Proteases give cancer a defining characteristic of being able to break through extracellular matrix barriers and invade into other tissues in response to chemotactic signals. Recently, the cell surface protease-activated receptor (PAR)-1 has been shown to act as a chemokine receptor in inflammatory cells, and its expression is tightly correlated with metastatic propensity of breast cancer cells. The aim of the present study was to determine whether activation of PAR1 or the other known PARs (PAR2–4) can regulate migration and invasion of breast cancer cells. We found that the highly invasive MDAMB231 breast cancer cell line expressed very high levels of functional PAR1, PAR2, and PAR4, whereas minimally invasive MCF7 cells had trace amounts of PAR1 and low levels of PAR2 and PAR4. Despite the differences in expression, PAR2 and PAR4 acted as chemokine receptors in both invasive and minimally invasive breast cell lines. Quite unexpectedly, we found that activation of PAR1 with thrombin or the peptide agonist SFLLRN markedly inhibited invasion and migration of MDAMB231 cells when applied as a concentration gradient in the direction of cell movement. Additionally, we demonstrated that inhibition of chemotaxis was mediated through a Gi/phosphoinositide-3-OH kinase-dependent pathway. Activation of Gi signaling with epinephrine or wasp venom mastoparan also inhibited invasion and migration of the breast cancer cells. These findings suggest that therapeutics targeted toward Gi-couplers that are selectively expressed in breast cancer cells could prove beneficial in halting the progression of invasive breast cancer.

The blood coagulation protease thrombin can elicit many cellular responses such as platelet aggregation, inflammation, chemotaxis, mitogenesis, apoptosis, and angiogenesis (1, 2, 3). Thrombin can also enhance adhesion of cancer cells to endothelium, platelets, and the ECM4(4, 5). Thrombin mediates its cellular effects by activating the PARs. Four different PARs have been identified to date: PAR1, PAR2, PAR3, and PAR4 (6, 7, 8, 9, 10). PAR1 and PAR3 are activated by thrombin, PAR2 is activated by tryptase/trypsin and PAR4 is activated by thrombin and tryptase/trypsin. PAR1 was originally discovered on platelets and serves as the prototype for this subfamily of G-protein coupled receptors (6). PAR1 is activated when thrombin cleaves the peptide backbone between residues R41–S42 located within the receptor NH2-terminal extracellular domain. Proteolytic cleavage exposes a new NH2 terminus that activates the receptor in an intramolecular mode. Synthetic peptides that correspond to the first few amino acids of the freshly cleaved NH2 terminus (i.e., SFLLRN), can function as intermolecular agonists to PAR1 (6, 11). Other serine proteases found in tumors and blood, such as plasmin, can also activate PAR1, albeit with lower efficiency (12).

PAR1 has also been proposed to play a role in the pathological invasion processes of breast cancer (13, 14). Like the majority of cancers, invasive breast cancers originate from epithelial tissues, which are enveloped by the ECM. Normal breast epithelial cells do not have the capacity to migrate efficiently in response to chemotactic signals. Examination of preinvasive breast carcinoma biopsy specimens reveals high expression levels of the RANTES chemokine both in the carcinoma cells and the infiltrating leukocytes (15). Application of RANTES, MIP-1α, MIP-1β, or MCP-1 chemokines that are secreted by leukocytes (16) stimulates chemotaxis of MCF-7 cells. Thus, breast carcinoma cell lines including MCF7, T47D, and ZR-75–1 have somehow acquired or reacquired the ability to migrate but still lack invasive capabilities (16). To be invasive, the cancer cell must produce matrix metalloproteases and have signaling pathways that can coordinate the cyclic cytoskeletal motions required for attachment and traction through integrin-ECM contacts (17, 18). Because proteases are essential for cancer cells to invade through the ECM, there has been a concerted effort (13, 14, 19, 20) to determine whether PARs are also involved in the invasion and metastasis processes. For instance, studies (21) have shown that thrombin activation of PAR1 promotes cell adhesion to vitronectin, whereas trypsin activation of PAR2 stimulates α5β1-dependent adhesion to fibronectin in human gastric carcinoma cells.

