PAR2-induced inflammatory responses in human kidney tubular epithelial cells

PAR2-induced inflammatory responses in human kidney tubular epithelial cells

David A. Vesey,1,2 Jacky Y. Suen,3 Vernon Seow,3 Rink-Jan Lohman,3 Ligong Liu,3 Glenda C. Gobe,1

David W. Johnson,1,2 and David P. Fairlie3

1Centre for Kidney Disease Research, The University of Queensland Department of Medicine at the Princess AlexandraHospital, Queensland, Australia; 2Department of Nephrology, Princess Alexandra Hospital, Queensland, Australia;and 3Institute for Molecular Bioscience, The University of Queensland, Brisbane, Queensland, Australia

Submitted 1 October 2012; accepted in final form 30 December 2012

Vesey DA, Suen JY, Seow V, Lohman RJ, Liu L, Gobe GC,Johnson DW, Fairlie DP. PAR2-induced inflammatory responsesin human kidney tubular epithelial cells. Am J Physiol RenalPhysiol 304: F737–F750, 2013. First published January 2, 2013;doi:10.1152/ajprenal.00540.2012.—Protease-activated receptor-2(PAR2) is a G protein-coupled receptor abundantly expressed in thekidney. The aim of this study was to profile inflammatory gene andprotein expression induced by PAR2 activation in human kidneytubular epithelial cells (HTEC). A novel PAR2 antagonist, GB88, wasused to confirm agonist specificity. Intracellular Ca2! (iCa2!) mobi-lization, confocal microscopy, gene expression profiling, qRTPCR,and protein expression were used to characterize PAR2 activation.PAR2 induced a pronounced increase in iCa2! concentration that wasblocked by the PAR2 antagonist. Treatment with SLIGKV-NH2 at theapical or basolateral cell surface for 5 h induced expression of a rangeof inflammatory genes by greater than fourfold, including IL-1",TRAF1, IL-6, and MMP-1, as assessed by cDNA microarray andqRTPCR analysis. Using antibody arrays, GM-CSF, ICAM-1,TNF-#, MMP-1, and MMP-10 were among the induced proteinssecreted. Cytokine-specific ELISAs identified three- to sixfold in-creases in GM-CSF, IL-6, IL-8, and TNF-#, which were blocked byGB88 and protein kinase C inhibitors. Treatment of cells at thebasolateral surface induced more potent inflammatory responses, withrelease of MCP-1 and fibronectin to the apical and basolateral com-partments; apical treatment only increased secretion of these factors tothe apical compartment. PAR2 activation at the basolateral surfacedramatically reduced transepithelial electrical resistance (TEER)whereas apical treatment had no effect. There was very little leakage($5%) of peptides across the cell monolayer (liquid chromatography-mass spectrometry). In summary, SLIGKV-NH2 induced robust pro-inflammatory responses in HTEC that were antagonized by GB88.These results suggest that PAR2 antagonists could be useful disease-modifying, anti-inflammatory agents in kidney disease.

protease-activated receptor-2; kidney tubule; inflammation; cytokine

INFLAMMATION IS A COMMON FEATURE in all forms of kidney disease,irrespective of the mechanism of initiation. It is an importantprotective response that drives wound healing and repair pro-cesses. As in other tissues, however, protracted or uncontrolledinflammation in the kidney is detrimental, promoting tissue de-struction and fibrosis, which, over time, can lead to organ failure.Mitigating ongoing inflammatory responses would be of thera-peutic value. Serine proteases, such as thrombin, tryptase, andtrypsin, can cause inflammation by disrupting tissue architectureor by activating certain cell surface receptors, including protease-activated receptors (PARs) (31, 39). The present study investi-

gates the role of the second member of this class of receptor,PAR2, in inflammatory responses in kidney cells.

Protease-activated receptor 2 (PAR2) is a particular class AG protein-coupled receptor that is activated mainly by trypsin-like proteases including trypsin, mast cell tryptase, and coag-ulation factors Xa and VIIa (3). Proteolytic cleavage of theNH2-terminal extracellular domain of PAR2 exposes a newNH2 terminus, referred to as a tethered ligand, which triggersreceptor activation. Short synthetic peptides, corresponding tothe human tethered ligand sequence, Ser-Leu-Ile-Gly-Lys-Val-NH2 (SLIGKV-NH2), can also activate PAR2 in the absence ofproteolysis albeit at micromolar instead of nanomolar concen-trations. Such hexapeptides have been widely used experimen-tally as exogenous agonists to tease out roles for PAR2 inphysiological and pathophysiological conditions (3, 32).

PAR2 is expressed in a wide range of human tissues, but itis especially prominent in epithelial cells at the interfacebetween the external environment and internal milieu, such asin the gastrointestinal tract, respiratory system, and kidneytubules (6, 15, 21). Although the precise functions of PAR2 inthese tissues are unclear, there is evidence for PAR2 regulationof inflammation, cell proliferation, protease sensing, and epi-thelial barrier function (8, 22, 23, 27, 50). In models of colitis,lung disease, and glomerulonephritis, PAR2 knockout mice(PAR2%/%) are protected to some extent from disease progres-sion (19, 34, 42). Direct infusion of PAR2-activating peptidesor serine proteases into the lung or colon leads to increasedinflammatory cytokine production and increased epithelial per-meability (7–9, 29, 30, 34, 42). Our recent studies havedemonstrated that a novel human PAR2 antagonist, GB88, canameliorate inflammatory disease progression in rat models ofpaw edema, colitis, and arthritis (29, 30, 43).

Within the human kidney, PAR2 is especially prominent inthe proximal tubule cells of the renal cortex and renal vascu-lature (21, 48, 49). Studies have shown PAR2 involvement inthe control of renal blood flow, ion transport, inflammation,and fibrosis (5, 17, 18, 21). Primary cultures of these corticaltubular epithelial cells express high levels of functional PAR2(49). Activation with the PAR2 agonist, SLIGKV-NH2, elicitsa rapid rise in intracellular calcium and subsequent productionof the proinflammatory mediator monocyte chemoattractantprotein-1 (MCP-1). In the present study, we sought to inves-tigate in more detail the proinflammatory responses in humanprimary kidney tubular epithelial cells (HTEC) to activation byPAR2 activating peptides, SLIGKV-NH2 and 2f-LIGRLO-NH2, using calcium mobilization assays, microarray analysis,antibody arrays, specific cytokine ELISAs, and with the aid ofa new PAR2 antagonist, GB88, the only compound currentlyknown at low micromolar concentrations to inhibit the actionsof all known PAR2 agonists.

Address for reprint requests and other correspondence: D. A. Vesey, Dept.of Nephrology, Level 2, Ambulatory Renal and Transplant Services Bldg.,Princess Alexandra Hospital, Ipswich Rd., Woolloongabba, Brisbane, Queens-land (e-mail: [email protected]).

Am J Physiol Renal Physiol 304: F737–F750, 2013.First published January 2, 2013; doi:10.1152/ajprenal.00540.2012.

1931-857X/13 Copyright © 2013 the American Physiological Societyhttp://www.ajprenal.org F737

mailto:[email protected]

MATERIALS AND METHODS

Tubule cell isolation and cell culture. Segments of macroscopicallyand histologically normal renal cortex (&10 g) were obtained asepti-cally from the noncancerous pole of adult human kidneys removedsurgically because of small renal clear cell carcinomas1. Patients wereotherwise healthy. Informed consent was obtained prior to eachoperative procedure and the use of human renal tissue for primary cellculture was reviewed and approved by the Princess Alexandra Hos-pital Research Ethics Committee, Brisbane, Australia. The method forisolation and culture of human kidney tubular epithelial cells isdescribed in detail elsewhere (49). Briefly, the cortical tissue wasminced finely, washed several times, and agitated for 20 min at 37°Cin Hanks’ Balanced Salt Solution (HBSS) containing collagenase typeII (1 mg/ml) and calcium. Cold HBSS was added and the solutioncontaining tubular fragments passed through a 100-'m sieve. Afterwashing three times, the tubular fragments were resuspended in 45%Percoll in phosphate-buffered saline (PBS) and centrifuged at 20,000 g. Ahigh-density band, previously shown to be tubule fragments, wasremoved and cultured in a serum free, hormonally defined DMEM/F12 medium containing 10 ng/ml epidermal growth factor, 5 'g/mlinsulin, 5 'g/ml transferrin, 50 nM hydrocortisone, 50 'M prosta-glandin E1, 50 nM selenium, 5 pM triiodothyronine, penicillin (50U/ml), and streptomycin (50 'g/ml). Cells were routinely cultured inthis medium.

