Kallikrein-related Peptidase 4 (KLK4) Initiates Intracellular Signaling via Protease-activated Receptors (PARs)

Kallikrein-related peptidase 4 (KLK4) is one of the 15 members of the human KLK family and a trypsin-like, prostate cancer-associated serine protease. Signaling initiated by trypsin-like serine proteases are transduced across the plasma membrane primarily by members of the protease-activated receptor (PAR) family of G protein-coupled receptors. Here we show, using Ca2+ flux assays, that KLK4 signals via both PAR-1 and PAR-2 but not via PAR-4. Dose-response analysis over the enzyme concentration range 0.1–1000 nm indicated that KLK4-induced Ca2+ mobilization via PAR-1 is more potent than via PAR-2, whereas KLK4 displayed greater efficacy via the latter PAR. We confirmed the specificity of KLK4 signaling via PAR-2 using in vitro protease cleavage assays and anti-phospho-ERK1/2/total ERK1/2 Western blot analysis of PAR-2-overexpressing and small interfering RNA-mediated receptor knockdown cell lines. Consistently, confocal microscopy analyses indicated that KLK4 initiates loss of PAR-2 from the cell surface and receptor internalization. Immunohistochemical analysis indicated the co-expression of agonist and PAR-2 in primary prostate cancer and bone metastases, suggesting that KLK4 signaling via this receptor will have pathological relevance. These data provide insight into KLK4-mediated cell signaling and suggest that signals induced by this enzyme via PARs may be important in prostate cancer.

Kallikrein-related peptidase 4 (KLK4) 2 is a trypsin fold serine protease from the 15-member human KLK family (1)(2)(3). Consistent with its predicted substrate specificity, which is based on the presence of an aspartate six residues before the catalytic serine (1), KLK4 cleaves peptide substrates following arginine or lysine residues (4 -6). In addition, several macromolecular substrates of potential pathophysiological relevance have been identified from in vitro studies, including another KLK family member, pro-prostate-specific antigen (also known as KLK3) and pro-urokinase-type plasminogen activator (4,7) as well as fibrinogen (6) and the urokinase-type plasminogen activator receptor (7).
KLK4 is highly expressed in normal prostate (1,3,8) and recently this protease has been associated with prostate cancer progression. For example, stable overexpression of KLK4 in prostate cancer PC-3 cells resulted in an increased ability of these cells to migrate, accompanied by a transition from an epithelial morphology to a fibroblastic shape and, consistently, a significant decrease in E-cadherin protein levels and an increase in vimentin expression (11). In addition, using an inducible expression system, it has been demonstrated that overexpression of KLK4 results in significantly increased colony formation, migration, and proliferation of PC-3 cells and another prostate cancer cell line, DU145 (12). Furthermore, KLK4 protein levels are elevated in malignant prostate compared with normal tissue (12,13), while prostate cancer patient serum contains antibodies that bind recombinant KLK4 (14). Most recently, using co-culture systems, it has been shown that KLK4 is a potential mediator of cellular interactions between prostate cancer cells and osteoblasts (bone-forming cells) in bone metastases (15).
Members of the G protein-coupled receptor (GPCR) subfamily comprising protease-activated receptor (PAR)-1 to PAR-4, in contrast to other GPCRs, which are activated by docking of soluble ligands, are irreversibly activated by the action of proteases (16 -19). Indeed, PAR activation is almost exclusively mediated by trypsin fold serine proteases with substrate specificity for cleavage following arginine or lysine residues. Cleavage-inducing activation occurs at a unique site within the amino-terminal exodomain of the receptor, generating a new amino terminus that serves as a tethered ligand that binds intramolecularly, causing allosteric changes within the PAR, followed by receptor coupling to heterotrimeric G proteins and signal transduction (16). Importantly, cleavage downstream of the activation site fails to mobilize Ca 2ϩ and results in unresponsiveness to protease agonists (20,21). Recently, several members of the KLK family have been shown to initiate trans-plasma membrane signal transduction via PARs. Oikonomopoulou et al. (22) demonstrated that KLK14 activates PAR-2 and PAR-4 but inactivates (or "disarms") PAR-1. This group also showed that KLK5 and KLK6 activate PAR-2 (22). In addition, Angelo et al. (23) have shown that KLK6 is capable of cleaving a peptide spanning the PAR-2 activation site but not peptides spanning the activation site of the other PARs. KLK5 and KLK14 signaling via PAR-2 has been demonstrated independently in a study that also showed that KLK7 and KLK8 are not capable of signaling through this receptor (24).
Since the mechanism by which KLK4 mediates its effects on cells is not known, we have examined the ability of this protease to initiate cell signaling via members of the PAR family by examining changes in intracellular calcium ion concentration. We show that KLK4 initiates Ca 2ϩ mobilization via PAR-1 and PAR-2 but not via PAR-4. Focusing on PAR-2, we have also examined the ability of KLK4 to activate extracellular signalregulated kinase 1/2 (ERK1/2), including siRNA knockdown approaches, to demonstrate that KLK4 signaling is specifically mediated by this receptor. The potential physiological relevance of KLK4 signaling via PAR-2 was explored by immunohistochemical analysis of agonist and receptor expression in primary prostate cancer and bone metastasis lesions. We have also examined cellular consequences of KLK4-mediated signaling via PAR-2 in prostate cancer PC-3 cells.
Expression Constructs and Transfections-An insect cell KLK4 expression construct was generated by ligating the human KLK4 open reading frame, including the prepro region from the previously published pDNA3.1:KLK4 construct (11), into the pIB/V5-His vector (Invitrogen). This construct generates the complete KLK4 amino acid sequence followed by V5 (GKPIPNPLLGLDST) and His 6 tags. Insect Sf9 cells were transfected using Cellfectin (Invitrogen). A mammalian expression construct encoding PAR-2 with green fluorescent protein (GFP) at the COOH terminus was generated in the pEGFP-N1 vector (Clontech, Mountain View, CA). A PCR employing Pfu DNA polymerase (Invitrogen) was used to amplify the PAR-2 coding region. Following amplification and purification, the PAR-2 PCR product was restriction-digested and then ligated into pEGFP-N1. Mammalian cells were transiently transfected using Lipofectamine (Invitrogen). The sequence of both constructs was verified.