Recently, Even-Ram et al.(13) demonstrated that PAR1 expression levels are directly correlated with degree of invasiveness in both primary breast tissue specimens and established cancer cell lines. High levels of PAR1 mRNA were found in infiltrating ductal carcinoma and very low amounts in normal and premalignant atypical intraductal hyperplasia. Transfection of the invasive breast cancer cells with antisense PAR1 DNA abolished invasion implying that PAR1 expression was somehow involved in the invasion process. However, the effect of activation of PAR1 by thrombin or the contribution of the other PARs, if any, in the breast cancer cell invasion process was not evaluated.

Here, we determined that invasive breast cancer cells express high levels of PAR1, PAR2, and PAR4. We present the first evidence that PAR2 and PAR4 can act as chemokine receptors. Moreover, we make the unanticipated observation that thrombin and PAR1 peptide ligands strongly inhibit migration and invasion of MDAMB231 breast cancer cells. This PAR1-mediated inhibition requires that the agonist is placed in the direction of cell migration. Analysis of G-protein dependent signaling pathways lead us to conclude that the inhibitory effects of PAR1 on invasion are mediated by a Gi/PI(3)kinase-dependent pathway. Furthermore, selective stimulation of Gi-signaling by other pathways can also prevent cell migration and invasion. These findings suggest a novel strategy to interdict migration and invasion of breast cancer cells.

Materials and Reagents.

Human α-thrombin was obtained from Hematological Technologies (Essex Junction, VT). The PAR agonists T-(pf-F)LLRN (TFLLRN), SFLLRN, SLIGKV, GYPGKF, and AYPGKF were synthesized as COOH-terminal amides at the Tufts University peptide core facility. Recombinant hirudin, PMA, GFX, wortmannin, LY, genistein, mastoparan M5280, and trypsin were obtained from Calbiochem (La Jolla, CA). MTT, PTx, and epinephrine were from Sigma Chemical Co. (St. Louis, MO). Fura2/AM was from Molecular Probes (Eugene, OR), PPACK and TLCK were from Boehringer Mannheim (Indianapolis, IN).

Cell Culture.

MDAMB231 and MCF7 cells were obtained from the National Cancer Institute (NIH, Bethesda, MD). MDAMB231 cells were maintained in RPMI 1640 supplemented with 10% FBS, 0.15% sodium bicarbonate, and 1% penicillin and streptomycin under conditions of 5% CO2 at 37°C. MCF7 cells were cultured in RPMI 1640 supplemented with 10% FBS and 1% penicillin and streptomycin. NIH3T3 fibroblast cells were maintained in DMEM supplemented with 10% FBS and 1% penicillin and streptomycin.

Flow Cytometry.

Rabbit polyclonal PAR1-Ab (raised against residues S42FLLRNPNDKYEPF55C), PAR3-Ab (raised against residues T42FRGAPPNSFEEFP55C), and PAR4-Ab (raised against residues G42YPGQVSANDSDTLELP58C) were generated from keyhole limpet hemocyanin conjugates and purified on the corresponding antigen peptide-conjugated columns by affinity chromatography as described previously (12). Goat anti-PAR2 antibody and FITC-conjugated donkey antigoat antibody were purchased from Santa Cruz Biotechnology (Santa Cruz, CA). Flow cytometry was performed on MDAMB231 and MCF7 cells as before (12). The reactivity of each of the PAR-antibodies was confirmed using tissue culture cells that expressed the individual PAR receptors. The PAR antibodies were not cross-reactive.

Invasion Assays.

