While the role of G-protein–coupled receptors (GPCR) in cancer is acknowledged, their underlying signaling pathways are understudied. Protease-activated receptors (PAR), a subgroup of GPCRs, form a family of four members (PAR1–4) centrally involved in epithelial malignancies. PAR4 emerges as a potent oncogene, capable of inducing tumor generation. Here, we demonstrate identification of a pleckstrin-homology (PH)-binding motif within PAR4, critical for colon cancer growth. In addition to PH–Akt/PKB association, other PH-containing signal proteins such as Gab1 and Sos1 also associate with PAR4. Point mutations are in the C-tail of PAR4 PH-binding domain; F347 L and D349A, but not E346A, abrogate these associations. Pc(4–4), a lead backbone cyclic peptide, was selected out of a mini-library, directed toward PAR2&4 PH-binding motifs. It effectively attenuates PAR2&4–Akt/PKB associations; PAR4 instigated Matrigel invasion and migration in vitro and tumor development in vivo. EGFR/erbB is among the most prominent cancer targets. AYPGKF peptide ligand activation of PAR4 induces EGF receptor (EGFR) Tyr-phosphorylation, effectively inhibited by Pc(4–4). The presence of PAR2 and PAR4 in biopsies of aggressive breast and colon cancer tissue specimens is demonstrated. We propose that Pc(4–4) may serve as a powerful drug not only toward PAR-expressing tumors but also for treating EGFR/erbB-expressing tumors in cases of resistance to traditional therapies. Overall, our studies are expected to allocate new targets for cancer therapy. Pc(4–4) may become a promising candidate for future therapeutic cancer treatment.

This article is featured in Highlights of This Issue, p. 1369

The growing appreciation of G-protein–coupled receptor (GPCR) in cancer pathogenesis is emerging, yet GPCR-based therapy in cancer is rare (1, 2). Mammalian protease-activated receptors (PAR) are a subgroup of GPCRs that form a family of four members, PAR1–4 (3, 4). Within the dynamically energetic and flexible tumor microenvironment, both soluble and matrix-immobilized proteases are engaged to maintain tumor growth and progression. Indeed, a robust cross-talk between proteases and their cell surface receptors such as the PAR family members is actively taking place. We have previously shown that PAR1 and PAR2 oncogenes are potent activators of β-catenin stabilization recruiting LRP5/6 coreceptors in malignant cancer as also in the physiologic invasion of placenta anchorage to the uterus decidua, in the first trimester of pregnancy (5–8). Both PAR1 and PAR2 have essential assignments in a wide range of epithelial tumors (9–11). Formerly, we reported the identification of signal-binding motifs within PAR1&2 C-tails, critical for breast cancer growth (12). These motifs are the binding sites of pleckstrin homology (PH) domain residing in signal proteins such as Akt/PKB forming a complex of PAR–Akt/PKB. Other PH-domain signal proteins, Etk/Bmx and Vav3, also associate with PAR1 and PAR2 through their PH domains. The PH domain is common among protein modules that drive cellular signaling in intermolecular interactions (13). PH-binding modules serve nonetheless as recognition sites connecting target proteins to implement discrete signaling network. PAR1 and PAR2 play a fundamental part in cancer advancement, allocating a dominant role for PAR2 over PAR1 (7, 14), while PAR3 functions mainly as a coreceptor. PAR4 is a receptor for thrombin-induced cellular responses and is often coexpressed with PAR1. In fact, thrombin activation of human platelets is carried out by PAR1 and PAR4 (15, 16). PAR4 function appears to be essential for the later stages, while PAR1 controls the early stages of platelet activation (16). Unexpectedly, PAR4 emerges as a potent oncogene (17–21). High-throughput RNA sequencing survey of selected GPCR transcriptional profile revealed the expression of 195 GPCRs that were either up- or downregulated during reprogramming to cancer stem cell (CSC) sphere formation (22). Among these, PAR4 (f2rl3) and PAR2 (f2rl1) GPCRs are considerably upregulated in CSC sphere formation.

Targeted therapy has been one of the most promising treatment options in cancer. For example, small molecules targeting the phosphorylated tyrosine sites of the EGFR/erbB/HER family members, such as lapatinib, erlotinib, gefitinib, afatinib, and neratinib, are in use as effective anti–breast cancer targeted therapy. In addition, monoclonal antibodies directed to the extracellular portion of erbB family members, namely pertuzumab and trastuzumab, are used in combination to prevent heterodimerization of Her2/Her3 (23). Upon resistance to these drugs, an inhibitor of CDK12 (cyclin-dependent kinase 12) termed dinaciclib has been proved effective in alleviating resistance to trastuzumab (24). We aimed to inhibit the interactions of PH-signal proteins with the PAR4 by transforming the linear PH-binding motif into an active backbone-cyclic peptide with improved drug properties. Backbone cyclization (BC) has been proved a valuable tool in the rational conversion of active regions of proteins to cyclic peptidomimetic drug leads (25–27). Based on this strategy, we have selected from a mini-library, a potent backbone cyclic peptide PAR4 Pc(4–4) with “drug-like” properties (e.g., in terms of solubility, permeability, metabolic stability affecting oral bioavailability, metabolism, clearance, and toxicity).

In the current article, we have identified a PH-binding motif within PAR4 and assigned Grb2-associated binding proteins (Gab1) and Son of Sevenless 1 (Sos1) association with PAR4. Pc(4–4) compound inhibits PAR4- and PAR2–Akt/PKB association as recapitulated by inhibition of tumor growth in vivo. Stimulation of PAR4 is sufficient to activate EGFR/erbB, whereas Pc(4–4) potently inhibits it. The current study suggests that PAR family member inhibition could be an effective therapy in certain subsets of cancers, especially those displaying high dependence on AKT and/or EGFR.

Animal models

Animals used in experiments were performed in accordance with the guidelines of the institution ethics committee (AAALAC standard). Mice (HSD: athymic nude-Foxn1Nu Nu/Nu mice) were kept under specific pathogen-free (SPF) conditions at the Hadassah Medical Center animal facility unit of the Hebrew University and were regularly screened for standard pathogens. All animal experiments were approved by the animal committee of the Hebrew University (MD-20–15924–5).

Par4 mutant generation

PAR4 C-tail mutation were constructed by replacing E to A (E/A), D to A (D/A), or F to L (F/L) using site-directed mutagenesis strategy. In brief, to mutate E346 with A, the par4-flag wild-type plasmid template was amplified with Q5 high-fidelity polymerase (NEB), using mutant-specific forward primer (5′ TCG GCC GCG TTCAGG GAC 3′) and reverse primer (5′ GTCCCTGAACGCGGCCGA 3′; point mutated codons are underlined in both the primers). The amplified product was purified and subjected to 10 units Dpn-1 treatment (NEB) at 37°C for 1 hour to digest the parental strands completely. Next, Dpn-1 treated amplified product (100 ng) was transformed using XL10-Gold Ultracompetent Cell (Agilent). The transformants colonies were screened by colony PCR using Par4-specific primers and then fidelity of the cloned product was verified by sequencing. Similar approach was applied to generate F347 L mutant, using mutant-specific forward primer (5′ GCC GAG CTC AGG GAC AAG 3′) and reverse primer (5′ CTT GTC CCTGAGCTC GGC 3′). For D349A-mutant, forward primer (5′ GAG TTC AGG GCC AAG GTG CGG 3′) and reverse primer (5′ CCG CAC CTT GGC CCTGAACTC 3′) were used.

Cells and culture conditions

All cell lines were routinely authenticated based on morphology and growth characteristics. HEK293, HCT116, HT29, Lovo, MDA-MB-231, and RKO (obtained from the American Type Culture Collection) were grown in DMEM, supplemented with 1 mmol/L l-glutamine, 50 μg/mL streptomycin, 50 U/mL penicillin (GIBCO-BRL), and 10% fetal calf serum (Biological Industries). Cells were maintained in a humidified incubator with 8% CO2 at 37°C.

Generation of Par4 construct and production of viral particles

To prepare stable clones of PAR4, the human(h)Par4 gene was amplified with Q5 high-fidelity polymerase (NEB), using forward primer (5′GAATTCGCCGCCACCATG TGGGGGCG ACTGCTCC′3) having EcoR1 restriction site (underlined) with Kozak sequences (bold) and the reverse primer (5′ACTAGTTCACTGGAGCAAAGA GGAGTGGG′3) having Spe1 restriction site and cloned it into pLVX- EF1 α-IRES-Puro viral vector. For the generation of viral particles, HEK293 cells were transfected with three plasmids system that includes, pCMVDR8.9, pMD2.VSVG, and Par4-pLVX-EF1α-IRES-Puro vector using PEI as a transfection reagent. Medium was replaced with fresh medium 24 hours later. On day 3 after transfection, medium was collected, and the viral particles were concentrated 10-fold by centrifuging for 1 hour at 40,000 rpm.

Preparation of RKO stable clones expressing Par4

RKO (0.5 × 106) cells were infected with hPar4 10× viral particles along with Polybrene infection reagent. After 72 hours posttransduction, cells were subjected to puromycin selection (0.5 μg/mL). Cells with puromycin resistance were grown, collected, and were either used to isolate RNA or protein lysate preparation.

Cell lysate preparations, immunoprecipitation, and Western blot analysis

To prepare protein cell lysates for immunoprecipitation, cells were solubilized in CelLytic buffer (Sigma-Aldrich). In general, lysis buffer contains 10 mmol/L Tris-HCl (pH 7.4), 150 mmol/L NaCl, 1 mmol/L EDTA, 1% Triton X-100. All utilized lysis buffers were supplemented with protease inhibitor cocktail, 1 mmol/L phenylmethylsulfonylfluoride, PMSF, and 1 mmol/L Na-orthovanadate (Sigma) to prevent protein degradation. Protein cell lysate used for IP, were incubated at 4°C for 20 minutes, and then disrupted by sonication. Finally, soluble supernatant was collected after centrifugation 12,000 rpm for 20 minutes at 4°C.