Generation and Purification of Recombinant KLK4 (rKLK4)-Following transfection with the KLK4-pIB/V5-His construct, stable Sf9 cells were selected with 50 g/ml Blasticidin (Invivo-Gen, San Diego, CA). KLK4 was purified from conditioned media from these cells using Ni 2ϩ -nitrilotriacetic acid superflow resin by following the instructions of the manufacturer (Qiagen, Doncaster, Australia). Eluted fractions containing KLK4, identified by analysis of a Coomassie-stained polyacrylamide gel, were pooled and then concentrated. Following dialysis against phosphate-buffered saline (PBS; pH 7.4) at 4°C overnight, KLK4 was aliquoted and stored at Ϫ80°C.
Activation of rKLK4-Recombinant zymogen KLK4 was incubated with the metalloendopeptidase thermolysin in PBS at pH 7.4 for 1 h at 37°C (KLK4/thermolysin ratio 80:1). The amount of active enzyme produced was quantified by active site titration using the pseudosuicide inhibitor 4-methylumbelliferyl p-guanidinobenzoate (27). Zymogen KLK4 and thermolysin-treated KLK4 were examined on a Coomassie-stained polyacrylamide gel, protein bands were excised, and the amino terminus of the excised treated KLK4 was sequenced by Edman degradation at the Australian Proteome Analysis Facility (North Ryde, Australia). Thermolysin activity was stopped by the addition of phosphoramidon (10 M).
Kinetic Measurements of Activated rKLK4-Determination of kinetic parameters was performed using 40 nM active rKLK4 against the tripeptide substrate benzyl-FVR-p-nitroanilide. Assays were performed in 50 mM Tris-HCl (pH 7.4), 20 mM CaCl 2 , 0.01% (v/v) Tween 20 at 28°C by following p-nitroanilide release photometrically at 405 nm. Experiments to determine the kinetics of cleavage of a fluorescence-quenched peptide with sequence spanning the PAR-2 serine protease activation site were performed at 37°C in 50 mM Tris-HCl (pH 7.4), 20 mM CaCl 2 , 0.01% (v/v) Tween 20 by continuously measuring fluorescence at 440 nm following excitation at 330 nm. Emissions were monitored over 15 min using a Polarstar Optima plate reader (BMG Labtech, Melbourne, Australia). Enzyme activity was determined from a standard curve generated from the fluorescence obtained following complete cleavage by trypsin of known amounts of peptide using an absorption coefficient of 10 4 M Ϫ1 cm Ϫ1 . To obtain k cat and K m values, initial velocity data at each substrate concentration were fitted to the Michaelis-Menten equation by nonlinear regression analysis using GraphPad Prism 4.0 (GraphPad Software, San Diego, CA). Experiments were performed in triplicate on three independent occasions with results displayed as means Ϯ S.E.
Measurement of Changes in Intracellular Ca 2ϩ -Cells grown to 80% confluence were washed with PBS, detached nonenzymatically, resuspended (4 ϫ 10 6 cells/ml) in extracellular medium (121 mM NaCl, 5.4 mM KCl, 0.8 mM MgCl 2 ⅐6H 2 O, 1.8 mM CaCl 2 , 5.5 mM glucose, 25 mM HEPES (pH 7.4)) containing 0.2% (w/v) bovine serum albumin (Sigma) and then loaded with the fluorescence indicator Fura-2 acetoxymethyl ester (1.0 M; Invitrogen) at room temperature for 60 min. Cells were then pelleted followed by resuspension in extracellular medium (without bovine serum albumin) at a concentration of 2 ϫ 10 6 cells/ml for fluorescence measurements. The ratio of fluorescence at 510 nm after excitation at 340 and 380 nm was monitored using a Polarstar Optima fluorescent plate reader. Single agonist treatments were KLK4 (300 nM), trypsin (10 nM), thrombin (10 nM), PAR-1 AP and PAR-2 AP (100 M), and PAR-4 AP (500 M). Displayed data are representative of experiments performed in triplicate and repeated on three independent occasions. Dose-response experiments were performed over the KLK4 concentration range of 0.1-1000 nM in triplicate and performed on three independent occasions with results displayed as means Ϯ S.E.
Knockdown of PAR-2 Expression-The mammalian siRNA expression vector pSilencer 3.1-H1 puro (Ambion, Austin, TX) was used to reduce expression of PAR-2. Candidate PAR-2 siRNA target sequences were designed as previously described (28) and then aligned against the human genome data base using the BLAST algorithm to eliminate those with significant homology to other genes. Three sequences selected were 5Ј-GATCCAGGAAGAAGCCTTATTGGTTTCAAGAGAACC-AATAAGGCTTCTTCCTTTTTTTGGAAA-3Ј, 5Ј-GATCC-AGTAGACTTGGTGTGAAGATTCAAGAGATCTTCACA-CCAAGTCTACTTTTTTTGGAAA-3Ј, and 5Ј-GATCCGTA-GTCGTGAATCTTGTTCATTCAAGAGATGAACAAGAT-TCACGACTATTTTTTGGAAA-3Ј. The sequences were synthesized (Sigma) and inserted into the pSilencer 3.1-H1 puro vector (Ambion) according to the instructions of the manufacturer. Fibroblasts from Par-1 null mice stably expressing PAR-2 (PAR-2-LMF) were transfected with the PAR-2 pSilencer 3.1-H1 constructs or the supplied pSilencer 3.1-H1 negative control using Lipofectamine. After 48 h, 2 g/ml puromycin was added to the medium to select for stable transfected cells. PAR-2 expression levels and agonist-induced induction of ERK phosphorylation were examined by Western blot analysis.
