|Home | About | Journals | Submit | Contact Us | Français|
Serine proteases generated during injury and inflammation cleave protease-activated receptor 2 (PAR2) on primary sensory neurons to induce neurogenic inflammation and hyperalgesia. Hyperalgesia requires sensitization of transient receptor potential vanilloid (TRPV) ion channels by mechanisms involving phospholipase C and protein kinase C (PKC). The protein kinase D (PKD) serine/threonine kinases are activated by diacylglycerol and PKCs, and can phosphorylate TRPV1. Thus, PKDs may participate in novel signal transduction pathways triggered by serine proteases during inflammation and pain. However, it is not known whether PAR2 activates PKD, and the expression of PKD isoforms by nociceptive neurons is poorly characterized. Using HEK293 cells transfected with PKDs, we found that PAR2 stimulation promoted plasma membrane translocation and phosphorylation of PKD1, PKD2 and PKD3, indicating activation. This effect was partially dependent on PKCε. Using immunofluorescence and confocal microscopy, with antibodies against PKD1/PKD2, PKD3 and neuronal markers, we found that PKDs were expressed in rat and mouse dorsal root ganglia (DRG) neurons, including nociceptive neurons that expressed TRPV1, PAR2 and neuropeptides. PAR2 agonist induced phosphorylation of PKD in cultured DRG neurons, indicating PKD activation. Intraplantar injection of PAR2 agonist also caused phosphorylation of PKD in neurons of lumbar DRG, confirming activation in vivo. Thus, PKD1, PKD2 and PKD3 are expressed in primary sensory neurons that mediate neurogenic inflammation and pain transmission, and PAR2 agonists activate PKDs in HEK293 cells and DRG neurons in culture and in intact animals. PKD may be a novel component of a signal transduction pathway for protease-induced activation of nociceptive neurons and an important new target for anti-inflammatory and analgesic therapies.
Injury and inflammation result in the release and activation of multiple serine proteases that can regulate cells by cleaving protease-activated receptors (PARs), a family of four G protein-coupled receptors (Ossovskaya and Bunnett, 2004; Ramachandran and Hollenberg, 2008). Proteases from the circulation (e.g., thrombin, factors VIIa, Xa), inflammatory cells (e.g., mast cell tryptase), and the epithelium (e.g., trypsin I, II and IV) cleave PARs within the extracellular amino-terminus to expose a tethered ligand, which binds to and activates the receptor. PAR2 is prominently expressed by a subpopulation of primary spinal afferent neurons containing the neuropeptides substance P (SP) and calcitonin gene-related peptide (CGRP) (Steinhoff et al., 2000). PAR2 agonists stimulate the release of SP and CGRP from these nerves to cause neurogenic inflammation of peripheral tissues and hyperalgesia to thermal and mechanical stimuli (Knecht et al., 2007; Nguyen et al., 2003; Steinhoff et al., 2000; Vergnolle et al., 2001; Vergnolle et al., 1999). PAR2-induced hyperalgesia requires sensitization of transient receptor potential vanilloid (TRPV) ion channels TRPV1 and TRPV4, non-selective cation channels that permit nociceptors to detect thermal and mechanical stimuli. This sensitization occurs by mechanisms dependent on phospholipase C (PLC) and protein kinase C epsilon (PKCε (Amadesi et al., 2006; Amadesi et al., 2004; Grant et al., 2007). Although protein kinase D (PKD) is also regulated by PLC and PKCε, it is not known whether PAR2 activates PKD or whether PKD also mediates PAR2-induced sensitization of TRPV1 or TRPV4.
The three isoforms of PKD, PKD1 (also known as PKCμ), PKD2 and PKD3, comprise a new family of serine/threonine kinases within the Ca2+/calmodulin-dependent protein kinase (CaMKs) group. PKDs play a unique role as direct downstream targets of diacylglycerol (DAG) and PKCs (Rozengurt et al., 2005). Structurally, PKDs are characterized by a regulatory domain that is homologous with the DAG/phorbol ester-sensitive PKCs, and a catalytic domain that is homologous with CaMKs (Rozengurt et al., 2005). PKDs are activated by phosphorylation of serine residues within the activation loop of the catalytic domain through a rapid PKC-dependent transduction pathway (Waldron and Rozengurt, 2003). However, a PKC-independent activation of PKDs has also been described (Jacamo et al., 2008; Rozengurt et al., 2005). Pharmacological agents, such as the DAG analog phorbol esters (Zugaza et al., 1996), and physiological stimuli that elevate intracellular DAG, such as agonists of G protein-coupled receptors (e.g., bradykinin and thrombin) and growth factors, activate PKD through a PKC-dependent mechanism (Poole et al., 2008; Tan et al., 2004; Tan et al., 2003; Zugaza et al., 1997). DAG induces translocation of PKD from the cytosol to the plasma membrane where PKC can phosphorylate and activate PKD. Active PKD dissociates from the plasma membrane and translocates back to the cytosol and, in the case of PKD3, to the nucleus (Rozengurt et al., 2005).