Protein cell lysates were separated on 8% to 10% SDS-PAGE followed by transfer to Immobilon-P membrane (Millipore). Membranes were blocked and probed with the appropriate primary antibodies accordingly. Primary antibodies that include anti-GFP (Cell Signaling Technology), anti-Akt (Cell Signaling Technology), anti–β-actin (A5441; Sigma-Aldrich), anti-PAR4 (ab5787; Abcam), anti-HA (901503; BioLegend), anti-phosphotyrosine (05–321; Sigma-Aldrich), Gab1 (3232; Cell Signaling Technology), Sos1 (140621; Abcam), anti β-Arrestin (ab206972; Abcam), and anti-EGFR (4405; Cell Signaling Technology) were suspended in 3% BSA in 10 mmol/L Tris-HCl pH 7.5, 100 mmol/L NaCl, and 0.1% Tween-20. After washes, blots were incubated with secondary antibodies conjugated to horseradish peroxidase. Immunoreactive bands were detected by the enhanced chemiluminescence (ECL) reagent (Pierce). Protein cell lysates (400–800 μg) were used for immunoprecipitation (IP) analysis. For IP, anti-flag (sc-166355; Santa Cruz Biotechnology), anti-PAR4 (sc-1666; Santa Cruz Biotechnology), and anti-PAR2 (sc-13504; Santa Cruz Biotechnology) antibodies were added to the cell lysates and processed as described previously (5). The Santa Cruz anti-PAR2 antibody (sc-13504) was used for pulldown, and the Abcam anti-PAR2 antibody (ab180953) was used for detection of PAR2 protein on Western blot.

Quantitative real-time PCR

To determine the overexpression of Par4 in Par4-RKO stable clone, RNA was extracted from wt and Par4-RKO stable clone using GenElute RNA Kit (Sigma-Aldrich). To prepare the cDNA, 1 μg of RNA was reverse transcribed using reverse transcriptase (Promega). Quantitative real-time PCR (qRT-PCR) was conducted using Par4-specific forward primer (5′CCCAGCGTCTACGACGAGA′3) and the reverse primer (5′ CACAGACTTGGCCTGG TAG′3). Hprt gene was used as a house keeping gene for the normalization. In qRT-PCR, triplicates of 6 ng cDNA template was used with 500-nmol/L gene-specific primers using 2× PerfeCTa SYBR Green mix (Agentek) on an automated rotor gene system RG-3000A (Corbett Research). Data obtained from three independent qRT-PCR experiments were analyzed by the 2−ΔΔCt method as described in the manufacturer's instructions and were expressed as fold change over the indicated controls.

Ectopic tumor xenograft mice model study

Par4-RKO stable clones or HCT116 cells (1 × 106 cells) were suspended in medium and inoculated subcutaneously into the right flank of a groups (n = 6) of 6- to 8-week-old Hsd: Athymic Nude-Foxn1nu mice (referred to as nude mice). Pc(4–4) were dissolved in Dulbecco's Phosphate Buffer (Biological Industries) at 100 μmol/L (50 μL) and injected subcutaneous either simultaneously or 4 days after tumor cell injection or 3 weeks after RKO/hPar4 clone/s inoculation. Tumor volumes were monitored twice a week by caliper measurements of each dimension and calculated using the following formula: V  =  4/3π (length/2)(width/2)(depth/2). Mice were terminated by cervical dislocation under aesthetic condition when tumor volumes attained the level of the Institutional Animal Committee approval or when animals showed distress, to avoid unnecessary suffering.

Scratch wound healing assay

Scratch wound healing assay was performed as described previously with some modifications (28). In brief, Lovo cells (3 × 106/well) were seeded in a 6-well plate and cultured overnight to form confluent monolayers. The cell monolayer was replaced with serum-free medium overnight, and a wound was introduced into the monolayer using a pipette tip. Next, cells were treated with different concentrations of cyclic Pc(4–4) peptide (150–300 μmol/L) followed by PAR4 AYPGKF activation (200 μmol/L). Untreated Lovo cells were used as a control. Wound images were photographed using Nikon light microscope.

Matrigel invasion assay

Blind-well chemotaxis chambers with 13-mm diameter filters were used for this assay. Polyvinylpyrrolidone-free polycarbonate filters, 8 μm pore size (Costar Scientific Co.), were coated with basement membrane Matrigel (50 μL of 1 mg/mL Matrigel applied per blind well; 50 μg/filter) as described previously (11). Matrigel was diluted to the desired final concentration with cold distilled water, applied onto the filters, and dried under a hood. Lovo cells (1.8 × 105), suspended in DMEM containing 0.1% bovine serum albumin (BSA), were added to the upper chamber. Conditioned medium of 3T3 fibroblasts was applied as chemoattractant and placed in the lower compartment of the Boyden chamber. Next, cells were treated with either 200 μmol/L of AYPGKF alone or 200 μmol/L of AYPGKF with varying concentration of cyclic Pc(4–4) peptide (50 nmol/L–150 μmol/L). Cells without any treatment were used as control. The Matrigel chambers were incubated overnight in a humidified incubator with 8% CO2 at 37°C. At the end of the incubation, the cells on the upper surface of the filter were removed by wiping with a cotton swab. The filters were fixed and stained using hematoxylin and eosin (Sigma-Aldrich). Cells from various areas of the bottom side membrane surface were counted. The assay was performed in triplicates and were carried out three times.

Cell transfections and PAR activation

Cells grown to 70% to 80% confluency were transfected with 0.5 to 1 μg amount of plasmid DNA using PEI transfection reagent (Polysciences) according to the manufacturer's instructions. Cells were collected 48 hours after transfection and protein lysates/RNA were prepared. To activate PAR2, a ligand of hexapeptide SLIGKV (GenScript) was used. To activate PAR4, a ligand of hexapeptide AYPGKF (GenScript) was applied.

PI3K inhibition

Wortmanin (Sigma Aldrich cat#W1638), a PI3K inhibitor, was used.

Immunohistochemistry

Paraffin-embedded slides derived from tumors generated either by RKO/PAR4 stable clones or HCT116 cells following Pc(4–4) treatment or not as also human breast tissue section [stages of Her2/Neu and triple-negative (TN); Prof. Uziely, Head Oncology Ambulatory Services Unit, Hadassah Medical Center, Jerusalem; Helsinki 0163–16-HMO] and human colon tissue sections (Prof. Ayala Hubert) were used for IHC. After deparaffinization and rehydration, the slides were incubated with 3% H2O2 prior to antigen retrieval. Antigen unmasking was carried out by heating (20 minutes) in a microwave oven with 1× antigen retrieval citrate buffer (Cat# ab93678, Abcam). After blocking with CAS-Block (Cat# 008120, Invitrogen), slides were incubated with the following primary antibodies: Ki-67 (cat# ab16667; Abcam), cleaved caspase-3 (cat# ab2302; Abcam), PAR4 (cat# ab5787; Abcam), and PAR2 (Cat #APR-032; Alomone Labs). Next, following washing, the slides were incubated with peroxidase conjugated antibodies (Abcam). Color was developed using DAB Substrate Kit (Cat# 34002, Thermo Fisher Scientific), followed by counter staining with Mayer's hematoxylin (Cat# 3801582E, Leica Biosystems). Controls using only secondary abs (with no primary antibodies) showed low to background staining in all cases.

Plasmids and reagents

Human PAR2 (hPar2/f2rl1) plasmid was kindly provided by Dr. Morley D. Hollenberg (Faculty of Medicine, University of Calgary, Alberta, Canada). pLVX-EF1α-IRES-Puro, pCMVDR8.9, and pMD2.VSVG plasmids were kind gift from Dr. Tzvi Granot (Hebrew University). pBJ-FLAG-hPar4 (cat #53231), PH-Akt-GFP (cat#51465), PH-Akt (R25C)-GFP (cat #51466), HA-Sos1 (cat #32920), and EGFR-GFP (cat #32751) plasmids were purchased from Addgene. All mentioned plasmids were sequenced. Details of plasmids are available on request.

Statistical analysis

All the experiments were carried out in triplicates, whereby the data are represented as mean ± SD. The significant difference of tested samples in comparison with control was determined by performing either Student t test or analysis of variance (ANOVA) with Tukey multiple comparison posttest (GraphPad Prism 6.0) wherever required. The criterion for statistical significance was as follows: P < 0.05 was considered significant, P < 0.01 as highly significant, and P < 0.001 as very highly significant.

Data availability

The data generated in this study are available within the article and its Supplementary Data files. Kaplan–Meier plots are publicly available at Kaplan–Meier plotter server (http://kmplot.com/analysis/index.php?p=background) with Affymetrix ID F2RL1 (206429_at), F2RL3 (207221_at), and (297221_at).

PAR4 associates with Akt/PKB: identification of PAR4-binding region

Previously, we demonstrated PH-binding motifs within PAR1 and PAR2 C-tails as sites for drug development (12); we now ask whether this domain also resides in PAR4. Akt/PKB, a serine/threonine protein kinase that plays a pivotal role in tumor cell survival, proliferation, and invasiveness (29–31), was found capable of associating with PAR4, as shown by immunoprecipitation (IP) analyses. This association takes place regardless of whether analyzed in HEK293 cells (Fig. 1A), following transient transfection of flg-hPar4 or via endogenous association between PAR4 and Akt, subsequent to AYPGKF, PAR4 activation in a colon cancer cell line HCT116 (Fig. 1B). The synthetic hexapeptide, AYPGKF, mimics the thrombin-unmasked tethered ligand of PAR4 to cause receptor activation (32, 33). IP of cell lysates using either anti-flg or anti-PAR4 abs, respectively, analyzed by Western blots for PAR4 and Akt/PKB coassociation were carried out. This association was seen between 5 and 30 minutes of PAR4 activation (Fig. 1A and B). Next, we aimed to identify the region within PAR4 C-tail that mediates this association. Three consecutive synthetic peptides representing PAR4 C-tail were prepared whereby each peptide contained two overlapping amino acids of the upstream peptide. The peptides were designated, C-tail PAR4-peptides (CTP4) A-C; CTP4-A; CTP4-B; CTP4-C. MDA-MB-231 cells were analyzed for Matrigel invasion assay following AYPGKF PAR4 activation, in the presence or absence of the peptides, CTP4A-C. It showed unequivocally that only CTP4-A peptide effectively inhibited PAR4-induced Matrigel invasion, while CTP4-B and CTP4-C had no effect (Fig. 1C and D). This result was supported by IPs between PAR4 and Akt/PKB in the presence and absence of CTP4-A peptide and CTP4-B (Fig. 1E). Specific association between PAR4 and Akt/PKB was seen either in wt PAR4 or in the presence of CTP-B. In contrast, peptide CTP4-A inhibited PAR4–Akt/PKB association is (Fig. 1E). Certainly, as in PAR1 and PAR2, the PH-binding motif of PAR4 resides in the membrane proximal region of PAR4 C-tail (Fig. 1F). PAR2 and PAR4 play a significant role in cancer of gastrointestinal tract (GI) as manifested by the statically significant Kaplan–Meier curve. A lower survival time is seen in the presence of high PAR2/f2rl1 and PAR4/f2rl3 as compared with low expression of either PAR (Fig. 1G).

Figure 1.

