Introduction

Top Abstract Introduction Materials & Methods Results Discussion References

The G-protein-coupled receptors stimulated by thrombin (proteinase-activated receptor-1: PAR₁) or by trypsin (PAR₂) are activated by the proteolytic unmasking of anchored N-terminal cryptic receptor-activating sequences (SLIGKV and SLIGRL for human and rodent PAR₂; SFLLR and SFFLR for human and rodent PAR₁) (Vu et al., 1991; Coughlin et al., 1992; Nystedt et al., 1994, 1995; Al-Ani et al., 1995; Böhm et al., 1996; Saifeddine et al., 1996). The G-protein-coupled receptors for thrombin are distinct from the GPIb/IX-V platelet binding site for thrombin, which may also play a role in thrombin action (Harmon and Jamieson, 1986; Jamieson, 1997). Strikingly, short synthetic peptides based on these revealed N-terminal activating sequences of the G-protein-coupled PARs (so-called PAR₁- or PAR₂-activating peptides, or PAR-APs) can, in isolation, activate either PAR₁ or PAR₂ (Vu et al., 1991; Nystedt et al., 1994; Hollenberg et al., 1996, 1997). The PAR₂-activating peptides, derived from either the human (SLIGKV-NH₂) or rodent (SLIGRL-NH₂) receptor sequences, can mimic the action of trypsin in activating PAR₂, but they are unable to activate the PAR₁ thrombin receptor because of a lack of an essential aromatic amino acid substituent at position 2 of the activating peptide (Hollenberg et al., 1993; Natarajan et al., 1995). In contrast, thrombin receptor-activating peptides derived from the human PAR₁ receptor sequence (e.g., SFLLR-NH₂) have been observed by us and by others to activate both PAR₁ and PAR₂ (Blackhart et al., 1996; Hollenberg et al., 1997). Nonetheless, we and others have shown that peptides such as TFLLR-NH₂ (TF-NH₂) (Hollenberg et al., 1997) or TFLLRNPNDK-NH₂ (Blackhart et al., 1996) can selectively activate PAR₁ but not PAR₂. One approach to evaluating the selectivity (or lack thereof) of agents targeted to either PAR₁ or PAR₂ has employed a xenopus oocyte receptor expression system, in which only one of the two receptors are expressed (Blackhart et al., 1996). As an alternative approach, we have used a receptor desensitization paradigm, employing an intact contractile tissue bioassay to illustrate, for instance, the selective activation of PAR₂ but not PAR₁ by the PAR₂ -AP, SLIGRL-NH₂ (Saifeddine et al., 1996). We sought to extend the receptor-desensitization paradigm for use with a cultured cell system, in which an intracellular calcium signal rather than a contractile response might be used as an index of receptor activation. Furthermore, we reasoned that, provided receptor cross-desensitization did not occur, it would be advantageous to assess PAR-targeted ligands in a cell that expressed both receptors. With both receptors present in the same cell, the selectivity or nonselectivity of a variety of compounds that would affect PAR₁ and/or PAR₂ could be efficiently evaluated in a single experiment. To this end, we developed a calcium-signaling assay, employing cultured human embryonic kidney cells (HEK293) in which the action of PAR₁ and PAR₂ agonists and antagonists could be evaluated simultaneously. Using this assay, we expected to evaluate the PAR₁/PAR₂ selectivity of the compounds listed in Table 1. Many of these agents have been previously described either as potent thrombin-receptor ligands (Feng et al., 1995) or as thrombin (PAR₁) receptor antagonists (Doorbar and Winter, 1994; Seiler et al., 1995; Bernatowicz et al., 1996). We also wished to evaluate the PAR₁/PAR₂ selectivity of the originally-described thrombin receptor-activating peptides SFLLRNPNDKYEPF-NH₂ (SF₁₄-NH₂) and SFLLR-NH₂(SF-NH₂) (so-called TRAPs: Vu et al., 1991; Hollenberg et al., 1992; Hui et al., 1992; Sabo et al., 1992; Scarborough et al., 1992; Vassallo et al., 1992; Chao et al., 1993).

	Materials and Methods

Top Abstract Introduction Materials & Methods Results Discussion References

Cell Culture and Fluorescence Measurements. Human embryonic kidney cells (HEK293) that express the SV40 T-antigen were kindly provided by Dr. Jonathan Lytton, University of Calgary, Faculty of Medicine, Calgary, AB Canada. Cells were propagated without the use of trypsin at 37^oC under an atmosphere of 5% CO₂ in room air in 80 cm² T-flasks using Dulbecco's minimal medium supplemented with 10% v/v fetal calf serum. Confluent cell monolayers to be used for calcium-signaling measurements were rinsed free of growth medium and disaggregated with calcium-free isotonic phosphate-buffered saline. Disaggregated cells were pelleted by centrifugation and were resuspended in 1 ml Dulbecco's minimal medium /10% fetal calf serum for loading with the calcium indicator, fluo-3 (Molecular Probes Inc., Eugene, OR) at a final concentration of 22 µM (25 µg/ml) of fluo-3 acetoxymethyl (AM) ester. Indicator uptake was allowed to proceed for 20-25 min at room temperature in the presence of 0.25 mM sulfinpyrazone, after which cells were washed two times by centrifugation and resuspension with the buffer described below, so as to remove excess dye. Cells loaded with fluo-3 were then resuspended to yield a stock suspension of about 6 × 10⁶ cells/ml in a buffer, pH 7.4, of the following composition: NaCl (150 mm), KCl (3 mM), CaCl₂ (1.5 mM), HEPES (20 mM), glucose (10 mM), and sulfinpyrazone (0.25 mM). Fluorescence measurements, reflecting elevations of intracellular calcium, were conducted at 24^oC, using a Perkin-Elmer fluorescence spectrophotometer, with an excitation wavelength of 480 nM and an emission recorded at 530 nM. Cell suspensions (about 2 ml at a concentration of approximately 3 × 10⁵ cells/ml) were maintained in suspension in a stirred (magnetic flea bar) thermostatted plastic cuvette (total volume 4 ml), and peptide stock solutions were added directly to the cuvette to monitor peptide-induced changes in fluorescence. To construct concentration-effect curves for fluorescence yield, the signals caused by a test peptide were expressed as a percentage (% A23187) of the fluorescence peak height yielded by replicate cell suspensions, when treated with 2 µM of ionphore, A23187 (Sigma, St. Louis, MO). This concentration of A23187 was at the plateau of its concentration-effect curve for a fluorescence response. Under the assay conditions, we established by high performance liquid chromatography analysis (as in the past, Tay-Uyboco et al., 1995) that peptide degradation did not occur, and we determined that the presence of proteinase inhibitors (e.g., amastatin) did not potentiate the action of peptides in the assay.