PKDs are expressed in multiple tissues and cell types, including in the nervous system (Song et al., 2007; Wang et al., 2004), and PKDs regulate diverse cellular processes, including signal transduction, membrane trafficking and protein transport (Rozengurt et al., 2005). Recent studies suggest that PKD also regulates inflammation and pain transmission. For example, PKC-dependent activation of PKD in spinal neurons has been observed during endotoxemia and inflammation (Song et al., 2007). PKD also directly interacts with TRPV1, a multimodal cation channel the mediates neurogenic inflammation and pain (Caterina et al., 1997), in cell lines and primary sensory neurons, suggesting that PKD regulates TRPV1 and thereby controls sensory perception and hyperalgesia (Wang et al., 2004). Thus, PKD may be a novel component of a signal transduction pathway, which is activated in primary sensory neurons during inflammation, that sensitizes ion channels that regulate neurogenic inflammation and hyperalgesia. However, the expression and the distribution of PKD isoforms in primary spinal sensory neurons is poorly characterized. It is not known if PKD isoforms are expressed by nociceptive neurons that also express PAR2, or whether inflammatory agents, including agonists of PAR2, activate PKD in these neurons.
In this study, we initially determined if stimulation of PAR2 induces sub-cellular redistribution and phosphorylation of PKD isoforms, consistent with their activation. To do so, we studied a model system of HEK293 cells, which naturally express PAR2, overexpressing PKD isoforms tagged with green fluorescent protein (GFP). This system allowed us to specifically examine PAR2-induced activation of PKD in an isoform-specific manner by examining membrane translocation and phosphorylation by immunostaining, confocal microscopy and Western blotting. We then sought to determine if similar activation occurs in primary sensory neurons. We first examined the expression and the distribution of PKD isoforms in neurons of dorsal root ganglia (DRG). We localized PKD isoforms by immunostaining and confocal microscopy, using markers of different populations of neurons to characterize those expressing PKD. To assess whether agonists of PAR2 and other proinflammatory mediators activate PKD in these neurons, we incubated cultured DRG neurons with a PAR2-selective agonist or a mixture of proinflammatory mediators, and used an antibody to phosphorylated (activated) PKD to assess activation. To determine whether this activation can also occur in the intact animal, we administered a PAR2 agonist by intraplantar injection, and studied phosphorylation of PKD in neurons of the lumbar DRG. Our results show that PKD isoforms are expressed by primary spinal afferent neurons, including those that express PAR2, TRPV1, SP and CGRP, which mediate protease-induced neurogenic inflammation and hyperalgesia. PAR2 agonists activate all isoforms of PKD in model systems, cultured neurons and the intact animal. These results support the hypothesis that PKD isoforms mediate protease-induced neurogenic inflammation and pain.
C57Bl6 mice (male, 6–8 weeks) and Sprague Dawley rats (male, 200–250 g) were from Charles River Laboratories (CA). The Institutional Animal Care and Use Committee of the University of California, San Francisco approved the study. Animals were humanely killed using sodium pentobarbital (200 mg/kg i.p.) and bilateral thoracotomy.
Primary antibodies are shown table 1. Secondary antibodies were: goat anti-rabbit, -mouse or -guinea pig IgG conjugated to fluorescein isothiocyanate, rhodamine red or Texas Red (Jackson Immuno-Research, West Grove, PA, 1:200); rabbit anti-horseradish peroxidase (Jackson Immuno-Research, 1:500); goat anti-mouse or rabbit conjugated to Alexa Fluor 680 (Invitrogen, 1:10,000) or IRDyeTM 800 (Rockland Immunochemicals, Gilbertsville, PA, 1:10,000). Streptavidin conjugated to Alexa-488 was from Molecular Probes Invitrogen (Carlsbad, CA, 1:2,000).
The antibody to PKD1/2 (anti-PKD1/2) has been previously characterized (Rey et al., 2001; Rey et al., 2003b). PKD1/2 immunoreactivity was eliminated by preincubation of the diluted antiserum (1:100) with the synthetic peptide used for the immunization (Santa Cruz Biotechnology, sc-639P), suggesting specificity (see supplemental Figure 1 A). Moreover, in rat DRG lysates analyzed by Western blotting, the anti-PKD1/2 antibody recognizes a band of ≥100 KDa, consistent with the size of rat PKD1 (~102 KDa). Also in this case, the positive immunoreaction was prevented by preincubation with the peptide used for immunization (Zhu et al., 2008). The PKD1/2 antibody also recognized a faint band of ≤100 KDa, similar to the size of rat PKD2 (~96.5 KDa). This finding can be probably explained since the antibody is not selective for PKD1 but also reacts with PKD2. The antibody to PKD3 (anti-PKD3) has been previously characterized (Rey et al., 2001; Rey et al., 2003b). The peptide used for the immunization and production of PKD3 antibody was not available. However, there is only one residue that is shared between PKD3 and PKD1 within the sequence of the immunogen peptide. Moreover, when the immunogen peptide sequence is subjected to Blast analysis, neither PKD1 nor PKD2 are identified as homologous. We further investigated specificity of the PKD3 antibody using HEK293 cells transfected with PKD1-, PKD2- and PKD3-GFP. PKD3-like immunoreactivity (LI) was detected only in cells expressing PKD3-GFP, confirming specific detection of PKD3 and not PKD1 