Akt/PKB associates with PAR4. A, Immunoprecipitation (IP) analysis of PAR4 and Akt/PKB. HEK293 cells were transfected with wt flg-hPar4, and IP was performed following PAR4 AYPGKF activation using anti-flg antibodies and immunoblotting with anti-Akt antibodies. B, HCT116 cells were activated by AYPGKF for PAR4, and IP was performed by anti-PAR4 Abs and immunoblotting with anti-Akt antibodies. This is a representative of three times performed experiments. C, Matrigel invasion in the presence and absence of the C-tail–derived PAR4 peptides; CTP4 A-C were carried out in MDA-MB-231 cells. While AYPGKF activation of PAR4 induces Matrigel invasion of MDA-MB-231 cells, it was attenuated in the presence of CTP-A. In contrast, no effect was observed in the presence of peptides CTP-B or C, similar to nontreated activated parental MDA-MB-231 cells. D, Histograms represent quantification of the cells/HPF that invaded the Matrigel layer. Unpaired Student t test was used. This experiment is a representative of three independent experiments performed in triplicates. E, The PAR4-derived peptide CTP4-A potently inhibits PAR4–Akt association. HEK293 cells were transfected with wt flg-hPar4 followed by AYPGKF activation of PAR4. IP was performed using anti-flg antibodies of cells treated or not with CTP-A peptide and immunoblotting with anti-Akt antibodies. This experiment is a representative of three independent experiments performed. F, Schematic presentation. PAR1, PAR2, and PAR4 C-tails showing the site of the PH-binding domain. G, Kaplan–Meier curves for PAR2 and PAR4 in GI cancer. Survival time in high versus low expression levels of PAR2/f2rl1 and PAR4/f2rl3.

Figure 1.

Akt/PKB associates with PAR4. A, Immunoprecipitation (IP) analysis of PAR4 and Akt/PKB. HEK293 cells were transfected with wt flg-hPar4, and IP was performed following PAR4 AYPGKF activation using anti-flg antibodies and immunoblotting with anti-Akt antibodies. B, HCT116 cells were activated by AYPGKF for PAR4, and IP was performed by anti-PAR4 Abs and immunoblotting with anti-Akt antibodies. This is a representative of three times performed experiments. C, Matrigel invasion in the presence and absence of the C-tail–derived PAR4 peptides; CTP4 A-C were carried out in MDA-MB-231 cells. While AYPGKF activation of PAR4 induces Matrigel invasion of MDA-MB-231 cells, it was attenuated in the presence of CTP-A. In contrast, no effect was observed in the presence of peptides CTP-B or C, similar to nontreated activated parental MDA-MB-231 cells. D, Histograms represent quantification of the cells/HPF that invaded the Matrigel layer. Unpaired Student t test was used. This experiment is a representative of three independent experiments performed in triplicates. E, The PAR4-derived peptide CTP4-A potently inhibits PAR4–Akt association. HEK293 cells were transfected with wt flg-hPar4 followed by AYPGKF activation of PAR4. IP was performed using anti-flg antibodies of cells treated or not with CTP-A peptide and immunoblotting with anti-Akt antibodies. This experiment is a representative of three independent experiments performed. F, Schematic presentation. PAR1, PAR2, and PAR4 C-tails showing the site of the PH-binding domain. G, Kaplan–Meier curves for PAR2 and PAR4 in GI cancer. Survival time in high versus low expression levels of PAR2/f2rl1 and PAR4/f2rl3.

Close modal

The PH-domain of Akt/PKB associates with PAR4

To elucidate whether Akt/PKB binds via its PH-domain, we have applied constructs of either wt Akt/PKB PH-domain alone or an Akt/PKB-PH domain mutant R25C, impaired in its lipid binding capability of Akt/PKB-PH domain. Transient transfections of flg-hPar4 and either wt GFP-PH-Akt or PH-domain R25C mutant were carried out in HEK293 cells. Cells were activated by the PAR4 AYPGKF ligand for the indicated periods of time and the lysates were further processed for IP analysis. Binding of GFP-Akt/PKB-PH domain module alone with the PAR4 was obtained (Fig 2A). In contrast, the R25C Akt-PH domain mutant, failed to associate with PAR4 (Fig. 2B), indicating the requirement of a lipid moiety for PAR4–Akt association. Indeed, the application of a PI3K-inhibitor; wortmannin at 5 μmol/L (34), inhibited the otherwise observed association between PAR4 and Akt/PKB (Fig. 2C). Hence, the phosphatidylinositol (3, 4, 5)-trisphosphate lipid is essential for the complex formation. We conclude that the Akt/PKB–PAR4 binding association involves membrane lipid anchoring via the PAR4-PH–binding domain.

Figure 2.

Akt/PKB associates with PAR4 via its PH domain. A, PH-Akt module binds PAR4 C-tail but not mutant R25C. HEK 293 cells were transiently transfected with flg-hPar4 construct and either with GFP-PH-Akt domain alone or GFP-R25C. IP of cell lysates following PAR4 AYPGKF activation was carried out using anti-flg antibodies. Detection of either Akt-PH domain alone associated with PAR4 was performed with anti-GFP. GFP-PH domain alone was shown to bind PAR4; no binding was obtained when the mutant R25C of low lipid-binding affinity was present (B). This experiment is a representative of three independent experiments performed. C, The PI3K inhibitor wortmannin inhibits binding of Akt to PAR4. Treatment of HEK293 cells with wortmannin (5 μmol/L) for various time points or not, following transient transfection with flg-hPar4 and AYPGKF activation was performed. Cell lysates were immunoprecipitated with anti-flg antibodies, and anti-Akt was used to assess the association of Akt with the PAR4 C-tail. A potent inhibition of Akt-PAR4 association was observed in the presence of wortmannin. This experiment is a representative of three independent experiments performed in triplicates. D, The F&D amino acids as part of the "FRD" sequence within PAR4–PH binding domain are essential for PAR4-Akt association. Amino acid residues of PAR4- and PAR2 -PH binding motifs. E, Mutations inserted to flg-hPar4 generated flg-hPar4MutE346A. HEK293 cells were transiently transfected with either wt flg-hPar4 or flg-hPar4MutE346A. IP of cell lysates following PAR4 activation was carried out using anti-flg antibodies. Detection of PAR4–Akt association was performed by anti-Akt Abs. While association with wt hPar4 is seen after 5- and 10-minute AYPGKF PAR4 activation, upon using flg-hPar4 E346A, no inhibition is seen and the association with Akt is observed at 30-minute activation. This experiment is a representative of three independent experiments performed in triplicates. F, Mutations inserted to flg-hPar4 generated flg-hPar4Mut F347A and flg-hPar4Mut D349A. HEK293 cells were transiently transfected with either wt flg-hPar4 or flg-hPar4Mut F347A or flg-hPar4Mut D349A. IP of cell lysates following PAR4 activation was carried out using anti-flg antibodies. Detection of PAR4–Akt association was performed by anti-Akt Abs. While association with wt hPar4 is seen after 30 minutes of AYPGKF PAR4 activation, no such binding is obtained when using either flg-hPar4 F347A or flg-hPar4D349A. This experiment is a representative of three independent experiments performed in triplicates.

Figure 2.

Akt/PKB associates with PAR4 via its PH domain. A, PH-Akt module binds PAR4 C-tail but not mutant R25C. HEK 293 cells were transiently transfected with flg-hPar4 construct and either with GFP-PH-Akt domain alone or GFP-R25C. IP of cell lysates following PAR4 AYPGKF activation was carried out using anti-flg antibodies. Detection of either Akt-PH domain alone associated with PAR4 was performed with anti-GFP. GFP-PH domain alone was shown to bind PAR4; no binding was obtained when the mutant R25C of low lipid-binding affinity was present (B). This experiment is a representative of three independent experiments performed. C, The PI3K inhibitor wortmannin inhibits binding of Akt to PAR4. Treatment of HEK293 cells with wortmannin (5 μmol/L) for various time points or not, following transient transfection with flg-hPar4 and AYPGKF activation was performed. Cell lysates were immunoprecipitated with anti-flg antibodies, and anti-Akt was used to assess the association of Akt with the PAR4 C-tail. A potent inhibition of Akt-PAR4 association was observed in the presence of wortmannin. This experiment is a representative of three independent experiments performed in triplicates. D, The F&D amino acids as part of the "FRD" sequence within PAR4–PH binding domain are essential for PAR4-Akt association. Amino acid residues of PAR4- and PAR2 -PH binding motifs. E, Mutations inserted to flg-hPar4 generated flg-hPar4MutE346A. HEK293 cells were transiently transfected with either wt flg-hPar4 or flg-hPar4MutE346A. IP of cell lysates following PAR4 activation was carried out using anti-flg antibodies. Detection of PAR4–Akt association was performed by anti-Akt Abs. While association with wt hPar4 is seen after 5- and 10-minute AYPGKF PAR4 activation, upon using flg-hPar4 E346A, no inhibition is seen and the association with Akt is observed at 30-minute activation. This experiment is a representative of three independent experiments performed in triplicates. F, Mutations inserted to flg-hPar4 generated flg-hPar4Mut F347A and flg-hPar4Mut D349A. HEK293 cells were transiently transfected with either wt flg-hPar4 or flg-hPar4Mut F347A or flg-hPar4Mut D349A. IP of cell lysates following PAR4 activation was carried out using anti-flg antibodies. Detection of PAR4–Akt association was performed by anti-Akt Abs. While association with wt hPar4 is seen after 30 minutes of AYPGKF PAR4 activation, no such binding is obtained when using either flg-hPar4 F347A or flg-hPar4D349A. This experiment is a representative of three independent experiments performed in triplicates.

Close modal

Critical roles for the PAR4 C-tail amino acids; F347L and D349A within the PH-binding motif

Analysis of both PAR2 and PAR4 PH-binding domains reveals an identical core of three amino acids; Phe-Arg-Asp (FRD) which appears in both (Fig. 2D). This sequence is critical for the association between PAR4 PH-binding motif and PH-signal proteins. A mutation inserted in the amino acid E346A outside the “FRD” core sequence did not affect the interaction of PAR4-Akt/PKB (Fig. 2E), yet delayed it as compared with the wt PAR4–Akt/PKB interaction with any functional consequence. In contrast, mutations introduced within the “FRD” core sequence abrogated the association with PAR4-PH-Akt/PKB. Mutant constructs of hPar4 F347 L and hPar4 D349A were prepared and transiently transfected to HEK293 cells. While PAR4–Akt form a complex, neither mutants showed association with Akt/PKB (Fig. 2F). It is concluded that both F and D amino-acids are necessary and required for PAR4 and PH–Akt association.