Peptides and Other Reagents. All peptides were synthesized by solid phase methods at the Peptide Synthesis Facility, The University of Calgary, Faculty of Medicine, Calgary, AB Canada, (Director, Dr. D. McMaster), or were provided through the courtesy of Dr. L. Leblond, via the peptide synthesis facility at BioChem Therapeutic, Laval, PQ Canada. The composition and purity of all peptides were confirmed by high performance liquid chromatography analysis, mass spectral analysis, and amino acid analysis. Stock solutions, prepared in 25 mM HEPES buffer pH 7.4, were analyzed by quantitative amino acid analysis to verify peptide concentrations and purity. Thapsigargin (TG) was from Sigma (St. Louis, MO).

Evaluation of Receptor Desensitization and Cross-Desensitization of PAR₁ and PAR₂ The receptor desensitization assay made use of the principle that repeated exposure of a tissue to an agonist that is receptor-selective leads to a diminution/desensitization of the receptor's response to that agonist, but not to the tissue's response to a second agonist, which activates a distinct receptor system. The desensitization approach that we employed previously to demonstrate the independent activation of PAR₁ and PAR₂ in a gastric longitudinal muscle strip bioassay (Saifeddine et al., 1996) was expanded upon, using the calcium signal yielded by receptor activation in HEK cells as an index of agonist activity. The key to the analysis of the PAR₁-targeted ligands that we evaluated in this study lies in the use of the PAR₁-selective agonist, TF-NH₂ and the PAR₂-selective agonist, SLIGRL-NH₂ (SL-NH₂). In the desensitization protocol described in detail in the following paragraph, both of these receptor-selective agonists were used at concentrations in the mid-range of their concentration-effect curves (10-30 µM). These concentrations selectively activate either PAR₁ (TF-NH₂) or PAR₂ (SL-NH₂). At these concentrations of receptor-selective agonists, the pretreatment of cells with any test compound that activated/desensitized or blocked one of the receptors would result in a diminution of the calcium signal subsequently generated by adding the receptor-selective agonist (TF-NH₂ or SL-NH₂) cumulatively, without removing the test compound from the cuvette.

View larger version (13K): [in this window] [in a new window]   Fig. 1.   Homologous desensitization of PAR1 by thrombin and PAR2 by trypsin, as well as lack of heterologous desensitization either of PAR2 by PAR1 activation due to thrombin or Cit-NH2, or of PAR1 by SL-NH2. HEK cell suspensions were first exposed twice to receptor-desensitizing concentrations of thrombin (Thr, , 10 U/ml or 100 nM, to desensitize PAR1: tracings A and C), trypsin (, 5 U/ml or 10 nM, to desensitize PAR2: tracing B), Cit-NH2 (, 10 µM to desensitize PAR1: tracing D), or SL-NH2 (, 100 µM, to desensitize PAR2: tracings E and F), with continuous recording of the fluorescence tracing (E530). At 10 min after the addition of the first agonist to the cuvette, a test concentration of the second agonist was added (e.g., thrombin, followed by a test concentration of the PAR1-AP, Cit-NH2: tracing A). The fluorescence signal caused by the sequential addition of the test compound after receptor-specific desensitization was compared with the signal caused by the same concentration of test compound without earlier desensitization of the cells (right-hand portion of tracings A to F.)

Preparation of HEK Cell RNA and Detection of PAR₁, PAR₂, and PAR₃ mRNA by the Reverse Transcriptase-Polymerase Chain Reaction (RT-PCR). HEK cell monolayers were grown to confluence, and total RNA was prepared using the TRI-reagent (Molecular Research Center, Cincinnati, OH). The RNA was reverse transcribed (RT) with a first strand cDNA synthesis kit using pd(N)6 primer (Pharmacia LKB Biotechnology, Uppsala, Sweden) according to manufacturers recommendations at 37^oC for 60 min; 3 µl of this solution was used with primer pairs targeted to human (h)PAR₁, PAR₂, and PAR₃. For (h)PAR₁, the primer pairs were: forward primer, PAR₁ F:5' AAAAGCTTCCCGCTCATTTTTTCTCAGGAA 3', and reverse primer PAR₁ R: 5' GGGAATTCAATCGGTGCCGGAGAAGT3'. For (h)PAR₂, the primer pairs were: forward primer, PAR₂ F:5' CACCACCTGTCACGATGTGCT3' and reverse primer, (h)PAR₂ R: 5`CCCGGGCTCAGTAGGAGGTTTTAACAC3'. For (h)PAR₃, the primer pairs were: forward primer, PAR₃ F: 5'TTT(T/G)TCAT(A/C)CTGAAGCAGGA3'; and for the reverse primer, PAR₃ R: 5'CTATTTTGTAAGGTAAGCAG3'. The signal yielded by the three sets of PAR primer pairs were normalized to the polymerase chain reaction (PCR) signal generated concurrently by an actin primer pair (Watson et al., 1992) that spanned an actin intron: forward primer: 5'CGT GGG CCG CCC TAG GCA CCA3'; reverse primer: 5'TTG GCC TTA GGG TTC AGG GGG3'. The detection of an intron-free 243 base-pair product using this primer pair can confirm the absence of DNA-derived intron sequences in the RT product obtained from tissue RNA. Routinely, amplification was obtained using 2.5 units of Taq DNA polymerase (Promega, Madison, WI) in a 10 mM Tris-HCl buffer, pH 9.0 (50 µl, final volume) containing MgCl₂ (1.5 mM), KCl (50 mM), 0.1% v/v Triton X-100, and 0.2 mM each of deoxynucleotide triphosphates. The amplification reaction was allowed to proceed for 40 cycles, beginning with a 1-min denaturing period at 94°C, followed by a 1-min reannealing time at 55°C, then a 1-min primer extension period at 72°C. The PCR products were separated by 1.5% agarose gel electrophoresis and visualized by ethidium bromide. PCR products were subcloned into the pGEM-T Vector (Promega, Madison WI) for sequencing by the dideoxynucleotide chain termination method (Sanger et al., 1977), employing a T₇DNA sequencing kit (Pharmcia). Sequencing was done in both the 5' and 3' directions.