Peptides, enzymes, and chemicals. The PAR2 activating peptide,SLIGKV-NH2, and a nonactivating peptide, reverse sequence peptide,VKGILS-NH2, were synthesized as carboxyl-terminal amides andpurified to (95% via reversed-phased high-performance liquid chro-matography by either Auspep (Melbourne, Australia) or the Divisionof Chemistry and Structural Biology at the Institute for Molecular

Bioscience, The University of Queensland. The PAR2 antagonistGB88, PAR2 peptide agonist 2f-LIGRLO-NH2, and synthetic PAR2agonist were synthesized as previously described (4, 43). Two com-mercially available protein kinase C (PKC) inhibitors bisindolylma-leimide 1 (1 'M) and Gö6983 (5 'M) [Merck (Victoria, Australia)]were used to assess the role PKC in PAR2-induced inflammatorycytokine production.

Cell treatments. All experiments were performed on confluentpassage 2 HTEC in 96- (black walled), 48-, or 12-well plates (Corn-ing, NY). Cells on transwell filters were cultured in 12-well plates.Before experimentation, cells were made quiescent by two washesfollowed by incubation for 24 h in basic media (DMEM/F12 mediumwith antibiotics). Effects of the PAR2 activating peptide, SLIGKV-NH2,2 or control peptide, VKGILS-NH2, on cytokine production weremeasured by cDNA microarray analysis, inflammation antibody ar-rays, and cytokine-specific ELISAs (MCP-1, GM-CSF, IL-6, IL-8,TNF-#, and fibronectin). Analysis was at 5 h or 24 h posttreatment. Insome experiments, treatment was either to the apical or basolateralmonolayer surface. 2f-LIGRLO-NH2 (a more potent PAR2 agonistthan SLIGKV-NH2), IL-1", and transforming growth factor-1"(TGF-1") were used in some experiments for comparison and con-firmation. Media conditioned by HTEC were harvested and stored at%80°C until assayed.

Quantitative RT-PCR. Cells were grown to confluence and totalRNA isolated using a RNeasy Mini Kit (Qiagen) according to man-ufacturer’s instructions. RNA was reverse-transcribed using Super-script III (Invitrogen) and an oligo(dT) primer. Relative gene expres-sion was quantified by real-time PCR (qRTPCR) using SYBR GreenPCR master mix (Applied Biosystems, Foster City, CA) on an

1 While the tissue used in this study was taken as “normal,” it is possible thatit has been influenced at a molecular level by the adjacent cancerous tissue orhave some characteristic of precancerous cells, despite appearing histologicallynormal.

2 In initial experiments only SLIGKV-NH2 was available for use. In laterexperiments we had the option to use 2f-LIGRLO-NH2, which is more potentand apparently more selective. In all experiments SLIGKV-NH2 and 2f-LIGRLO-NH2 produced similar responses, but the latter was about 50-foldmore potent.

Fig. 1. Functional expression of protease-activated receptor 2 (PAR2) by human kid-ney tubular epithelial cells (HTEC). Cellswere grown to confluence as described inMATERIALS AND METHODS. A: PAR2 mRNAmeasurements by qRTPCR in HT29, HTEC,and HEK293 cells (n ) 7). B: maximalintracellular calcium (iCa2!) mobilization inHTEC, HT29, and HEK293 cells in responseto 2f-LIGRLO-NH2 (2 'M). C: concentra-tion-dependent iCa2! release by HTECtreated with PAR2 agonists trypsin (red),GB110 (black), or 2f-LIGRLO-NH2 (blue;n ) 7). D: antagonism of PAR2-inducediCa2! release by GB88. Trypsin (10 nM),GB110 (1 'M), 2f-LIGRLO-NH2 (1 'M)(n ) 7). Data points are means * SE fromtriplicate measurements. iCa2! release is ex-pressed as a percentage of the maximal re-sponse elicited by the Ca ionophoreA-23187. In C and D each trace is an aver-age response from cells in 8 different wells.Error bars represent the SE. *P $ 0.05(HTEC vs. HT29 and HEK293).

F738 PAR2 AND INFLAMMATION IN KIDNEY TUBULE CELLS

AJP-Renal Physiol • doi:10.1152/ajprenal.00540.2012 • www.ajprenal.org

Applied Biosystems Prism 7000 sequence detector. Amplificationcycle proceeded as follows: 50°C for 2 min and 95°C for 10 min,follows by 40 cycles of 95°C for 15 s and 50°C for 1 min. cDNAlevels at the linear phase of amplification were compared with hypo-xanthine guanine phosphoribosyl transferase (HPRT) levels, and ex-pressed as a relative expression of HPRT. The sequence of the primersused in this study were HPRT, forward primer 5=-TCA GGC AGTATA ATC CAA AGA TGG T-3= and reverse primer 5=-AGT CTGGCT TAT ACT CAA CAC TTC G-3=; PAR2, forward primer

5=-GGG TTT GCC AAG TAA CGG C-3= and reverse primer 5=-GGGAAC CAG ATG ACA GAG AGG-3=; IL-6 forward primer 5=-GCCCAC CGG GAA CGA AAG AGA-3= and reverse primer 5=-GACCGA AGG CGC TTG TGG AGA AG-3=; IL-8, forward primer5=-ACC ACC GGA AGG AAC CAT CTC ACT-3= and reverseprimer 5=-CTT GGC AAA ACT GCA CCT TCA CAC-3=; CCL2,forward primer 5=-AGC TCG CAC TCT CGC CTC CAG-3= andreverse primer 5=-GGC ATT GAT TGC ATC TGG CTG AGC-3=; F3,forward primer 5=-CCG GCG CTT CAG GCA CTA CAA A-3= and

Fig. 2. PAR2 activation on the surface ofHTEC. A: inhibition of SLIGKV-NH2 in-duced iCa2! release by 1 and 10 'M GB88.Data points are means * SE from triplicatemeasurements. *P $ 0.005 vs. control, #P $0.005 vs. SLIGKV-NH2 alone. B and C: con-focal imaging of tight junction protein 1(ZO-1) and PAR2 in HTEC grown on poly-ester transwell membranes (maximum inten-sity projection with a +63 objective). Mouseanti-ZO-1 with Texas Red-labeled second-ary antibody and rabbit anti-PAR2 with Al-exa fluor 488-labeled secondary antibody. Atleft the PAR2 antibody was preincubatedwith the control peptide antigen (1 'g/'l)prior to the primary antibody incubation.The ZO-1 antibody was omitted. C: an or-thogonal projection of confocal z stacks.

F739PAR2 AND INFLAMMATION IN KIDNEY TUBULE CELLS


reverse primer 5=-CGG GTT TGG GTT CCC ACT CCA AAA T-3=;IL-11, forward primer 5=-AGC TGA CGG GGA CCA CAA CCT-3=and reverse primer 5=-GTC AGC ACA CCT GGG AGC TGT AG-3=;CSF2, forward primer 5=-CCG GCG TCT CCT GAA CCT GAG T-3=and reverse primer 5=-GGT CGG CTC CTG GAG GTC AAA C-3=;MMP-1, forward primer 5=-CCT AGC TGG GAT ATT GGA GCAGCA-3= and reverse primer 5=-TCC GCT TTT CAA CTT GCC TTTGTC T-3=; TNF-#, forward primer 5=-CCC AGG GAC CTC TCTCTA ATC-3, and reverse primer 5=-ATG GGC TAC AGG CTT GTCACT-3=; IL-1", forward primer 5=-CCT CTT CGA GGC ACA AGGCAC AA-3= and reverse primer 5=-TGGCTGCTTCAGACACTT-GAGCAAT-3=; and TRAF 1, forward primer 5=-GCA GTG CTGCCC AGG ATC ACA-3= and reverse primer 5=-ACA GAC TGTGGG CTT CCC TTG A-3=.