Gab1 and SOS1 associate with PAR4

Throughout evaluations of other PH-signal proteins that play a role in pro-tumor pathways, we examined, Gab1 for possible association with PAR4. Grb2-associated binding proteins (Gab) forms a family of three members; Gab1, Gab2, and Gab3 in mammals. These are well-known scaffold proteins engaged in protein–protein interactions and considered as signal “amplifiers” in transduction of multiple signal pathways in cancer (35–38). An association between PAR4 and Gab1is seen after 30 minutes of PAR4 activation (Fig. 3A). In parallel, we have assessed the potential binding properties between Sos1; Son of Sevenless 1 and PAR4. Sos1 is a guanine nucleotide exchange factor (GEF). The Sos1 protein is composed of; a Dbl homology (DH) domain and a PH-domain. When sos1 and flg-hPar4 constructs were introduced to HEK293 cells followed by PAR4 AYPGKF activation, a robust association between PAR4 and SOS1 was observed (Fig. 3B). Mutation of either essential PAR4 amino-acid (e.g., D349A and F347L), failed to bind Sos1 (Fig. 3C). To confirm PAR4 and Sos1 association, endogenous PAR4 and Sos1 immune-complex detection was performed in HT29 colon cancer cell line (Fig. 3D). HT29 cells were serum starved for 24 hours, followed by AYPGKF PAR4 activation. A strong association between PAR4 and Sos1 was observed within 45 minutes of activation up to 1 hour analyzed (Fig. 3D). The possibility that a guanine nucleotide exchange adaptor protein associates directly with the PAR4 C-tail, may illuminate on the direct cross-talk between receptor tyrosine kinases (RTK), EGFR/erbB and PAR4.

Figure 3.

Other PH-binding signal proteins associate with PAR4. A, PAR4–GAB1 association. HEK293 cells were transiently cotransfected with flg-hPar4 and gab1. IP of cell lysates following PAR4 AYPGKF activation was carried out using anti-flg. Detection of Gab1 was performed using anti-Gab1 antibodies. This experiment is a representative of three independent experiments performed in triplicates. B, SOS1–PAR4 association. HEK293 cells were transiently cotransfected with flg-hPar4 construct and sos1. IP of cell lysates following PAR4 AYPGKF activation was carried out using anti-flg and detected by anti-SOS1 abs. This experiment is a representative of three independent experiments performed in triplicates. C, PAR4 PH-binding domain mutants, D349A and F347L, abolish association with Sos1. HEK293 cells were transiently cotransfected with either flg-hPar4 or flg-hPar4D349A or flg-hPar4F347 L with sos1 plasmid. IP of cell lysates following PAR4 AYPGKF activation was carried out using anti-flag and detected by anti-HA abs. This experiment is a representative of three independent experiments performed in triplicates. This experiment is a representative of three independent experiments performed in triplicates. D, Endogenous SOS1 associates with PAR4. HT29 colon cancer cell line exhibiting high PAR4 levels was AYPGKF activated for PAR4. IP was carried out using anti-PAR4, and SOS1 detection was performed using anti-SOS1 abs.

Figure 3.

Other PH-binding signal proteins associate with PAR4. A, PAR4–GAB1 association. HEK293 cells were transiently cotransfected with flg-hPar4 and gab1. IP of cell lysates following PAR4 AYPGKF activation was carried out using anti-flg. Detection of Gab1 was performed using anti-Gab1 antibodies. This experiment is a representative of three independent experiments performed in triplicates. B, SOS1–PAR4 association. HEK293 cells were transiently cotransfected with flg-hPar4 construct and sos1. IP of cell lysates following PAR4 AYPGKF activation was carried out using anti-flg and detected by anti-SOS1 abs. This experiment is a representative of three independent experiments performed in triplicates. C, PAR4 PH-binding domain mutants, D349A and F347L, abolish association with Sos1. HEK293 cells were transiently cotransfected with either flg-hPar4 or flg-hPar4D349A or flg-hPar4F347 L with sos1 plasmid. IP of cell lysates following PAR4 AYPGKF activation was carried out using anti-flag and detected by anti-HA abs. This experiment is a representative of three independent experiments performed in triplicates. This experiment is a representative of three independent experiments performed in triplicates. D, Endogenous SOS1 associates with PAR4. HT29 colon cancer cell line exhibiting high PAR4 levels was AYPGKF activated for PAR4. IP was carried out using anti-PAR4, and SOS1 detection was performed using anti-SOS1 abs.

Close modal

Design and synthesis of the PAR4 backbone cyclic peptide mini-library

The design of the PAR4 backbone cyclic peptide mini-library followed the Cycloscan scheme that was described previously (27). Analysis of the PAR2 and PAR4 PH-domain linear peptides led to assume that amino acids Ala and Val in the PAR4 sequence are non-pharmacophoric and can be replaced by the N-Alkylated-Gly (Gly building units used for the backbone cyclization).

  • PAR2: Ser-His-Asp-Phe-Arg-Asp-His-Ala

  • PAR4 Ser-Ala-Glu-Phe-Arg-Asp-Lys-Val

The structure of the designed library is shown in Supplementary Fig. S1. It is composed of three backbone cyclic peptides of PAR4, Pc(2–2), Pc(4–4), and Pc(6–6). All members of the library have the same pharmacophoric sequence. They differ in the ring size (26, 30, and 34 atoms, respectively). We assumed that the different ring size will alter the bioactive conformation, and thus will allow to select the most active inhibitory peptide.

Selection of a lead backbone cyclic peptide from the mini-library, directed to inhibit PAR4 PH-binding domain

The mini-library components are illustrated (Supplementary Fig. S1A and S1B). Application of Pc(2–2) inhibitor of smaller ring size showed no inhibition of the otherwise noticeable association between PAR4 and PH-Akt/PKB, even at 200 μmol/L concentration (Supplementary Fig. S1C). Similarly, as shown in Supplementary Fig. S1D, application of Pc(6–6) inhibitor, did not affect the association between PAR4 and Akt/PKB. When we applied different concentrations of Pc(4–4) prior to the IP analysis, we observed a strong inhibition of Akt/PKB–PAR4 association between 50 nmol/L to 150 μmol/L (Fig. 4A; Supplementary Fig. S1E).

Figure 4.

PAR4c(4–4) inhibits PAR4–Akt/PKB, PAR2–Akt/PKB association, Matrigel invasion, and migration. A, PAR4–Akt association. HEK293 cells were transiently transfected with flg-hPar4 followed by AYPGKF PAR4 activation in the presence and absence of Pc(4–4) within the range of 50 nmol/L to 150 μmol/L. IP of cell lysates following different periods of activation was carried out using anti-flg abs. Detection of Akt was performed by anti-Akt abs. β-Actin serves as a control gene for loading. The experiment was carried out three times. B, Matrigel invasion of Lovo colon cancer cells. Matrigel invasion before and after AYPGKF PAR4 activation and in the presence and absence of PAR4c(4–4) inhibitor was carried out. A range between 50 nmol/L and 150 μmol/L PAR4c(4–4) inhibitor was applied. While ample Matrigel invasion is seen following PAR4 activation, this was attenuated in the presence of the inhibitor at all the concentrations analyzed, similar to control prior to PAR4 activation. This experiment is a representative of three independent experiment performed in triplicates. C, Histograms represent quantification of the cells/HPF that invaded the Matrigel layer. Unpaired Student t test was used. D, Wound-scratch migration assay. Lovo cancer cell line was grown to confluence. Then, an equal wound scratch was introduced to the cell monolayers. Monolayers were serum starved overnight, and PAR4 was activated or not with AYPGKF. One hour prior to activation of PAR4c(4–4), either 150 or 300 μmol/L concentrations was added to the monolayers. While the gap in the scratch started to be filled 24 hours after PAR4 activation, in the presence of Pc(4–4) inhibitor, a marked inhibition is observed. This experiment is a representative of three independent experiment performed in triplicates. E, PAR2–Akt association is inhibited by Pc(4–4). HEK293 cells were transiently transfected with hPar2 followed by SLIGKV PAR2 activation in the presence and absence of PAR4c(4–4) at 150 μmol/L. IP of cell lysates following different periods of activation was carried out using anti-PAR2 abs. Detection of Akt was performed by anti-Akt abs. β-Actin serves as a control gene for loading. The experiment was carried out three times. F, Endogenous PAR4–Akt association is inhibited by Pc(4–4). HCT116 colon cancer cells were AYPGKF PAR4 activated for different periods of time, in the presence and absence of Pc(4–4) inhibitor. IP of cell lysates was carried out using anti-PAR4 abs. Detection of Akt (cell signaling Abs) showed association with PAR4. PAR4 in the immune-complex indicates equal amounts immunoprecipitated. The experiment was carried out three times. G, Pc(4–4) inhibits endogenous PAR4–β-arrestin association. HCT116 colon cancer cells were AYPGKF PAR4 activated for different periods of time, in the presence and absence of Pc(4–4) inhibitor. IP of cell lysates was carried out using anti-PAR4 abs. Detection of β-arrestin (abcam Abs) showed levels of association with PAR4. PAR4 in the immune-complex indicates equal amounts immunoprecipitated. This is a representative of three times experiments performed.

Figure 4.

PAR4c(4–4) inhibits PAR4–Akt/PKB, PAR2–Akt/PKB association, Matrigel invasion, and migration. A, PAR4–Akt association. HEK293 cells were transiently transfected with flg-hPar4 followed by AYPGKF PAR4 activation in the presence and absence of Pc(4–4) within the range of 50 nmol/L to 150 μmol/L. IP of cell lysates following different periods of activation was carried out using anti-flg abs. Detection of Akt was performed by anti-Akt abs. β-Actin serves as a control gene for loading. The experiment was carried out three times. B, Matrigel invasion of Lovo colon cancer cells. Matrigel invasion before and after AYPGKF PAR4 activation and in the presence and absence of PAR4c(4–4) inhibitor was carried out. A range between 50 nmol/L and 150 μmol/L PAR4c(4–4) inhibitor was applied. While ample Matrigel invasion is seen following PAR4 activation, this was attenuated in the presence of the inhibitor at all the concentrations analyzed, similar to control prior to PAR4 activation. This experiment is a representative of three independent experiment performed in triplicates. C, Histograms represent quantification of the cells/HPF that invaded the Matrigel layer. Unpaired Student t test was used. D, Wound-scratch migration assay. Lovo cancer cell line was grown to confluence. Then, an equal wound scratch was introduced to the cell monolayers. Monolayers were serum starved overnight, and PAR4 was activated or not with AYPGKF. One hour prior to activation of PAR4c(4–4), either 150 or 300 μmol/L concentrations was added to the monolayers. While the gap in the scratch started to be filled 24 hours after PAR4 activation, in the presence of Pc(4–4) inhibitor, a marked inhibition is observed. This experiment is a representative of three independent experiment performed in triplicates. E, PAR2–Akt association is inhibited by Pc(4–4). HEK293 cells were transiently transfected with hPar2 followed by SLIGKV PAR2 activation in the presence and absence of PAR4c(4–4) at 150 μmol/L. IP of cell lysates following different periods of activation was carried out using anti-PAR2 abs. Detection of Akt was performed by anti-Akt abs. β-Actin serves as a control gene for loading. The experiment was carried out three times. F, Endogenous PAR4–Akt association is inhibited by Pc(4–4). HCT116 colon cancer cells were AYPGKF PAR4 activated for different periods of time, in the presence and absence of Pc(4–4) inhibitor. IP of cell lysates was carried out using anti-PAR4 abs. Detection of Akt (cell signaling Abs) showed association with PAR4. PAR4 in the immune-complex indicates equal amounts immunoprecipitated. The experiment was carried out three times. G, Pc(4–4) inhibits endogenous PAR4–β-arrestin association. HCT116 colon cancer cells were AYPGKF PAR4 activated for different periods of time, in the presence and absence of Pc(4–4) inhibitor. IP of cell lysates was carried out using anti-PAR4 abs. Detection of β-arrestin (abcam Abs) showed levels of association with PAR4. PAR4 in the immune-complex indicates equal amounts immunoprecipitated. This is a representative of three times experiments performed.