Measurement of Platelet Aggregation. Washed platelets were isolated from citrate-anticoagulated plasma obtained from healthy volunteers who denied the use of nonsteroidal anti-inflammatory agents or other platelet-targeted drugs over at least the preceding 2 weeks. Washed platelet suspensions were prepared as outlined previously (Mustard et al., 1972), and aggregation was monitored by light scattering measurements done at 37°C using stirred platelet suspensions (400 µl) in a Bio Data aggregometer. Peptides in a volume of 50 µl were added directly to the stirred platelet suspension, and the degree of aggregation observed after a 3-min time period was expressed as a percentage of the maximum aggregation caused by each agonist. The use of the washed platelet suspension eliminated the need to add an aminopeptidase inhibitor (e.g., amastatin), as required for the use of platelet-rich plasma samples. Concentration-effect curves for the aggregating activity of peptides were thereby constructed, and the potency of each peptide was expressed as a concentration for which aggregation was half-maximal (EC₅₀). To assess the activity of putative PAR₁ inhibitors, each compound (in 50 µl) was added to the platelet suspension 1 min before the addition of either thrombin (in 50 µl of buffer: 0.015-0.1 U/ml, final concentration) or SFLLR-NH₂, (final concentration 3 µM). The concentration of thrombin used for platelets derived from different donors was based on the minimum thrombin concentration required to cause maximal platelet aggregation. Once determined, this minimum effective thrombin concentration was used to estimate the IC₅₀ for the putative PAR₁ inhibitors.

	Results

Top Abstract Introduction Materials & Methods Results Discussion References

RT-PCR Detection of PARs in HEK Cells. We wished to obtain a semiquantitative estimate of the abundance of PAR₁ and PAR₂ receptors in the HEK cells by a biochemical method. To this end, we used an RT-PCR approach, with amplimers targeted to these receptors; we also used primer pairs targeted to the more recently-described thrombin receptor, PAR₃ (Ishihara et al., 1997) to assess its abundance in the HEK cells, relative to PAR₁ and PAR₂. Relative to the PCR signal for actin, the relative intensities of the PCR signals for the three PARs were: PAR₃ > PAR₂ > PAR₁ (data not shown). It was not possible, with anti-receptor antibodies, to detect receptor protein for any of the PARs in the HEK cell line. Because PAR₃ can be activated only by thrombin and not by a variety of receptor-derived activating peptides related to SF-NH₂ or Cit-NH₂ (Ishihara et al., 1997), we anticipated that the presence of PAR₃ in the HEK cells would not be an interfering factor in the desensitization assay, which is described in the sections to follow.

Validation of the HEK Assay: Lack of Heterologous Receptor Cross-Desensitization and Adequacy of Calcium Stores for Repetitive Cell Activation. Before proceeding with the HEK calcium signaling assay, which depends upon the principle of homologous receptor activation/desensitization and upon a receptor-mediated elevation of intracellular Ca⁺⁺, it was necessary to answer two main questions with regard to the system: 1) Did the protocol of sequential cell stimulation by one agonist/antagonist followed by another allow sufficient time for the refilling of intracellular calcium stores, such that the calcium signal yielded by the addition of the first agonist would not, simply because of inadequate calcium stores, diminish the calcium signal yielded by the sequential addition of the second agonist? 2) Even if calcium stores proved adequate, was there heterologous receptor cross-desensitization between PAR₂ and PAR₁ in the HEK cell system, as has been documented in the endothelial cell, wherein PAR₂ activation can cross-desensitize PAR₁ (Mirza et al., 1996)?

Homologous Versus Heterologous PAR Desensitization in HEK Cells. Because a cross-desensitization between PAR₂ and PAR₁ had been observed in human endothelial cell cultures (Mizra et al., 1996), it was important to establish in the HEK cell system that the activation of PAR₁ did not simultaneously cross-desensitize PAR₂ and vice versa. To deal with this issue, we first used thrombin and SL-NH₂ as selective activators of PAR₁ and PAR₂, respectively. We then used a new PAR₁-selective receptor-activating peptide (Ala-parafluoroPhe-Arg-Cha-Cit-Tyr-NH₂) (Cit-NH₂) as well as the previously described PAR₁-selective agonist, TF-NH₂ (Hollenberg et al., 1997), and the PAR₂-selective agonist, SL-NH₂, to assess the cross-desensitization of PAR₁ and PAR₂ in the HEK cell system. The principle of the desensitization assay is illustrated by the data in Fig. 1. As expected, activation of PAR₁ by thrombin desensitized the subsequent HEK cell response to the selective PAR₁ agonist, Cit-NH₂ (tracing A, Fig. 1). Activation of PAR₂ by trypsin similarly desensitized the subsequent HEK cell response to the selective PAR₂ agonist, SL-NH₂ (tracing B, Fig. 1). However, as shown in tracing C of Fig. 1, activation of PAR₁ by thrombin did not affect the PAR₂ signal caused by the PAR₂-selective agonist, SL-NH₂, nor did activation of the thrombin receptor by the PAR₁ receptor-selective agonist, Cit-NH₂, diminish the response to SL-NH₂ (tracing D, Fig. 1). Furthermore, activation of PAR₂ by SL-NH₂ did not cause cross-desensitization of the response to thrombin; nor did PAR₂ activation cross-desensitize the HEK cell response to Cit-NH₂ (tracings E and F, Fig. 1). In view of the results described in the above paragraphs, we were able to arrive at the standard cross-desensitization protocol described in detail in the Materials and Methods section. The compounds that we wished to evaluate for their ability to affect either PAR₁ or PAR₂ are summarized in Table 1.