or PKD2 (Supplemental figure 1 B-O). Using a similar approach, the lack of cross-reaction between the PKD3 antibody and PKD1 and PKD2 was confirmed in a recent study that characterized the expression of PKD3 in the developing mouse (Ellwanger et al., 2008). In this study, Western blotting analysis indicated that the anti-PKD3 recognizes a band of a ~100 KDa in HEK293 cells, consistent with the size of PKD3. The phospho-PKD/PKCmu antiserum has been previously characterized (Jacamo et al., 2008; Waldron et al., 2001). The antibody recognizes the phosphorylated state of the equivalent serines within the activation loop of the three PKD isoforms (PKD1: Ser744-Ser748; PKD2: Ser706-Ser710; PKD3: Ser731-Ser735) (Waldron et al., 2001). The peptide used for the immunization and production of the antibody was not available. However, we detected immunoreactivity only in cells transfected with PKD1-, PKD2- and PKD3-GFP when stimulated with phorbol ester (not shown), suggesting specificity. Moreover, in rat DRG lysates analyzed by Western blotting, the anti-phospho-PKD/PKCmu antibody recognized a band of ~100 KDa, consistent with the size of rat PKDs (Besirli and Johnson, 2006). The CGRP antibody has been characterized (Wong et al., 1993), including in enteric, DRG and spinal cord neurons of the rat (Cottrell et al., 2005). The pattern of CGRP immunoreactivity that we detected confirms the results of previous studies of peptidergic sensory neurons in rats (Steinhoff et al., 2000). Moreover, the lack of positive immunoreactivity in histological sections of DRG from alpha-CGRP knockout mice confirms specific detection of CGRP (Zhang et al., 2001). The SP antibody has been used to characterize peptidergic sensory neurons, and the pattern of expression we detected confirms previous studies (Steinhoff et al., 2000). Moreover, the lack of positive immunoreactivity in primary sensory neurons from preprotachykinin-A knockout mice confirms specific detection of SP (Sanderson Nydahl et al., 2004). The antibody to PAR2 has been previously characterized and staining in vascular tissues and in pancreatic stellate cells was prevented by preadsorption with the antigen used for immunization (Masamune et al., 2005) and there was no staining when the antibody was replaced with normal murine IgG2a (Molino et al., 1998). Moreover, we detected PAR2-LI only in rat kidney epithelial cells (KNRK) transfected with rat or human PAR2 but not in untransfected cells suggesting specific detection of PAR2 (Bunnett, unpublished observations). The pattern of PAR2-LI we detected in DRG is identical to previous reports (Grant et al., 2007), and is similar to the distribution found using a different PAR2 antiserum and by in situ hybridization (Steinhoff et al., 2000). The antibody to TRPV1 has been previously characterized in rat DRG (Amadesi et al., 2004), and the pattern of TRPV1-LI we detected was similar to previous studies that used a different TRPV1 antiserum (Caterina et al., 1997). Immunoreactivity was abolished by preadsorption with the antigen blocking peptide (not shown) (Steinhoff et al., 2000). Moreover, the lack of TRPV1-LI in DRG from TRPV1 knockout mice confirms specific detection of TRPV1 (Baiou et al., 2007). The antibody to neurofilament 200 kDa (N52) recognizes the phosphorylated and dephosphorylated H-chain of neurofilament 200 kDa, one of the three intermediate filament proteins that serve as major supporting elements of the neuronal cytoskeleton. As neurofilaments are typically restricted to neurons, N52 is used as neuronal markers to identify cells of neuronal origin (Shaw et al., 1986).
PAR2-activating peptide (AP) (SLIGRL-NH2) and the inactive PAR2-reverse peptide (RP) (LRGILS-NH2) were from SynPep Corp. (Dublin, CA). Phorbol-12, 13-dibutyrate (PDBu) and phorbol-12-myristate-13-acetate (PMA) were from Calbiochem. Bradykinin and SP were from Bachem (Torrance, CA). 5-Hydroxytryptamine and histamine were from Sigma (St. Louis, MO). The membrane-permeable PKCε translocation inhibitor peptide, conjugated to an 11-amino-acid protein-transduction domain from the human immunodeficiency virus TAT protein (YGRKKRRQRRRC-disulfide bond-CEAVSLKPT-COOH, TAT-PKCεI) (Schwarze et al., 1999), and the inactive scrambled sequence (TAT-PKCεI-sc) were from SynPep. Biotin-conjugated isolectin IB4 from Bandeiraea simplicifolia was from Sigma (L3759). It was incubated with histological sections of DRG (1:100) to detect nociceptive non-peptidergic, small diameter, unmyelinated neurons (Silverman and Kruger, 1990). IB4 binds to alpha-D-galactose expressed on sensory ganglion cells and their central terminals in the spinal cord.
HEK-FLP cells (Invitrogen) were maintained in DMEM, 10% FBS and Zeocin (100 μg/ml). Cells were transiently transfected with 2–3 μg of PKD1, PKD2 and PKD3 (GFP tagged), using Lipofectamine 2000™ (Invitrogen). At 48 h after transfection, cells were incubated in low-serum medium (0.5% FBS) for 16–18 h (Poole et al., 2008). Cells were treated at different time points with test substances, for then processed for immunofluorescence or Western blot analysis.
Cultures of DRG neurons from thoracic and lumbar spinal regions were prepared as described (Amadesi et al., 2004). Neurons were incubated in low-serum medium (0.2% FBS) for 16–18 h. Cells were treated with PAR2-AP or PAR2-RP (100 μM, 1, 5, 15 min), PDBu (1 μM, 5 min), or with an inflammatory mixture of bradykinin (1 μM), SP (1 μM), 5-hydroxytryptamine (10 μM) and histamine (10 μM) (1–5 min), and then processed for immunofluorescence.