Invasion of cells through Matrigel was determined using a Transwell system (6.5 mm diameter, 8-μm pore size with polycarbonate membrane; Corning Costar) as described previously (22) with the following modifications. Cells were starved overnight in serum-free medium containing 0.1% BSA. Matrigel (40 μg) was polymerized in the upper well at 37°C for 2 h. Cells (25,000) were added to the upper well of each chamber in 200 μl of serum-free medium containing 0.1% BSA. CM (600 μl) from NIH3T3 fibroblasts was added to the lower well with or without various concentrations of agonists or inhibitors. The cells were incubated for 40 h at 37°C. Cells in the top well were completely removed by swiping with cotton swabs and the cells on the underside of the membrane were stained using the Diff-Quik (Dade Boehringer) stain kit. Cells on 15–30% of the membrane area were counted at ×20–40 and extrapolated to 100% of the membrane surface. Percentage of basal invasion was defined as number of cells migrating in the presence of Matrigel/number of cells migrating in the absence of Matrigel × 100, and invasion in the presence of agonist was defined as (number of cells migrating in the presence of Matrigel + agonist) number of cells migrating in the absence of Matrigel × 100. Statistical analyses were performed using a two-tailed Student’s t test. Statistical significance was assumed to occur when P ≤ 0.05.

Migration Assays.

Transwell migration assays were performed essentially as the invasion assays above except that Matrigel was omitted. Cells (50,000/200 μl in serum-free medium) were added to the upper wells and agonists (thrombin, SFLLRN, TFLLRN, SLIGKV, GYPGKF, or AYPGKF) were added to the upper and/or lower wells of the transwell chambers for the “checkerboard” (23) migration studies. Cells were incubated for 5 h at 37°C before being stained and counted and statistical analyses conducted as in the invasion studies.

To determine what role the individual PARs play in breast cancer invasion, we first measured the expression levels of all of the known PARs in breast cancer cell lines with well-characterized metastatic propensity (24). Two breast cancer cell lines were examined in detail: MCF7 and MDAMB231. MCF7 cells were originally isolated from a pleural effusion of a postmenopausal woman with breast cancer (24). MCF7 cells are estrogen receptor- and progesterone receptor-positive and have low metastatic potential. The MDAMB231 cells were from a pleural effusion of a 51-year-old woman with breast cancer. Unlike MCF7 cells, MDAMB231 cells are estrogen receptor- and progesterone receptor-negative, are highly invasive, and produce lung metastases from primary mammary fat pad tumors in nude mice (24).

Flow cytometry with PAR-specific antibodies was used to determine the surface expression profiles of PAR1, PAR2, PAR3, and PAR4 in the breast cancer cell lines. As shown in Fig. 1,A, MDAMB231 cells express relatively high levels of PAR1 and PAR4 and lower levels of PAR2. The MCF7 breast cell line expresses low levels of PAR2 and PAR4 (Fig. 1,B) and trace levels of PAR1 (Fig. 1 B). Little or no PAR3 was present in the MCF7 or MDAMB231 cells. A second noninvasive breast carcinoma cell line, T47D, gave essentially identical PAR expression profiles as MCF7 (data not shown) and was not studied further.

Functional activity assays of PAR1, PAR2, and PAR4 were conducted by monitoring intracellular calcium fluxes in response to agonist. On activation by cognate protease or soluble peptide ligand, the PARs stimulate phospholipase C-β via Gq and/or βγ(Gi), which causes production of inositol triphosphate and mobilization of Ca2+ from intracellular stores. The SFLLRN peptide agonist activates both PAR1 and PAR2; TFLLRN activates PAR1; SLIGKV activates PAR2; GYPGKF and AYPGKF activate PAR4; and thrombin activates PAR1, PAR3, and PAR4. Synthetic agonists for PAR3 have not yet been described. As shown in Fig. 2,A, the MDAMB231 cells exhibit robust intracellular calcium fluxes when stimulated with PAR1, PAR2, or PAR4 agonists. Thrombin and SFLLRN cause sharp increases in intracellular Ca2+ that quickly fall back to baseline levels. The PAR1-selective peptide TFLLRN gives a Ca2+ response similar in profile to that elicited by thrombin. The PAR2-selective agonist SLIGKV also gives a sharp Ca2+ response, and the PAR4-selective ligand GYPGKF gives a robust Ca2+ response. The Ca2+ responses were markedly different in the MCF7 cells (Fig. 2 B). Addition of TFLLRN gave a nearly undetectable Ca2+ response consistent with the flow cytometry data, which showed trace levels of PAR1 expression. A larger Ca2+ response was elicited with SLIGKV and SFLLRN indicating that PAR2 is functional in the MCF7 cells. The PAR4 agonists GYPGKF and thrombin both gave very slow but appreciable rises in intracellular Ca2+ demonstrating that PAR4 is functionally active in the MCF7 cells.