Selective Desensitization of PAR₁ by Thrombin. Having established the working protocol for the assay of the PAR₁/PAR₂ selectivity of peptide agonists, we sought to deal with the following questions: 1) What type of desensitization of PAR₁ and PAR₂ was caused by the enzymes, thrombin (Fig. 2) and trypsin (Fig. 3)? 2) Could we obtain evidence for the activation of PAR₃ in the HEK cells? 3) What was the selectivity/activity of previously described thrombin receptor antagonists (Figs. 4-8)? 4) What was the relative PAR₁/PAR₂ selectivity of previously described PAR₁-targeted agonists (Fig. 9)? 5) What was the relative PAR₁/PAR₂ selectivity of the originally described thrombin receptor-activating peptides (referred to in the literature as TRAPs), SF-NH₂, and SF₁₄-NH₂ (Fig. 10)? Finally, we wished to determine if the desensitization protocol that we had so far developed (see above sections) might depend on the nature of the receptor-selective agonist used to generate the standard PAR₁ or PAR₂ signal (steps 1 and 2 described in the preceding paragraph.) (See Fig. 11).

View larger version (20K): [in this window] [in a new window]   Fig. 2.   Concentration-Desensitization (, ) and concentration-stimulation () curves for thrombin-mediated activation of HEK cells. As illustrated in Fig. 2, cells were first stimulated by the addition of thrombin (concentrations ranging from 0.5-100 nM; 0.05-10 U/ml), and the magnitude of the calcium signal (% A23187, ) relative to that caused by 2 µM A23187 in an aliquot of the indicator HEK cells was monitored. At 10 min after the addition of thrombin, a test concentration of either the PAR1-AP, TF-NH2 (, 10 µM), or the PAR2-AP, SL-NH2 (, 30 µM) was added to the same cells that had been pretreated with thrombin, and the fluorescence signal caused by the second addition of agonist was monitored. The residual fluorescence signal measured after thrombin pretreatment was expressed as a percentage [% residual PAR1 () or PAR2 ()] of the control fluorescence signal observed by activating either PAR1 or PAR2 in HEK cell aliquots with the addition of either 10 µM TF-NH2 or 30 µM SL-NH2 to cells without thrombin pretreatment. Data points represent average values for measurements done with three or more separate cell suspensions, for which the standard errors of the mean fell within the size of the symbols shown. The figure is representative of experiments done on three or more independently grown crops of HEK cells. Thrombin pretreatment did not affect the subsequent activation of PAR2 by SL-NH2 (), but desensitized 70% of the subsequent activation of PAR1 by TF-NH2 ().

View larger version (19K): [in this window] [in a new window]   Fig. 3.   Concentration-desensitization (, , ) and concentration-stimulation () curves for trypsin-mediated activation of HEK cells. As for thrombin (Fig. 2), HEK cell suspensions were first stimulated by the addition of trypsin (0.2-500 nM; 0.1-250 U/ml), and the calcium signal (% A23187, ) relative to that caused by 2 µM A23187 was measured. Following trypsin treatment, cells were exposed to either (10 nM; 1 U/ml) thrombin () or 10 µM TF-NH2 () to monitor PAR1 activation, and to 30 µM SL-NH2 for activating PAR2 (). The residual fluorescence signal measured after trypsin pretreatment was expressed as a percentage [% residual PAR1 monitored by TF-NH2 () or thrombin () stimulation; or % residual PAR2 monitored by SL-NH2 () stimulation] of the control fluorescence signal measured without earlier trypsin treatment, by activating PAR1 with either thrombin () or TF-NH2 () and by activating PAR2 with SL-NH2 (). Values represent the means (±S. E. M., bars, except where error bars were smaller than the size of the symbols) for measurements done with three or more cell aliquots. The data are representative of experiments done with two or more separately grown crops of HEK cells.

View larger version (17K): [in this window] [in a new window]   Fig. 4.   Diminished calcium signal in response to the addition of either PAR1- () or PAR2- () activating peptides by pretreatment of HEK cells with the putative PAR1 antagonist, Mpr-NH2. As illustrated for thrombin in Fig. 2, HEK cell suspensions were exposed first to Mpr-NH2 ( , 1-300 µM) and the calcium signal (E530, % A23187, ), relative to 2 µM A23187, was measured. At 10minutes after the earlier addition of Mpr-NH2, the residual PAR1 () or PAR2- () mediated calcium fluorescence response (% control) relative to that observed in identical cell suspensions that had not been pretreated with Mpr-NH2, was monitored by the addition of either the PAR1-AP, TF-NH2 (10 µM, ) or the PAR2-AP, SL-NH2 (30 µM, ). Values represent the means (±S. E. M., bars) for measurements done with triplicate cell suspensions derived from two or more separately grown crops of HEK cells.

View larger version (14K): [in this window] [in a new window]   Fig. 5.   Desensitization of the Mpr-NH2-mediated calcium signal by desensitization of PAR2 but not by desensitization of PAR1. HEK cell suspensions were first exposed to receptor-desensitizing concentrations of either the PAR2-AP, SL-NH2 (, 100 µM, tracing A) or the PAR1-AP, TF-NH2 (, 40 µM, tracing B). After 10 min, a test concentration of Mpr-NH2 (*, 15 µM) was added to the cell suspensions, with a continuous recording of calcium-mediated fluorescence (E530, scale for time and fluorescence shown between tracings A and B). The calcium signal caused by 15 µM Mpr-NH2 () in a cell suspension that had not been previously exposed to either SL-NH2 (tracing A) or TF-NH2 (tracing B) is shown to the right of each tracing. The tracings are representative of three or more independently conducted experiments with replicates derived from different HEK cell crops.