Mice under light anesthesia (2.5% isoflurane) received an intraplantar injection of saline, PAR2-AP or PAR2-RP (10 μg in 10 μl saline per paw). After 30 min, animals were transcardially perfused with 4% paraformaldehyde (PFA) in 100 mM PBS, pH 7.4 and lumbar DRG (L4 – L6) were collected and processed for immunofluorescence.
Animals were transcardially perfused with 4% PFA in 100 mM PBS, pH 7.4. DRGs were removed and fixed in 4% PFA for 4 h at room temperature, then washed, incubated in 25% sucrose in PBS overnight at 4°C, and embedded in OCT. Sections of mouse and rat DRG (10–16 μm) were cut and mounted on slides coated with poly-L-lysine. Histological sections were washed and incubated in 10 mM PBS, pH 7.4, containing 5% normal goat or donkey serum and 0.3% Triton X-100 for 1 h at room temperature. Specimens were incubated with primary antibodies for 16 h at 4°C, then washed and incubated with secondary antibodies: goat anti-rabbit, -mouse or -guinea pig IgG conjugated to fluorescein isothiocyanate, rhodamine red, Texas Red (2 h, room temperature), or streptavidin conjugated to Alexa-488 (2 h, room temperature). TRPV1 signals were amplified using a Fluorescein Tyramide Signal Amplification System (Perkin Elmer, Boston, MA). To determine the specificity of the PKD1/2 antibody, we pre-absorbed the antibody with the peptide used for immunization. To determine the specificity of the PKD3 antibody, as the immunizing peptide is not available, we localized PKD3 in HEK293 cells transfected with PKD1-, PKD2- or PKD3-GFP and activated with PDBu. Histological sections were mounted in Prolong Gold (Invitrogen).
HEK293 cells transiently transfected with PKD1-, PKD2- or PKD3-GFP and cultured rat DRG were fixed with 4% PFA for 20 min at 4°C. Cells were washed and incubated with 10 mM PBS, pH 7.4 containing 1% normal goat serum and 0.1% saponin for 30 min at room temperature. HEK293 cells were incubated with rabbit anti-pSer744/748 (16 h, 4°C), washed, and incubated with goat anti-rabbit IgG conjugated to rhodamine red-X (1 h, room temperature). DRG were treated with the Fluorescein Tyramide Signal Amplification System (Perkin Elmer) according to the manufacturer's instructions. DRG were incubated with the rabbit anti-pSer744/748 (16 h, 4°C), washed, and incubated with rabbit anti-horseradish peroxidase (1 h, room temperature). Cultured cells were mounted in Vectashield (Vector Laboratories, Burlingame, CA).
Specimens were observed with a Zeiss Axiovert and an LSM 510 META laser scanning confocal microscope with Zeiss EC Plan-Neofluar 40x/1.3 or Zeiss Plan-Apo 100x/1.4 oil objectives. Images of 512 × 512 pixels were collected. For colocalization, single optical sections at the same focal plane were taken separately and the corresponding channels were merged. Images were collected, colored to represent the appropriate fluorophores, and processed to adjust contrast and brightness using Adobe Photoshop CS2 9.0 (Adobe Systems, Mountain View, CA). The pixel intensity of images was quantified using the Zeiss LSM Image Browser, Version 126.96.36.199. Neurons were considered positive for the expression of the proteins of interest if the immunoreactive signal was above the average background fluorescence signal used as minimum threshold. The activation of PKD in HEK293 cells was expressed as the ratio between the pixel intensity of the pSer744/748 and GFP signals. The activation of PKD in cultured DRG neurons was expressed as the ratio between the intensity of the pSer744/748 signal per area analyzed (pixel intensity/μm). Ratios were then normalized to the appropriate control. The activation of PKD in histological sections of mouse DRG was expressed as average pixel intensity of the pSer744/748 signal. Neuronal sub-populations were distinguished on the basis of cell diameter (large neurons, >30 μm; medium/small neurons, ≤30 μm). Values were compared to controls within the same neuronal sub-population.
Cells were washed and lysed in 50 mM Tris-HCl pH 7.4, 50 mM NaF, 1 mM Na3VO4, 1% SDS and protease inhibitors, boiled, and centrifuged. Lysates (40 μg) were separated by SDS-PAGE (8% acrylamide), transferred to polyvinylidene difluoride membrane (Immobilon FL, Millipore Corp.), and blocked in blocking buffer (LI-COR, Lincoln, NE), and incubated with rabbit anti-pSer744/748 (1:1,000) and mouse anti-GFP (1:10,000) (overnight, 4°C). Membranes were washed and then incubated with secondary antibodies coupled to either Alexa Fluor 680 (1:10,000) or IRDyeTM 800 (1:10,000) (1 h, room temperature). Immunoreactive proteins were detected using an Odyssey Infrared Imaging System (LI-COR). For the densitometric analysis, the pSer744/748 signal and the GFP-signal (used as loading control) were calculated, then expressed as a ratio and compared to non-treated cells.
Total RNA from rat DRG was isolated using Trizol (Invitrogen). RNA (5 μg) was reverse transcribed using standard protocols with random hexamers and TaqMan reverse transcription reagents (Applied Biosystems, Foster City, CA). Subsequent PCR reactions used primers specific for rat PKD1 (forward: 5′-ggagactggacagcaaaagc-3′ and reverse 5′-gtccagaacccagaacctca-3′), PKD2 (forward: 5′-gatccgtttcctcaggtgaa-3′ and reverse: 5′-tggcgctctgatacaaactg-3′) and PKD3 (forward 5′-aaatgtgctgcttgcatctg-3′ and reverse 5′-acgttctccaatgcgagttt-3′). Control reactions omitted reverse transcriptase. PCR products were separated by electrophoresis (2% agarose gel), and visualized using ethidium bromide. Product identity was confirmed by sequencing.