Cell Surface Biotinylation-PC-3 cells were washed three times with cold PBS, and then plasma membrane proteins were biotinylated by incubation with 1.22 mg/ml EZ-link NHS-SS-Biotin at 4°C for 1 h with gentle agitation. The cells were then washed in PBS prior to lysis in a buffer containing 20 mM HEPES, 150 mM NaCl, 1 mM EDTA, and 1% Triton X-100. After centrifugation, bead immobilized streptavidin was added into the supernatant and incubated for 15 min on ice to capture biotinylated proteins. The streptavidin beads were pelleted by centrifugation, and the supernatant was recovered for analysis of cytoplasmic (nonbiotinylated) proteins. The beads were then washed thoroughly, and associated cell surface (biotinylated) proteins were eluted into Laemmli sample buffer. Cytoplasmic and cell surface fractions were examined for PAR-2 by Western blot analysis.
Western Blot Analysis-Whole cell lysates were collected in a buffer containing Triton X-100 (1%, v/v), 50 mM Tris-HCl (pH 7.4), NaCl (150 mM), and protease inhibitor mixture (Roche Applied Science). Fractions containing membrane proteins and soluble proteins were isolated as previously described (31). Protein concentrations were determined by microbicinchoninic acid assay (Pierce). For ERK1/2 phosphorylation, cells were grown to 50% confluence and serum-deprived for 24 h before treatment, and then collection of whole cell lysates was as above, except the lysis buffer also contained 1 mM sodium orthovanadate and 50 mM NaF (Sigma). Equal amounts of lysates (20 g) or biotinylated fractions were separated by SDS-PAGE and transferred to a nitrocellulose membrane that was blocked with Odyssey blocking buffer (LI-COR, Lincoln, NE) containing 0.1% (v/v) Tween 20. Membranes were incubated overnight at 4°C with an anti-PAR-1 (1:1000), anti-PAR2 (1:1000), anti-GAPDH (1:1000), anti-ERK (1:1,000), or antiphospho-ERK (1:2,000) antibody and then washed before incubation with species-appropriate fluorescently conjugated secondary antibodies for 1 h at room temperature. Membranes were analyzed using an Odyssey Infrared Imaging System (LI-COR; Millennium Science, Surrey Hills, Australia) and where relevant signal intensity determined using LI-COR imaging software and exported to Microsoft Excel for graphical representation as mean Ϯ S.E. Significance was examined using Student's t test with a p value of Ͻ0.05 considered significant.
Flow Cytometry-Cells grown in serum-free media overnight were detached nonenzymatically, washed in PBS, and then resuspended at 2 ϫ l0 6 cells/ml. Cells were subjected to nonpermeabilizing fixation in 1% formaldehyde on ice for 5 min, washed with PBS, and then resuspended in staining buffer (PBS, pH 7.4, 0.2% bovine serum albumin). Following incubation with an anti-PAR-2 antibody (SAM11; 2 g/10 6 cells) on ice for 60 min, cells were washed and then incubated with fluorescently tagged anti-mouse secondary antibody. Cell surface PAR-2 was assessed using an FC500 flow cytometer (Beckman Coulter, Gladesville, Australia). In parallel, to assess the levels of total PAR-2, following fixation, cells were permeabilized by incubation with 0.01% Triton X-100 in PBS for 5 min at room temperature.
Confocal Microscopy-Cells plated on sterile poly-L-lysine (Sigma)-coated glass coverslips were allowed to adhere overnight then transfected with either the PAR-2-GFP expression construct or vector pEGFP-N1. For agonist treatments, cells were incubated with either KLK4 (100 nM) or PAR-2 AP (100 M) for 10 min at 37°C before fixation with 4% (v/v) formaldehyde. Nuclei were stained by incubating cells for 5 min at room temperature with 4Ј,6-diamidino-2-phenylindole (1:1500) in PBS. Coverslips were mounted on slides, and cells were imaged with a Leica SP5 confocal microscope (Leica Microsystems, Sydney, Australia). Images were processed using Adobe Photoshop CS3 and displayed using CorelDraw. The amount of PAR-2 on the cell surface was quantified by examining the fluorescence of randomly selected untreated and KLK4-treated cells (n ϭ 15) using ImageJ software (National Institutes of Health, Bethesda, MD), adapting the approach of Scherrer et al. (32) for quantifying GPCR cell surface expression. Briefly, three cellular regions of interest were defined: the whole cell, the intracellular region, and the nucleus. The signal obtained for the whole cell and the intracellular region was corrected for background signal by subtracting nuclear fluorescence. Cell surface fluorescence was then obtained by subtracting the corrected value for the intracellular region from the corrected value for the whole cell. These values were then divided by the number of pixels contained within each region to give fluorescence density values for the cell surface (D i surface) and cytoplasm (D i cytoplasm). The ratio of D i surface to D i cytoplasm was determined to normalize data across the counted cell population. Results are displayed graphically as mean Ϯ S.E., and significance was examined using Student's t test with a p value of Ͻ0.05 considered significant.