View larger version (13K): [in this window] [in a new window]   Fig. 6.   Inhibition by Met-OH of PAR1 activation by thrombin but not by TF-NH2. HEK cell suspensions were first exposed to a concentration of Met-OH (, 500 µM) above that previously reported to inhibit thrombin-mediated platelet activation (Doorbar and Winter, 1994). After 10 min, the residual PAR1-mediated calcium fluorescence signal (E530) generated by the addition of either thrombin (, 5 nM; 0.5U/ml, tracing A) or the PAR1-AP, TF-NH2 (, 2.5 µM, tracing B) was monitored. The calcium signal caused in an identical cell aliquot by the addition of either thrombin (, tracing A) or TF-NH2 (, tracing B) to cells that had not been pretreated with Met-OH is shown in the right-hand portion of each tracing. The tracings are representative of two independently conducted experiments. The scale for time and calcium signal is shown between the tracings.

View larger version (16K): [in this window] [in a new window]   Fig. 7.   Desensitization of PAR1 () and PAR2 () by pretreatment of HEK cells with the putative PAR1 antagonist, trans-Cinnamoyl-(f)F-R-L-R-O-NH2. As illustrated for Mpr-NH2 in Fig. 4, HEK cell suspensions were first exposed to tc-fF-NH2 (, 1-400 µM) and the calcium signal thereby generated (E530) was measured as a percentage (% A23187, ) relative to 2 µM A23187. At 10 min after the earlier addition of tc-fF-NH2, the residual PAR1- () or PAR2- () mediated calcium fluorescence response was measured as a percentage (% control , ) of the control fluorescence response measured upon adding the test concentrations of either the PAR1-AP, TF-NH2 (, 10 µM), or the PAR2-AP, SL-NH2 (, 30 µM) to identical cell suspensions that had not been pretreated with tc-fF-NH2. Values represent the means for measurements done with triplicate cell suspensions (±S. E. M. smaller than size of symbols) obtained from two or more HEK cell crops.

View larger version (13K): [in this window] [in a new window]   Fig. 8.   Partial desensitization of the HEK cell calcium signal caused by tc-fF-NH2 caused by pretreatment with the PAR1- and PAR2-selective agonists, TF-NH2 and SL-NH2, as well as full desensitization by pretreatment with the nonselective PAR1/PAR2 agonist, SF-NH2. HEK cell suspensions were first pretreated twice with desensitizing concentrations of either the PAR-selective agonists, TF-NH2 (, 40 µM for PAR1, tracing A) and SL-NH2 (, 100 µM for PAR2, tracing B) or the nonselective PAR1/PAR2 agonist, SF-NH2 (, 40 µM, tracing C). After 10 min, a test concentration of tc-fF-NH2 close to its EC50 for causing a calcium signal ([star], 50 µM) was added to the cell suspension, with continuous monitoring of all the calcium fluorescence (E530). The calcium signal caused by 50 µM tc-fF-NH2 () in the agonist-pretreated cells (left-hand set of tracings A to C) was compared with the signal caused in identical cell suspensions that had not been desensitized by agonist pretreatment (right-hand set of tracings A to C). The experiment shown is representative of measurements done with two sets of independently prepared HEK cell suspensions. The scale for time and fluorescence is shown between tracings A and B.

View larger version (17K): [in this window] [in a new window]   Fig. 9.   Desensitization of PAR1 () and PAR2 () by PAR1-selective agonists: concentration-desensitization curves. As illustrated for thrombin in Fig. 2, HEK cell suspensions were first exposed to different concentrations of three PAR1-selective agonists: hArg-NH2 (solid lines), Cit-NH2 (dash lines), and TF-NH2 (dash/dot lines). After 10 min, the cell suspension was challenged by the sequential addition of test concentrations of either the PAR1-selective agonist, TF-NH2 (, 10 µM) or the PAR2-selective agonist, SL-NH2 (, 30 µM). The residual PAR1 () or PAR2 () calcium signal (E530) observed in the agonist-pretreated cells was measured as a percentage (% control) of the calcium signal caused by either 10 µM, TF-NH2 () or 30 µM, SL-NH2 () in identical cell suspensions that had not been pretreated with the PAR1-selective agonists. Values represent the means (±S.E.M., bars) for measurements done with three or more replicate cell suspensions coming from two or more independently grown HEK cell crops. Error bars smaller than the symbol size are not shown.

View larger version (22K): [in this window] [in a new window]   Fig. 10.   Desensitization of PAR1 () and PAR2 () by the originally described thrombin receptor-activating peptides SFLLR-NH2 and SFLLRNPNDKYEPF-NH2: concentration-desensitization curves. As illustrated for the PAR1-selective agonists in Fig. 9, concentration-desensitization curves were obtained for the ability of the originally described thrombin receptor-activating peptides, SF14-NH2 (dashed lines) and SF-NH2 (solid lines) to desensitize either PAR1 () or PAR2 (). Values represent the means of measurements done with triplicate cell suspensions obtained from two or more separately grown HEK cell crops; standard error bars were smaller than the symbols shown.

View larger version (13K): [in this window] [in a new window]   Fig. 11.   Concentration-desensitization curves for TF-NH2 using test concentrations of either Cit-NH2 (, 1 µM) or TF-NH2 () to monitor PAR1 desensitization. As outlined for the agonists shown in Figs. 9 and 10, a concentration-desensitization curve was determined for pretreatment of HEK cell suspensions with different concentrations of TF-NH2 (0.5-40 µM), followed by monitoring the residual calcium signal (E530) yielded by the subsequent addition of a test concentration of either Cit-NH2 (, 1 µM) or TF-NH2 (, 10 µM). The residual PAR1-stimulated fluorescence signal (E530) observed for the two different PAR1-APs was expressed as a percentage (% control) of the signal observed upon adding either Cit-NH2 () or TF-NH2 (, 10 µM) to identical HEK cell suspensions that had not been pretreated with the different concentrations of TF-NH2 (x-axis). Values represent the means of measurements done with triplicate cell suspensions obtained from two independently grown HEK cell crops. Error bars were smaller than the symbols shown.