Calcium mobilization assay. Cells were grown to confluence in96-well, black-walled, clear-bottomed plates and then made quiescentby two washes followed by incubation for 24 h in basic media. On theday of experimentation, the medium was removed and cells wereincubated in a dye loading buffer [HBSS with 4 'M Fluo-3, 25 'lpluronic acid, 1% fetal bovine serum (FBS), and 2.5 mM probenecid]for 1 h at 37°C. Cells were washed twice with HBSS and transferredto a Polarstar spectrofluorimeter (BMG, Durham, NC) for agonistinjection and fluorescence measurements. Various concentrations ofSLIGKV-NH2 were added 10 s after reading commenced, and fluo-rescence was measured in real time from the bottom of the plate usingexcitation at , ) 480 nm and emission at , ) 520 nm. HBSS wasprepared in-house while all other reagents were purchased from

Invitrogen (Carlsbad, CA). Plates were purchased from DKSH (Zu-rich, Switzerland). Calcimycin (A23187, Invitrogen) was used tomeasure maximum fluorescence, with individual results normalizedaccordingly.

Measurement of transepithelial resistance. Transepithelial electri-cal resistance (TEER) was measured using a Millicell-ERS epithelialvoltohm-meter (Millipore) fitted with chopstick electrodes. The elec-trical resistance was recorded as the mean of three consecutivemeasurements. All values were corrected for background resistance.

Immunofluorescence staining. Cells were grown on transwell mem-brane inserts (Corning), fixed with 4% paraformaldehyde for 15 min,and stored in phosphate-buffered saline at 4°C prior to staining. Cellswere permeabilized with 0.2% Triton X-100, aldehyde groupsquenched with 50 mM ammonium chloride, and then blocked with 1%bovine serum albumin in PBS. The primary antibodies used were arabbit anti-PAR2 (APR-032, Alomone Labs) and mouse anti-zonaoccludins (ZO)-1 (61–7300, Invitrogen). They were applied overnightat 4°C (1/500 dilution). For negative controls either the primaryantibody was omitted (ZO-1) or a blocking peptide was used topreabsorb the antibody (PAR2). Following washing with PBS, thesecondary antibody, (goat anti-rabbit/mouse Alexa Fluor 488/594,Invitrogen; or goat anti-mouse Texas red), together with 4=,6=-di-amidino-2-phenylindole (DAPI) nuclear stain, was applied overnightat 4°C. The membranes were excised and mounted onto glass slides inglycerol and capped with a glass coverslip.

Secreted proteins. A human inflammatory antibody array (cat. no.AAH-INF-3, RayBiotech, Norcross, GA) was used to measure cyto-

Table 1. Expression of genes in response to basolateral and apical treatment of human tubular epithelial cells.

Fold Expression Over VKGILS-NH2 Treatment

Gene Name Gene Symbol Basolateral Apical

Tumor necrosis factor TNF 3.1 1.9Interleukin 6 IL6 4.0 3.5Heparin-binding EGF-like growth factor HBEGF 4.1 2.6Fatty acid 2-hydroxylase FA2H 4.2 1.7Interleukin 1 receptor, type II IL1R2 4.2 1.4ST3 beta-galactoside alpha-2,3-sialyltransferase 1 ST3GAL1 4.4 2.0Chemokine (C-C motif) ligand 20 CCL20 4.4 5.3SHC (Src homology 2 domain containing) family, member 4 SHC4 4.4 2.9Galanin prepropeptide GAL 4.6 1.4Transmembrane prostate androgen-induced protein TMEPAI 3.3 1.7High mobility group AT-hook 2 HMGA2 4.6 1.8Serpin peptidase inhibitor, clade B (ovalbumin), member 9 SERPINB9 4.9 1.9Dual specificity protein phosphatase 5 DUSP5 4.9 2.6Sodium channel, nonvoltage-gated 1, delta SCNN1D 5.2 3.4Frizzled homolog 10 (Drosophila) FZD10 5.3 1.5Interleukin-1 receptor-associated kinase-like 2 IRAK2 5.5 2.8Colony stimulating factor 2 (granulocyte-macrophage) CSF2 5.6 6.3Angiopoietin-like 4. ANGPTL4 5.6 2.0Matrix metalloproteinase 10 MMP10 5.7 7.0Tissue factor F3 5.9 2.4Neuron navigator 3 NAV3 5.9 2.7Epiregulin EREG 5.9 2.9Interleukin 23, alpha subunit p19 IL23A 5.9 3.4Rho GTPase activating protein 22 ARHGAP22 6.3 2.0Bone morphogenic protein 2 BMP2 6.5 2.3A kinase (PRKA) anchor protein 12 AKAP12 6.6 2.7Interleukin 8 IL8 7.2 5.0Intercellular adhesion molecule 1 ICAM1 7.4 4.1Serine proteinase 22 PRSS22 9.1 3.4Tumor necrosis factor, alpha-induced protein 3 TNFAIP3 9.9 3.9G0/G1 switch 2 G0S2 10.2 5.4Pentraxin 3 PTX3 10.3 2.8Interleukin 1 Alpha IL1A 12.2 9.9Matrix metalloproteinase 1 MMP1 13.4 9.2Interleukin 11 IL11 17.8 7.2Interleukin 1 Beta IL1B 19.9 11.4TNF receptor-associated factor 1 TRAF1 24.6 7.6



kines secreted by HTEC treated with SLIGKV-NH2. In some exper-iments, cells were treated apically or basolaterally with SLIGKV-NH2

(100 'M) with or without GB88 for 24 h, and conditioned mediumwas collected from the apical and basolateral compartments and

stored at %80°C until assayed. Conditioned cell medium (0.5–1 ml)was used in each array. GM-CSF, MCP-1, IL-8, IL-6, and tumornecrosis factor-# (TNF-#) were measured using a commercial ELISAkit (RnD Systems, Minneapolis, MN) according to the manufacturer’s

Fig. 3. qRT-PCR analysis of SLIGKV-NH2-induced inflammatory gene expression. HTECwere grown to confluence [transepithelial electri-cal resistance (TEER) ( 600 - · cm2] on transwellinserts and then treated with or without SLIGKV-NH2 (100 'M) from the apical or basolateralaspect for 5 h. RNA was extracted and analyzed byqRT-PCR using specific primers for IL-8, IL-1",IL-6, TNF#, CCL2, CSF2, IL-11, tissue factor F3,MMP-1, and TNF receptor-associated factor-1(TRAF-1). Data represent the average of 3 inde-pendent experiments using cells from 3 separatedonors. Bars: white, no treatment (control); black,basolateral (Baso) treatment; gray, apical treat-ment. **P $ 0.05.



instructions. Matrix metalloproteinase (MMP) secretion by cells wasalso measured using a MMP antibody array (AAH-MMP-1, RayBio-tech, Norcross, GA) and fibronectin concentrations were measuredusing a specific fibronectin ELISA (Dako, Noble Park, Australia) aspreviously described (49).

cDNA microarray experiments. Microarray analysis was performedon RNA extracted from primary cultures of HTEC (grown on trans-well inserts), following treatment for 5 h with SLIGKV-NH2 (100'M) or control peptide VKGILS- NH2 (100 'M) at either the apicalor basolateral surface. Total RNA was extracted using a QiagenRNeasy mini kit (Qiagen, Hilden, Germany), and its concentration,purity, and integrity measured using a NanoDrop spectrophotometerand Agilent 2100 Bioanalyzer (Agilent Laboratories, Palo Alto, CA).Biotinylated cRNA was prepared from 200 ng of total RNA usingAmbion Illumina RNA amplification kit and hybridized to a SentrixHuman-6 Expression BeadChips ((46,000 gene transcript targets,

Illumina, San Diego, CA) using the manufacturer’s hybridizationsolution. All reagents and protocols for washing, detection, andscanning were used according to the BeadStation 5003 system proto-cols (Illumina, San Diego, CA).

Genes were considered differentially expressed between agonist-treated, (apical and basolateral), and control peptide treated if theydisplayed a minimum fourfold change in expression. Additionally, toavoid genes with close to background levels of expression, only a rawsignal of (15 times the control sample was considered.

