPCRs and disrupt their association with heterotrimeric G proteins to terminate the signal. PAR₁ desensitization is diminished in fibroblasts lacking β-arrestins [51]. Similarly, a PAR₂ mutant that is unable to interact with β-arrestins shows sustained signaling and diminished desensitization [47].

Receptor endocytosis followed by trafficking to lysosomes for degradation permanently arrests PAR signaling. Agonists promote endocytosis and lysosomal trafficking of PAR₁ and PAR₂ [52–54]. Both receptors internalize by a clathrin-mediated process involving the GTPase dynamin, which mediates detachment of clathrin-coated pits [55, 56]. β-arrestins also mediate endocytosis of many GPCRs by acting as adaptors that link receptors to clathrin. However, β-arrestins do not mediate agonist- induced endocytosis of PAR₁, which proceeds normally in β-arrestin-deficient fibroblasts [51]. Endocytosis of PAR₂ does requires β-arrestins. PAR₂ agonists induce rapid translocation of β-arrestins from the cytosol to the plasma membrane, followed by a prolonged colocalization of PAR₂ and β-arrestins in early endosomes [47, 54]. Compared to our understanding of endocytosis, less is known about the molecular mechanisms of post-endocytic sorting of GPCRs to lysosomes. Sorting nexin 1, a membrane-associated protein that interacts with the EGF receptor, is required for lysosomal trafficking of PAR₁ [57]. A large amount of the E3 ligase c-Cbl is necessary to sort PAR₂ from early endosomes to lysosomes [58].

Expression of PARs in the Nervous System

PARs_1–4 are expressed in the brain, spinal cord, and associated ganglia, where they have been detected in both neurons and astrocytes. PAR₁ mRNA is present in the rat neocortex, cingulate/retrosplenial cortex, subiculum, hypothalamic and thalamic nuclei, and in layers of the hippocampus, cerebellum, and olfactory bulb, where it is present in neurons and astroglia [59]. Immunoreactive PAR₁ is found in the human cerebellum, temporal and frontal lobes, and the striatum, where it is prominently detected in astrocytes [60]. PAR₁ is found in dopaminergic neurons in the substantia nigra, GABA-containing granule cells of the olfactory bulb, and glutaminergic neurons of the anterior thalamus. In rats, PAR₁ is expressed in spinal cord motoneurons, preganglionic autonomic neurons, and dorsal root ganglia where it colocalizes with the neuropeptides substance P (SP) and calcitonin gene-related peptide (CGRP) [61]. PAR₂ is also expressed by neurons and glial cells throughout the brain, including the hypothalamus [62, 63], and by primary spinal afferent neurons containing SP and CGRP [32]. Functional experiments confirm the widespread expression of PARs in astrocytes and neurons. Astrocytes express all four PARs [64], and thrombin increases [Ca²⁺]_i in a glioma cell line and in astrocytes by activating both PAR₁ and PAR₄ [65]. Astrocytes also express PAR₂ and respond to PAR₂ agonists [66]. PAR₁ and PAR₂ agonists also signal to dorsal root ganglia neurons to increase [Ca²⁺]_i, providing functional evidence for expression of these receptors [32, 46, 61].

PARs are also expressed in the peripheral nervous system, notably by enteric neurons. A large proportion of myenteric neurons respond to agonists of PAR₁, PAR₂, and PAR₄ [46, 67, 68]. Agonists of these receptors depolarize myenteric neurons and increase their excitability [67, 68]. Approximately 50–60 percent of AH-type neurons (sensory neurons) and slightly fewer S neurons (inter and motor neurons) respond to PAR₂ agonists. The PAR₂-responsive S neurons also express nitric oxide (NO) synthase and project axons in an anal direction. These are inhibitory motor neurons, and activation would be expected to suppress motility. Submucosal neurons also express PARs, where activation may affect fluid and electrolyte secretion in the intestine. PAR₂ is expressed in submucosal neurons of the guinea pig small intestine, some of which also express vasoactive intestinal polypeptides and are thus secretomotor neurons that play an important role in controlling fluid and electrolyte transport [69]. PAR₂ agonists evoke transient depolarization of submucosal neurons followed by a prolonged hyperexcitability that can last for several hours. This long-term hyperexcitability mimics that observed after degranulation of mast cells, suggesting that mast cell proteases such as tryptase may induce long-term excitability of sub-mucosal neurons through PAR₂ activation, which could result in enhanced fluid and electrolyte secretion from the mucosa [70, 71].

Proteases That Activate PARs in the Nervous System

Proteases from a variety of sources (including the circulatory system, inflammatory cells, and even the neurons themselves) have the potential to activate PARs in the central and peripheral nervous systems (Table 31.1). Although many of these enzymes activate PARs on neurons both in vitro and in vivo, in general the proteases that activate neuronal PARs under physiological and pathophysiological conditions have not been unequivocally identified.