Results were normalized to the appropriate control and expressed as mean ± SEM. Groups were compared using ANOVA and Bonferroni tests or Student's t test analysis. Differences were considered significant when p was ≤0.05.
To investigate whether stimulation of PAR2 activates PKD1, PKD2 and PKD3, we studied trafficking and phosphorylation of GFP-tagged kinases expressed in HEK293 cells. By immunostaining and confocal microscopy, we studied redistribution of kinases and the levels of phosphorylated kinases as a measure of activation. To monitor the activation of the PKD isoforms we used an antibody raised against a peptide phosphorylated on serines equivalent to Ser744 and Ser748 of PKD but that predominantly detects the phosphorylated state of Ser744, as originally shown (Waldron et al., 2001) and verified in a recent study (Jacamo et al., 2008). This antibody (pSer744/748) detects the phosphorylated state of the equivalent serines within the activation loop of the three PKD isoforms (PKD1: Ser-744; PKD2: Ser-706; and PKD3: Ser-731). In non-treated cells (0 min), PKD1- (Figure 1, A1-C1) PKD2- (Figure 1, A2-C2) and PKD3-GFP (Figure 1, A3-C3) were localized in the cytosol and a pSer744/748-LI was not detected. PAR2-AP (100 μM, 1 min) induced translocation of each PKD isoform from the cytosol to the plasma membrane and increased the pSer744/748-LI for all the three PKDs (PKD1, figure 1, D1-F1; PKD2, figure 1, D2-F2; PKD3, figure 1, D3-F3). After 5 min, PKDs started to return to the cytosol, but remained phosphorylated, suggesting a sustained activation (PKD1, figure 1, G1-I1; PKD2, figure 1, G2-I2; PKD3, figure 1, G3-I3). In a similar manner, PMA, a non-selective PKC activator used as a positive control (1 μM, 5 min), induced a robust translocation of PKD1-, PKD2- and PKD3-GFP from the cytosol to the plasma membrane and a large increase of the pSer744/748 signal for all the three PKDs (PKD1, figure 1, A4-C4; PKD2, figure 1, D4-F4; PKD3, figure 1, G4-I4).
To investigate the role of PKCε in PAR2-mediated activation of PKDs, we treated cells with the PKCε inhibitor (TAT-PKCεI) or with the inactive scrambled sequence used as control (TAT-PKCεI-sc) (10 μM, 15 min). In cells pretreated with TAT-PKCεI-sc, PAR2-AP (5 min) increased the levels of phosphorylated PKDs-GFP (fold-increases compared to non-treated cells: PKD1, 1.03; PKD2, 1.47; PKD3, 0.48; figure 2A–C). After 15 min, phosphorylation of PKDs was still sustained (fold-increases compared to control: PKD1, 0.46; PKD2, 0.76; PKD3, 0.47; figure 2A–C). Pretreatment with TAT-PKCεI reduced PKDs phosphorylation induced by PAR2-AP applied for 5 min (phosphorylation reduction % compared to control: PKD1, 13.7; PKD2, 16.3; PKD3, 14.7; figure 2A–C) and PKD3 phosphorylation induced by PAR2-AP applied for 15 min (phosphorylation reduction %: 15.0, figure 2C) compared to cells treated with TAT-PKCεI-sc for the same time.
Activation of PKDs was confirmed by Western blot analysis using the pSer744/748 antibody and densitometry. PAR2-AP (1–30 min) increased the level of phosphorylated PKD1-GFP (1.5-, 4.2-, 4.1- and 7.8-fold increases at 1, 5, 15, 30 min. Figure 3A, B), PKD2-GFP (6.2-, 8.8-, 7.2- and 3.5-fold increases at 1, 5, 15, 30 min. Figure 3C, D), and PKD3-GFP (2-, 2.5-, 2.8- and 2.4-fold increases at 1, 5, 15, 30 min. Figure 3E, F) when compared to non-treated cells. As expected, the PKD activator PMA also increased the level of phosphorylated PKDs-GFP, whereas PAR2-RP was completely inactive (Figure 3A–F). Thus, stimulation of PAR2 induces translocation and a sustained activation of PKD1-, PKD2- and PKD3-GFP in HEK293 cells. PAR2-mediated PKD phosphorylation depends on PKCε activation.