Liquid chromatography (LC)-mass spectrometry (MS). After 24-htreatment, medium was collected from both the basolateral and apicalcompartments of the HTEC monolayers and stored at %80°C untilanalyzed. For analysis, 100-'l samples were diluted 1:2 in acetoni-trile, vortexed, sonicated, and centrifuged (13 K rpm, 5 min). Standardcurves of 2f-LIGLRO-NH2 and SLIGKV were constructed using thesame method, where stock concentrations were diluted 1:2 in aceto-

Fig. 4. Human inflammation antibody array of inflam-matory factors secreted by HTEC treated for 24 h withPAR2 agonist SLIGKV-NH2. HTEC were grown toconfluence in serum-free defined medium, washed twicewith buffer, and transferred to basic medium for 24 h,before treatment with PAR2 agonist SLIGKV-NH2 (100'M). Medium was analyzed in an inflammation anti-body. The experiment was repeated 3 times with cellsisolated from 3 different patients. A: representativearray. B: densitometric analysis of array. Arrows indi-cate proteins, CMCSF and TNF-#, which showed thegreatest changes in expression following SLIGKV-NH2

treatment.



nitrile (final concentration 0.12, 1.2, and 12 'M). All samples wereanalyzed by LCMS/MS (ABSCIEX 4000 QTRAP Triple Quadrupole,Linear Ion-Trap LC/MS/MS mass spectrometer). Chromatographywas carried out on a C18 column (Phenomenex, 5 'm, 2.1 + 50 mm)using a linear gradient (5–80% buffer B in 8 min, flow rate 0.35ml/min). Buffer A was 0.1% formic acid (aq), and buffer B was 90/10acetonitrile/0.1% formic acid (aq). Retention times were 6.39 and 6.89min for SLIGKV and 2-furoyl-LIGRLO, respectively, in positive-ionmode.

Statistical analysis. All studies were performed in triplicate fromHTEC cultures obtained from at least three separate human donorsunless otherwise indicated. Each experiment contained internal con-trols originating from the same culture preparation. For the purposesof analysis, each experimental result was expressed as a change fromthe control value, which was regarded as 100%, and analyzed inde-pendently. Results are expressed as means * SE. Statistical compar-isons between two groups were made using unpaired t-tests. Multiplegroup comparisons were made by ANOVA. GraphPad Prism version5.03 was used to construct graphs and statistical analysis. P values $0.05 were considered significant. Antibody array densitometric anal-ysis was made using ImageJ 1.45s.

RESULTS

HTEC express high levels of PAR2 mRNA and functionalPAR2 protein at their cell surface. We have previously mea-sured PAR2 mRNA expression in 22 established human celllines by qRTPCR and found that HT29 colon carcinoma cellsand HEK293 human embroyonic kidney cells expressed rela-tively high levels of PAR2 transcript (43). Here we examinedkidney HTECs established from normal adult human kidneycortex, which had been cultured under serum-free conditions(passage 2). PAR2 mRNA was threefold more highly ex-pressed than in HT29 or HEK293 cells (Fig. 1A, n ) 7). Anintracellular calcium (iCa2!) mobilization assay was used tomeasure iCa2! flux in these three cell types in response to thePAR2 agonist, 2f-LIGRLO-NH2, to gauge whether the cells

differentially expressed PAR2 functional protein on the cellsurface. The signal intensity for each of the three cell types wasvery similar (Fig. 1B), suggesting comparable cell surfaceexpression of PAR2 despite different intracellular levels ofPAR2 mRNA. The PAR2 peptide agonist 2f-LIGRLO-NH2

(EC50 1.5 * 0.5 'M) and the synthetic PAR2 nonpeptideagonist GB110 (EC50 1.7 * 0.8 'M) induced a rapid concen-tration-dependent increase in intracellular calcium (Fig. 1C).Trypsin-induced responses were approximately two log unitsmore potent than for these synthetic agonists in this particularcell type (EC50 30 * 23 nM, Fig. 1C). The PAR2 antagonistGB88 effectively inhibited iCa2! release induced by trypsin,SLIGKV-NH2, GB110, and 2f-LIGRLO-NH2 (Fig. 1D andFig. 2A) with a similar IC50 (10 'M) in each case in this celltype.

HTEC form monolayers that express PAR2. HTEC grown onpolyester transwell membrane inserts formed tight cobblestonemonolayers that expressed tight junction protein, ZO-1, anddeveloped TEER. Recent studies (27) have indicated that thereis differential signaling via PAR2 at apical and basolateral cellmembranes, so we wished to investigate possible differentialinflammatory responses to PAR2 activation at these two dif-ferent cell surfaces. Experiments using cells grown in thisformat were performed once cells reached a TEER of (600- ·cm2. Confocal microscopy revealed that whereas ZO-1 wasexclusively localized to the cell membrane, prominent punctatePAR2 staining was present in the cytoplasm and to a lesserextent at the cell membrane (Fig. 2B). By construction of Zstacks, PAR2 appeared to be present both apically and baso-laterally with no preferential polarized localization (Fig. 2C).

Inflammatory gene expression induced by SLIGKV-NH2. Toprofile inflammatory gene expression induced in HTEC by PAR2activation, a cDNA microarray screening approach was initiallyused. Cells were treated with SLIGKV-NH2 (100 'M) on their

Fig. 5. Secretion of cytokines by HTECtreated with SLIGKV-NH2 (SLI) and antago-nism by the PAR2 antagonist, GB88. A: TNF-#.B: GMCSF. C: IL-6. D: IL-8. HTEC weregrown to confluence in a defined medium,followed by 24 h in basic medium and thenwith basic medium containing SLIGKV-NH2

(200 'M) with or without GB88 (10 'M). Theexperiment was repeated 3 times using cellstaken from different patients. The results fromcells of one patient are shown. ***P $ 0.05.



apical or basolateral surface for 5 h, and RNA was extracted,amplified, labeled, and examined using a Sentrix Human-6 Ex-pression BeadChip. For this study, only selected genes thatshowed an increased transcript expression in response to treatmentwere reported. Genes were only considered to exhibit differentialexpression when the raw signal for treated cells was 15 timesgreater than background and exhibited a !4-fold change inexpression compared with control cells treated with the non-PAR2-activating control peptide VKGILS-NH2.

A total of 36 genes showed increased expression when cellswere treated from the basolateral side, whereas only 11 wereinduced !4-fold when treated from the apical side. None of thegenes expressed by apical treatment was different from thoseexpressed in response to basolateral treatment (Table 1). The dataindicated that for many genes there was greater induction whentreatment was from the basolateral surface. The raw signal wasgreatest for the genes IL-6, IL-8, and CCL-20.

When cells were treated with SLIGKV-NH2 at the basolateralsurface the expression of at least 10 genes associated with proin-flammatory signaling were significantly induced. This includedgenes in the TNF gene family [TNF receptor-associated factor 1(TRAF1) and TNF alpha-induced protein 3 (TNFAIP3)], the ILgene family (IL-8, IL-6, IL-1#, IL-1", IL-11 and IL-23A), and

IL-1 receptor associated protein genes, namely IL-1 receptor-associated kinase 2 (IRAK2) and IL-1 type 2 receptor (IL-IR2).Other similarly upregulated inflammatory genes included thechemokine CCL20, tissue factor (F3), colony stimulating factor 2(CSF2), pentraxin 3 (PTX3), and intercellular adhesion molecule1 (ICAM-1). The cyclooxygenase-2 (COX-2) prostaglandin syn-thase gene PTGS2 was not upregulated following PAR2 activa-tion, as has been reported recently for HEK293 cells and othercells (16, 33, 44). A number of growth factors, including platelet-derived growth factor subunit B (PDGFB), heparin binding epi-dermal growth factor (HBEGF), epiregulin (EREG), bone mor-phogenic protein 2 (BMP2), leukemia inhibitory factor (LIF), andangiopoietin-like-4b, (ANGPTL4), were upregulated. Proteasegenes, including MMP-1 and -10, serine protease 22, PRSS22,and protease inhibitor SERPINB9, were induced. Genes for otherMMPs were not induced by SLIGKV-NH2 at 5 h.

RTPCR analysis of upregulated inflammatory genes. UsingqRTPCR the expression of 10 of the key inflammatory genesupregulated in the microarray experiment was more closely ex-amined (Fig. 3). All showed a significant upregulation of mRNAin response to SLIGKV-NH2 treatment (3- to 26-fold), which wasin a similar induction range to that observed in the microarrayexperiment. However, whereas the microarray analysis showed

Fig. 6. Human MMP antibody array of conditioned me-dium from HTEC treated for 24 h with or without PAR2agonists SLIGKV-NH2. HTEC were treated with PAR2agonist SLIGKV-NH2 (100 'M) for 24 h and MMPsecretion into the conditioned medium measured using anMMP antibody array. The experiment was repeated 3times with cells isolated from 3 different patients. A: rep-resentative array. B: densitometric analysis.



significantly more induction following basolateral than apicaltreatment, the qRTPCR analysis showed that only genes forTNF-# and TRAF1 were significantly different. The gene encod-ing MCP-1, CCL2, was included in the qRTPCR analysis as it hadbeen previously shown, at the protein level, to be induced bySLIGKV-NH2 (49).