Injury and inflammation generate several proteases of the coagulation cascade that can activate multiple PARs. Disruption of the integrity of the blood–brain barrier during pathological events such as head trauma and stroke may allow coagulation proteases to enter neural tissue, where they could activate PARs on neurons and astrocytes. During injury these proteases may also activate PARs in peripheral neurons, although this possibility has not been examined. Many inflammatory cells are recruited to sites of injury and inflammation, where they could release multiple proteases that activate neuronal PARs. In this regard, mast cells may be of particular importance because they are closely associated with sensory nerve fibers in peripheral tissues that contain SP and that are thus nociceptive neurons [35] (Figure 31.1). Indeed, tryptase released from mast cells can cleave and activate PAR₂ on primary spinal afferent neurons to trigger the release of SP and CGRP in peripheral tissues, where they cause neurogenic inflammation, and within the dorsal horn of the spinal cord, where they cause pain transmission [6, 27, 28, 32, 72]. This mechanism could explain the poorly understood pain of irritable bowel syndrome, where there is a marked influx of tryptase-containing mast cells in close proximity to sensory nerve fibers in the colon, which is accompanied by hyperalgesia to distension of the bowel [73]. In addition, neuronal tissue itself expresses proteases that may activate PARs. In the olfactory bulb, mRNA coding for prothrombin and PAR₁ colocalize in the same cell populations, suggesting that neuronally derived thrombin may activate neurons by cleaving PAR₁ [59]. Moreover, several trypsinogens, including trypsinogen IV, are also expressed by neural tissues and, if released, could activate PARs [25, 29, 74]. However, nothing is known about the mechanisms that control the release of these neuronal proteases, thus their physiological role remains to be defined.

PAR Regulation of Neuronal Excitability

Accumulating evidence indicates that certain proteases can regulate the excitability of neurons, with important functional consequences. For example, trypsin and tryptase increase the excitability of primary spinal afferent neurons to induce hyperalgesia to thermal and mechanical stimuli [6, 72] and also to induce hyperexcitability of enteric neurons, with consequences for intestinal secretion and motility [67–69]. This increased activity of neurons depends on protease regulation of the function of ion channels.

Proteases could regulate ion channels by two general mechanisms: a direct mechanism by which proteases cleave the channel protein itself, and an indirect mechanism by which proteases cleave and activate PARs, which couple to signaling cascades that regulate the activity, localization, or level of expression of the channel. There is good evidence that certain proteases interact directly with ion channels to alter their activity, affecting both ionic transport and electrical excitability of target cells. Channel-activating proteases (CAPs) cleave certain epithelial sodium channels (ENaCs) to increase channel open probability sixfold, an activity mimicked by trypsin, although the precise target of enzymatic cleavage remains obscure [75]. Proteases can also regulate the acid-sensing ion channels (ASICs), members of the ENaC family that are found on neurons and that open in response to decreasing extracellular pH [76]. ASICs are expressed by nociceptive fibers and may transduce pain signals in response to local acidity [77]. ASIC1a and ASIC1b are cleaved by several serine proteases (e.g., trypsin, which only affects ASIC1a, chymotrypsin, and proteinase K), shifting the pH of channel opening and steady-state inactivation to a more acidic pH [78]. Proteases can also modify the activity of another neuronal ion channel, the NMDA receptor for the transmitter glutamate. Tissue-plasminogen activator, which is released by cortical neurons upon depolarization, interacts with NMDA receptors on the same neurons to increase their sensitivity to glutamate, leading to an increase in excitotoxic cell death [79]. The altered activity occurs as a result of tissue-plasminogen activator cleavage of the NR1 subunit of the NMDA receptor. However, although there is an increasing amount of evidence that proteases can regulate the activity of sensory ion channels, cleavage of TRPV channels by proteases has not been identified to date. The major known mechanism by which proteases regulate TRPV channels on sensory nerves is through the activation of PARs.

PAR Sensitization of TRPV Channels

Proteases that derive from mast cells (e.g., tryptase) and epithelial tissues (e.g., trypsins) can cleave PAR₂ on the fibers of primary spinal afferent neurons to trigger the release of SP and CGRP in peripheral tissues and in the dorsal horn of the spinal cord [6, 27, 28, 32, 72]. In a similar manner, thrombin activates PAR₁ on sensory nerves in peripheral tissues to cause neurogenic inflammation [61]. However, in contrast to PAR₂ agonists, thrombin causes analgesia by mechanisms that are not fully understood [80]. The important roles played by TRPV1 and TRPV4 in detecting noxious thermal and mechanical stimuli, respectively, suggests that PAR₂-induced thermal and mechanical hyperalgesia may be mediated by sensitization of these ion channels. This sensitization could occur through increased channel activity, decreased channel inhibition, insertion of additional channels into the cell membrane, or a combination of the three. Indeed, agonists of both GPCRs and receptor tyrosine kinases can sensitize TRPV1 by these mechanisms and thereby induce thermal hyperalgesia and inflammatory pain.

