Protease-activated Receptor 2 (PAR2) Protein and Transient Receptor Potential Vanilloid 4 (TRPV4) Protein Coupling Is Required for Sustained Inflammatory Signaling*

Background: Receptors activate channels of sensory nerves to cause inflammation and pain by unknown mechanisms. Results: Protease-activated receptor 2 (PAR2) stimulated transient receptor potential vanilloid 4 (TRPV4) by generation of channel agonists. This required a key TRPV4 tyrosine and induced inflammation. Conclusion: PAR2 opens TRPV4 by functional coupling. Significance: Antagonism of PAR2-TRPV4 coupling could alleviate inflammation and pain. G protein-coupled receptors of nociceptive neurons can sensitize transient receptor potential (TRP) ion channels, which amplify neurogenic inflammation and pain. Protease-activated receptor 2 (PAR2), a receptor for inflammatory proteases, is a major mediator of neurogenic inflammation and pain. We investigated the signaling mechanisms by which PAR2 regulates TRPV4 and determined the importance of tyrosine phosphorylation in this process. Human TRPV4 was expressed in HEK293 cells under control of a tetracycline-inducible promoter, allowing controlled and graded channel expression. In cells lacking TRPV4, the PAR2 agonist stimulated a transient increase in [Ca2+]i. TRPV4 expression led to a markedly sustained increase in [Ca2+]i. Removal of extracellular Ca2+ and treatment with the TRPV4 antagonists Ruthenium Red or HC067047 prevented the sustained response. Inhibitors of phospholipase A2 and cytochrome P450 epoxygenase attenuated the sustained response, suggesting that PAR2 generates arachidonic acid-derived lipid mediators, such as 5′,6′-EET, that activate TRPV4. Src inhibitor 1 suppressed PAR2-induced activation of TRPV4, indicating the importance of tyrosine phosphorylation. The TRPV4 tyrosine mutants Y110F, Y805F, and Y110F/Y805F were expressed normally at the cell surface. However, PAR2 was unable to activate TRPV4 with the Y110F mutation. TRPV4 antagonism suppressed PAR2 signaling to primary nociceptive neurons, and TRPV4 deletion attenuated PAR2-stimulated neurogenic inflammation. Thus, PAR2 activation generates a signal that induces sustained activation of TRPV4, which requires a key tyrosine residue (TRPV4-Tyr-110). This mechanism partly mediates the proinflammatory actions of PAR2.

G protein-coupled receptors of nociceptive neurons can sensitize transient receptor potential (TRP) ion channels, which amplify neurogenic inflammation and pain. Protease-activated receptor 2 (PAR 2 ), a receptor for inflammatory proteases, is a major mediator of neurogenic inflammation and pain. We investigated the signaling mechanisms by which PAR 2 regulates TRPV4 and determined the importance of tyrosine phosphorylation in this process. Human TRPV4 was expressed in HEK293 cells under control of a tetracycline-inducible promoter, allowing controlled and graded channel expression. In cells lacking TRPV4, the PAR 2 agonist stimulated a transient increase in [Ca 2؉ ] i . TRPV4 expression led to a markedly sustained increase in [Ca 2؉ ] i . Removal of extracellular Ca 2؉ and treatment with the TRPV4 antagonists Ruthenium Red or HC067047 prevented the sustained response. Inhibitors of phospholipase A 2 and cytochrome P450 epoxygenase attenuated the sustained response, suggesting that PAR 2 generates arachidonic acid-derived lipid mediators, such as 5,6-EET, that activate TRPV4. Src inhibitor 1 suppressed PAR 2 -induced activation of TRPV4, indicating the importance of tyrosine phosphorylation. The TRPV4 tyrosine mutants Y110F, Y805F, and Y110F/Y805F were expressed normally at the cell surface. However, PAR 2 was unable to activate TRPV4 with the Y110F mutation. TRPV4 antagonism suppressed PAR 2 signaling to primary nociceptive neurons, and TRPV4 deletion attenuated PAR 2 -stimulated neurogenic inflammation. Thus, PAR 2 activation generates a signal that induces sustained activation of TRPV4, which requires a key tyrosine residue (TRPV4-Tyr-110). This mechanism partly mediates the proinflammatory actions of PAR 2 .