Immunohistochemical Analysis-Archival formalin-fixed paraffin-embedded blocks from primary prostate cancers (n ϭ 6; Gleason scores 3 ϩ 4 to 4 ϩ 5) and prostate cancer bone metastases (n ϭ 2) were obtained from Sullivan Nicolaides Pathology (Taringa, Australia) and the Royal Prince Alfred Hospital (Sydney, Australia), respectively, following institutional ethics approval. Immunohistochemistry was performed as previously described (33). Briefly, serial sections (4 m) were deparaffinized and then rehydrated, and after antigen retrieval in urea (5% w/v) in 0.1 M Tris buffer (pH 9.5), serial sections were incubated in H 2 O 2 (3%, v/v) to quench endogenous peroxidase. Sections were then blocked in normal goat serum (10%) and incubated overnight at 4°C with either an anti-PAR-2 monoclonal (SAM11; 1:1000 dilution) or an anti-KLK4 rabbit polyclonal (1:250 dilution) antibody. As negative controls, mouse and rabbit immunoglobulins replaced the SAM11 and anti-KLK4 primary antibodies, respectively. The EnVision ϩ peroxidase polymer detection system (Dako, Botany, Australia) was used with 3,3Ј-diaminobenzidine (Sigma) as the chromogen. The sections were counterstained with Mayer's hematoxylin, visualized by microscopy (Leitz, Laborlux S, Germany), and photographed using a Nikon OXM1200 digital camera. Images were processed using Adobe Photoshop CS3 and displayed using CorelDraw.

Examination of KLK4-initiated Changes in Intracellular
Ca 2ϩ Ion Concentration via PARs-For these studies, rKLK4 was purified from the conditioned media of Sf9 insect cells stably transfected with an expression construct containing the complete human KLK4 coding sequence in frame with V5 and polyhistidine-encoding sequences. Amino-terminal sequencing indicated that thermolysin-activated rKLK4 had the predicted NH 2 terminus of mature active KLK4 ( 31 IINED). The kinetic parameters obtained for thermolysin-activated rKLK4 against the tripeptide substrate benzyl-FVR-p-nitroanilide (K m ϭ 46.9 M, k cat ϭ 1.8 s Ϫ1 , and k cat /K m ϭ 38.3 s Ϫ1 ⅐mM Ϫ1 ) agreed well with the findings of Matsumura et al. (K m ϭ 48.3 M, k cat ϭ 1.28 s Ϫ1 , and k cat /K m ϭ 38.3 s Ϫ1 ⅐mM Ϫ1 ) using the same substrate and rKLK4 generated in Drosophila melanogaster S2 cells (34).
Signaling via PARs is initiated almost exclusively by arginine/ lysine-specific serine proteases and induces transient changes in intracellular Ca 2ϩ ion concentration (17). PAR activation leading to calcium mobilization requires proteolytic cleavage at defined activation sites; cleavage downstream of this site fails to mobilize Ca 2ϩ and results in receptor inactivation (18,20,21). Since the ability of the arginine/lysine-specific serine protease KLK4 to initiate intracellular signaling has not previously been examined, we studied the ability of this protease to induce changes in intracellular [Ca 2ϩ ] in PAR-1-LMF, PAR-2-LMF, or PAR-4-LMF cells (generated from mouse embryonic fibroblasts from Par-1 knock-out mice stably expressing either PAR-1, PAR-2, or PAR-4 (26)). Cells stably expressing PAR-3 are not available, and this receptor in mice does not mediate transmembrane signaling, leading to the proposal that PAR-3 acts as a co-receptor for signaling via PAR-4 (35). As shown in Fig. 1, A and B, activated rKLK4 initiated a prompt, transient Ca 2ϩ mobilization in PAR-1-and PAR-2-expressing cells. In these experiments, the maximum calcium signal initiated by rKLK4 was higher via cells expressing PAR2 than those expressing PAR1 (compare the peak heights in Fig. 1, A and B). The specificity of the response via PAR-1 and PAR-2 was indicated by the lack of Ca 2ϩ flux in cells expressing PAR-4 in response to activated rKLK4 (Fig. 1C). Positive control agonists for PAR-1 (thrombin and PAR-1 AP), PAR-2 (trypsin and PAR-2 AP), and PAR-4 (trypsin and PAR-4 AP) each induced changes in Ca 2ϩ concentration in the respective PAR-expressing cells (Fig. 1, D-I) (36 -39), confirming that the receptor expressed by each cell line was functional. In addition, PAR-1 AP did not signal via PAR-2 or PAR-4, PAR-2 AP did not signal via PAR-1 or PAR-4, and PAR-4 AP did not signal via PAR-1 or PAR-2 (data not shown), confirming that Ca 2ϩ mobilization in these cells was mediated only via the respective PAR. In other controls, pro-rKLK4 and phosphoramidon-inhibited thermolysin did not initiate PAR signaling, while PAR-4 cells were shown to have intact PAR following treatment with activated rKLK4 as the PAR-4 AP, applied 3 min after protease treatment, induced transient Ca 2ϩ mobilization (data not shown).
Concentration Dependence of KLK4-initiated Signaling via PAR-1 and PAR-2-The concentration-effect curves for KLK4mediated activation of PAR-1 and PAR-2 Ca 2ϩ signaling are shown in Fig. 2. PAR-1-LMF and PAR-2-LMF cells were incubated with increasing concentrations of enzyme over the range 0.1 nM to 1 M. As shown in Fig. 2, A and B, increasing concentrations of activated rKLK4 stimulated increased Ca 2ϩ mobilization via both PARs. The maximum response via PAR-1 was achieved at ϳ300 nM activated rKLK4 with a concentration to stimulate half-maximal response (EC 50 ) of ϳ30 nM. In contrast, a maximum dose response via PAR-2 was not achieved at the maximum concentration of rKLK4 currently available to us. However, using the dose response achieved at 1 M of rKLK4 as the lower limit for the maximum response, we estimated that the lowest value for the EC 50 via PAR-2 was ϳ300 nM, which suggests that KLK4 has higher potency (inverse of EC 50 ) for PAR-1 than PAR-2.
To compare the efficacy of KLK4-initiated Ca 2ϩ mobilization via PAR-1-and PAR-2-expressing cells, fluorescence values obtained at each concentration of rKLK4 were divided by the highest response observed for each of these PARs, which was via PAR-2 at 1 M rKLK4. As shown in Fig. 2C, this analysis suggests that KLK4 has greater efficacy for initiating Ca 2ϩ signaling via PAR-2 than PAR-1.