Next, we determined the effects of activation of the PARs on the invasion of MDAMB231 cells through Matrigel using an in vitro invasion system. Matrigel is prepared from a mouse sarcoma tumor extract that contains basement membrane components including laminin, collagen IV, and heparan sulfate proteoglycans (25). Cells were seeded onto the upper well of a transwell apparatus layered with Matrigel, and CM from NIH3T3 fibroblasts was added to the bottom well. The CM contains chemokines that stimulate very high basal levels (10–12%) of MDAMB231 invasion through the Matrigel (Fig. 3). By comparison, CM elicits extremely low levels of invasion (≤1.5%) in the MCF7 cells, which is consistent with previous studies (13, 14). Neither MDAMB231 nor MCF7 cells invade the Matrigel (<0.01%) when CM is omitted from the bottom wells in the presence or absence of any of the PAR agonists (data not shown). Thrombin and SFLLRN had no effect on cell migration when added to the top well (Fig. 3, ▪).

Quite strikingly, however, thrombin and SFLLRN cause marked inhibition of invasion of the MDAMB231 cells through Matrigel when added as a gradient in the direction of cell movement. As shown in Fig. 3,A, low concentrations of thrombin added to the bottom well blocks 80% of cell invasion with an IC50 value of ∼5 nm. Addition of hirudin, a direct thrombin inactivator, reversed the inhibition of invasion by exogenously added thrombin (Fig. 4). The PAR1 peptide agonist SFLLRN inhibits 93% of invasion with an IC50 of 200 nm (Fig. 3,B). However, high amounts of the serine protease-alkylating agents PPACK or TLCK, hirudin alone, or a PAR1-blocking antibody (3) did not inhibit basal invasion of the MDAMB231 cells (Fig. 4). This effectively rules out the possibility (13, 14) that activation of PAR1 by endogenously produced serine proteases, including thrombin, is necessary for the invasion of MDAMB231 cells through the ECM.

Stimulation of PAR2 and PAR4 with their respective ligands, SLIGKV and GYPGKF (Fig. 3,B), or trypsin (data not shown) did not inhibit basal invasion of the MDAMB231 cells but slightly stimulated invasion (20–30%) at higher concentrations of peptide ligand. Identical results (data not shown) were obtained for all of the PAR agonists from independently maintained MDAMB231 cells (gift of Vimla Band, NEMC, Boston, MA). Next, we determined whether the inhibitory effects of thrombin or SFLLRN on the invasion of the MDAMB231 cells were attributable to impaired growth or apoptosis (2, 26). As shown in Table 1, 0.1–100 nm of thrombin or 0.1–100 μm of SFLLRN had little or no effect on the viability of the MDAMB231 breast cancer cells during the time period of the invasion experiments (40–48 h).

As anticipated by the invasion data, we showed that activation of PAR1 inhibits migration of the MDAMB231 cells. Checkerboard analysis (23) was conducted in which the PAR1 agonist was placed in either the top and/or bottom wells of the migration chamber. By varying the direction of the concentration gradient, we determined that thrombin and SFLLRN inhibited cell migration only when placed in the bottom well along with the chemokines present in the CM. As shown in Fig. 5, A and B, thrombin inhibits migration of the MDAMB231 cells when present at ≥1 nm concentration in the bottom well. Addition of thrombin to the upper well (Fig. 5, C and D) did not inhibit cell migration. Similar results were observed with the PAR1 ligand, SFLLRN. Placement of SFLLRN (≥1 μm) in the lower well inhibited ≤80% of migration of the MDAMB231 cells (Fig. 5, E and F). Placement of SFLLRN in the upper well had little or no effect on migration (Fig. 5, G and H). Addition of the PAR2 or PAR4 ligands SLIGKV and GYPGKF to the upper or lower wells did not inhibit migration of the MDAMB231 cells (data not shown).