Exposure of sensory neurons to other inflammatory mediators (e.g., ATP, bradykinin, NGF), which are released in response to tissue damage or other stresses, induces thermal nociception by sensitizing TRPV1. The mechanisms of this sensitization have been thoroughly investigated. ATP and bradykinin, acting via metabotropic P2Y₁ receptors and B₂ receptors, respectively, activate PLCβ and consequently PKCɛ [81, 82]. Activated PKCɛ translocates to the plasma membrane and phosphorylates TRPV1 at S⁵⁰² and S⁸⁰⁰ to enhance its open probability and inhibit its desensitization [83, 84]. Binding of the membrane phospholipid PIP₂ to TRPV1 has been suggested to inhibit channel gating [85], and exposure to NGF sensitizes TRPV1 by removing this inhibition [86]. NGF is thought to bind to its tyrosine kinase receptor TrkA to activate PLCγ, which hydrolyzes PIP₂ and removes its inhibitory effects on TRPV1. However, a recent study disputes this mechanism of NGF sensitization of TRPV1, suggesting instead that the most significant effect occurs through activated TrkA receptors that insert TRPV1-containing vesicles into the plasma membrane [87]. Activated TrkA phosphorylates phosphatidylinositol 3-kinase, with a downstream activation of Src kinase leading to phosphorylation of vesicular TRPV1 molecules on Y²⁰⁰, targeting the vesicles to the membrane.

Several observations suggest that PAR₂ may regulate TRPV1 channels in sensory neurons to induce hyperalgesia to thermal stimuli (Figures 31.1 and 31.2). PAR₂ is coexpressed in primary sensory neurons, alongside CGRP, SP, and TRPV1, which are all important components of nociceptive pathways [6, 32]. Activation of PARs induces neurogenic inflammation and nociception and, at subinflammatory doses, both mechanical and thermal hyperalgesia [6, 32, 72]. Furthermore, other inflammatory agents, such as bradykinin, ATP, and NGF, acting through protein kinases, can sensitize TRPV1 [81, 82, 86]. The first report of an interaction between PAR₂ and TRPV1 was that PAR₂ agonists enhance both capsaicin-induced release of CGRP from dorsal root ganglia and fos expression in spinal neurons [27]. Capsazepine, a TRPV1 antagonist, also inhibits the hyperalgesia [88] and gastric mucus secretion [89] induced by PAR₂ activation in vivo, providing further evidence for TRPV1’s role in PAR₂ regulation of neuronal function. PAR₂ agonists also sensitize TRPV1-mediated calcium mobilization and TRPV1 currents in HEK cells expressing TRPV1 and in primary spinal afferent neurons [6, 7]. PAR₂-induced thermal hyperalgesia, measured as a decrease in the paw withdrawal latency to a radiant heat, is abolished in mice lacking TRPV1 [6, 7]. Furthermore, nonalgesic doses of PAR₂ agonists and capsaicin act synergistically to produce hyperalgesia [6]. Together, these results suggest that proteases released during trauma or inflammation can potentially activate PAR₂ expressed in primary sensory neurons to sensitize TRPV1, resulting in exacerbated responses to TRPV1 activators, exemplified by thermal hyperalgesia. However, there are certain gaps in our understanding of this process.

FIGURE 31.2

Potential mechanisms by which PAR₂ sensitizes TRPV1. (1) Activated PAR₂ couples to G_αq11 and activates PLCβ. (2) This pathway leads to the activation of PKCɛ, which translocates to the plasma membrane. (3) PKCɛ can phosphorylate (more...)