KLK4 Rapidly Cleaves a Peptide Spanning the PAR-2 Activation Site-Since concentration-effect curves indicated that KLK4 initiated Ca 2ϩ mobilization displayed higher efficacy via PAR-2 than via PAR-1 (Fig. 2C), KLK4 activation of PAR-2 was analyzed further. A fluorescence-quenched peptide spanning the PAR-2 activation site (ortho-aminobenzoic acid-SKGR2SLIGK-N-(2,4-dintrophenyl)ethylenediamine), which is activated by low concentrations of trypsin, tryptase, and acrosin but not by high concentrations of thrombin (40), was used to examine the kinetics of direct PAR-2 cleavage by activated rKLK4 (40 nM). The activated protease was incubated with varying concentrations of substrate, and fluorescence was monitored over time with initial velocities fitted to the Michaelis-Menten equation to determine kinetic parameters. As shown in Fig. 3, KLK4 was able to efficiently cleave the PAR-2 peptide with kinetic parameters of K m 8.0, k cat ϭ 7.9 s Ϫ1 , and k cat /K m ϭ 1.0 s Ϫ1 ⅐M Ϫ1 , which are similar to the values observed for KLK6 against a similar PAR-2 peptide (ortho-aminobenzoic acid-SSKGR2SLIGQ-N-2,4-dintrophenyl)ethylenediamine) of K m ϭ 5.6 M, k cat ϭ 7.9 s Ϫ1 , and k cat /K m ϭ 1.4 s Ϫ1 ⅐M Ϫ1 (23).
KLK4 Initiates Phosphorylation of ERK via PAR-2-The ability of KLK4 to signal via PAR-2 was further examined by analyzing ERK1/2 activation, another intracellular pathway downstream of PAR activation (17). We performed Western blot analysis to detect phosphorylated ERK1/2 in PAR-2-LMF cells stimulated with activated rKLK4 (10 nM) using total ERK1/2 as a control for relative expression of the kinase. Trypsin and the PAR-2 AP were used as positive controls for ERK1/2 activation  via PAR-2. As shown in Fig. 4A, KLK4 induced ERK1/2 phosphorylation within 5 min, with reduced levels of activation apparent at 15 min. PAR-2 AP and trypsin induced ERK1/2 phosphorylation in a similar time-dependent manner (Fig. 4, B and C, respectively). The dose response of KLK4-initiated signaling via PAR-2 was examined by incubating PAR-2-LMF cells with increasing amounts of enzyme (0.1-300 nM) and examining changes in ERK1/2 activation. Consistent with Ca 2ϩ mobilization assays, as shown in Fig. 4D, increasing concentrations of activated rKLK4 stimulated increased ERK1/2 activation.
We also examined the specificity of KLK4 signaling via PAR-2 by using siRNA to reduce the expression levels of this receptor in PAR-2-LMF cells. We generated three PAR-2 siRNA constructs and stably transfected these into mouse PAR-2-LMF cells. As shown in Fig. 5A (left) by Western blot analysis, it was apparent that construct 2 effectively reduced PAR-2 levels. Densitometric analysis indicated that this construct reduced PAR-2 levels by up to ϳ80% (mean ϳ60%), whereas siRNA constructs 1 and 3 failed to consistently reduce PAR-2 expression levels (Fig. 5A, right). The effect of PAR-2 knockdown on KLK4-initiated intracellular signaling was examined by assessing levels of activated and total ERK1/2 in lysates from controls and cells stably transfected with siRNA construct 2. Densitometric analysis was used to determine a value for phosphorylated ERK at each time point relative to untreated cells and a value for total ERK at each time point relative to untreated cells; the ratio of these values is the -fold increase of ERK activation over basal levels (41). When treated with KLK4, cells transfected with construct 2 showed ϳ70% reduction in ERK1/2 phosphorylation after 5 min relative to cells transfected   with a nonspecific control (Fig. 5B). A similar reduction in ERK1/2 phosphorylation after 5 min was also apparent in cells with reduced PAR-2 levels that had been treated with PAR-2 AP (Fig. 5C). Consistent with results of Ca 2ϩ flux assays, these data indicate that KLK4 initiates intracellular signaling via PAR-2.
PAR-1 and PAR-2 in Prostate-derived Cell Lines-We examined PAR-1 and PAR-2 expression in five prostate cell lines derived from a range of pathologies, by reverse transcription-PCR and Western blot analysis using PAR-1-LMF and PAR-2-LMF cells as positive controls for PAR-1 and PAR-2 mRNA and protein. The cell lines were derived from the following sources: RWPE1, virus-transformed normal epithelium (42); RWPE2, Ki-ras-transformed RWPE1 cells (42); LNCaP, lymph node metastasis (43); PC-3, bone metastasis (44); and DU145, brain metastasis (45). As shown in Fig. 6A, PAR-1 and PAR-2 mRNA was detected in each of these cell lines, as indicated by the presence of a specific 125-and 583-bp product, respectively. Similarly, PAR-1 and PAR-2 protein was detected at high levels in each of the analyzed cell lines (Fig. 6, B and C). Consistent with a previous report (46), PAR-1 was predominantly expressed as an ϳ65-kDa protein in each cell line, which reduced to ϳ40 kDa upon treatment with N-glycosidase F (data not shown). Higher PAR-1 molecular weight bands present in lysates from RWPE-1, RWPE-2, PC-3, and D145 cells are thought to be multimeric forms of the protein. In each prostate cell line, PAR-2 was apparent as bands at ϳ55, ϳ65, ϳ85, and ϳ105 kDa. Consistently, PAR-2 bands at ϳ65, ϳ85, and ϳ105 kDa reduced to the ϳ55 kDa band on tunicamycin treatment of breast cancer-derived cell lines, whereas the ϳ85 kDa was also phosphorylated in these cells (47). Interestingly, for PAR-2 only the 55 kDa band was apparent in PAR-2-LMF cells, potentially indicating differential post-translational modification (e.g. glycosylation) of PAR-2 as previously proposed (31).