Close modal

When we analyzed the impact of Pc(4–4) on the capability of activated PAR4 in Lovo colon cancer cell line to invade Matrigel, a robust invasion is seen (Fig. 4B), inhibited in the presence of Pc (4–4). This is observed within the range of 50 nmol/L to 150 μmol/L (Fig. 4B and C). Migration was evaluated by “wound scratch assay” applied. An equal wide scratch was introduced to the monolayer cultures after overnight “cell starvation”. Application of the Pc(4–4) or not were carried out on AYPGKF PAR4 activated monolayers. Consequently, after 24 hours, the wound site is nearly filled up. In contrast, in the presence of 150 and 300 μmol/L of Pc(4–4), a marked inhibition of migration/proliferation is seen and the scratch remained nearly intact (Fig. 4D). Based on the similarity between PAR4 and PAR2 PH-binding domains (Fig. 2E), we analyzed whether Pc(4–4) is capable of inhibiting also the PAR2–Akt/PKB association. HEK293 cells were transiently transfected with hPar2 and were SLIGKV PAR2 activated, in the presence and absence of the Pc(4–4) compound. A strong inhibition of PAR2–Akt/PKB association was obtained as shown in Fig 4E. Hence, the Pc(4–4) is a potent inhibitor also for PAR2, the dominant member of PAR family (7, 14). Importantly, Pc(4–4) effectively inhibits endogenous PAR4–Akt association as evaluate in HCT116 cells (Fig. 4F). It appears that Pc(4–4) effectively inhibits also PAR4–β-arrestin associations (Fig. 4G), although β-arrestin binding site in PAR4 (e.g., RAGLSQRS) resides immediately downstream to the PH-binding motif. It acts similar to “RAG8” peptide that blocks PAR4–β-arrestin associations (39).

Effect of Pc(4–4) on tumor growth, in vivo

Having demonstrated the inhibitory effect of Pc(4–4), as preventing the association between Akt/PKB and PAR2&4, leading to inhibited migration and invasion in vitro, we next set out to evaluate its effect on tumor growth, in vivo. The physiologic significance of PAR4 PH-binding motif in vivo is demonstrated using a xenograft mouse model of subcutaneously inoculated RKO/Par4 clone/s overexpressing PAR4 (Fig. 5A and B). Toward this, we have generated stable clones (sc) overexpressing PAR4 in RKO cells, a colon epithelial cell-line (expressing wt p53). Levels of PAR4 in the generated RKO stable clones are shown (Fig. 5C). For xenograft tumor formation, we have utilized two types of cell lines, the HCT116 cells, an aggressive colon cancer cell line, overexpressing oncogenes among of which is also PAR4, and clone/s overexpressing hPar4, generated in RKO line; RKO/hPar4 cells. HCT116 (1 × 106) and RKO/hPar4 (1 × 106) clone/s were inoculated subcutaneously in nude mice. Pc(4–4) cyclic peptide was injected at the site of the tumor (100 μmol/L; 0.25 mg/30-gram mouse), either at the time of tumor cell injection or 4 days after. The Pc(4–4) cyclic peptide was applied repeatedly 3 times per week. The HCT116 cell inoculation generated large tumors and were terminated after 20 days, whereas the RKO/hPar4 inoculation developed tumors on a slower pace and were terminated after 32 days. Large tumors were generated in both inoculated lines, and significantly smaller tumors were observed in the presence of Pc(4–4) (Fig. 5A and B; Supplementary Data; Supplementary Fig. S2). The inhibition of tumor growth was seen regardless whether the Pc(4–4) was applied at the same time of tumor inoculation or 4 days later. This was obtained irrespective of using an aggressive colon cancer cell line (Supplementary Data; Supplementary Fig. S2), or stable RKO/hPar4 clone/s, overexpressing PAR4 (Fig. 5A and B).

Figure 5.

PAR4c(4–4) potently inhibits RKO/hPar4 clone/s induced tumor generation. A, A clone of RKO/PAR4 overexpressing cells (1 × 106) was injected subcutaneously into nude mice. PAR4c(4–4) (100 μmol/L; 0.25 mg/30 gm) was subcutaneously injected at the site of the tumor either on the same day of RKO/PAR4 clone inoculation or 4 days later. The inhibitor was applied repeatedly 3 times/week for 3 weeks. Mice were weighted every 3 days, and tumor size was measured. Mice were sacrificed after 32 days. B, Tumor volume measurements of RKO/PAR4 clone inoculated into nude mice. Tumors were weighed and measured for size at the indicated time points, and tumor volume was calculated. Error bars, SD; *, P < 0.05. C, qPCR of a representative RKO stable clone (sc) of PAR4. Error bars, SD; ***, P < 0.001. This experiment is a representative of three independent experiment performed in triplicates. Each group of mice treatment represents n = 7 mice. D, IHC of RKO/PAR4–derived tumors. Immunohistochemistry of RKO compared with small tumors in the presence of PAR4c(4–4) inhibitor. Levels of proliferation in the large and small tumors were indicated by applying anti-ki67 abs (a and b). For apoptosis, application of active caspase-3 Abs were applied (c and d). While high levels of proliferation are seen in the large tumors (a) and little in the small tumors (b), high levels of active caspase-3 in the small tumors (c) were obtained as compared with the very little in the large tumors. This experiment is a representative of three independent experiment performed in triplicates. E, Histograms representing positive IHC in each treatment group. Error bars, SD; *, P < 0.05; ***, P < 0.001. F, H&E staining. Hematoxylin and eosin staining of large (left) and small (right) tumors distinctly indicate to the appearance of a wide basement membrane in the small tumors, whereas a thinner and cell-invaded basement membrane is observed in large tumors.

Figure 5.

PAR4c(4–4) potently inhibits RKO/hPar4 clone/s induced tumor generation. A, A clone of RKO/PAR4 overexpressing cells (1 × 106) was injected subcutaneously into nude mice. PAR4c(4–4) (100 μmol/L; 0.25 mg/30 gm) was subcutaneously injected at the site of the tumor either on the same day of RKO/PAR4 clone inoculation or 4 days later. The inhibitor was applied repeatedly 3 times/week for 3 weeks. Mice were weighted every 3 days, and tumor size was measured. Mice were sacrificed after 32 days. B, Tumor volume measurements of RKO/PAR4 clone inoculated into nude mice. Tumors were weighed and measured for size at the indicated time points, and tumor volume was calculated. Error bars, SD; *, P < 0.05. C, qPCR of a representative RKO stable clone (sc) of PAR4. Error bars, SD; ***, P < 0.001. This experiment is a representative of three independent experiment performed in triplicates. Each group of mice treatment represents n = 7 mice. D, IHC of RKO/PAR4–derived tumors. Immunohistochemistry of RKO compared with small tumors in the presence of PAR4c(4–4) inhibitor. Levels of proliferation in the large and small tumors were indicated by applying anti-ki67 abs (a and b). For apoptosis, application of active caspase-3 Abs were applied (c and d). While high levels of proliferation are seen in the large tumors (a) and little in the small tumors (b), high levels of active caspase-3 in the small tumors (c) were obtained as compared with the very little in the large tumors. This experiment is a representative of three independent experiment performed in triplicates. E, Histograms representing positive IHC in each treatment group. Error bars, SD; *, P < 0.05; ***, P < 0.001. F, H&E staining. Hematoxylin and eosin staining of large (left) and small (right) tumors distinctly indicate to the appearance of a wide basement membrane in the small tumors, whereas a thinner and cell-invaded basement membrane is observed in large tumors.

Close modal

IHC of large and small tumors have indicated increased proliferation in the large tumors as compared with little in the small tumors generated in the presence of Pc(4–4), evaluated by ki67 staining (Fig. 5Da, b; E). Accordingly, induced levels of active caspase-3 for apoptosis is seen in the small tumors while nearly none or very little in the large tumors obtained (Fig. 5Dc, d; E). H&E staining shows basement membrane in large and small tumors (Fig. 5F).

Pc(4–4) inhibited significantly preformed tumors

In this experimental design, we have implanted RKO/hPar4 clone/s (1 × 106 cells) in nude mice (Par4 levels in the clones, Supplementary Fig. S3A–S3C). Tumors were slowly formed (during 3 weeks). At the time, tumors were distinctly observed (∼1 cm length, Supplementary Fig S3C), Pc(4–4) was injected subcutaneously in the vicinity of the tumor site (3× week; 5 mg/kg). The experiment was terminated when we observed large RKO/hPar4 tumors. In contrast, distinctly smaller tumors were seen in the presence of Pc(4–4) (Supplementary Fig. S3A and S3D).

Activation of PAR4 leads to the phosphorylation of EGFR

AYPGKF activation of PAR4 leads to pTyr-EGFR. This was shown by cotransfection of HEK293 cells with hPar4 and egfr followed by AYPGKF PAR4 activation. Western blot detection of EGFR Tyr (Y)-phosphorylation (pY-EGFR) was carried out by anti phospho-Tyr abs (Fig. 6A). The result indicated that activation of PAR4 alone is sufficient to initiate activation of EGFR directly, as recapitulated by pY-EGFR. It may take place via the association of Sos1 with PAR4 PH-binding domain, linking Grb1 and Shc for the direct association with the C-tail of EGFR. This describes an inside activation of EGFR via initiation of signaling events introduced by PAR4 activation. Indeed, when we have applied or not Pc(4–4), following activation of PAR4 in HEK293 cells that were cotransfected with hPar4 and egfr, potent inhibition of the PAR4 induced pY-EGFR (Fig. 6B) was observed. Notably, PAR4 mutants F347 L and D349A failed to induce EGFR1 phosphorylation (Fig. 6C). Figure 6D schematically depict our proposed cross-talk between PAR and RTKs. To assess the clinical relevance of PAR2/4, we stained tissue from breast and colon cancer patients and found it can be applicable also to Her2/Neu+ patients that develop resistance to conventional therapy that express high levels of PAR4/f2rl3 (Fig. 6E and F) and EGFR/erbB (40), in addition to colon tumors expressing high levels of PAR4/f2rl3 and PAR2/f2rl1 (Fig. 6G; Supplementary data Fig. S4 and S5).