SLIGKV-NH2 induces secretion of inflammatory cytokines.A membrane-based human inflammation antibody array was usedto determine whether any of the inflammatory genes highlightedin the microarray and qRTPCR experiments were translated intosecreted proteins. A previous study had demonstrated thatSLIGKV-NH2 markedly induces secretion of MCP-1 by proximaltubule cell cultures (49). Figure 4 shows enhanced secretion of arange of inflammatory proteins in response to SLIGKV-NH2, themost prominent differentially secreted cytokines being ICAM-1,GM-CSF, and TNF-#. Other secreted cytokines and chemokinesdetected included IL-6, IL-8, MCP-1, regulated on activation,normal T cell expressed and secreted (RANTES), tissue inhibitorof metalloproteinase 2 (TIMP-2), and macrophage inflammatoryprotein-1 " (MIP-1"), but using this antibody array, there did notappear to be enhanced secretion by treated cells. The mRNAlevels of most of these proteins increased within 5 h (Table 1).Although the expression of IL-1# and IL-1" mRNA was mark-edly induced by SLIGKV-NH2 (Table 1), no secreted IL-1 wasdetectable using this antibody array (positions 5 and 6, Fig. 4). Ofthe three CSF proteins present on the array, CSF1, -2, and -3, atpositions 1, 2, and 11, only CSF2 (GM-CSF) appeared to besignificantly induced by SLIGKV-NH2.

To confirm these results, the cytokines were measured usingspecific ELISAs. IL-6, IL-8, GM-CSF, and TNF-# were allsignificantly enhanced by SLIGKV-NH2 (Fig. 5, A–D). Usingthe samples in the array described above, the ELISA indicateda (10-fold difference in the levels IL-6 and IL-8 betweencontrols and treatment. Release of each of the four cytokinesmeasured was antagonized by the PAR2 antagonist, GB88(Fig. 5, A–D). As previously described, SLIGKV-NH2 inducedsecretion of significant amounts of fibronectin and MCP-1,which were also antagonized by GB88 (not shown).

Induction of matrix metalloproteinases (MMP) by SLIGKV-NH2.A membrane-based human MMP antibody array analysis ofmedium from HTEC treated with SLIGKV-NH2 for 24 h dem-onstrated that secretion of MMP-1 and -10, but not MMP-2, -3, -8,-9, and -13, was upregulated by 4 and (60-fold, respectively(Fig. 6). This correlated with observations from the microarrayand qRTPCR experiments. Tissue inhibitors of metalloproteinases(TIMPs)-1 and -3 were strongly expressed by HTEC, but it isunclear from this array whether levels were affected by treatment.

Secretion of TNF-" and GM-CSF by HTEC is PKC dependent.To investigate whether PAR2-induced cytokine secretion in-volved signaling via PKC, HTEC cells were grown to confluencein 48-well tissue culture plates and then pretreated for 1 h withPKC inhibitors bisindolylmaleimide 1 (1 'M) or Gö6983 (5 'M)prior to 24 h exposure to 2f-LIGRLO-NH2. Conditioned mediumwas taken for analysis of GM-CSF and TNF-#. Both of theseinhibitors almost completely blocked the secretion of these cyto-kines ((95% for TNF-# and (90% for GM-CSF, Fig. 7). Theconcentrations of inhibitors used were similar to those usedpreviously and were nontoxic (MTT assay and lactate dehydro-genase assay; data not shown) (28).

Differential secretion of cytokines in response to apical orbasolateral treatment. Using cells grown to confluence on trans-well inserts, the secretion of MCP-1 and fibronectin was exam-ined in response to treatment with SLIGKV-NH2 at the basolat-eral or apical compartments. These responses were compared withthose elicited by TGF-1" (5 ng/ml), a known inducer of fibronec-tin production by HTEC, and by IL-1", a known inducer of bothfibronectin and MCP-1 (11, 47, 49). Under control conditions,significantly more fibronectin and MCP-1 were secreted into theapical compartment (&3:1). As expected, treatment of the cellswith SLIGKV-NH2 from the basolateral or apical surface in-creased secretion of both fibronectin and MCP-1. However,whereas apical treatment with SLIGKV-NH2 only increased se-cretion of these molecules from the apical surface, basolateraltreatment appeared to increase secretion from both membraneaspects. TGF-1" only increased secretion of fibronectin from thesurface to which it was added. IL-1" (1 ng/ml), on the other hand,appeared to induce secretion into both compartments irrespectiveof the treatment surface (Fig. 8).

SLIGKV-NH2 reduces transepithelial resistance (TEER) onlywhen cells are treated from the basolateral surface. BasolateralPAR2 activation using SLIGKV-NH2 and 2f-LIGRLO-NH2

has previously been shown to compromise the integrity of lungand intestinal epithelial cell monolayers (27, 50). AlteredHTEC monolayer integrity was thus considered as a possibleexplanation for enhanced levels of fibronectin and MCP-1 in

Fig. 7. PAR2-induced TNF# and GMCSF secretion is antagonized by GB88and broad-spectrum protein kinase C (PKC) inhibitors. HTEC grown toconfluence on plastic were treated with or without GB88 (10 'M) or PKCinhibitors [bisindolylmaleimide-1 (Bis-1), 1 'M; Go-6983, 5 'M], for 1 hprior to treatment with PAR2 agonist 2f-LIGRLO-NH2 (1 'M). Cell condi-tioned media was taken at 24 h for TNF# and GMCSF. Values are means *SE (n ) 3). ***P $ 0.001, **P $ 0.005 vs. vehicle control.



the apical compartment following basolateral treatment. To testthis idea, the TEER across the HTEC monolayers was mea-sured at various times following treatment. Basolateral treat-ment of HTEC with SLIGKV-NH2 induced a concentration-and time-dependent loss of TEER. At 100 'M SLIGKV-NH2,TEER was decreased by (70% within 1 h, and it remained sofor the remainder of the experiment (24 h). At 24 h, 25 and12.5 'M SLIGKV-NH2 decreased TEER by approximately40% and 25%, respectively. When cells were treated with 100'M SLIGKV-NH2 from the apical surface there was no alter-ation in TEER (Fig. 9). IL-1" (1 ng/ml) added either to thebasolateral or apical compartment greatly reduced TEER by(70% within 1 h. TGF-" (1 ng/ml) added at the apical orbasolateral compartment failed to influence TEER. GB88 at 10'M failed to antagonize the SLIGKV-NH2-induced reductionin TEER. In fact GB88 alone reduced TEER by (50% over thecourse of the experiment. The PKC inhibitor Bis-1 failed toreduce the SLIGKV-NH2-induced reduction in TEER. Alone ithad no action on TEER.

PAR2 activating peptides fail to cross epithelial monolayers.One possibility to account for the enhanced secretion of MCP-1and fibronectin when cells were treated from the basolateralsurface is that the activating peptides are able to pass through theepithelial monolayer when the TEER is reduced and then alsoactivate apical PAR2. Using LC-MS we were unable to detectsignificant passage of the peptides ($5%) from one side of thecell layer in either direction even after 24 h of basolateral treat-ment (Fig. 10).

DISCUSSION

PAR2 was originally cloned in 1994 and was shown to bewidely expressed by human tissues, notably the skin, pancreas,liver, gut, and kidneys (6, 15, 35). Epithelial cells lining the

gut, where trypsin and other PAR2-activating proteases arepresent at relatively high concentrations, express especiallyhigh levels of PAR2 at both the apical and basolateral mem-branes. PAR2 here is known to have important roles in iontransport, barrier protection, and inflammatory responses (12,26, 30, 32, 36, 49, 50), but its physiological importance inkidney, where serine proteases are not normally present inappreciable concentrations, is unknown (48). This study hasbeen directed toward examining inflammatory responses elic-ited by PAR2 activation in tubular epithelial cells derived fromthe cortex of human kidneys.