PKD1 is expressed in rat DRG (Wang et al., 2004). However, the distribution of the PKD isoforms within the different neuronal types has not been characterized. By RT-PCR, we found that rat DRG neurons expressed mRNA of the sizes predicted for PKD1 (448 bp) and PKD2 (479 bp) and PKD3 (488 bp) (Fig. 4). By immunofluorescence and confocal microscopy, we localized PKD-LI in histological section of rat DRG neurons. Using an antibody that recognizes both PKD1 and PKD2 (PKD1/2, figure 5) we detected PKD1/2-LI in 64% of small, medium and large diameters neurons (255/401 neurons). In some neurons, PKD1/2-LI co-localized with CGRP (Figure 5, A–C) and TRPV1 (Figure 5, D–F), which are markers for peptidergic small-diameter neurons considered to be nociceptors, and also with PAR2 (Figure 5, G–I) that is known to be expressed by a subset of nociceptive neurons (Steinhoff et al., 2000). PKD1/2-LI also co-localized with IB4, a marker for nociceptive non-peptidergic, small diameter, unmyelinated neurons (IB4, figure 5, J–L), and with N52, a marker for large diameter neurons with myelinated fibers (Figure 5, M–O). Table 2 summarizes the expression pattern for PKD1/2-LI and neuronal markers. Using an antibody that recognizes PKD3, we detected PKD3-LI with a distribution pattern similar to that for PKD1/2-LI (Figure 6). PKD3-LI was detected in 77% of small, medium and large diameter neurons (210/273 neurons), in peptidergic nociceptive neurons (CGRP and TRPV1 positive neurons. Figure 6, A–C and D–F), in non-peptidergic unmyelinated neurons (IB4 positive. Figure 6 J–L), and in myelinated neurons (N52 positive. Figure 6, M–O). In some neurons, PKD3-LI co-localized with PAR2 (Figure 6, G–I). Table 2 summarizes the expression pattern for PKD3-LI and neuronal markers. Incubating the PKD1/2 antibody with the peptide used for the immunization or omitting primary antibodies prevented staining, suggesting antibody specificity (Supplemental figure 1A). Since the peptide used for the immunization and production of PKD3 antibody was not available, we further confirmed specificity of the PKD3 antibody using HEK293 cells transfected with PKD1-, PKD2- and PKD3-GFP (Supplemental figure 1, B–O). Cells were treated with PDBu (1 μM, 5 min), a non-selective PKC/PKD activator, to induce kinase activation (Rey et al., 2004). In non-treated cells, PKDs-GFP were localized in cytosolic vesicles, but when treated with PDBu, PKDs-GFP moved to the plasma membrane. The anti-PKD3 antibody showed positive immunoreactivity only in cells transfected with PKD3-GFP (Supplemental figure 1, J–O), where it was colocalized with the GFP-signal, suggesting specific interaction of the antibody only with PKD3. There were no detectable signals in cells expressing PKD1- or PKD2-GFP.
Thus, PKDs are expressed in rat DRG neurons, localized in a subset of peptidergic nociceptive neurons, and, in some of these neurons, are co-expressed with PAR2.
We similarly examined the localization of PKD1/2 and PKD3 in mouse DRG. We detected PKD1/2-LI in 59% of small, medium and large diameters neurons (72/123 neurons) (Figure 7, A–F) and PKD3-LI in 77% of small, medium and large diameters neurons (59/77 neurons) (Figure 7, G–L). Moreover, PKDs were co-localized in a subset of neurons with CGRP (PKD1/2, CGRP: figure 7, A–C; PKD3, CGRP: figure 7, G–I) and IB4 (PKD1/2, IB4: figure 7, D–F; PKD3, IB4: figure 7, J–L). Table 2 summarizes the expression pattern for PKD1/2-LI, PKD3 and neuronal markers in histological section of mouse DRG. Thus, PKD1/2 and PKD3 are expressed in mouse peptidergic nociceptive neurons.
We next investigated whether inflammatory mediators activate PKDs in rat DRG in culture. To monitor the activation of the PKD isoforms we used the pSer744/748 antibody. In non-treated cells (treated with vehicle, 0.2% DMSO) the pSer744/748-LI was detected at low levels, usually in cytosolic vesicles (Figure 8 A, J). PDBu (1 μM, 5 min, used as a positive control, induced a 1.5-fold increase of the pSer744/748-LI in the whole cell when compared to cells treated with vehicle (pixel intensity/μm: from 1.00±0.08 to 2.55±0.34), and also induced a redistribution of PKD from the cytosol to the plasma membrane (Figure 8 B, J). A mixture of inflammatory mediators applied for 1–5 min, also induced a ≥0.85-fold increase of the pSer744/748-LI in the whole cell (pixel intensity/μm: from 1.00±0.08 to 1.85±0.26 at 1 min, and 2.80±0.52 at 5 min) (Figure 8 C, D and J). Similarly, PAR2-AP, the selective PAR2 agonist, induced a small but significant increase of the pSer744/748-LI in the whole cell when applied for 5 min (pixel/μm: from 1.00±0.07 to 1.42±0.18 at 5 min) (Figure 8 F, G and J) when compared to cells treated with the inactive control peptide PAR2,-RP (Figure 8 E, J); PKD phosphorylation returned to basal levels 15 min after the treatment (Figure 8 H, J). There was almost no detectable signal when the primary antibody was omitted (Figure 8 I, J). After pre-treatment with inflammatory mixture and PAR2-AP, pSer744/748-LI was increased at the plasmamembrane in only a few cells, which may reflect different levels of expression of PKDs, PAR2 and other receptors. For this reason, together with the finding that PKD1/2-LI and PKD3-LI were also detected at the plasma membrane of a few non-treated cells (Supplemental Figure 2), we were unable to examine whether PAR2 agonists or inflammatory mediators caused membrane translocation of PKD1/2-LI and PKD3-LI in cultured neurons. However, our results suggest that inflammatory mediators and the agonist of PAR2 can phosphorylate and activate PKD in primary sensory neurons in culture.