Figure 6.

Cross-talk between PAR4 and EGFR. A, Activation of PAR4 induces phosphorylation of EGFR. HEK293 cells were cotransfected with 0.4 μg wt flg-hPar4 and 0.8 μg gfp-egfr. Cells were serum starved over night following AYPGKF PAR4 activation (200 μmol/L) for 5 to 45 minutes. Detection by Western blot analyses of pY-EGFR and EGFR was performed using either antiphosphotyrosine or anti-EGFR antibodies, respectively. PAR4 activation induces EGFR tyrosine phosphorylation within 15 minutes. This experiment is a representative of three independent experiment performed in triplicates. B, Pc(4–4) inhibits EGFR phosphorylation. HEK293 cells were cotransfected with 0.8 μg wt flg-hPar4 and 1 μg gfp-egfr. Cells were serum starved overnight following AYPGKF PAR4 activation (200 μmol/L) for 5 to 45 minutes in the presence and absence of Pc(4–4) (100 μmol/L). Detection by Western blot analyses of pEGFR and EGFR was performed using either antiphosphotyrosine or anti-EGFR antibodies (1:1,000 dilution), respectively. PAR4 activation induces EGFR tyrosine phosphorylation, Pc(4–4), a PAR4 inhibitor potently inhibited. This experiment is a representative of three independent experiment performed in triplicates. C, PAR4 mutants F347 L and D349A abolish pTyr-EGFR. HEK293 cells were transiently cotransfected with either flg-hPar4 or flg-hPar4D349A or flg-hPar4F347 L with egfr1 plasmid. Cells were serum starved overnight following AYPGKF PAR4 activation (200 μmol/L) for 5 to 45 minutes. Detection by Western blot analyses of pY-EGFR and EGFR was performed using either antiphosphotyrosine or anti-EGFR antibodies, respectively. In the presence of flg-hPar4 potent pTyr-EGFR1 is observed within 15′ and 30′ AYPGKF activation. This is not seen in the presence of the PAR4 mutants F347 L or D349A. This experiment is a representative of three independent experiment performed in triplicates. D, Scheme depicting cross-talk between GPCRs and RTK. The activation of PAR4 leads to cross-talk with the EGFR. AYPGKF activation of PAR4 recruits also cell signal proteins that possess PH-domain, for example, Akt, Gab1 and Sos1. E–G, Expression of PAR4 in breast and colon cancer tissue biopsy specimens. IHC of PAR4. Representative sections of breast tumor tissues (E and F) and human colon tissue sections (G) IHC staining, using anti-PAR4 (1:50 dilution) antibody. All images were taken using Nikon light microscope at 10× and 20× magnification. Scale bar, 50 μm. PAR4 is abundantly expressed in human breast Her2/Neu+ tissue sections (E), as also in human breast triple-negative (TN) tumor section (F). PAR4 IHC staining of human colon cancer tissue sections with metastatic invasion (G). In all cases, control with no primary antibody showed very little to no staining. The experiment was carried out three times.

Figure 6.

Cross-talk between PAR4 and EGFR. A, Activation of PAR4 induces phosphorylation of EGFR. HEK293 cells were cotransfected with 0.4 μg wt flg-hPar4 and 0.8 μg gfp-egfr. Cells were serum starved over night following AYPGKF PAR4 activation (200 μmol/L) for 5 to 45 minutes. Detection by Western blot analyses of pY-EGFR and EGFR was performed using either antiphosphotyrosine or anti-EGFR antibodies, respectively. PAR4 activation induces EGFR tyrosine phosphorylation within 15 minutes. This experiment is a representative of three independent experiment performed in triplicates. B, Pc(4–4) inhibits EGFR phosphorylation. HEK293 cells were cotransfected with 0.8 μg wt flg-hPar4 and 1 μg gfp-egfr. Cells were serum starved overnight following AYPGKF PAR4 activation (200 μmol/L) for 5 to 45 minutes in the presence and absence of Pc(4–4) (100 μmol/L). Detection by Western blot analyses of pEGFR and EGFR was performed using either antiphosphotyrosine or anti-EGFR antibodies (1:1,000 dilution), respectively. PAR4 activation induces EGFR tyrosine phosphorylation, Pc(4–4), a PAR4 inhibitor potently inhibited. This experiment is a representative of three independent experiment performed in triplicates. C, PAR4 mutants F347 L and D349A abolish pTyr-EGFR. HEK293 cells were transiently cotransfected with either flg-hPar4 or flg-hPar4D349A or flg-hPar4F347 L with egfr1 plasmid. Cells were serum starved overnight following AYPGKF PAR4 activation (200 μmol/L) for 5 to 45 minutes. Detection by Western blot analyses of pY-EGFR and EGFR was performed using either antiphosphotyrosine or anti-EGFR antibodies, respectively. In the presence of flg-hPar4 potent pTyr-EGFR1 is observed within 15′ and 30′ AYPGKF activation. This is not seen in the presence of the PAR4 mutants F347 L or D349A. This experiment is a representative of three independent experiment performed in triplicates. D, Scheme depicting cross-talk between GPCRs and RTK. The activation of PAR4 leads to cross-talk with the EGFR. AYPGKF activation of PAR4 recruits also cell signal proteins that possess PH-domain, for example, Akt, Gab1 and Sos1. E–G, Expression of PAR4 in breast and colon cancer tissue biopsy specimens. IHC of PAR4. Representative sections of breast tumor tissues (E and F) and human colon tissue sections (G) IHC staining, using anti-PAR4 (1:50 dilution) antibody. All images were taken using Nikon light microscope at 10× and 20× magnification. Scale bar, 50 μm. PAR4 is abundantly expressed in human breast Her2/Neu+ tissue sections (E), as also in human breast triple-negative (TN) tumor section (F). PAR4 IHC staining of human colon cancer tissue sections with metastatic invasion (G). In all cases, control with no primary antibody showed very little to no staining. The experiment was carried out three times.

Close modal

Kaplan–Meier analysis of PAR expressing members in breast cancer

Combined analyses of patients with Her2/Neu+–positive breast cancer with PAR1/f2rl or PAR2/f2rl1 and PAR4/f2rl3 significantly correlated with a worse prognosis, expressed as overall survival (OS) time (Supplementary Fig. S6). A significantly worse prognosis is seen in the presence of PAR2 and PAR4. These results underscore the significance of PAR family members especially PAR2 and PAR4, as aggressive genes impacting on the etiology of the disease. The prospect for cancer intervention via novel PH-binding motif namely by Pc(4–4) (purification details in Supplementary Fig. S7) is presented.

Here, we demonstrate for the first time the role of PAR4 PH-binding motif in colon cancer and introduce a novel PAR-based therapy. Identification and localization of the amino acid residues involved in the PH-binding motif are shown, as also additional PH-signal proteins that associate with PAR4. A lead cyclic-peptide Pc(4–4) was selected out of a mini-library, directed toward the PH-binding motif, found as a powerful inhibitor of PAR4- and PAR2–Akt/PKB interactions. Pc(4–4) inhibits PAR4 induced cell migration and Matrigel invasion in vitro. It strongly reduces tumor growth of either RKO/PAR4 clone/s, or tumors generated by HCT116 cells of high PAR2&4 levels (among other oncogenes), implanted in nude mice and notably, inhibits preformed tumors. Importantly, we demonstrate the activation of EGFR exclusively through AYPGKF PAR4 ligand application, potently inhibited by Pc(4–4). These results pave the way and open a broader horizon for PAR-based therapy.

Despite the fact that there is no primary sequence identity between various PH domains, their tertiary structures are strikingly similar (13). The PH domain-containing proteins are essential constituents of signal transduction pathways regulating cell growth and survival and are powerful targets and regulators of key oncogenes (37, 38). A number of cytosolic PH-signal proteins bind phosphorylated phosphatidylinositide (PtdIns) lipids and therefore are bound to cell membranes. On the other hand, direct protein–protein interactions, independent of lipid association, is emerging. We previously demonstrated that while the PH-Akt/PKB association with PAR1&2 is lipid dependent, the PH–Etk/Bmx interaction is independent, mediated directly via protein–protein interactions (12). As such is also the PH domain of Dbl that directly targets cytoskeletal matrix and was found necessary for oncogenic transformation in a lipid independent manner (41). Likewise, PH–Etk/Bmx interaction with focal adhesion kinase (FAK), the main signal protein of integrin-mediated adhesion, is mediated by protein–protein interaction through the association with the FERM domain of FAK (42). Therefore, both manners of protein–protein and protein–lipid binding interactions play a role in PH-domain module signaling interactions. In the current study, we have identified a specific PH-binding motif within the C-tail of PAR4 and assigned in addition to Akt/PKB-PH domain also the association of Gab1 and Sos1 signal proteins. The oncogenic potential of PAR4 is demonstrated by generation of large tumors in a xenograft mouse model driven by PAR4 epithelial clones of RKO colon epithelial cells, in vivo. Consequently, evidence is brought showing a marked inhibition of tumor growth in the presence of the lead cyclic peptide, Pc(4–4), directed toward the PAR2&4 PH-binding region. Pc(4–4) potently also inhibited the generation of large tumors instigated by HCT116 cell line overexpressing among others, high levels of PAR2 and PAR4. Noticeably, this takes place also when distinctly large tumors were preformed prior to the injection of Pc(4–4) compound. It remains to be seen if other nonpeptide PAR4 inhibitors developed over the past decade or so (43) to inhibit platelet function might also block the signal pathways affected by Pc(4–4). Importantly, we demonstrate the inhibition of PAR4- and possibly PAR2-mediated EGFR transactivation by Pc(4–4). This result paves the way to a broader horizon for PAR-based therapy.

Akt/PKB is a central protein kinase that drives cancer proliferation and is activated by its C-terminal phosphorylation, it contains a kinase-site and a PH-domain motif. An auto inhibition of Akt/PKB takes place upon the intramolecular interaction between the C-tail phosphorylation site and the N-lobe PH-kinase domain linker which is released upon phosphorylation at residue Ser473 of Akt/PKB (44). Other examples for targeting the PH-domain of cancer signal proteins is the inhibition of the PH-domain of Cnksr1 (connector enhancer of kinase suppressor of Ras 1), a multi-domain scaffold protein, important for cell proliferation, survival, and migration, which inhibits mutated K-Ras proto-oncogenes and tumor growth (45). Similarly, small molecules directed to the PH-domain of Gab1 were found effective in breast cancer growth and development (46). In the clinics, effective therapeutic results were obtained by inhibitors of Sos1, the guanine nucleotide exchange factor (GEF) that binds to Ras for activation (47). Whereas mutants of Ras are powerful oncogenes, a KrasG12C inhibitor has been approved by the FDA and a G12D inhibitor is being currently under development.