Cross-Desensitization of PAR₁ by Trypsin: Quantitative Estimate of Receptor Selectivity. In contrast with thrombin, which cannot activate PAR₂, elevated concentrations of trypsin (greater than about 20 U/ml or 40 nM) can also activate PAR₁ (Vu et al., 1991). Therefore, with the cross-desensitization assay, we used trypsin to illustrate quantitatively the dual specificity of its ability to activate/desensitize PAR₂ at concentrations lower than about 5 U/ml (10 nM) and to activate/desensitize PAR₁ at concentrations higher than about 20 U/ml (>40 nM) (Fig. 3). Although trypsin was able to desensitize almost completely the response of PAR₂ to SL-NH₂, (Fig. 3), this protease was not able to desensitize more than about 35% of the PAR₁ response to TF-NH₂ (, Fig. 3). Nonetheless, prior treatment of the cells with concentrations of trypsin that were lower than those required to desensitize the response to TF-NH₂, "disarmed" completely the ability of thrombin (, Fig. 3) to activate PAR₁. In Fig. 3, the response of the HEK cells to trypsin, in terms of the fluo-3 calcium signal, could be the result of the combined activation of PAR₁ and PAR₂.

Test for Thrombin-Activated PAR₃ Signaling in HEK Cells. Because PAR₃ is evidently insensitive to activation by PAR-activating peptides (Ishihara et al., 1997), we reasoned that the peptide, SFLLR-NH₂ (SF-NH₂), which can activate/desensitize both PAR₁ and PAR₂ (Hollenberg et al., 1997), would be able to desensitize both PAR₁ and PAR₂ without activating PAR₃. If at these concentrations of SF-NH₂ (about 40 µM), PAR₃ were not activated, we reasoned that, although a selective PAR₁-activating peptide would no longer yield a calcium signal, thrombin should still be able to activate PAR₃ to yield an increase in fluo-3 fluorescence. We found that, at concentrations of SF-NH₂ that desensitized the HEK cell response to both TF-NH₂ and SL-NH₂ (i.e., both PAR₁ and PAR₂ were desensitized: not shown), no further fluorescence response to thrombin (10 U/ml; 100 nM) was observed. Similarly, after fully desensitizing PAR₁ with Cit-NH₂, thrombin (10 U/ml; 100 nM) failed to cause a fluorescence response (not shown). This concentration of thrombin (10 U/ml) would have been more than sufficient to activate PAR₃ (Ishihara et al., 1997).

Relative Selectivity of Putative PAR₁-Targeted Antagonists for PAR₁ Compared with PAR₂. Previous work had described a PAR₁ receptor antagonist, Mpr-Phe-Cha-Cha-Arg-Lys-Pro-Asn-Asp-Lys-NH₂ (Mpr-NH₂) (Seiler et al., 1995). We assessed the PAR₁/PAR₂ selectivity of Mpr-NH₂ using the HEK cell assay, as shown in Fig. 4. Prior treatment of HEK cells with Mpr-NH₂ led to a reduction in the subsequent ability of both TF-NH₂ and SL-NH₂ to activate PAR₁ and PAR₂ respectively, with comparable IC₅₀s for blocking or desensitizing both receptors. Mpr-NH₂ was able to block PAR₂ completely (, Fig. 4), but was only able to block about 60% of the HEK cell PAR₁ response to TF-NH₂ (, Fig. 4).

Relative Selectivity of PAR₁-Targeted Agonists for PAR₁ Compared with PAR₂. In previous work, the peptide, Ala-parafluoroPhe-Arg-Cha-hArg-Tyr-NH₂ (hArg-NH₂) had been synthesized as a high potency PAR₁ agonist for use as a PAR₁ receptor binding probe (Feng et al., 1995; Ahn et al., 1997). In accord with these studies, we synthesized the peptide Ala-parafluoroPhe-Arg-Cha-Cit-Tyr-NH₂ (Cit-NH₂) as a potential alternative to hArg-NH₂ for use as a PAR₁ receptor binding probe; and we had already synthesized the peptide, TF-NH₂, which in preliminary work appeared to be a PAR₁-selective agonist (Hollenberg et al., 1997). We first assessed the relative PAR₁/PAR₂ specificity of hArg-NH₂ and, in parallel, evaluated the receptor selectivity of Cit-NH₂ and TF-NH₂ (Fig. 9), although as expected, both hArg-NH₂ and Cit-NH₂were able to activate/desensitize PAR₁, hArg-NH₂ at concentrations between 2 and 40 µM was also able to activate/desensitize PAR₂ by more than 60%. In contrast, Cit-NH₂, at concentrations as high as 50 µM did not affect PAR₂ (Fig. 9). In the HEK cell assay, the relative selectivity of Cit-NH₂ for the PAR₁ receptor, compared with PAR₂ was about 280:1, whereas the PAR₁-selectivity of hArg-NH₂ was lower by about 2-fold (about 120:1, Fig. 9 and Table 2). Both hArg-NH₂ and Cit-NH₂ were found to be full agonists for the PAR₁ receptor (not shown). In comparison with hArg-NH₂ and Cit-NH₂, the much simpler peptide TF-NH₂, composed entirely of naturally occurring amino acids, demonstrated a PAR₁/PAR₂-selectivity of about 220:1 (Fig. 9 and Table 2).

Relative Receptor Selectivity of Receptor-Activating Peptides Derived from Human PAR₁. Because many previous studies, including our own (Yang et al., 1992; Tay-Uyboco et al., 1995), have used peptides derived from the human PAR₁ activating sequence as surrogates for the action of thrombin (previously termed thrombin receptor activating peptides, or TRAPs), we wished to evaluate the PAR₁/PAR₂ receptor selectivity of TRAPs that had been used extensively in previous work without the knowledge that such agonists could affect both PAR₁ and PAR₂: SFLLR-NH₂ (SF-NH₂) and SFLLRNPNDKYEPF-NH₂ (SF₁₄-NH₂) (originally described by Vu et al., 1991). As shown in Fig. 10, both SF₁₄-NH₂ and SF-NH₂ were somewhat selective for PAR₁ over PAR₂ (about 2- to 4-fold), but neither of the originally described TRAPs were anywhere near as PAR₁-selective as hArg-NH₂ (120:1), TF-NH₂ (220:1), or Cit-NH₂ (280:1). Table 2 summarizes the activities of all peptides studied in terms of their PAR₁/PAR₂ selectivity and their IC₅₀s for attenuating the calcium signal in HEK cells caused by subsequent activation of either PAR₁ (TF-NH₂) or PAR₂ (SL-NH₂)