KLK4 Initiated Signaling in a Prostate-derived Cell Line-PC-3 cells were selected for analysis of KLK4-initiated signaling in a prostate-derived cell line. For serine protease-mediated PAR activation to occur, the receptor must be located on the cell surface (36). PC-3 cell surface expression of PAR-1 has been previously reported (48). We assessed plasma membrane localization of PAR-2 in these cells by cell surface biotinylation and flow cytometry approaches.
In biotinylation experiments, plasma membrane proteins were isolated by treating intact PC-3 cells with a biotinylating agent, and biotinylated (cell surface) proteins were purified using streptavidin-conjugated beads. As shown in Fig. 7A by Western blot analysis using a PAR-2-specific monoclonal antibody, PAR-2 was detected in both the cytoplasm and plasma membrane fractions. Reprobing this membrane with an anti-GAPDH antibody indicated that the plasma membrane fraction purified by biotinylation was largely uncontaminated by intracellular proteins. Consistent data were obtained from our flow cytometry analyses. In these experiments, cell surface-localized PAR-2 was assessed by examining PC-3 cells, which had been subjected to nonpermeabilizing fixation. In parallel, total PAR-2 was determined by examining PC-3 cells permeabilized with the detergent Triton X-100. As shown in Fig. 7B, the anti-PAR-2 antibody detected both plasma membrane-localized PAR-2 (left) and total PAR-2 (right) well above background levels of fluorescence obtained when the cells were stained only with the secondary antibody.
The ability of KLK4 to initiate signaling in these cells was assessed by examining transient changes in Ca 2ϩ ion mobilization. PAR-1 AP and PAR-2 AP were used as positive controls for signaling via the respective PARs. KLK4 initiated a prompt, transient Ca 2ϩ mobilization in these cells as did PAR-1 AP and PAR-2 AP (Fig. 7C). The dose response of KLK4-initiated Ca 2ϩ mobilization was examined by incubating PC-3 cells with increasing concentrations of enzyme (0.1-1000 nM). As shown in Fig. 7D, increasing concentrations of KLK4 stimulated increased Ca 2ϩ mobilization consistent with levels observed from PAR-2-LMF cells (Fig. 2B). These data indicate that PAR-2 is expressed in a range of prostate cell lines, that PC-3 cells express functional PAR-1 and PAR-2, and that KLK4 initiates Ca 2ϩ mobilization in these cells, consistent with signaling via PAR-1 and PAR-2.
KLK4 Induces Loss of PAR-2 from the Surface of Prostate Cancer-derived PC-3 Cells-It is known that following activation, PAR-2 is processed by receptor uncoupling from G proteins and endocytosis followed by ubiquitination-mediated lysosomal degradation (49 -52). We employed confocal microscopy to quantitatively examine internalization of PAR-2 from the cell surface of KLK4-stimulated PC-3 cells. Cells transiently expressing PAR-2-GFP were used to facilitate localization of the receptor. As shown in Fig. 8A, in unstimulated PC-3 cells, PAR-2 was detected at the plasma membrane (arrowheads) and in cytoplasmic vesicles, including within structures previously identified as the Golgi apparatus (50,53). Stimulation of PAR-2-GFP-expressing cells with activated rKLK4 (100 nM) for 10 min resulted in internalization of PAR-2 from the cell surface and endocytosis (Fig. 8A, arrow) (54). In controls, PAR-2 AP treatment also caused PAR-2 internalization and redistribution throughout PC-3 cells in a manner similar to KLK4 (data not shown). As described under "Experimental Procedures," quantitative analysis of these data employing a modification of a previously described approach (32) was used to determine the amount of PAR-2 on the cell surface before and after KLK4 stimulation. Cellular regions of interest were defined for the whole cell, the intracellular region, and the nucleus, (marked on the example unstimulated cell shown in Fig. 8B (top) as red, yellow, and blue, respectively (middle)). As shown in Fig. 8B (bottom), graphical representation of the values obtained from this analysis indicated that PAR-2 cell surface levels decreased by ϳ80% following stimulation with KLK4. These data indicate that KLK4 induces loss of PAR-2 from the cell surface and receptor internalization.
KLK4 and PAR-2 Are Co-expressed in Primary and Bone Metastatic Prostate Cancer-Since we demonstrated that the prostate cancer-associated protease KLK4 initiates intracellular signaling via PAR-2, efficiently cleaves a peptide spanning the PAR-2 activation site, and induces PAR-2 cellular processing, we next examined the potential pathological relevance of these observations by performing immunohistochemical analysis of six primary prostate cancer tissue specimens. The expression pattern of receptor and agonist in consecutive tissue sections was examined using a PAR-2-specific monoclonal antibody and our previously described anti-KLK4 antibody (55). Representative images are shown in Fig. 9. PAR-2 and KLK4 had similar expression patterns, with both expressed by glandular epithelial cells with little evidence of stromal staining (Fig. 9, panels A and B and panels D and E, respectively). PAR-2 was detected in benign glands (BNG) of benign prostatic hyperplasia (BPH), prostatic intraepithelial neoplasia (PIN), and cancer (Ca). Expression levels were higher in regions of PIN and Ca than in BNG regions of BPH (comparing BNG and PIN in Fig. 9A and BNG and Ca in Fig. 9B). A similar staining pattern was apparent for KLK4. As published previously (13), prostate cancer tissue samples showed little KLK4 staining in benign glands, with stronger staining in regions of PIN (comparing BNG and PIN in Fig. 9D) and Ca (comparing BNG and Ca in Fig. 9E). Negative controls were free of staining (Fig. 9, D and H).