HTEC cells express relatively high levels of PAR2 mRNA,threefold higher than in HEK293 and HT29 cells, which wereamong the highest PAR2 expressing cells in a recent study of 22human cell lines commonly used in PAR2 studies (32). Confocalmicroscopy demonstrated that PAR2 protein was located both atthe cell surface and within the cytoplasm, with no apparentpreference for apical or basolateral membrane expression. Thiswas consistent with comparable iCa2! release induced by PAR2agonist 2f-LIGRLO-NH2 in all three cell types, and indicates thatHTEC have significantly greater cytosolic reserves of PAR2mRNA on hand for translation to protein and rapid mobilization tothe cell surface as needed. This may be a consequence of a needfor the kidney to maintain high surveillance and sensitivity to thepresence and action of proteases that can potentially degradeproteins leading to tissue damage and renal dysfunction. In linewith this protective measure, a number of studies have shownupregulation of PAR2 by inflammatory mediators such as IL-1,TNF-#, and lipopolysaccharide (LPS) and, interestingly, alsoPAR2 agonists (38, 44). Although not reported here, because itwas below the fourfold induction cut-off, SLIGKV-NH2 induceda 1.5-fold increase in PAR2 mRNA expression.

Fig. 8. Basolateral and apical secretion of fibronectin andMCP-1 by HTEC treated with SLIGKV-NH2 (SL), trans-forming growth factor-" (TGF), or interleukin-1" (IL).Cells were grown to confluence on membrane transwells indefined DMEM/F12 medium and then for 24 h in basicmedium. The TEER of the cells before treatment was &600- · cm2. The cell monolayers were treated either from theapical or basolateral surface with the indicated factors:SLIGKV-NH2 (SL 100 'M), TGF-1" (TGF, 5 ng/ml), IL-1(IL, 1 ng/ml). A or B indicates the treatment surface: A )Apical, B ) Basolateral. MCP-1 levels in the medium fromthe apical and basolateral compartments were assayed usinga specific MCP-1 ELISA according to the manufacturer’sinstructions. Data presented are means * SE for 3 or 4independent experiments using cells from different donors.*P $ 0.05 vs. control.



HTEC respond to a number of functional modulators ofPAR2, such as the endogenous agonist trypsin, the synthetichexapeptide agonist 2f-LIGRLO-NH2, the synthetic nonpep-tide agonist GB110, and the nonpeptide antagonist GB88, in asimilar manner to many other cell types (43). However, slightlyhigher concentrations of these agonists seem to be required toelicit comparable agonist and antagonist effects in HTEC thanother cell types (43). The iCa2! response of HTEC to trypsinshowed considerable variability (EC50 30 * 23 nM, n ) 7), incontrast to that seen with the synthetic agonists where the EC50

was relatively tight (&1.6 * 0.7 'M, n ) 7). This variabilitymay relate to nonspecific receptor activation by trypsin con-tributing some Ca2! release, or to PAR2 polymorphisms, or todifferent glycosylated PAR2 isoforms, which can alter or maskthe protease cleavage/activation site (13, 14). The PAR2 an-tagonist used in these studies, GB88, was effective at blockingiCa mobilization induced by all three agonist types used inthese studies on HTEC cells.

Using cDNA microarray analysis and qRTPCR we profiledthe proinflammatory genes induced in HTEC within 5 h byactivation of PAR2 with the human PAR2 tethered ligandsequence SLIGKV-NH2, and also assessed protein expressionafter 24 h using cytokine protein arrays and individual ELISAs.Other studies have consistently shown that activation of PAR2

induces expression of proinflammatory proteins, such asCOX-2, TNF-#, IL-8, and IL-6 in a variety of cell typesincluding lung, intestinal, and gastric epithelia and bloodmonocytes (24, 37, 44). A recent report by Suen et al. com-pared the expression of genes induced by the PAR2 agonists,2f-LIGRLO-NH2 and trypsin, in the human embryonic kidneycell line HEK239 (44). Proinflammatory genes, COX-2, IL-8,and TNF-9, were among those induced by both agonists andcould be therefore attributed specifically to PAR2 induction. Inthe present study, gene profiling of HTEC exposed basolater-ally to SLIGKV-NH2 for 5 h induced 36 genes by fourfold ormore, many of which have previously been associated withproinflammatory responses. Treatment at the apical surfaceinduced a similar array of genes but the induction level wasmuch reduced. Ten of the genes examined by qRT-PCRsupported the finding of enhanced inflammatory gene expres-sion following PAR2 activation. However, only the genes forTRAF1 and TNF-# were more consistently, strongly, anddifferentially induced by treatment at different cell surfaces.These findings are consistent with other studies, where baso-lateral PAR2 activation elicited a more pronounced response(27, 46). Using antibody cytokine arrays it was shown that anumber of proinflammatory genes expressed after PAR2 acti-vation are then converted to protein and secreted. The most

Fig. 9. Basolateral but not apical treatment of HTEC withSLIGKV-NH2 causes loss of TEER (A) and is not blocked bya PAR2 antagonist, GB88 (B). Cells were grown to confluenceon membrane transwells. The TEER before treatment was&600 - · cm2. They were then treated either on the apical orbasolateral surface with SLIGKV-NH2 at the indicated con-centrations, and 24 h after treatment, the TEER was againmeasured. Data presented are means * SE for representativeexperiments (n ) 4). Measurements were made with a Milli-cell-ERS (Millipore). Each measurement was the mean of 3separate measurements. *P $ 0.05.



prominent of these proteins were the known proinflammatoryagents GM-CSF, TNF-#, MMP-1, and MMP-10. The produc-tion of IL-6, IL-8, GM-CSF, and TNF-# was suppressed byGB88, consistent with their PAR2-dependent expression.

Like other GPCRs, PAR2 activation is known to elicit biolog-ical responses via activation of various G proteins of which theGq/11 subunit is known to be central. Gq/11 activates phospholipaseC, which in turn cleaves phosphatidylinositol 4,5-bisphosphate(PIP2) to produce diacylglycerol (DAG) and inositol 1,4,5-trisphosphate (IP3). These mediators then lead to activation ofPKC and iCa2! release. To determine if PKC is involved inPAR2-mediated cytokine production, two commercial PKC in-hibitors were used to block PKC activity. Without causing toxic-

ity, the secretion of GM-CSF and TNF-# were almost completelyabolished, suggesting the importance of PKC in PAR2-mediatedGM-CSF and TNF-# secretion. Recent studies have also shownthe important involvement of PKC in PAR2-induced inflamma-tion (2, 10).

The reduction in TEER following basolateral, but not apicaltreatment of polarized HTEC with PAR2 agonists SLIGKV-NH2 and 2f-LIGRLO-NH2, may indicate loss of epithelialmonolayer integrity, such that proteins and peptides (includingthe PAR2 agonists) can pass from one side of the membrane tothe other. The apparent increase in fibronectin and MCP-1secretion from the apical cell surface following basolateraltreatment may also reflect this. We found no evidence thateither SLIGKV-NH2 or 2f-LIGRLO-NH2 could pass throughthe epithelial monolayer. The fact that GB88 failed to reducethe SLIGKV-NH2-induced reduction in TEER may be relatedto the fact that GB88 is a PAR2 pathway-specific antagonist(43). It specifically blocks Gq/11-mediated intracellular signal-ing pathways that lead to iCa2! release, but fails to antago-nize ERK1/2 activation. The inability of Bis-1 to block theSLIGKV-NH2 induced reduction in TEER suggests thatPAR2-induced cytokine production and loss of TEER aremediated by different PAR2 signaling pathways.

This finding may be important. Inflammatory responses inkidney tubules may be mediated by serine proteases arisingfrom sources within the organ parenchyma (i.e., the basolateralside), and not the lumen of the tubular organ (i.e., external tothe tissue). This would be consistent with chronic inflamma-tory conditions in the kidney being driven by endogenousproteases (such as tryptase produced by mast cells), and notthose present in lumen (apical portion) of the tubular tissue,such as those that could be produced by bacteria, viruses, orother pathogens. This finding may help guide drug treatmentregimes for kidney diseases, whereby drugs which targetPAR2 basolaterally may be more effective than those whichtarget apical PAR2. This viewpoint parallels current think-ing about PAR2 activation and PAR2 antagonist treatmentin the colon (27).

A large number of serine proteases ((30) have now beenshown to activate PAR2 in vitro (1). The clotting factors X andVII, when activated, are able to form PAR2-activating com-plexes with tissue factor (TF-FXa-FVIIa). Clotting disordersthat affect the kidney, such as thrombotic microangiopathy,could lead to basolateral PAR2 activation via this mechanism.Mast cells, which produce PAR2-activating proteases chymaseand tryptase, have been shown to accumulate in the diseasedkidney and promote inflammation and fibrosis and circulatingtryptase levels increase with renal dysfunction (20, 40, 41, 45).Trypsinogen has also been shown to be produced by kidneytubule cells (25). Its function in the kidney, however, isunclear. Its activation following injury or infection may welllead to PAR2 activation.