Injury and inflammation trigger the activation of proteases from the circulation, immune cells, and epithelial tissues that regulate cells by cleaving protease-activated receptors (PARs) 3 , members of a family of four G protein-coupled receptors (GPCRs) (1,2). By cleaving PARs at specific sites within the extracellular N-terminal domains, proteases reveal tethered ligands that bind to and activate the cleaved receptors. Synthetic peptides that mimic the tethered ligand domains (PARactivating peptides, APs) can directly activate PARs and are useful tools to probe receptor functions. Once activated, PARs regulate multiple pathophysiological processes, including inflammation, pain, hemostasis, and healing. PAR 2 is coexpressed with substance P and calcitonin generelated peptide by a subpopulation of primary spinal-afferent neurons that control neurogenic inflammation and pain transmission (3,4). Activation of PAR 2 on sensory nerve endings evokes the local release of these neuropeptides, which stimulate extravasation of plasma proteins, infiltration of neutrophils, and vasodilation (neurogenic inflammation). PAR 2 activation also promotes the central release of neuropeptides that activate second-order spinal neurons that transmit pain. These mechanisms contribute to painful inflammation of the intestine (5, 6), pancreas (7,8), and joints (9). Therefore, it is of considerable interest to understand the mechanisms by which PARs regulate the activity of nociceptive neurons.
In addition to sensitization, emerging evidence suggests that GPCR signaling can directly activate TRP channels. Responses of dorsal root ganglion (DRG) neurons to bradykinin and histamine are largely dependent on Ca 2ϩ influx through TRPV1 (28,42). Products of phospholipase A 2 (PLA 2 ) and lipoxygenase can directly activate TRPV1, including N-arachidonoylethanolamide, which is derived from membrane phospholipids, and hydroperoxy or hydroxy eicosatetraenoic acids and leukotriene B4 products of the lipoxygenase-dependent metabolism of arachidonic acid (43). Diacylglycerol directly activates TRPV1 through a kinase-independent mechanism that underlies cellular responses to M 3 muscarinic and glutamate mGluR5 agonists (44,45). 5-Hydroxytryptamine and acetylcholine activate TRPV4 by a mechanism that is dependent on generation of epoxyeicosatrienoic acid (5Ј,6Ј-EET and 8Ј,9Ј-EET) products of cytochrome P450 epoxygenase-dependent catalysis of arachidonic acid (46 -48).
We have identified a new mechanism by which PAR 2 activates TRPV4 channels. Although TRPV1, TRPV4, and TRPA1 mediate the pronociceptive actions of PAR 2 (12)(13)(14)(15)(16)(17)(18)(19), the mechanisms underlying the functional interactions between TRPs and PAR 2 are not fully understood. HEK293 cells endogenously express PAR 2 (49) and respond robustly to PAR 2 stimuli, including trypsin or the tethered ligand mimetic peptide, SLIGRL-NH 2 , with a rapid, transient increase in intracellular calcium ([Ca 2ϩ ] i ). By modulating the expression of TRP channels, we report the novel finding that TRPV4, but not TRPA1, contributes to the cellular response to PAR 2 agonists in HEK cells. This mechanism involves PLA 2 -and cytochrome P450 epoxygenase-dependent catalysis of arachidonic acid and requires the phosphorylation of a key TRPV4 tyrosine residue (Tyr-110) that has been implicated in sensitization of TRPV4 (50). We propose that the activation of TRPV4 serves to prolong PAR 2 signaling and to amplify its proinflammatory and pronociceptive actions. These findings have implications for the consequences of protease-dependent PAR 2 activation during injury and inflammation.
Animals-Rats (Sprague-Dawley, 200 -250 g, male) and mice (trpv4 Ϫ/Ϫ and trpv4 ϩ/ϩ littermates, 20 -30 g, male) (52) were studied. Procedures involving animals were approved by the Institutional Animal Care and Use Committee of the University of California, San Francisco or the University of Melbourne Animal Experimentation Ethics Committee.
Measurement of [Ca 2ϩ ] i in DRG Neurons-Rats were anesthetized with isoflurane and killed by bilateral thoracotomy. DRG were collected from all spinal levels and cultured as described (13,53). Neurons were plated onto glass coverslips coated with poly-L-lysine and laminin (100 g/ml) and cultured for 48 -72 h. Neurons were loaded with Fura2-AM ester (2 M, 30 min, 37°C) in HBSS, washed, and incubated in HBSS for 30 min before study. Responses of individual neurons to agonists were measured by microscopy as described for HEK cells. Neurons were challenged sequentially with trypsin (300 units/ml), GSK1016790A (10 nM, TRPV4 agonist), capsaicin (100 nM, TRPV1 agonist), and KCl (50 mM). In some experiments cells were pretreated with HC067047 (10 M, TRPV4 antagonist) 30 min prior to addition of agonists or were assayed in Ca 2ϩ -free HBSS.
Analysis of Ca 2ϩ Signals-The transient Ca 2ϩ response induced by PAR 2 -AP challenge of HEK cells was calculated as the difference between the basal 340/380-nm fluorescence emission ratio (average of four readings immediately prior to application of PAR 2 -AP) and the maximal fluorescence that was measured 10 -20 s after PAR 2 -AP application. The sustained Ca 2ϩ response was calculated as the difference between the basal fluorescence and the fluorescence measured at 50 -60 s after PAR 2 -AP application. The results were expressed as a ratio of the transient to the sustained response. A diagram illustrating these measurements is presented in Fig. 2B. Representative traces were obtained by averaging values recorded from three different wells with paired vehicle and treated groups. At least three technical repeats were performed for each experiment.