Bone metastasis is the cause of significant morbidity and mortality in prostate cancer patients (56). Recently, we have shown that KLK4 is expressed by both prostate cancer cells and osteoblasts in the in vivo metastatic bone environment and that osteoblast-like SaOs2 cells induce KLK4 expression in co-culture systems with prostate cancer-derived LNCaP and PC-3 cells (15). Accordingly, we performed immunohistochemistry to examine the expression pattern of KLK4 and PAR-2 in consecutive sections from prostate cancer bone metastases from two patients. As shown in Fig. 10, receptor and agonist had significant overlap in expression in prostate cancer bone metastasis. In regions of prostate cancer lesions, PAR2 and KLK4 were highly expressed by prostate cancer cells (Fig. 10, A and C,  respectively). In regions containing prostate cancer lesions and bone, strong staining of osteoblasts lining the bone surface was also apparent for both PAR2 and KLK4 (arrows in Fig. 10, B and  D, respectively). Negative controls were free of staining (data not shown).

DISCUSSION
Intracellular signaling, initiated via activation of members of the protease-activated receptor family of GPCRs, by the action of trypsin-like serine proteases is important in physiology and Fractions were analyzed for PAR-2 using antibody SAM11 and for purity using an antibody against a cytoplasmic marker (GAPDH). B, flow cytometry analysis for PAR-2, cell surface (fixed, nonpermeabilized cells; left) and total (fixed, permeabilized cells; right) PAR-2. Cells were incubated with either an anti-PAR-2 SAM11 antibody followed by a fluorescently tagged secondary antibody (black line) or a fluorescently tagged secondary antibody only (gray line). C, KLK4 (left), PAR-1 AP (middle), and PAR-2 AP (right) initiated changes in intracellular Ca 2ϩ ion concentration in the prostate cancer-derived PC-3 cell line. Cells were incubated with either KLK4 (300 nM) or activating peptide (100 M) as indicated, and fluorescence at 510 nm was measured following alternating excitation at 340 and 380 nm (Em510 (340/380) ) using a fluorescent plate reader. The ratio of Em510 (340/380) is proportional to intracellular Ca 2ϩ ion concentration. Experiments were performed in triplicate and repeated three times. D, concentration dependence of KLK4-induced changes in intracellular Ca 2ϩ ion concentration in prostate-derived PC-3 cells. Cells were incubated with KLK4 (0.1 nM to 1 M), and fluorescence was measured as described in D. Experiments were performed in triplicate and repeated three times.
Our data indicate that KLK4 induces Ca 2ϩ mobilization via PAR-1 and PAR-2 but not via PAR-4. This ability to signal via more than one PAR is known for other serine proteases. These include thrombin, which signals via PAR-1 and PAR-4 and also cleaves PAR-3 (16), trypsin, which activates PAR-2 (37, 53) and PAR-4 (39), activated Factor X, which, in complex with activated Factor VII and tissue factor, signals via PAR-1 and PAR-2 (59,60), and KLK14, which activates both PAR-2 and PAR-4 and inactivates PAR-1 (22). Interestingly, our analysis of Ca 2ϩ mobilization in response to KLK4 suggests that although signaling via PAR-1 is more potent than via PAR-2, KLK4 displayed greater efficacy for signaling via the latter PAR. This contrasts with trypsin IV, which had similar potency as well as efficacy for inducing Ca 2ϩ mobilization via these PARs in receptor-overexpressing cells (61). It is possible that the maximum signaling via PAR-1 observed by us occurred at lower KLK4 concentration either because of lower levels of expression of this receptor or because at higher enzyme concentration, KLK4 cleaves PAR-1 other than at the activation site, thereby inactivating or "disarming" the receptor. Although these proposals remain to be tested, the latter has been suggested to explain the observations that Ca 2ϩ mobilization via PAR-2 in response to KLK6 reached a maximum at a much lower concentration than KLK14, and this maximum response to KLK6 was much lower than observed for KLK14 (22). It was proposed that the responses to KLK6 were due to a receptor activating ability at lower enzyme concentrations via cleavage of PAR-2 at the consensus activation site and to a "disarming" ability at higher concentration via cleavage downstream of the activation site. Thus, it is possible that KLK4 regulates cellular activity by differential activation of PAR-1 and PAR-2, at lower enzyme concentration signaling via both PARs and as KLK4 concentration increases via PAR-2 as PAR-1 is disarmed. Alternatively, it is possible that efficacy was lower via PAR-1, because this receptor is expressed at levels lower than PAR-2 on the surface of the cells used in this study.
Due to the ability of KLK4 to signal via PAR-2 over a broader enzyme concentration range, we analyzed signaling via this receptor-agonist system in greater detail and also examined the expression of these proteins during prostate cancer progression. Our enzymatic studies, using a fluorescence-quenched peptide indicated that KLK4 efficiently cleaves a peptide spanning the PAR-2 activation site, suggesting that intracellular signaling mediated by KLK4 occurs via direct proteolytic activation of this receptor. The specificity of KLK4 signaling via PAR-2 was assessed by Western blot analysis of ERK1/2 activation. We observed that KLK4-induced phosphorylation of ERK1/2 could be substantially reduced by siRNAmediated knock-down of PAR-2 levels. Our confocal microscopy analyses indicated that KLK4-induced intracellular signaling via PAR-2 is accompanied by loss of this GPCR from the cell surface and receptor internalization.