Evaluation of the Use of Different PAR-Selective Standard Agonists in the HEK Cell Assay. Because the desensitization assay depends primarily on the activity of the unknown compound added to the cells before the PAR-selective test compound (routinely, TF-NH₂, in the PAR₁ assay we have developed), the desensitization profile for a given unknown compound should not, in principle, depend on the PAR-selective receptor probe that is used to generate the standard signal (step 2 in the procedure outlined above). To test this principle, we measured the PAR₁ desensitization profile for TF-NH₂, (as an unknown), using either the standard concentration of TF-NH₂ itself (10 µM) or, the more potent PAR₁-selective agonist, Cit-NH₂ (1 µM). As shown in Fig. 11, the PAR₁ desensitization profile for TF-NH₂ was virtually superimposable, irrespective of which selective PAR₁ receptor probe was used to generate the standard calcium signal.

Activity of Peptides in a Human Platelet Aggregation Assay. We wished to compare the PAR_1. selective activities of Mpr-NH2, Met-OH, tc-fF-NH₂, hArg-NH₂, Cit-NH₂, TF-NH_2, SF-NH₂, and SF₁₄-NH₂ determined in the HEK cell assay, in terms of attenuating the PAR₁ calcium signal (Table 2), with their activities in a platelet aggregation assay (results summarized in Table 3). Platelets do not possess PAR₂, and thus the potential cross-reactivity of the peptides at PAR₂ was not an issue in the platelet assay. All peptides except Mpr-NH₂ (up to 20 µM) and Met-OH were platelet agonists, with EC₅₀ s in the range of 0.1 to 10 µM (Table 3). Mpr-NH₂, as previously described, proved to be an inhibitor of both thrombin-mediated and SF-NH₂-mediated platelet aggregation. However, the IC₅₀ for Mpr-NH₂ depended heavily on the concentration of either thrombin or SF-NH₂, which was used as a platelet agonist. For instance, at a concentration of 0.015 U/ml thrombin, the IC₅₀ for Mpr-NH₂ was 0.7 µM (Table 3), whereas at a concentration of 0.04 U/ml thrombin, the IC₅₀ was about 6 µM (Table 2). Both Mpr-NH₂ and Met-OH proved to be poor antagonists of SF-NH₂ in the platelet aggregation assay (IC₅₀ > 200 µM), and Met-OH was a weak antagonist of thrombin-mediated platelet aggregation (Table 3, and data not shown). Surprisingly, tc-fF-NH₂ on its own was a weak platelet agonist, causing a reversible aggregation at concentrations up to 20 µM. At concentrations lower than 20 µM, tc-tF-NH₂ was able to inhibit SF-NH₂-induced platelet aggregation, with an IC₅₀ of about 2 µM (Table 2). This inhibition, however, was observed only by adding tc-fF-NH₂ to the cuvette first and by waiting until the light scattering caused by this peptide had returned to baseline, before adding the full agonist (3 µM SF-NH₂ or 0.05 U/ml thrombin) that caused aggregation.

	Discussion

Top Abstract Introduction Materials & Methods Results Discussion References

Our data, obtained with the newly described desensitization assay, considerably extended our preliminary observations describing TF-NH₂ as a PAR₁-selective agonist (Hollenberg et al., 1997). The principal finding of our study was that ligands originally developed as PAR₁-targeted ligands can also be seen to activate PAR₂, with a greater or lesser selectivity for PAR₁. The results with the PAR₁ antagonist, Mpr-NH₂ were particularly instructive as, with comparable potencies, this peptide proved to be an antagonist for PAR₁ and full agonist for PAR₂ (Figs. 4 and 5). Confirming earlier reports (Seiler et al., 1995), Mpr-NH₂ was a PAR₁ antagonist in the platelet, where PAR₂ was absent; this compound was also an antagonist for the HEK cell PAR₁ receptor. Thus, previous work in which this PAR₁ antagonist or in which nonselective PAR₁-APs were used to evaluate tissues such as the vascular endothelium (Lum et al., 1993; Zimmerman et al., 1994), wherein PAR₁ and PAR₂ coexist (Al-Ani et al., 1995), may need to be reevaluated. The data obtained with the Mpr-NH₂ peptide also point to differences whereby this peptide docks with PAR₁ and PAR₂, on the one hand leading to receptor activation (PAR₂), on the other to inhibition (PAR₁).

The principle of the PAR₁/PAR₂ desensitization assay depends on a density of PAR₁ and PAR₂ receptors that is low enough so that the activation/desensitization of receptors by a test compound will diminish the calcium signal generated by the subsequent addition of the PAR-selective test compound. Thus, the assay we describe will only work efficiently for agonists if the proportion of spare receptors for the calcium signal response is low, relative to the total numbers of receptors present in the system. That our assay works suggests that the proportion of spare PAR₁ and PAR₂ receptors in the HEK cells is indeed appropriately low for the purposes of our new assay. However, the absolute values of the IC₅₀s for the various test compounds obtained in the desensitization assay for interacting with either PAR₁ or PAR₂ must be interpreted with caution, relative to the potencies that the peptides we have studied may exhibit in other responsive systems. For instance, differences in receptor numbers between cell types could shift the concentration-effect curves to the left (higher receptor numbers) or to the right (lower receptor numbers). Nonetheless, the relative potencies within a series of compounds that interact with either PAR₁ or PAR₂ (or both) would be expected to be the same, irrespective of the tissue in which the compounds were assayed. It is, therefore, important to note that the relative potencies of the PAR₁-targeted agonists in the PAR₁ desensitization assay (hArg-NH₂ > Cit-NH₂ > tc-fF-NH₂  TF-NH₂  SF-NH₂ > SF₁₄-NH₂) (Table 2) were in good accord with the relative EC₅₀s of the same compounds in the platelet aggregation assay (Table 3). Because of the complex dependence of the IC₅₀s for antagonists on the thrombin concentration in the platelet aggregation assay, a correlation with the HEK cell assay was problematic.