For analysis of DRG, Ca 2ϩ responses were included of cells that responded to K ϩ stimulation with an increase in 340/ 380-nm fluorescence emission ratio of Ͼ 0.1 units above the initial base line. The trypsin-responsive population was subdivided further on the basis of the magnitude of responses to GSK1016790A, with lower and upper quartiles designated as "low" and "high" responders, respectively. At least 640 neurons were analyzed (n ϭ 5 independent cultures) per treatment group.
Assessment of Inflammation-Mice were anesthetized with isoflurane (2%), and baseline paw thickness was measured using a digital caliper (Mitutoyo, Aurora, IL). PAR 2 -AP (50 g/paw) or 0.9% NaCl (50 l) were administered by intraplantar injection. The paw thickness was measured from 30 -180 min after injection. In some experiments, mice were treated with 17-ODYA (5 mg/kg, 150 l, intraperitoneal) or vehicle (25% DMSO, 0.9% NaCl, 150 l, intraperitoneal) 30 min before the intraplantar injections. The paw thickness was normalized to base line (0 min). Mice were killed 6 h after the injection. Paws were collected, snap-frozen in liquid nitrogen, and assessed for tissue myeloperoxidase (MPO) activity as described (55). MPO was solubilized with hexadecyltrimethylammonium bromide, and MPO activity was measured with a dianisidine-H 2 O 2 assay. Changes in absorbance at 450 nm over a 15-min period were determined using a microplate reader (Molecular Devices).
Data were expressed as MPO activity relative to total protein (units/mg) and normalized to controls.
Statistical Analysis-Results were expressed as the mean Ϯ S.E. and were compared by Student's or one-sample t test (onetailed) or one-way analysis of variance and Newman-Keuls test, as indicated, using GraphPad Prism (v5.0). Differences were considered significant when p Ͻ 0.05. To examine the effect of TRPV4 expression on PAR 2 -evoked Ca 2ϩ signaling, we generated HEK cell lines expressing TRPV4 with or without an HA tag. TRPV4 was expressed under control of a tetracycline-inducible promoter, which enabled controlled expression of the channel. The expression of TRPV4 was examined by measuring changes in [Ca 2ϩ ] i in response to the TRPV4 activator 4␣-PDD and by immunofluorescence and confocal microscopy using an antibody to the HA.11 epitope. In HEK-TRPV4 cells not exposed to tetracycline, 4␣-PDD had no effect on [Ca 2ϩ ] i (Fig. 1C), and immunoreactive TRPV4 was undetectable (D). Incubation of HEK-TRPV4 cells with graded concentrations of tetracycline (0.0001-0.1 g/ml, 16 h) induced graded expression of functional and immunoreactive TRPV4. In tetracycline-treated cells, 4␣-PDD elicited a gradual increase in [Ca 2ϩ ] i that was sustained for the period of observation (200 s) and was graded with the concentration of tetracycline (Fig.  1C). Responses to 4␣-PDD were detected after incubation with 0.001 g/ml tetracycline and were maximal after 0.1 g/ml tetracycline. The basal [Ca 2ϩ ] i was also elevated by the highest expression levels of TRPV4 (0.01, 0.1 g/ml), although the cells were microscopically normal and remained responsive to TRPV4 and PAR 2 agonists. Immunoreactive TRPV4 was detected in some cells after incubation with 0.001 g/ml tetracycline and present at the plasma membrane of all cells after incubation with 0.1 g/ml tetracycline (Fig. 1D).

PAR
In HEK control cells (expressing the empty vector without a TRPV4 insert), PAR 2 -AP (50 M) evoked a transient increase in [Ca 2ϩ ] i that was maximal after 10 -20 s and that declined to base line after ϳ75 s of stimulation ( Fig. 2A). In tetracyclinetreated (0.1 g/ml, 16 h) HEK-TRPV4 cells, PAR 2 -AP evoked a similar rapid increase in [Ca 2ϩ ] i that was markedly sustained ( Fig. 2A). The sustained response was quantified by determining the ratio of the [Ca 2ϩ ] i at the maximal point of the transient phase (10 -20 s) and during the sustained phase (50 -60 s) (Fig.  2B). This analysis revealed that the magnitude of the sustained phase was proportional to the concentration of tetracycline and, thus, the level of TRPV4 expression (Fig. 2C). TRPV4 expression similarly enhanced the sustained increase in [Ca 2ϩ ] i to trypsin (300 units/ml), a physiologically relevant PAR 2 agonist (Fig. 2D).