To investigate whether PAR2-stimulation activates PKD in vivo in the intact animal, we injected saline, PAR2-RP or PAR2-AP into the mouse paw (10 μg paw, 30 min). We monitored activation of PKD using the pSer744/748 antibody in histological sections of lumbar DRG (L4-6), that receive nervous projections from the paws (Takahashi et al., 1994). Intraplantar injection of PAR2-AP induced a ≥0.14-fold and ≥0.15-fold increase of the pSer744/748-LI in the total cell of medium/small diameter neurons (diameter ≤ 30 μm) but not in large diameter neurons (diameter > 30 μm), when compared to saline or to PAR2-RP (average pixel: PAR2-AP: 142.57±1.74, 161 neurons; saline: 124.51±1.64, 79 neurons; PAR2-RP: 123.37±2.28, 118 neurons) (Figure 9 A–D). Thus, stimulation of PAR2 can activate PKDs in primary sensory neurons, in the intact animal.
This study shows, for the first time, that stimulation of PAR2 promotes sub-cellular re-distribution and PKCε-dependent phosphorylation/activation of PKD1, PKD2 and PKD3 expressed in a heterologous system, HEK293 cells. Our results also show that PKDs are expressed in rat and mouse DRG neurons. Some of these neurons are peptidergic, nociceptive neurons that express PAR2 and the PKD1 target, TRPV1. An agonist of PAR2 and proinflammatory mediators activate PKDs in these neurons in culture and in the intact animals. Thus, PKD isoforms are present in nociceptive neurons and are activated by agents, notably PAR2 agonist, that cause neurogenic inflammation and hyperalgesia.
To determine if stimulation of PAR2 induces activation of PKD, we used HEK293 cells transfected with GFP-tagged PKDs. HEK293 cells are a model system that has been extensively used to characterize molecular mechanisms of signaling, including neuron-specific pathways (Campbell et al., 2005; Shaw et al., 2002). Although originally derived from human embryonic kidney epithelial cells, HEK293 cells resemble neurons in that they express neuronal proteins and display neuronal characteristics (Campbell et al., 2005; Shaw et al., 2002). Moreover, HEK293 cells are an established model system used to characterize PAR2 signaling and regulation of TRPV channels (Amadesi et al., 2004; Dai et al., 2007; Grant et al., 2007), including kinase activation (Amadesi et al., 2006).
We found that PAR2 stimulation promoted translocation to the plasma membrane and phosphorylation of all the three PKD isoforms expressed in HEK293 cells, indicative of activation. To our knowledge, this is the first report of a role of PKD as second messenger in the molecular pathway of PAR2-mediated responses. However, the contribution of different PKD isoforms to PAR2-mediated effects remains unknown. PKD1 directly interacts with TRPV1 in cell lines and in primary sensory neurons, thereby regulating TRPV1 functions, such as sensory perception and hyperalgesia (Wang et al., 2004). Thus, it is possible that PKD contributes to PAR2-induced sensitization of TRPV channels. We recently showed that PAR2 sensitizes TRPV1 and TRPV4 to induce hypersensitivity to thermal and mechanical stimuli, and that this sensitization requires activation of PKC, which phosphorylates TRP channels, enhancing activity (Amadesi et al., 2006; Grant et al., 2007). Further, PAR2-mediated sensitization of TRPV1 and TRPV4 is prevented by a non-selective PKC and PKD inhibitor, Gö6978 (Amadesi et al., 2006; Grant et al., 2007), and sensitization of TRPV1 is prevented by a selective PKCε antagonist (Amadesi et al., 2006; Grant et al., 2007). PKD is a known target of PKC, including PKCε (Brandlin et al., 2002; Haworth et al., 2007; Poole et al., 2008; Rey et al., 2004; Romero et al., 2006; Waldron and Rozengurt, 2003), PKC-alpha, -delta and -eta (Berna et al., 2007; Song et al., 2007; Tan et al., 2003; Tinsley et al., 2004; Waldron et al., 1999). In previous studies, we and others have shown that PAR2 stimulation activates PKCε in transfected cells, in primary spinal sensory neurons (Amadesi et al., 2006) and in enteric submucosal neurons (Poole et al., 2007). Here we showed that stimulation of PAR2 induced activation of PKD1, PKD2 and PKD3 and this effect was reduced in cells treated with a PKCε translocation inhibitor peptide. Together, these findings suggest that a PKCε-dependent activation of PKD may be a novel signal transduction pathway that mediates protease-induced responses. The finding that the PKCε inhibitor reduces PKD activation in particular at the early time point (5 min after PAR2 stimulation) is in agreement with a recent report of a rapid PKC-dependent followed by a PKC-independent PKD activation pathway as a novel mechanism of prolonged PKD activation induced by agonists of GPCR such as bombesin (Jacamo et al., 2008). The small reduction of PAR2-mediated activation of PKD obtained by the inhibition of PKCε suggests that the activation of other upstream PKC isoforms is maybe involved. PAR2 stimulation can activate other PKC isozymes such as PKC alpha (Ahamed and Ruf, 2004). However, it is also possible that, due to the heterologous system used in our study, it is difficult to prevent the activation of an over expressed kinase (PKD) while inhibiting an endogenously expressed kinase (PKCε).