One deficiency is that the signal transduction mechanism by which PAR₂ sensitizes TRPV1 is not fully understood. PAR₂ couples to Gα_q/11 and activates PLCβ, releasing IP₃ and DAG. PAR₂-induced activation of PLCβ could enhance TRPV1 by reducing the levels of its substrate, the membrane phospholipid PIP₂, consequently freeing TRPV1 from PIP₂-mediated inhibition [85, 86] (Figure 31.2). In addition, PLCβ-induced release of IP₃ and DAG could activate PKC, which phosphorylates TRPV1 to increase its open probability and decrease desensitization [83, 84]. Indeed, PAR₂-mediated potentiation of TRPV1 responses is diminished by PLC inhibitors and also by the nonselective inhibition of PKC inhibitors, supporting both hypotheses [6]. The identity of the PKC isozyme responsible for PAR₂ sensitization of TRPV1 remains to be determined. One particularly promising candidate is PKCɛ, which plays a major role in transmitting mechanical and thermal hyperalgesia [90]. Also, PKCɛ phosphorylates TRPV1 [82] to mediate bradykinin-induced sensitization of TRPV1 currents [81]. Indeed, selective antagonism of PKCɛ suppresses PAR₂-induced sensitization of TRPV1 currents in nociceptive neurons [7]. However, the contribution of other PKC isoforms cannot be excluded, because Gö6976, an inhibitor of conventional PKC isoforms (α, β1, β2, and γ), but not ɛ, also inhibits PAR₂-mediated potentiation of TRPV1 signals [6]. This implicates protein kinase D1 (PKD1), as Gö6976 is also a PKD1 inhibitor. Initially thought to be an atypical member of the PKC family (PKCμ), PKD1 is now recognized as a member of a new family of three PKD isozymes that resemble the CAM kinase family [91]. Certain novel PKC isoforms activate PKD (e.g., thrombin activation of PKD in platelets involves PKCδ [92]), which may mediate some of the effects of PKC activation. This possibility is particularly relevant in light of the finding that PKD1 binds and phosphorylates TRPV1 on S¹¹⁶, enhancing TRPV1-mediated responses, while anti-sense oligonucleotides directed against PKD1 decreased TRPV1 responses [93]. There is also evidence that sensitization of TRPV1 may also occur through activation of PKA. PKA is required for injury-induced hyperalgesia [94], and it phosphorylates TRPV1 on S¹¹⁶ to regulate its desensitization [95, 96], thus sensitizing TRPV1 currents in nociceptive neurons [97, 98].

Another gap in our understanding of the role of PARs in pain transmission is the mechanism of PAR₂-induced mechanical hyperalgesia, which is evident after both somatic and visceral mechanical stimuli [72, 99]. Although the channel that mediates responses to mechanical stimuli is not unequivocally identified, TRPV4 is a strong candidate. TRPV4 is the mammalian homologue of the C. elegans gene osm-9, which encodes a protein conferring sensitivity to osmotic pressure, touch, and odorants [100]. Originally called VR-OAC, this channel responds to small reductions in tonicity; in rats it is expressed in circumventricular organs, which detect changes in systemic osmolarity, and in neurosensory structures such as Merkel cells and sensory neurons [101]. Based on its structure, ionic selectivity (preference for Ca²⁺), and thermal sensitivity (>27°C), VR-OAC was placed in the same TRPV family as TRPV4 [102, 103]. The phorbol ester 4α-phorbol 12,13-didecanoate (4αPDD) is a synthetic TRPV4 agonist [104], and the cytochrome P450 product of arachidonic acid, 5’,6’-epoxyeicosatrienoic acid (5’,6’-EET), is a potential endogenous TRPV4 agonist [105] (hypoosmotic stimuli leads to cell swelling, phospholipase A₂ activation, and release of arachidonic acid [106]). The expression of TRPV4 in sensory neurons and its activation by changes in cell volume suggest that it detects osmotic and mechanical stimuli. TRPV4^−/− mice show abnormal osmotic regulation and decreased acid and pressure nociception (not impaired touch or heat sensation) [107, 108]. Although injection of hypotonic saline into rat hindpaws does not induce nociception, hypertonic saline causes pain, and pretreatment with an inflammatory mediator, PGE₂, sensitizes responses to both stimuli [101, 109]. These effects require TRPV4, because they are prevented by TRPV4 deletion or downregulation. In a similar manner to PGE₂, PAR₂ agonists also increase the magnitude of the Ca²⁺ mobilization in response to hypotonic saline and 4αPDD in both human bronchial epithelial cells, which constitutively express both TRPV4 and PAR₂, and in HEK cells heterologously expressing TRPV4 (Bunnett, unpublished). In slices of rat spinal cord, exposure to 4αPDD or hypotonic saline stimulates the release of SP and CGRP, which is potentiated by preexposure to PAR₂ agonist. Hypotonic saline injected into mice paws does not produce any nociceptive behavior, but the combination of PAR₂–AP and hypotonic saline increases the sensitivity to mechanical pain to a greater degree than PAR₂–AP alone, indicating a sensitization of TRPV4 by PAR₂ activation. TRPV4^−/− mice do not show any increased sensitivity with either injection, indicating that both the mechanical hyperalgesia induced by PAR₂ activation and the pain induced by hypotonic saline are mediated by TRPV4 (Bunnett, unpublished). Together, these findings suggest that PAR₂ agonists can sensitize TRPV4 to induce hyperalgesia to mechanical stimuli, possibly by a mechanism that resembles those that mediate PAR₂-induced sensitization of TRPV1.