To determine whether TRPV4 similarly affects signaling by other GPCRs, we examined the consequences of TRPV4 expression on responses to ATP, which mobilizes Ca 2ϩ in HEK293 cells by activating P2Y receptors (56). ATP (1 M) evoked a rapid and transient increase in [Ca 2ϩ ] i in HEK control cells (Fig. 2E). The magnitude and duration of the ATP-evoked Ca 2ϩ response were unaffected by TRPV4 expression (p ϭ 0.6 and p ϭ 0.7 to control cells, respectively). Thus, not all GPCRs can activate TRPV4.
To determine whether PAR 2 can activate other TRP channels, which would represent a more general mechanism of TRP regulation, we examined the effect of expression of TRPA1 on the responses to PAR 2 -AP. TRPA1 was selected because it is coexpressed with PAR 2 by nociceptive neurons and because pretreatment with PAR 2 agonists amplifies responses to TRPA1 agonists, which is indicative of channel sensitization. TRPA1 was expressed in HEK cells using a tetracycline-inducible promoter to allow for regulated expression. Tetracyclineinduced expression of TRPA1 was confirmed by responsiveness to allyl-isothiocyanate or cinnamaldehyde (data not shown). In contrast to TRPV4, TRPA1 expression did not affect the amplitude or duration of the response to PAR 2 -AP relative to identically treated control cells (Fig. 2F). Thus, TRPA1 does not contribute to PAR 2 -evoked Ca 2ϩ signaling. Addition of the HA epitope tag did not affect PAR 2 or TRPV4 responses to activating stimuli (data not shown). However, all subsequent experiments used untagged TRPV4.
To assess the cell-to-cell variability of PAR 2 -induced activation of TRPV4, we examined responses of individual HEK-TRPV4 cells using microscopy. In HEK-control cells, PAR 2 -AP caused a transient increase in [Ca 2ϩ ] i , with no sustained phase (Fig. 3, A, C, and E). In HEK-TRPV4 cells incubated with tetra-cycline (0.1 g/ml, 16 h), most cells exhibited a sustained plateau of [Ca 2ϩ ] i after treatment with PAR 2 -AP (Fig. 3, B, D, and  E). These results indicate that activation of PAR 2 results in a transient increase in [Ca 2ϩ ] i and that a more sustained response requires coexpression of TRPV4.
Our results suggest that PAR 2 can activate TRPV4, possibly by generating endogenous TRPV4 agonists or by activating signaling pathways that alter channel gating or localization. We refer to this activation as PAR 2 "coupling" to TRPV4. The mechanism of this coupling is not related to the rapid mobilization of intracellular Ca 2ϩ per se because ATP stimulation of P2Y receptors did not activate TRPV4 despite eliciting a substantial release of intracellular Ca 2ϩ ions.
PAR 2 Stimulates an Influx of Extracellular Ca 2ϩ through TRPV4-To determine the source of the sustained TRPV4-dependent Ca 2ϩ response and to further assess the involvement of TRPV4, we either removed extracellular Ca 2ϩ ions or treated cells with the non-selective TRPV inhibitor Ruthenium Red. was dependent on the tetracycline concentration used for induction (0.001-0.1 g/ml) and, thus, the level of TRPV4 expression. Tetracycline had no effect on control cells. *, p Ͻ 0.05; **, p Ͻ 0.01 compared with HEK control cells for each tetracycline condition (Student's t test). n Ն 3 experiments. D, trypsin induced a sustained phase only in tetracycline (0.1 g/ml)-induced HEK-TRPV4 cells. E, expression of TRPV4 did not affect the magnitude or duration of the Ca 2ϩ response to ATP compared with HEK control cells. F, TRPA1 expression (HEK-TRPA1, 0.1 g tetracycline) had no effect on the magnitude or duration of the PAR 2 -AP response compared with HEK control cells.