Having established that stimulation of PAR2 induced activation of the three PKD isoforms over-expressed in HEK293 cells, we sought to investigate if similar mechanisms occur in primary sensory neurons that play an essential role on protease-induced neurogenic inflammation and transmission of pain. We first determined if PKD are expressed in rat DRG. The RT-PCR products and the positive immunoreactivity to previously characterized PKD1/PKD2 and PKD3 antibodies (Rey et al., 2001; Rey et al., 2003b), suggest that PKDs are expressed in rat DRG neurons. The expression of PKD1/2 in DRG neurons has been reported using Western blotting (Wang et al., 2004). However our study, using RT-PCR and immunostaining to different antibodies, further characterizes the expression of the PKD isoforms in DRG neurons. In addition, we also used markers to define the pattern of PKD expression within different neuronal cell types. To our knowledge this is the first study to confirm expression of the three PKD isoforms and characterize PKD distribution in rat DRG neurons. We found that PKDs are expressed in IB4 positive cells, and co-expressed with TRPV1 and CGRP, all markers for small-diameter nociceptive and peptidergic neurons. These findings suggest that PKDs can potentially regulate neuronal functions including neurogenic inflammation and nociception.
Our study shows that PKDs are co-expressed with PAR2, supporting the possibility that PKDs may regulate PAR2-mediated neurogenic effects including hyperalgesia and inflammation. Proteases activate PAR2, which is expressed in small-diameter nociceptive neurons, inducing the release of SP and CGRP to promote neurogenic inflammation and hyperalgesia to mechanical and thermal stimuli (Amadesi et al., 2006; Amadesi et al., 2004; Grant et al., 2007; Knecht et al., 2007; Nguyen et al., 2003; Steinhoff et al., 2000; Vergnolle et al., 2001; Vergnolle et al., 1999). Stimulation of PAR2 activates PLC, leading to production of inositol 1,4,5-triphosphate and DAG, which increases [Ca2+]i and activates protein kinases, respectively (Mule et al., 2002; Swystun et al., 2005). Agonists of GPCRs, including bradykinin and thrombin, activate PKD1, PKD2 and PKD3 in many different cell types (Rey et al., 2004; Rey et al., 2003a; Tan et al., 2003; Yuan et al., 2005; Zugaza et al., 1997). We found that bradykinin, other inflammatory compounds and a selective PAR2 agonist induced phosphorylation and activation of PKDs in cultured rat DRG. The less robust phosphorylation and activation of PKDs induced by the different agonists in DRG compared to HEK293 cells is expected as HEK293 overexpressed PKDs, and probably express higher level of endogenous PAR2, as suggested by intracellular calcium imaging assay and immunostaining (Amadesi, unpublished observations).
Neuronal expression of PKD is not unique to rat. We found that PKDs are also expressed in mouse small-diameter nociceptive and peptidergic DRG neurons mouse. The proportion of PKD1/2-LI and PKD3-LI positive neurons is similar in rat and mice. Nevertheless, in mice, the neuronal expression pattern between PKDs appeared different, suggesting a higher expression of PKD1/2 in peptidergic nociceptive neurons compared to PKD3. However, this result remains to be confirmed using different approaches other then immunostaining.
More importantly, we showed that intraplantar injection of PAR2-AP in mice, increased phosphorylation of PKDs in lumbar DRG. Thus, agonists of PAR2, including proteases released during tissue injury and inflammation, can activate PKDs in the intact animal. In support of our findings, Song et al. (2007) observed that LPS treatment activates PKD in primary spinal neurons in culture, and that peripheral inflammation induced by carrageenan activates PKD in the spinal dorsal horn in rats. Together these findings suggest that PKD may be an important mediator in the signal transduction pathway that underlies responses produced in primary sensory neurons by different inflammatory mediators including serine proteases that activate PAR2.
In conclusion, we characterized the expression and the distribution of PKD isoforms in DRG neurons and identified PKDs as a novel molecular pathway of PAR2-mediated responses. Our findings suggest that stimulation of PAR2, which is co-expressed with PKDs in a sub-population of medium-small diameter nociceptive sensory neurons, can promote phosphorylation and activation of PKDs. Thus, PKDs may participate in a novel signal transduction pathway that underlies protease-mediated effects. PKD stimulation maybe involved in PAR2-induces phosphorylation and sensitization of TRP channels, such as TRPV1 and TRPV4. Our study adds to our understanding of the molecular mechanisms regulating responses induced by proteases that activate PAR2, including inflammation and hyperalgesia. Proteases that activate PAR2 are released and generated during pathological conditions, such as irritable bowel syndrome (Cenac et al., 2007). Proteases can directly stimulate sensory neurons and generate hypersensitivity symptoms through the activation of PAR2 (Cenac et al., 2007) and stimulation of PAR2 produces visceral hyperalgesia that is mediated by TRPV4 (Cenac et al., 2008). Therefore, our study contributes to our understanding of the molecular mechanisms of PAR2-mediated effects and of the role of proteases in human disease. However, the function of PKDs in PAR2-mediated effects needs to be elucidated. Further studies are required to understand the role of the PKD pathway in PAR2-mediated neurogenic inflammation, hyperalgesia and sensitization of TRP channels. The lack of selective PKD inhibitors and the possible overlapping roles of different PKD isoforms present challenges to understanding the roles of PKD isoforms in the nervous system.
We thank Lorna Divino for technical assistance and Pamela Derish for editing the manuscript.
Grant sponsor: National Institutes of Health DK57840, DK43207 and DK39957 (N.W.B.); the Wellcome Trust 071987 (A.D.G.); and a CJ Martin Fellowship NHMRC 454858 (DPP).