Our observation of increased PAR-2 expression during prostate cancer progression is consistent with a recent report showing elevated PAR-2 in regions of prostate cancer compared with adjacent normal glandular epithelial cells in 42% of patient samples (n ϭ 40) (62). These workers also showed that PAR-2 levels increased with increasing Gleason score (62). In addition, our finding of PAR-2 expression by osteoblasts in prostate cancer bone metastases is consistent with studies in mice demonstrating that murine osteoblasts also express this receptor in vitro and in vivo (63). The key observation from our immunohistochemical analysis of PAR-2 and KLK4 co-expression in primary prostate cancer as well as bone metastasis lesions suggests that KLK4 will contribute to PAR-2 activation in these settings. This is potentially significant, since recent reports have demonstrated cancer-associated functional consequences of PAR-2 activation in prostate cancer-derived cell lines. For example, in LNCaP cells, PAR-2 AP induced activation of members of the Rho GTPase family (62,64), proteins that are of critical importance in cytoskeletal reorganization and cancer (65). In addition, AP mediated activation of PAR-2-induced activity of the collagenases matrix metalloprotease 2 and 9 by the prostate cancer cell lines LNCaP, PC-3, and DU145 (66). Further, it has been suggested that activation of PAR-2 in bone mediates inhibition of osteoclast differentiation and bone lytic activity, and this effect is probably mediated by osteoblasts (67). This is potentially significant, since prostate adenocarcinomas have primarily an osteoblastic phenotype in bone metastasis lesions (68). Similarly, cancer-associated functional consequences of KLK4 expression have been noted in prostate cancer-derived cell lines. For example, KLK4 overexpression induced an increased ability of PC-3 cells to migrate, accompanied by a transition from an epithelial morphology to a fibroblastic shape and, consistently, a significant decrease in E-cadherin protein levels and an increase in vimentin (11). In addition, inducible overexpression of KLK4 resulted in significantly increased colony formation, migration, and proliferation of both PC-3 and DU145 prostate cancer-derived cells (12). Recently, we have shown that KLK4 levels increased in LNCaP and PC-3 cells co-cultured with osteoblast-like SaOs2 cells, whereas KLK4overexpressing PC-3 cells had increased migration toward SaOs2 cell conditioned media (15). These reports are suggestive that KLK4 activation of PAR-2 will be important in prostate cancer progression in both primary and bone lesions.
It is possible that KLK4-initiated signaling via PAR-1 will also be relevant in prostate cancer. Recently, two reports have consistently identified an increase in expression at mRNA and protein levels of this receptor during prostate cancer progression, although the expression pattern reported by these groups differed (62,69). Kaushal et al. (69) noted predominant expression in endothelial cells in cancerous regions with cancer cells also showing PAR-1 expression in some early and advanced stage specimens. In contrast, using a different antibody, Black et al. (62) reported PAR-1 expression in cancer cells in 45% of prostate cancer samples and in periglandular stromal cells in 55% of higher grade cancers. Thus, although the identity of the cell types expressing PAR-1 in cancerous prostate is contentious, loss of prostate organ structure during cancer progression will expose PAR-1 to KLK4. In addition, of relevance, PAR-1 expression has also been reported in human osteoblast-like cells (70) and mouse osteoblasts (71), and this GPCR mediates thrombin-induced proliferation of primary rat osteoblasts (72). The action of KLK4 on PAR-1 and PAR-2 may be particularly relevant in more advanced primary prostate cancer and bone metastatic lesions, since the level of Zn 2ϩ , an efficient inhibitor of KLK4 proteolytic activity (IC 50 ϭ 16 M (58)) in normal prostate, is known to drop 10 -20-fold in prostatic tissue and fluid from prostate cancer patients (73,74).
Of importance to the regulation of PAR-2 activation in normal and diseased prostate, in addition to KLK4, three other prostate-expressed serine proteases can initiate intracellular  signaling via PAR-2 in in vitro systems. These include the type II transmembrane serine proteases (75) TMPRSS2 (76), and matriptase/MT-SP1 (77) and another KLK serine protease, KLK14 (24). The ability of four prostatic serine proteases to signal via PAR-2 indicates that activation of this receptor in prostate will require both tight regulation of the activating proteases (for example by the presence of activators and inhibitors) and spatial and temporal expression of receptor and agonist. In this respect, although protein or mRNA expression has been reported for TMPRSS2, matriptase/MT-SP1 and KLK14 by normal and cancerous prostate epithelial cells (13, 78 -80), coexpression of these enzymes with PAR-2 in normal or diseased prostate, or bone metastases has not yet been addressed. Another aspect of regulation of PAR-2 signaling in prostate is the possibility that these serine proteases will most efficiently activate PAR-2 via a proteolytic cascade reminiscent of the blood coagulation (81) and wound healing (82) serine protease cascades.
PAR-1 and PAR-2 constitute a growing number of KLK4 substrates, including pro-prostate-specific antigen (4), fibrinogen (6), pro-urokinase-type plasminogen activator (4), and the urokinase-type plasminogen activator receptor (7) as well as type I and type IV collagen at lower efficiency (6). In this study, we observed that PAR-2-mediated activation of ERK1/2 was initiated at a lower KLK4 concentration than Ca 2ϩ mobilization (compare Figs. 2B and 4D). Of relevance, ERK1/2 activation and Ca 2ϩ mobilization can both be initiated via disparate events at the cell surface, including GPCR and receptor tyrosine kinase ligand docking (83)(84)(85). Thus, it is possible that KLK4 is able to cleave/activate other cell surface proteins in addition to PAR-1 and PAR-2 to initiate the differential induction of Ca 2ϩ ion flux and ERK1/2 phosphorylation observed by us.
This study has shown for the first time that the receptors PAR-1 and PAR-2 are activated by the serine protease KLK4 and that PAR-2 and KLK4 are co-expressed during prostate cancer progression. In addition to a putative role in this disease, KLK4 is thought to be critical in dental enamel formation (86). Also, cell culture approaches (9) and immunohistochemical analysis (10) indicate potential involvement in ovarian cancer. Thus, KLK4-mediated cell signaling via PARs may be important in prostate cancer and other pathophysiological processes.