Omission of extracellular Ca 2ϩ abolished the sustained phase of the PAR 2 -AP response of tetracycline-induced HEK-TRPV4 cells (Fig. 4, A and B; p ϭ 0.002). Ruthenium Red (10 M) also abolished the sustained phase of the PAR 2 -AP response (Fig. 4,  C and D). However, removal of extracellular Ca 2ϩ or addition of Ruthenium Red had no effect on the transient phase of the response to PAR 2 -AP. These results suggest that PAR 2 couples to TRPV4, which mediates an influx of extracellular Ca 2ϩ ions that comprise the sustained phase of the PAR 2 response. PLA 2 and Cytochrome P450 Epoxygenase Contribute to PAR 2 -induced Activation of TRPV4-PAR 2 couples to PLA 2 , which generates arachidonic acid (57), a substrate for synthesis of endogenous TRPV4 activators including 5Ј,6Ј-EET (25). We used a pharmacological approach to determine the contribution of PLA 2 and downstream enzymes to PAR 2 -dependent activation of TRPV4. The irreversible PLA 2 inhibitor MAFP (1, 10 M) inhibited both the transient and sustained phases of the response to PAR 2 -AP (Fig. 5, A and F), but lower concentrations of MFAP had no effect (data not shown). Thus, we could not ascribe a selective effect to inhibition of PLA 2 by this compound. MAFP also slightly inhibited TRPV4 activation by the synthetic agonist GSK1016790A (100 nM, Fig. 5A, p ϭ 0.045). The cytochrome P450 epoxygenase inhibitor 17-ODYA (50 M) inhibited the sustained phase of the response to PAR 2 -AP without reducing the transient phase of the response (Fig. 5, B and F; p ϭ 0.0001). 17-ODYA had a minor inhibitory effect on the response to the TRPV4 agonist 4␣PDD (100 nM), indicating that TRPV4 activity was mostly retained (Fig. 5B). Inhibition of cyclooxy-  genase by indomethacin (50 M) did not alter the responses to PAR 2 -AP (p ϭ 0.27; Fig. 5, C and F). Thus, PAR 2 coupling to TRPV4 appears to involve activation of cytochrome P450 epoxygenase, which can generate arachidonic acid-derived TRPV4 activators such as 5Ј,6Ј-EET.
Certain GPCRs, including PAR 2 , can sensitize TRP channels through activation of second messenger kinases, including PKC⑀, which can phosphorylate TRPs and regulate channel gating (12, 13, 16 -18). Therefore, we examined the effects of kinase inhibitors on PAR 2 coupling to TRPV4. The PKC inhibitor BIM-1 (100 nM) had no effect on the PAR 2 -AP responses (Fig. 5, D and F). However, the inhibitor of Src-family kinases, Src inhibitor 1 (Src1, 10 M), significantly reduced the sustained phase of the PAR 2 -AP [Ca 2ϩ ] i response without affecting the transient phase of the response (Fig. 5, E and F; p ϭ  0.0028). These data suggest the involvement of tyrosine kinase activity in the TRPV4-dependent sustained phase of the PAR 2 response.
Tyrosine 110 Is Required for TRPV4 Activation-To identify the putative sites of tyrosine phosphorylation, we mutated tyrosine residues within the N and C termini of TRPV4 (Y110F, Y805F, and Y110/805F). Cell surface biotinylation and Western blotting for TRPV4 indicated that wild-type TRPV4 and all TRPV4 mutants were expressed at similar levels at the cell surface (Fig. 6A). PAR 2 -AP induced a rapid and transient increase in [Ca 2ϩ ] i that was similar in cells expressing wild-type and mutant TRPV4 channels (Fig. 6B). However, the sustained phase of the response was markedly attenuated in cells expressing TRPV4-Y110F (Fig. 6, B and C; p Ͻ 0.0001 compared with the wild type). The sustained response to PAR 2 -AP of cells expressing TRPV4-Y805F was only slightly reduced (p ϭ 0.03 compared with the wild type), whereas the response of cells expressing the double mutant TRPV4-Y110/805F resembled that of cells expressing the single mutant TRPV4-Y110F (Fig. 6, B and C; p ϭ 0.001 compared with the wild type). Responses to the TRPV4 agonist GSK1016790A were unaffected by TRPV4 mutations (data not shown).
These data suggest an involvement of tyrosine kinases in the generation of the sustained phase of the PAR 2 response, presumably through modulation of TRPV4 by phosphorylation of Tyr-110, which we have identified as a key residue required for the TRPV4-dependent response to PAR 2 activation. TRP Channels Regulate PAR 2 -dependent Ca 2ϩ Signaling in DRG Neurons-To determine whether PAR 2 couples to TRPV4 in cells that naturally express these proteins, we examined responses of DRG neurons from rats to agonists of PAR 2 and TRP channels. DRG were first challenged with trypsin (300 units/ml), which was selected as a physiological agonist of PAR 2 that gave more robust responses than PAR 2 -AP (data not shown). DRG cultures were then challenged sequentially with the TRPV4 agonist GSK1016790A (10 nM) and the TRPV1 agonist capsaicin (100 nM). Cells were finally exposed to K ϩ (50 mM), which depolarizes neurons. Only those cells that responded to K ϩ and that were neurons were analyzed. Trypsin stimulated a rapid increase in [Ca 2ϩ ] i in neurons, with a peak within 5-15 s that gradually returned to prestimulation levels after 120 s (Fig. 7A). Approximately half of all PAR 2 -responsive neurons also responded to GSK1016790A (52.78 Ϯ 12.69%, 649 neurons, n ϭ 5 experiments) with a prompt and sustained elevation in [Ca 2ϩ ] i , although the magnitude of those responses was variable. Capsaicin elicited further responses in many of the neurons, which were generally sustained. Not every neuron that responded to GSK1016790A also responded to capsaicin and vice versa. These results indicate that PAR 2 and TRPV4 are coexpressed by Ͼ 50% of DRG neurons.