PAR4 was previously shown expressed in colon cancer tissue sections (48) as also in colon cancer cell lines (49) and absent in the nearby normal colonic epithelial cells. The presence of PAR4 has been demonstrated also in organoids isolated from fresh intestinal crypts and cultured in three dimensions (50). These organoids manifest various cell types present in the colonic epithelium, hence providing a good experimental system for colon tissues (51). Altogether, based on bioinformatics transcriptional profile of selected GPCRs (22) on one hand and organoid culture studies of normal colonic crypts on the other, it is suggested that PAR4/f2rl3 oncogene is expressed and plays a role in the colon stem cell progenitor niche. In addition to colon cancer, PAR4 has been demonstrated involved in a wide panel of epithelial malignancies (17–21).

Cyclic peptides often mimic the protein secondary structure and exhibit prolonged stability despite the presence of proteases. Backbone cyclization utilizes mainly atypical building blocks with an additional linker of customizable length covalently attached to a backbone functional group for peptide cyclization. By optimization of peptide cyclization in terms of ring size, endowing with enhanced conformational stability, the Pc(4–4) is showing longstanding effects on tumor growth in mice.

Another cyclic peptide named cilengitide directed to αvβ3/αvβ5 integrin-specific RGD site, has shown striking improvements in tumor management via the balanced regulation of the blood vessel network. While drastic reduction of blood vessels impairs proper drug delivery and hence should be carefully monitored, low-dose cilengitide has a significant impact on tumor reduction mainly by enabling improved drug delivery. Another orally available and biologically active cyclic peptide directed to αvβ3-RGD was developed by some of us (52, 53).

The FDA has approved several cyclic peptide drugs, out of which three, lanreotide, romidepsin, and pasireotide, are used in oncological treatments. While romidepsin is in fact a natural product generated from a gram-negative Chromabacterium violaceum and found useful in T-cell lymphoma (54), lanreotide and pasireotide are analogues of the endogenous cyclic peptide hormone somatostatin (55). Lanreotide and pasireotide block the GPCR somatostatin receptors. These cyclic peptide mimetics have a considerably longer plasma half-life than somatostatin and are utilized to treat endocrine tumors and acromegaly.

It is anticipated that Pc(4–4) cyclic peptide may successfully treat solid epithelial malignancies either by blocking directly GPCR/PARs as well as inhibiting EGFR/erbB induced tumors. Traditionally, it has been suggested that EGFR transactivation by GPCR stimulation involves a metalloproteinase that cleaves and release proHB-EGF (56, 57). As such, inhibitors of proHB-EGF processing or a mutant, which abrogates association of proHB-EGF, completely inhibit tyrosine phosphorylation of EGFR that is induced by lysophosphatidic acid LPA, a GPCR system. While this mode of transactivation addresses communications via the extracellular portion of EGFR, we propose an activation within the cell inside signaling. Whether PAR4 through PH-binding signal protein/s association induces EGFR1 phosphorylation via MMP thus releasing Hb-EG is yet an open issue. It can be answered, for example, via testing whether inhibition of metalloproteinases of AYPGKF activated PAR4 abrogates EGFR/erbB phosphorylation. Sos1 may serve as a junction protein mediating the direct cross-talk between GPCRs (for example, PAR4) with receptor tyrosine kinases, such as the EGFR/erbB. This can be achieved via further association of Sos1 with Grb2 and recruitment of Shc, which binds via its SH2 domain to Tyr phosphorylated erbB EGFR family members. It leads to initiation of intracellular signaling cascade events that activate also mutant/s of the EGFR/erbB members lacking the entire extracellular portion of the receptor. Altogether, Pc(4–4) inhibitor provides an additional line of therapy for Her2+ breast cancer that may develop resistance to current treatment as also colon cancer individuals, all of which express PAR4, PAR2 as also EGFR/erbB.

No disclosures were reported.

J.K. Nag: Investigation. H. Malka: Investigation. S. Sedley: Investigation. P. Appasamy: Investigation. T. Rudina: Investigation. T. Levi: Investigation. A. Hoffman: Investigation. C. Gilon: Investigation. B. Uziely: Investigation. R. Bar-Shavit: Supervision, investigation.

We thank Prof. A. Hubert for providing colon cancer tissue biopsies. This work is supported by the Israel Science Foundation grant 1420/16 to R. Bar-Shavit.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

Note: Supplementary data for this article are available at Molecular Cancer Therapeutics Online (http://mct.aacrjournals.org/).