Verification of the ability of the HEK cell signaling assay to identify receptor-selective agonists can be seen in the data obtained with thrombin, the PAR₁-selective peptide, Cit-NH₂, and the PAR₂-selective agonist, SL-NH₂. As has been observed in previous work (Nystedt et al., 1994), thrombin is unable to activate PAR₂, and SL-NH₂ is unable to activate PAR₁. The PAR₁ selectivity of Cit-NH₂ (almost 300:1) documented by our assay singles this compound out as an attractive, selective PAR₁ agonist for use in studies done in vivo. Based on our previous data and on the results with TF-NH₂, one can predict that the peptide Thr-parafluoroPhe-Arg-Cha-Cit-Tyr-NH₂ would be an even more selective PAR₁ agonist than Cit-NH₂. A key to our assay was that PAR₁ activation did not lead to heterologous desensitization of PAR₂, nor did PAR₂ activation desensitize PAR₁. These data appear to conflict with the observations of Mirza and colleagues (1996), who observed heterologous desensitization between PAR₁ and PAR₂ receptors in cultured human endothelial cells. That this cross-desensitization did not occur between PAR₁ and PAR₂ in the HEK cell assay may be due to two factors. First, our assay allowed for a time period of 10 min or more between the sequential addition of agonists to permit a complete refilling of intracellular calcium stores; this time period, which was longer than that used by Mirza et al. (1996), may have also allowed for a resensitization of receptors that had been cross-desensitized via a heterologous receptor mechanism. Alternatively, the enzymes responsible for heterologous receptor desensitization in cultured human endothelial cells and the relative abundance of PAR₁ and PAR₂ in the endothelial cells (Mirza et al., 1996) may differ considerably in comparison with the HEK cells, so as to account for the differences in heterologous receptor desensitization. This lack of heterologous receptor cross-desensitization, not only for the PAR₁/PAR₂ system, but also between the LPA receptor and the PAR receptors could allow for the simultaneous evaluation of multiple members of other receptor families that might be either present in HEK cells or coexpressed in these cells via transfection.

Because the HEK cells express sufficient wild-type human PAR₁ and PAR₂ receptors for the assay we describe, the system is far more convenient and precise than the use of PAR₁ and PAR₂ transfected xenopus oocytes, as described in work that appeared during the course of our studies (Blackhart et al., 1996). Although the data obtained with the xenopus oocyte receptor expression system can complement the results we describe in this report, the variability of response of the receptor-transfected oocytes makes difficult any precise assignment of the relative magnitude of PAR₁/PAR₂ ligand selectivity, as is possible with the HEK cell assay (Table 2). Unfortunately, the spectrum of PAR₁APs used in the oocyte expression assay (Blackhart et al., 1996) differed considerably from the compounds we studied with the HEK cell assay system, and a direct comparison of receptor selectivity was not possible. The calcium signaling assay employing the HEK cells can also be seen to complement the use of human platelets (Table 3), which represent a useful screening target for the evaluation of PAR₁ agonists and antagonists, but which are devoid of PAR₂ receptors. Furthermore, as opposed to the xenopus expression system, our assay was able to demonstrate simultaneously (Fig. 3) the activation/densensitization of PAR₁ by trypsin (20-100 nM), as reflected by TF-NH₂-mediated receptor activation and, at much lower concentrations (0.5-10 nM trypsin), the disarming of the ability of thrombin to activate PAR₁, presumably by trypsin cleavage of the receptor at a site downstream from the tethered receptor-activating ligand domain. Thus, the HEK assay will also prove of value for assessing the effects of a variety of proteinases on PAR₁ and PAR₂ activation.

We were puzzled to observe that the peptide tc-fF-NH₂ was an agonist at both PAR₁ and PAR₂ in the HEK cell assay (Figs. 7 and 8) and that this peptide was a partial agonist in the platelet aggregation assay, where antagonist activity could also be measured. Given the antagonist activity reported by Bernatowicz et al. (1996) for the peptide: trans-cinnamoyl-parafluoroPhe-Arg-Leu-Arg-NH₂ [IC₅₀ of about 1 µM for inhibiting PAR₁AP-induced platelet aggregation: Bernatowicz et al. (1996) compound 76, Table 9] and the antagonist activity of trans-cinnamoyl-parafluoroPhe-paraguanidinoPhe-Leu-Arg-Orn-NH₂ [IC₅₀, about 50 nM for inhibiting PAR₁AP-induced platelet aggregation: Bernatowicz et al. (1996), compound 88, Table 11], we expected tc-fF-NH₂ to be a full antagonist. The partial agonist activity in the peptide we synthesized (tc-fF-NH₂) points to the importance of the paraguanidino-phenyl group at position 3 in PAR₁ antagonist peptides. Furthermore, it is possible that tc-fF-NH₂ may nonspecifically desensitize the platelet to other agonists (e.g., ADP).

Finally, based on our PCR data, we expected to find evidence for the activation of PAR₃ by thrombin in the HEK cell assay. Because PAR₃ has been reported not to be activated by PAR-APs (Ishihara et al., 1997), we fully anticipated that the complete desensitization of PAR₁ by TF-NH₂ or by SF-NH₂ would still allow for the sequential activation of PAR₃ by thrombin. That this was not the case may be accounted for by several possibilities. First, activation of PAR₁ may cross-desensitize PAR₃. Second, PAR₁AP, such as Cit-NH₂, TF-NH₂, or SF-NH₂, which cannot activate PAR₃, may nonetheless prove to be antagonists of the PAR₃-tethered peptide. Third, PAR₃ may not, as do PAR₁ and PAR₂, couple to the calcium signaling mechanism in HEK cells. These possibilities represent interesting topics for further study. Notwithstanding, the presumed presence of PAR₃ (we were not able to assess the HEK cell content of PAR₃ protein) in the HEK cells appears not to interfere with the PAR₁/PAR₂ desensitization assay that we have developed. We anticipate that this fluorescence-based assay will prove of considerable use for screening of PAR₁- and PAR₂-targeted agents in future studies, not only for peptidomimetic compounds but also for other proteases that may affect either PAR₁ or PAR₂.