The omission of extracellular Ca 2ϩ ions reduced the basal [Ca 2ϩ ] i and blunted the peak and sustained phase of the Ca 2ϩ response to trypsin (Fig. 7B). The TRPV4 antagonist HC067047 (10 M) did not affect the basal [Ca 2ϩ ] i but markedly attenuated the peak and sustained Ca 2ϩ response to trypsin (Fig. 7B).
The variability in responsiveness to GSK1016790A probably reflects variable levels of TRPV4 expression in different neurons. Therefore, we sought to determine whether the magnitude of the PAR 2 responses correlated with the magnitude of the responses to GSK1016790A. To do so, we compared the FIGURE 6. PAR 2 -dependent activation of TRPV4 requires key tyrosine residues. A, TRPV4 WT and the TRPV4 Y110F, Y805F, and Y110F/Y805F mutants were expressed at similar levels at the cell surface of HEK cells and detected by cell surface biotinylation and TRPV4 Western blotting. B and C, the sustained phase of the Ca 2ϩ response to PAR 2 -AP was significantly reduced by mutation of the phosphorylation sites Tyr-110 and Tyr-805. A marked reduction in [Ca 2ϩ ] i for Y110F and the double mutant (Y110F and Y805F) was evident at 60 s post-PAR 2 -AP treatment (p Ͻ 0.0001) compared with the Y805F single TRPV4 mutant (expressed as percentage relative to the wild-type control, p ϭ 0.03 (C)). *, p Ͻ 0.05; ***, p Ͻ 0.0001 relative to the wild type; one sampled t test; n ϭ 9 experiments. Ca 2ϩ response to trypsin, which was quantified as the area under the curve measured up to 125 s after challenge with trypsin, for those neurons in the upper and lower quartiles of the GSK1016790A responses (Fig. 7C). The PAR 2 responses of the neurons in the upper quartile of GSK1016790A responsiveness were ϳ2-fold greater than the responses of the lower quartile (Fig. 7D, p ϭ 0.002). This finding is consistent with a role for TRPV4 in mediating the Ca 2ϩ responses of the neurons to PAR 2 activation. In contrast, when neuronal populations were similarly subdivided on the basis of their responsiveness to TRPV1 activation with capsaicin, there was no difference between the responsiveness of the upper and lower quartiles (Fig. 7D, p ϭ 0.06). Our results indicate that TRPV4 largely mediates the PAR 2 -induced Ca 2ϩ signals in DRG neurons.
TRPV4 Mediates PAR 2 -evoked Inflammation-PAR 2 agonists evoke neurogenic inflammation that depends on the local release of neuropeptides from primary spinal afferent neurons (3,5). To investigate the functional relevance of PAR 2 coupling to TRPV4 in the intact animal, we assessed the effects of PAR 2 -AP on peripheral inflammation in trpv4 ϩ/ϩ and trpv4 Ϫ/Ϫ mice. Intraplantar injection of PAR 2 -AP into trpv4 ϩ/ϩ mice resulted in an increase in paw thickness that was maximal within 30 min and maintained for 180 min, indicative of tissue edema and consistent with previous reports (Fig. 8A). The magnitude of PAR 2 -induced paw edema at early time points was reduced in trpv4 Ϫ/Ϫ mice compared with trpv4 ϩ/ϩ mice (p ϭ 0.006 at 30 min, p ϭ 0.052 at 60 min, n ϭ 10 mice). Paw thickness was similar in both groups at 120 and 180 min after PAR 2 -AP (Fig. 8A, p Ͻ 0.05 relative to NaCl-treated mice). No difference in paw thickness was detected between trpv4 ϩ/ϩ and trpv4 Ϫ/Ϫ mice after intraplantar injection of 0.9% NaCl. PAR 2 -AP also increased MPO activity in the paws of trpv4 ϩ/ϩ and trpv4 Ϫ/Ϫ mice (Fig. 8B, p ϭ 0.0028 and p ϭ 0.0088, respectively, compared with NaCl; n ϭ 8 -10). However, MPO activity was similar in trpv4 ϩ/ϩ and trpv4 Ϫ/Ϫ mice (p ϭ 0.8340).
These results indicate that the release of arachidonic acid metabolites and the activation of TRPV4 contributes to the initial phases of PAR 2 -evoked edema. However, TRPV4 is not involved in PAR 2 -mediated recruitment of inflammatory cells.