1.
Gutkind
JS
,
Kostenis
E
.
Arrestins as rheostats of GPCR signalling
.
Nat Rev Mol Cell Biol
2018
;
19
:
615
6
.
2.
Nag
JK
,
Bar-Shavit
R
.
Transcriptional landscape of PARs in epithelial malignancies
.
Int J Mol Sci
2018
;
19
:
3451
72
.
3.
Coughlin
SR
.
Thrombin signalling and protease-activated receptors
.
Nature
2000
;
407
:
258
64
.
4.
Rasmussen
UB
,
Vouret-Craviari
V
,
Jallat
S
,
Schlesinger
Y
,
Pagès
G
,
Pavirani
A
, et al
.
cDNA cloning and expression of a hamster alphathrombin receptor coupled to Ca2‏ mobilization
.
FEBS Lett
1991
;
288
:
123
8
.
5.
Turm
H
,
Maoz
M
,
Katz
V
,
Yin
Y
,
Offermanns
S
,
Bar-Shavit
R
.
Protease-activated receptor-1 (PAR1) acts via a novel Galpha13-dishevelled axis to stabilize beta-catenin levels
.
J Biol Chem
2010
;
285
:
15137
48
.
6.
Nag
JK
,
Kancharla
A
,
Maoz
M
,
Turm
H
,
Agranovich
D
,
Gupta
CL
, et al
.
Low-density lipoprotein receptor-related protein 6 is a novel coreceptor of protease-activated receptor-2 in the dynamics of cancer-associated β-catenin stabilization
.
Oncotarget
2017
;
8
:
38650
67
.
7.
Jaber
M
,
Maoz
M
,
Kancharla
A
,
Agranovich
D
,
Peretz
T
,
Grisaru-Granovsky
S
, et al
.
Protease-activated-receptor-2 affects protease-activated-receptor-1-driven breast cancer
.
Cell Mol Life Sci
2014
;
71
:
2517
33
.
8.
Grisaru-Granovsky
S
,
Nag
JK
,
Zakar
L
,
Rudina
T
,
Gupta
CL
,
Maoz
M
, et al
.
PAR1&2 driven placenta EVT invasion act via LRP5/6 as coreceptors
.
FASEB J
2020
;
34
:
15701
17
.
9.
Bar-Shavit
R
,
Turm
H
,
Salah
Z
,
Maoz
M
,
Cohen
I
,
Weiss
E
, et al
.
PAR1 plays a role in epithelial malignancies: transcriptional regulation and novel signaling pathway
.
IUBMB Life
2011
;
63
:
397
402
.
10.
Booden
MA
,
Eckert
LB
,
Der
CJ
,
Trejo
J
.
Persistent signaling by dysregulated thrombin receptor trafficking promotes breast carcinoma cell invasion
.
Mol Cell Biol
2004
;
24
:
1990
9
.
11.
Even-Ram
S
,
Uziely
B
,
Cohen
P
,
Grisaru-Granovsky
S
,
Maoz
M
,
Ginzburg
Y
, et al
.
Thrombin receptor overexpression in malignant and physiological invasion processes
.
Nat Med
1998
;
4
:
909
14
.
12.
Kancharla
A
,
Maoz
M
,
Jaber
M
,
Agranovich
D
,
Peretz
T
,
Grisaru-Granovsky
S
, et al
.
PH motifs in PAR1&2 endow breast cancer growth
.
Nat Commun
2015
;
6
:
8853
63
.
13.
Rebecchi
MJ
,
Scarlata
S
.
Pleckstrin homology doamains: a common fold with diverse functions
.
Ann Rev Biophys Biomol Struct
1998
;
27
:
503
28
.
14.
Sevigny
LM
,
Austin
KM
,
Zhang
P
,
Kasuda
S
,
Koukos
G
,
Sharifi
S
, et al
.
Protease-activated receptor-2 modulates protease-activated receptor-1-driven neointimal hyperplasia
.
Arterioscler Thromb Vasc Biol
2012
;
31
:
e100
6
.
15.
Connolly
TM
,
Condra
C
,
Feng
DM
,
Cook
JJ
,
Stranieri
MT
,
Reilly
CF
, et al
.
Species variability in platelet and other cellular responsiveness to thrombin receptor-derived peptides
.
Thromb Haemost
1994
;
72
:
627
33
.
16.
Kahn
ML
,
Zheng
YW
,
Huang
W
,
Bigornia
V
,
Zeng
D
,
Moff
S
, et al
.
A dual thrombin receptor system for platelet activation
.
Nature
1998
;
394
:
690
4
.
17.
Jiang
P
,
Li
SD
,
Li
ZG
,
Zhu
YC
,
Yi
XJ
,
Li
SM
.
The expression of protease-activated receptors in esophageal carcinoma cells: the relationship between changes in gene expression and cell proliferation, apoptosis in vitro and growing ability in vivo
.
Cancer Cell Int
2018
;
18
:
81
88
.
18.
Yu
G
,
Jiang
P
,
Xiang
Y
,
Zhang
Y
,
Zhu
Z
,
Zhang
C
, et al
.
Increased expression of protease-activated receptor 4 and Trefoil factor 2 in human colorectal cancer
.
PLoS One
2015
;
10
:
e0122678
.
19.
Wilson
SR
,
Gallagher
S
,
Warpeha
K
,
Hawthorne
SJ
.
Amplification of MMP-2 and MMP-9 production by prostate cancer cell lines via activation of protease-activated receptors
.
Prostate
2004
;
60
:
168
74
.
20.
Mußbach
F
,
Henklein
P
,
Westermann
M
,
Settmacher
U
,
Böhmer
FD
,
Kaufmann
R
.
Proteinase-activated receptor 1- and 4-promoted migration of Hep3B hepatocellular carcinoma cells depends on ROS formation and RTK transactivation
.
J Cancer Res Clin Oncol
2015
;
141
:
813
25
.
21.
Elste
AP
,
Petersen
I
.
Expression of proteinase-activated receptor 1–4 (PAR 1–4) in human cancer
.
J Mol Histol
2010
;
41
:
89
99
.
22.
Choi
HY
,
Saha
SK
,
Kim
K
,
Kim
S
,
Yang
GM
,
Kim
BW
, et al
.
G protein-coupled receptors in stem cell maintenance and somatic reprogramming to pluripotent or cancer stem cells
.
BMB Rep
2015
;
48
:
68
80
.
23.
Mishra
R
,
Patel
H
,
Alanazi
S
,
Yuan
L
,
Garrett
JT
.
HER3 signaling and targeted therapy in cancer
.
Oncol Rev
2018
;
12
:
355
.
24.
Choi
H-J
,
Jin
S
,
Cho
H
,
Won
H-Y
,
An
HW
,
Jeong
G-Y
, et al
.
CDK12 drives breast tumor initiation andtrastuzumab resistance via WNT andIRS1-ErbB-PI3K signaling
.
EMBO Rep
2019
;
20
:
e48058
.
25.
Hurevich
M
,
Swed
A
,
Joubran
S
,
Cohen
S
,
Freeman
NS
,
Britan-Rosich
E
, et al
.
Rational conversion of noncontinuous active region in proteins into a small orally bioavailable macrocyclic drug-like molecule: the HIV-1 CD4:Gp120 paradigm
.
Bioorganic Med Chem
2010
;
18
:
5754
61
.
26.
Hess
S
,
Linde
Y
,
Ovadia
O
,
Safari
E
,
Shalev
DE
,
Swed
A
, et al
.
Backbone cyclic peptidomimetic melanocortin-4 receptor agonist as a novel orally administrated drug lead for treating obesity
.
J Med Chem
2008
;
51
:
1026
34
.
27.
Talhami
A
,
Swed
A
,
Hess
S
,
Ovadia
O
,
Greenberg
S
,
Schumacher-Klinger
A
, et al
.
Cyclizing painkillers: development of backbone-cyclic TAPS analogs
.
Front Chem
2020
;
8
:
532577
-
532596
.
28.
Guha
P
,
Bandyopadhyaya
G
,
Polumuri
SK
,
Chumsri
S
,
Gade
P
,
Kalvakolanu
DV
, et al
.
Nicotine promotes apoptosis resistance of breast cancer cells and enrichment of side population cells with cancer stem cell-like properties via a signaling cascade involving galectin-3, α9 nicotinic acetylcholine receptor and STAT3
.
Breast Cancer Res Treat
2014
;
145
:
5
22
.
29.
Song
G
,
Ouyang
G
,
Bao
S
.
The activation of Akt/PKB signaling pathway and cell survival
.
J Cell Mol Med
2005
;
9
:
59
71
.
30.
Hanada
M
,
Feng
J
,
Hemmings
BA
.
Structure, regulation and function of PKB/AKT-a major therapeutic target
.
Biochim Biophys Acta
2004
;
1697
:
3
16
.
31.
Lu
J
,
Steeg
PS
,
Price
JE
,
Krishnamurthy
S
,
Mani
SA
,
Reuben
J
, et al
.
Breast cancer metastasis: challenges and opportunities
.
Cancer Res
2009
;
69
:
4951
3
.
32.
Vu
TK
,
Hung
DT
,
Wheaton
VI
,
Coughlin
SR
.
Molecular cloning of a functional thrombin receptor reveals a novel proteolytic mechanism of receptor activation
.
Cell
1991
;
64
:
1057
68
.
33.
Faruqi
TR
,
Weiss
EJ
,
Shapiro
MJ
,
Huang
W
,
Coughlin
SR
.
Structure-function analysis of protease-activated receptor 4 tethered ligand peptides. Determinants of specificity and utility in assays of receptor function
.
J Biol Chem
2000
;
275
:
19728
34
.
34.
Vanhaesebroeck
B
,
Leevers
SJ
,
Ahmadi
K
,
Timms
J
,
Katso
R
,
Driscoll
PC
, et al
.
Synthesis and function of 3-phosphorylated inositol lipids
.
Annu Rev Biochem
2001
;
70
:
535
602
.
35.
Gu
H
,
Neel
BG
.
The “gab” in signal transduction
.
Trends Cell Biol
2003
;
13
:
122
30
.
36.
Hoeben
A
,
Martin
D
,
Clement
PM
,
Cools
J
,
Gutkind
JS
.
Role of GRB2- associated binder 1 in epidermal growth factor receptor-induced signaling in head and neck squamous cell carcinoma
.
Int J Cancer
2013
;
132
:
1042
50
.
37.
Seiden-Long
I
,
Navab
R
,
Shih
W
,
Li
M
,
Chow
J
,
Zhu
CQ
, et al
.
Gab1 but not Grb2 mediates tumor progression in met overexpressing colorectal cancer cells
.
Carcinogenesis
2008
;
29
:
647
55
.
38.
Scheffzek
K
,
Welti
S
.
Pleckstrin homology (PH) like domains - versatile modules in protein-protein interaction platforms
.
FEBS Lett
2012
;
586
:
2662
73
.
39.
Ramachandran
R
,
Mihara
K
,
Thibeault
P
,
Vanderboor
CM
,
Petri
B
,
Saifeddine
M
,
Bouvier
M
,
Hollenberg
MD
.
Targeting a proteinase-activated receptor 4 (PAR4) carboxyl terminal motif to regulate platelet function
.
Mol Pharmacol
2017
;
91
:
287
95
.
40.
Cho
SY
.
Identification of ERBB pathway-activated cells in triple-negative breast cancer
.
Genomics Inform
2019
;
17
:
e3
.
41.
Zheng
Y
,
Zangrilli
D
,
Cerione
RA
,
Eva
A
.
The pleckstrin homology domain mediates transformation by oncogenic dbl through specific intracellular targeting
.
J Biol Chem
1996
;
271
:
19017
20
.
42.
Chen
R
,
Kim
O
,
Li
M
,
Xiong
X
,
Guan
JL
,
Kung
HJ
, et al
.
Regulation of the PH-domain-containing tyrosine kinase Etk by focal adhesion kinase through the FERM domain
.
Nat Cell Biol
2001
;
3
:
439
44
.
43.
Yu
X
,
Li
S
,
Zhu
X
,
Kong
Y
.
Inhibitors of protease-activated receptor 4 (PAR4): a review of recent patents (2013–2021)
.
Expert Opin Ther Pat
2022
;
32
:
153
70
.
44.
Chu
N
,
Salguero
AL
,
Liu
AZ
,
Chen
Z
,
Dempsey
DR
,
Ficarro
SB
et al
.
Akt kinase activation mechanisms revealed using protein semi synthesis
.
Cell
2018
;
174
:
897
907
.
45.
Indarte
M
,
Puentes
R
,
Maruggi
M
,
Ihle
NT
,
Grandjean
G
,
Scott
M
, et al
.
An inhibitor of the pleckstrin homology domain of CNK1 selectively blocks the growth of mutant KRAS cells and tumors
.
Cancer Res
2019
;
79
:
3100
11
.
46.
Chen
L
,
Du-Cuny
L
,
Moses
S
,
Dumas
S
,
Song
Z
,
Rezaeian
AH
, et al
.
Inhibitors induce large conformational changes of GAB1 pleckstrin homology domain and kill breast cancer cells
.
PLoS Comput Biol
2015
;
11
:
e1004021
.
47.
Hillig
RC
,
Sautier
B
,
Schroeder
J
,
Moosmayer
D
,
Hilpmann
A
,
Stegmann
CM
, et al
.
Discovery of potent SOS1 inhibitors that block RAS activation via disruption of the RAS–SOS1 interaction
.
Proc Natl Acad Sci USA
2019
;
116
:
2551
60
.
48.
Gratio
V
,
Walker
F
,
Lehy
T
,
Laburthe
M
,
Darmoul
D
.
Aberrant expression of proteinase-activated receptor 4 promotes colon cancer cell proliferation through a persistent signaling that involves Src and ErbB-2 kinase
.
Int J Cancer
2009
;
124
:
1517
25
.
49.
Zhang
H
,
Jiang
P
,
Zhang
C
,
Lee
S
,
Wang
W
,
Zou
H
.
PAR4 overexpression promotes colorectal cancer cell proliferation and migration
.
Oncol Lett
2018
;
16
:
5745
52
.
50.
Sébert
M
,
Denadai-Souza
A
,
Quaranta
M
,
Racaud-Sultan
C
,
Chabot
S
,
Lluel
P
, et al
.
Thrombin modifies growth, proliferation and apoptosis of human colon organoids: a protease-activated receptor 1- and protease activated receptor 4-dependent mechanism
.
Br J Pharmacol
2018
;
175
:
3656
68
.
51.
Sato
T
,
Stange
DE
,
Ferrante
M
,
Vries
RG
,
Van Es
JH
,
Van den Brink
S
, et al
.
Long-term expansion of epithelial organoids from human colon, adenoma, adenocarcinoma, and Barrett's epithelium
.
Gastroenterology
2011
;
141
:
1762
72
.
52.
Wong
PP
,
Demircioglu
F
,
Ghazaly
E
,
Alrawashdeh
W
,
Stratford
MR
,
Scudamore
CL
, et al
.
Dual-action combination therapy enhances angiogenesis while reducing tumor growth and spread
.
Cancer Cell
2015
;
27
:
123
37
.
53.
Weinmüller
M
,
Rechenmacher
F
,
Kiran Marelli
U
,
Reichart
F
,
Kapp
TG
,
Räder
AFB
, et al
.
Overcoming the lack of oral availability of cyclic hexapeptides: design of a selective and orally available ligand for the integrin αvβ3
.
Angew Chem Int Ed Engl
2017
;
56
:
16405
9
.
54.
Jain
S
,
Zain
J
.
Romidepsin in the treatment of cutaneous T-cell lymphoma
.
J Blood Med
2011
;
2
:
37
47
.
55.
Modlin
I
,
Pavel
M
,
Kidd
M
,
Gustafsson
BI
.
Somatostatin analogues in the treatment of gastro enteropancreatic neuroendocrin (carcinoid) tumours
.
Aliment Pharmacol Ther
2010
;
31
:
169
88
.
56.
Prenzel
N
,
Zwick
E
,
Daub
H
,
Leserer
M
,
Abraham
R
,
Wallasch
C
., et al
.
EGF receptor transactivation by G-protein coupled receptorstransactivation requires cleavage of proHB-EGF
.
Nature
1999
;
402
:
884
8
.
57.
Palanisamy
S
,
Xue
C
,
Shun Ishiyama
S
,
Sathyamangla Venkata
S
,
Naga Prasad
N
,
Kathleen Gabrielson
K
.
GPCR-ErbB transactivation pathways and clinical implications
.
Cell Signal
2021
;
86
:
110092
.

Supplementary data