In summary, there is increasing recognition of the roles ofproteolytic enzymes in cell signaling and inflammation. A keyreceptor for their actions is PAR2, which is expressed on thesurface of a range of human cell types. This study shows thatPAR2 expression is considerably higher in HTEC than in manyof the cell types previously recognized as among the mostimportant in PAR2 research, raising questions about the rolesfor PAR2, and its activating proteases, in the kidney. Here wehave profiled effects of PAR2 agonists and antagonists in

Fig. 10. PAR2 activating peptides do not cross the HTEC monolayers evenwhen TEER is reduced. HTEC were grown to confluence on transwellmembranes and then treated either on the apical or basolateral surface withSLIGKV-NH2 (A) or 2f-LIGRLO-NH2 (2 'M) (B). The TEER before treat-ment was (600 - · cm2. At 24 h the medium was collected from the apical andbasolateral sides and stored at %80°C before analysis. The concentration ofpeptides present in the medium was determined by LCMS as described inMATERIALS AND METHODS. The concentration of peptide on each side of themonolayer is expressed as a percentage difference from the concentrationinitially added to cells.



HTEC and discovered that proinflammatory gene and proteinexpression is mediated through PAR2 and can be downregu-lated by a PAR2 antagonist, GB88. Moreover, treatment ofHTEC with SLIGKV-NH2 at the basolateral cell surface wasmore effective than at the apical cell surface, and induced morepotent responses in expression of inflammatory genes andinflammatory proteins. These responses were substantiallyblocked by the PAR2 antagonist GB88 and by commerciallyavailable PKC inhibitors. These findings indicate that PAR2 isan important mediator of inflammatory responses in the kid-ney, and suggests that PAR2 antagonists could have promisinganti-inflammatory activity in chronic kidney diseases.

ACKNOWLEDGMENTS

Some of the work included in this manuscript has been presented in abstractform at past meetings of the American Society of Nephrology.

GRANTS

D. P. Fairlie holds an Australian Research Council (ARC) FederationFellowship.

This work was partly funded by the National Health and Medical ResearchCouncil of Australia (Grant 569595).

DISCLOSURES

J. Y. Suen and D. P. Fairlie are inventors on a patent AU20109033378covering PAR2 agonists and antagonists that is owned by The University ofQueensland.

AUTHOR CONTRIBUTIONS

Author contributions: D.A.V., D.W.J., and D.P.F. conception and design ofresearch; D.A.V., J.Y.S., V.S., and L.L. performed experiments; D.A.V.,J.Y.S., V.S., and L.L. analyzed data; D.A.V., J.Y.S., R.-J.L., L.L., G.C.G.,D.W.J., and D.P.F. interpreted results of experiments; D.A.V. and J.Y.S.prepared figures; D.A.V., D.W.J., and D.P.F. drafted manuscript; D.A.V.,R.-J.L., G.C.G., and D.P.F. edited and revised manuscript; D.A.V., J.Y.S.,V.S., R.-J.L., L.L., G.C.G., D.W.J., and D.P.F. approved final version ofmanuscript.

REFERENCES

1. Adams MN, Ramachandran R, Yau MK, Suen JY, Fairlie DP, Hol-lenberg MD, Hooper JD. Structure, function and pathophysiology ofprotease activated receptors. Pharmacol Ther 130: 248–282, 2011.

2. Amadesi S, Nie J, Vergnolle N, Cottrell GS, Grady EF, Trevisani M,Manni C, Geppetti P, McRoberts JA, Ennes H, Davis JB, Mayer EA,Bunnett NW. Protease-activated receptor 2 sensitizes the capsaicin re-ceptor transient receptor potential vanilloid receptor 1 to induce hyperal-gesia. J Neurosci 24: 4300–4312, 2004.

3. Barry GD, Le GT, Fairlie DP. Agonists and antagonists of proteaseactivated receptors (PARs). Curr Med Chem 13: 243–265, 2006.

4. Barry GD, Suen JY, Le GT, Cotterell A, Reid RC, Fairlie DP. Novelagonists and antagonists for human protease activated receptor 2. J MedChem 53: 7428–7440, 2010.

5. Bertog M, Letz B, Kong W, Steinhoff M, Higgins MA, Bielfeld-Ackermann A, Fromter E, Bunnett NW, Korbmacher C. Basolateralproteinase-activated receptor (PAR-2) induces chloride secretion in M-1mouse renal cortical collecting duct cells. J Physiol 521: 3–17, 1999.

6. Bohm SK, Kong W, Bromme D, Smeekens SP, Anderson DC, Con-nolly A, Kahn M, Nelken NA, Coughlin SR, Payan DG, Bunnett NW.Molecular cloning, expression and potential functions of the humanproteinase-activated receptor-2. Biochem J 314: 1009–1016, 1996.

7. Boitano S, Flynn AN, Sherwood CL, Schulz SM, Hoffman J, Gruzi-nova I, Daines MO. Alternaria alternata serine proteases induce lunginflammation and airway epithelial cell activation via PAR2. Am J PhysiolLung Cell Mol Physiol 300: L605–L614, 2011.

8. Cenac N, Coelho AM, Nguyen C, Compton S, Andrade-Gordon P,MacNaughton WK, Wallace JL, Hollenberg MD, Bunnett NW, Garcia-Villar R, Bueno L, Vergnolle N. Induction of intestinal inflammation inmouse by activation of proteinase-activated receptor-2. Am J Pathol 161:1903–1915, 2002.

9. Cenac N, Garcia-Villar R, Ferrier L, Larauche M, Vergnolle N,Bunnett NW, Coelho AM, Fioramonti J, Bueno L. Proteinase-activatedreceptor-2-induced colonic inflammation in mice: possible involvement ofafferent neurons, nitric oxide, and paracellular permeability. J Immunol170: 4296–4300, 2003.

10. Chang W, Chen J, Schlueter CF, Hoyle GW. Common pathways foractivation of proinflammatory gene expression by G protein-coupledreceptors in primary lung epithelial and endothelial cells. Exp Lung Res35: 324–343, 2009.

11. Chikaraishi A, Hirahashi J, Takase O, Marumo T, Hishikawa K,Hayashi M, Saruta T. Tranilast inhibits interleukin-1beta-induced mono-cyte chemoattractant protein-1 expression in rat mesangial cells. Eur JPharmacol 427: 151–158, 2001.

12. Chin AC, Lee WY, Nusrat A, Vergnolle N, Parkos CA. Neutrophil-mediated activation of epithelial protease-activated receptors-1 and -2regulates barrier function and transepithelial migration. J Immunol 181:5702–5710, 2008.

13. Compton SJ, Cairns JA, Palmer KJ, Al Ani B, Hollenberg MD, WallsAF. A polymorphic protease-activated receptor 2 (PAR2) displayingreduced sensitivity to trypsin and differential responses to PAR agonists.J Biol Chem 275: 39207–39212, 2000.

14. Compton SJ, Sandhu S, Wijesuriya SJ, Hollenberg MD. Glycosylationof human proteinase-activated receptor-2 (hPAR2): role in cell surfaceexpression and signaling. Biochem J 368: 495–505, 2002.

15. D=Andrea MR, Derian CK, Santulli RJ, Andrade-Gordon P. Differ-ential expression of protease-activated receptors-1 and -2 in stromalfibroblasts of normal, benign, and malignant human tissues. Am J Pathol158: 2031–2041, 2001.

16. Frungieri MB, Weidinger S, Meineke V, Kohn FM, Mayerhofer A.Proliferative action of mast-cell tryptase is mediated by PAR2, COX2,prostaglandins, and PPARgamma: possible relevance to human fibroticdisorders. Proc Natl Acad Sci USA 99: 15072–15077, 2002.

17. Grandaliano G, Pontrelli P, Cerullo G, Monno R, Ranieri E, Ursi M,Loverre A, Gesualdo L, Schena FP. Protease-activated receptor-2 ex-pression in IgA nephropathy: a potential role in the pathogenesis ofinterstitial fibrosis. J Am Soc Nephrol 14: 2072–2083, 2003.

18. Gui Y, Loutzenhiser R, Hollenberg MD. Bidirectional regulation ofrenal hemodynamics by activation of PAR1 and PAR2 in isolated perfusedrat kidney. Am J Physiol Renal Physiol 285: F95–F104, 2003.