DISCUSSION
Our results reveal an unexpected functional coupling between PAR 2 and TRP channels, which involves the generation of an arachidonic acid-derived TRPV4 activator and a tyrosine kinase signaling pathway (Fig. 9). We identified this coupling in a model HEK cell line and in primary nociceptive neurons. The coupling gives rise to long-lasting TRPV4-dependent Ca 2ϩ signals and demonstrates a direct involvement of TRPV4 in the cellular response to PAR 2 activation, which contributes to the proinflammatory effects of this receptor. We propose that this coupling represents a mechanism of TRP channel regulation that is distinct from the process of sensitization. This convergence of GPCR and TRP signaling may provide a mechanism through which the specificity and magnitude of cellular responses is conferred and controlled. This proposal FIGURE 8. PAR 2 -AP induced paw edema and granulocyte infiltration. A, intraplantar injection of PAR 2 -AP induced sustained paw edema (*, p Ͻ 0.05 compared with NaCl-treated mice; n ϭ 8 -10 mice/group). This effect was significantly attenuated in trpv4 Ϫ/Ϫ mice compared with trpv4 ϩ/ϩ mice at an early time point (#, p ϭ 0.006 at 30 min; n ϭ 10 mice/group). NaCl injection had no effect. B, intraplantar PAR 2 -AP also increased MPO activity in both trpv4 ϩ/ϩ and trpv4 Ϫ/Ϫ mice (p ϭ 0.0028 and p ϭ 0.0088 compared with NaCl-treated animals, n Ն 8 experiments), but there was no difference between genotypes (p ϭ 0.8340). C and D, pretreatment with 17-ODYA reduced PAR 2 -AP-induced paw edema at 30 min (**, p ϭ 0.0305 compared with PAR 2 -AP/vehicle-treated mice; #, p Ͻ 0.05 compared with NaCl-treated mice; n ϭ 5) without affecting MPO activity (p ϭ 0.019 compared with NaCl-treated mice; n ϭ 5) Analysis of variance and Newman-Keuls test were used.
is supported by other studies demonstrating functional coupling between GPCRs and TRP channels in native cells (28, 29, 37, 44 -48). However, the process of PAR 2 coupling to TRPV4 shows specificity because ATP acting through P2Y receptors did not activate TRPV4, and PAR 2 did not activate TRPA1. These observations suggest that the mechanism of PAR 2 coupling to TRPV4 is not simply due to elevated [Ca 2ϩ ] i , which was similarly stimulated by ATP. Two different modes of PAR 2 activation (PAR 2 -AP and trypsin) induced coupling between PAR 2 and TRPV4. Thus, coupling is not an artifact of using the synthetic tethered ligand-based activating peptide and can also occur in response to an endogenous activator.
Significance of PAR 2 -TRPV4 Coupling for Neurogenic Inflammation and Pain-PAR 2 agonists promote neurogenic inflammation and pain in multiple tissues (1,3,4). TRPV4 is coexpressed with PAR 2 in nociceptive spinal-afferent neurons, and sensitization of these channels following PAR 2 activation is a major mechanism through which both neurogenic pain and inflammation develop (13, 16 -18). In our proposed model, the activity of TRPV4 influences the duration of PAR 2 -dependent signaling, thereby augmenting the effects of acute PAR 2 activation. Thus, both coupling and sensitization of TRP channels may contribute to pathophysiological changes associated with PAR 2 activation.
Our experiments with cultured DRG neurons demonstrate that both the magnitude and duration of PAR 2 signaling are modulated by TRPV4 activity, whereas in HEK293 cells only the sustained phase of the PAR 2 -evoked Ca 2ϩ response was affected by TRPV4 expression. The PAR 2 response was greatly reduced in DRG neurons treated with the TRPV4 antagonist HC067047 or after omission of extracellular Ca 2ϩ ions. These findings are consistent with preliminary studies showing that Ruthenium Red also suppresses PAR 2 signaling in DRG neurons (data not shown). We conclude that PAR 2 -induced Ca 2ϩ signaling in DRG neurons depends in large part upon activation of TRPV4 and a TRPV4-mediated influx of extracellular Ca 2ϩ .
However, a substantial component of the increase in [Ca 2ϩ ] i also derives from intracellular Ca 2ϩ stores.
In this study, we have demonstrated a role for PAR 2 -TRPV4 coupling in the regulation of PAR 2 -mediated paw edema but not in granulocyte infiltration. These findings are consistent with a role for TRPV4 in PAR 2 -dependent signaling in DRG neurons, as the development of paw edema is neurogenic in origin, whereas granulocyte recruitment occurs independently of sensory innervation (3). PAR 2 -TRP Channel Coupling in Other Systems-Our observations may have important functional implications for other cell types in which PAR 2 and TRPV4 are coexpressed, such as bladder urothelium, colonic epithelium, and bronchial epithelial cells (18,(57)(58)(59)(60). Whether physiological and pathophysiological responses to PAR 2 activation are augmented by TRPV4 in these tissues has yet to be determined. PAR 2 activation in these cell types leads to increased barrier permeability and is associated with inflammation (61). Similarly, TRPV4 activation results in inflammation and altered cellular permeability of colonic epithelial cells (58). Thus, as with DRG neurons, TRP channels may also regulate PAR 2 -dependent signaling in these cells.
Signaling Pathways Involved in TRP Activation-PARs signal through phospholipase C and PLA 2 (57,62,63). Arachidonic acid derivatives are endogenous activators of TRPV1, TRPV4, and TRPA1 (24,25,43,64). PAR 2 coupling with TRPV4 involves production of arachidonic acid derivatives, as demonstrated using MAFP, an irreversible inhibitor of PLA 2 (65). This observation is consistent with evidence that PAR 2 activation leads to increased arachidonic acid release and prostaglandin production in enterocytes (57) and with other studies examining GPCR interaction with TRPV4 channels (47). Whether this is a common mechanism in cells that coexpress GPCRs and TRPV4 channels remains to be determined. PLA 2 -derived TRP channel activators contribute to both the transient and sustained phases of the PAR 2 response because MAFP inhibited FIGURE 9. Hypothesized mechanism of PAR 2 and TRVP4-dependent neurogenic inflammation. Proteases released during inflammation and injury cleave and activate PAR 2 on the terminals of nociceptive spinal afferent neurons (1). PAR 2 activates heterotrimeric G proteins, which leads to release of intracellular Ca 2ϩ (2) via the phospholipase C (PLC), phosphatidylinositol 4,5-bisphosphate (PIP 2 ), and inositol 1,4,5-trisphosphate (IP 3 ) pathways. PLA 2 hydrolyzes membrane phospholipids to form arachidonic acid (AA), and cytochrome P450 epoxygenase (P450) generates EETs (3). As endogenous TRPV4 agonists, EETs (most likely 5Ј,6Ј-EET) promote TRPV4 channel opening and influx of extracellular Ca 2ϩ (4). This requires tyrosine 110, a known target for phosphorylation (P) by tyrosine kinases (5). PAR 2 activation triggers calcium-dependent release of calcitonin gene-related peptide (CGRP) and substance P (SP), which mediate neurogenic inflammation (6). Question marks indicate that the tyrosine kinase remains to be identified and that PAR 2 -induced phosphorylation of TRPV4 has not been examined directly.
both components of the response. The inability of indomethacin to block TRPV4 activation excludes involvement of cyclooxygenase-derived prostaglandin in signaling to TRPV4. Lipoxygenase-derived arachidonic acid derivatives, such as hydroxyeicosatetraenoic acids, are endogenous TRPV1 agonists and are not reported to activate TRPV4 (28,43,66,67). In our experiments, lipoxygenase inhibition reduced cell viability and was not examined further (data not shown). We did not observe coupling between PAR 2 and TRPA1. This result is at variance with a recent study in which 5Ј,6Ј-EET dependent neuronal activation of TRPA1 was reported (68). The reason for this discrepancy is presently unknown but may relate to the differences in the studied receptors and the experimental systems.
TRPV4 is modulated by a range of serine/threonine and tyrosine kinases and contains target residues, including those for tyrosine kinases (50,69) and PKC (40). In addition, PKC-activating phorbol esters directly activate TRPV4 (41). We found that PKC-specific inhibition (BIM-1) had no effect on PAR 2 signaling, which excludes a role for PKC in coupling with TRPV4.
Two observations implicate tyrosine phosphorylation of TRPV4 in PAR 2 -TRPV4 coupling. First, the Src family kinase inhibitor, Src1, attenuated TRPV4-dependent PAR 2 signaling. Second, the magnitude of coupling was markedly attenuated by mutation of Tyr-110, a target residue for Src family kinases (50). A limitation of our work is that we have not identified the particular tyrosine kinase involved in PAR 2 -TRPV4 coupling. Although Src1 inhibited coupling, preliminary studies indicated that PP2 was less effective (data not shown). Thus, it is possible that other kinases are involved because nanomolar concentrations of PP2 block certain Src kinases (70). Another caveat of our study is that we did not directly demonstrate phosphorylation of TRPV4-Y110 in response to PAR 2 activation. Further studies using pharmacological and genetic approaches for manipulating kinase activity are required to identify the kinases that mediate tyrosine phosphorylation. Proteomic analyses are necessary to study PAR 2 -induced tyrosine phosphorylation of TRPV4. Once the specific kinase is identified, it will be possible to examine its contribution to protease-evoked inflammation and pain in mice.
Summary-In summary, we have demonstrated coupling between PAR 2 and TRPV4. This activation of TRP channels is mediated by production of endogenous activators and is dependent on key tyrosine residues of TRPV4. Our study has identified critical components of the intracellular signaling pathways underlying the activation of TRP channels. These may represent novel targets for therapeutics aimed at reducing augmented signaling under pathophysiological conditions while leaving normal responses intact. This selective inhibition of aberrant signaling is a more attractive target compared with global inhibition of TRP channels, which have important physiological roles in thermoregulation, osmoregulation, and nociception.