19. Hansen KK, Sherman PM, Cellars L, Andrade-Gordon P, Pan Z,Baruch A, Wallace JL, Hollenberg MD, Vergnolle N. A major role forproteolytic activity and proteinase-activated receptor-2 in the pathogenesisof infectious colitis. Proc Natl Acad Sci USA 102: 8363–8368, 2005.

20. Holdsworth SR, Summers SA. Role of mast cells in progressive renaldiseases. J Am Soc Nephrol 19: 2254–2261, 2008.

21. Hollenberg MD, Wijesuriya SJ, Gui Y, Loutzenhiser R. Proteinase-activated receptors and the kidney. Drug Dev Res 60: 36–42, 2003.

22. Hyun E, Andrade-Gordon P, Steinhoff M, Vergnolle N. Protease-activated receptor-2 activation: a major actor in intestinal inflammation.Gut 57: 1222–1229, 2008.

23. Jacob C, Yang PC, Darmoul D, Amadesi S, Saito T, Cottrell GS,Coelho AM, Singh P, Grady EF, Perdue M, Bunnett NW. Mast celltryptase controls paracellular permeability of the intestine. Role of pro-tease-activated receptor 2 and beta-arrestins. J Biol Chem 280: 31936–31948, 2005.

24. Johansson U, Lawson C, Dabare M, Syndercombe-Court NewlandAC, Howells GL, Macey MG. Human peripheral blood monocytesexpress protease receptor-2 and respond to receptor activation by produc-tion of IL-6, IL-8, and IL-1beta. J Leukoc Biol 78: 967–975, 2005.

25. Koshikawa N, Hasegawa S, Nagashima Y, Mitsuhashi K, Tsubota Y,Miyata S, Miyagi Y, Yasumitsu H, Miyazaki K. Expression of trypsinby epithelial cells of various tissues, leukocytes, and neurons in human andmouse. Am J Pathol 153: 937–944, 1998.

26. Kunzelmann K, Sun J, Markovich D, Konig J, Murle B, Mall M,Schreiber R. Control of ion transport in mammalian airways by proteaseactivated receptors type 2 (PAR-2). FASEB J 19: 969–970, 2005.

27. Lau C, Lytle C, Straus DS, DeFea KA. Apical and basolateral pools ofproteinase-activated receptor-2 direct distinct signaling events in theintestinal epithelium. Am J Physiol Cell Physiol 300: C113–C123, 2011.

28. Lee AW. The role of atypical protein kinase C in CSF-1-dependent Erkactivation and proliferation in myeloid progenitors and macrophages.PLos One 6: e25580, 2011.

29. Lohman RJ, Cotterell AJ, Barry GD, Liu L, Suen JY, Vesey DA,Fairlie DP. An antagonist of human protease activated receptor-2 atten-



uates PAR2 signaling, macrophage activation, mast cell degranulation,and collagen-induced arthritis in rats. FASEB J 26: 2877–2887, 2012.

30. Lohman RJ, Cotterell AJ, Suen J, Liu L, Do AT, Vesey DA, FairlieDP. Antagonism of protease-activated receptor 2 protects against experi-mental colitis. J Pharmacol Exp Ther 340: 256–265, 2012.

31. Ma L, Dorling A. The roles of thrombin and protease-activated receptorsin inflammation. Semin Immunopathol 34: 63–72, 2012.

32. Macfarlane SR, Seatter MJ, Kanke T, Hunter GD, Plevin R. Protei-nase-activated receptors. Pharmacol Rev 53: 245–282, 2001.

33. Moriyuki K, Sekiguchi F, Matsubara K, Nishikawa H, Kawabata A.Proteinase-activated receptor-2-triggered prostaglandin E(2) release, butnot cyclooxygenase-2 upregulation, requires activation of the phosphati-dylinositol 3-kinase/Akt/nuclear factor-kappaB pathway in human alveo-lar epithelial cells. J Pharmacol Sci 111: 269–275, 2009.

34. Moussa L, Apostolopoulos J, Davenport P, Tchongue J, Tipping PG.Protease-activated receptor-2 augments experimental crescentic glomeru-lonephritis. Am J Pathol 171: 800–808, 2007.

35. Nystedt S, Emilsson K, Larsson AK, Strombeck B, Sundelin J. Mo-lecular cloning and functional expression of the gene encoding the humanproteinase-activated receptor 2. Eur J Biochem 232: 84–89, 1995.

36. Ossovskaya VS, Bunnett NW. Protease-activated receptors: contributionto physiology and disease. Physiol Rev 84: 579–621, 2004.

37. Ramachandran R, Morice AH, Compton SJ. Proteinase-activated re-ceptor2 agonists upregulate granulocyte colony-stimulating factor, IL-8,and VCAM-1 expression in human bronchial fibroblasts. Am J Respir CellMol Biol 35: 133–141, 2006.

38. Ritchie E, Saka M, Mackenzie C, Drummond R, Wheeler-Jones C,Kanke T, Plevin R. Cytokine upregulation of proteinase-activated-recep-tors 2 and 4 expression mediated by p38 MAP kinase and inhibitory kappaB kinase beta in human endothelial cells. Br J Pharmacol 150: 1044–1054, 2007.

39. Rothmeier AS, Ruf W. Protease-activated receptor 2 signaling in inflam-mation. Semin Immunopathol 34: 133–149, 2012.

40. Sharma R, Prasad V, McCarthy ET, Savin VJ, Dileepan KN, Stech-schulte DJ, Lianos E, Wiegmann T, Sharma M. Chymase increasesglomerular albumin permeability via protease-activated receptor-2. MolCell Biochem 297: 161–169, 2007.

41. Sirvent AE, Gonzalez C, Enriquez R, Fernandez J, Millan I,Barber X, Amoros F. Serum tryptase levels and markers of renaldysfunction in a population with chronic kidney disease. J Nephrol 23:282–290, 2010.

42. Su X, Camerer E, Hamilton JR, Coughlin SR, Matthay MA. Protease-activated receptor-2 activation induces acute lung inflammation by neu-ropeptide-dependent mechanisms. J Immunol 175: 2598–2605, 2005.

43. Suen JY, Barry GD, Lohman RJ, Halili MA, Cotterell AJ, Le GT,Fairlie DP. Modulating human proteinase activated receptor 2 with anovel antagonist (GB88) and agonist (GB110). Br J Pharmacol 165:1413–1423, 2012.

44. Suen JY, Gardiner B, Grimmond S, Fairlie DP. Profiling gene expres-sion induced by protease-activated receptor 2 (PAR2) activation in humankidney cells. PLoS One 5: e13809, 2010.

45. Summers SA, Gan PY, Dewage L, Ma FT, Ooi JD, O’Sullivan KM,Nikolic-Paterson DJ, Kitching AR, Holdsworth SR. Mast cell activa-tion and degranulation promotes renal fibrosis in experimental unilateralureteric obstruction. Kidney Int 82: 676–685, 2012.

46. van der Merwe JQ, Hollenberg MD, MacNaughton WK. EGF receptortransactivation and MAP kinase mediate proteinase-activated receptor-2-induced chloride secretion in intestinal epithelial cells. Am J PhysiolGastrointest Liver Physiol 294: G441–G451, 2008.

47. Vesey DA, Cheung C, Cuttle L, Endre Z, Gobe G, Johnson DW.Interleukin-1beta stimulates human renal fibroblast proliferation and ma-trix protein production by means of a transforming growth factor-beta-dependent mechanism. J Lab Clin Med 140: 342–350, 2002.

48. Vesey DA, Hooper JD, Gobe GC, Johnson DW. Potential physiologicaland pathophysiological roles for protease-activated receptor-2 in the kid-ney. Nephrology 12: 36–43, 2007.

49. Vesey DA, Kruger WA, Poronnik P, Gobe GC, Johnson DW. Proin-flammatory and proliferative responses of human proximal tubule cells toPAR-2 activation. Am J Physiol Renal Physiol 293: F1441–F1449, 2007.

50. Winter MC, Shasby SS, Ries DR, Shasby DM. PAR2 activation inter-rupts E-cadherin adhesion and compromises the airway epithelial barrier:protective effect of beta-agonists. Am J Physiol Lung Cell Mol Physiol291: L628–L635, 2006.



PAR2-induced inflammatory responses in human kidney tubular epithelial cells

Documents

Transcript of PAR2-induced inflammatory responses in human kidney tubular epithelial cells