Checkerboard analysis was performed with different dilutions of CM derived from NIH3T3 cells to determine whether the chemokines in the CM cause chemotaxis (directed movement) or chemokinesis (random movement) of the MDAMB231 cells. As shown in Fig. 6,A, the CM is strongly chemotactic to the MDAMB231 cells, does not stimulate chemokinesis (Fig. 6 C), and, thus, must be added to the bottom well for migration to occur. Therefore, activation of PAR1 with a concentration gradient of thrombin or SFLLRN in the same direction as the chemotactic agent(s) inhibits chemotaxis in these highly invasive cells.

Similar to the effects seen in the MDAMB231 invasion experiment, PAR2 and PAR4 agonists stimulated migration of the MCF7 cells (Fig. 7). The PAR2- and PAR4-selective ligands SLIGKV and GYPGKF (and AYPGKF, data not shown) stimulate 1.5–2-fold increases in migration of the MCF7 cells indicating that these PARs can act as chemotactic receptors. PAR1 did not mediate the chemotactic effects of thrombin, because the PAR1-specific agonist TFLLRN had no effect on MCF7 migration (Fig. 7 B). It is important to stress that the basal levels of cell migration are quite different in MCF7 cells (2.5–5%) versus MDAMB231 (50–70%). Therefore, each of the PARs, depending on the cellular context, i.e., normal versus invasive, could conceivably operate as either chemoinhibitory or chemotactic receptors.

The next question to be addressed was which signaling pathways are involved in the PAR1-dependent inhibition of the breast cancer cell migration. Although the signaling pathways for cell migration have not been studied extensively in breast cancer cells, it is becoming clear that chemokine-dependent movement in inflammatory cells requires release of free βγ subunits from Gi-coupled chemokine receptors (23, 27). The free βγ subunits then stimulate PI(3)K and generate polarized lipid signals to direct cell movement (28, 29). We found that pretreatment of MDAMB231 cells with PTx, which inactivates Gi, results in complete inhibition of invasion in the presence or absence of thrombin (Fig. 8). Likewise, the specific PI(3)K inhibitor LY blocks ∼70% of the basal invasion of the MDAMB231 cells. Addition of thrombin to the LY-treated cells showed no additional inhibition of invasion, which is evidence that the inhibitory effects of thrombin require the activity of PI(3)K.

Stimulation of PKC with phorbol ester, PMA, greatly inhibits cell migration in the absence and presence of thrombin (30, 31). Surprisingly, blockade of PKC with GFX did not prevent thrombin inhibition of cell invasion (Fig. 8). Therefore, PAR1-mediated inhibition of MDAMB231 cell invasion is independent of PKC nor does the basal invasion process require PKC. Tyrosine kinases such as focal adhesion kinase and the src kinase family members fyn, yes, and lck have been shown to be essential for regulating the assembly and disassembly of focal adhesion complexes in migrating cells (18, 32). Accordingly, addition of the tyrosine kinase inhibitor genistein blocked 80–85% of cell invasion in the absence or presence of thrombin.

Next, we tested whether other activators of Gi-signaling besides PAR1 could inhibit cell migration and invasion. As shown in Fig. 9, addition of epinephrine, which activates the Gi-coupled α2-ARs (33) blocks 60% of MDAMB231 cell invasion. Expression of the Gi-coupled α2C-AR in MDAMB231 cells was confirmed by a search of the NCI60 cDNA microarray database (34). The wasp venom mastoparan which is a direct activator of Gi-dependent nucleotide exchange, also inhibits 60% of cell invasion and migration of the MDAMB231 cells (Fig. 9,A and data not shown). Moreover, mastoparan potently inhibits (IC50 ∼6 nm) essentially all migration of the poorly invasive breast cell line MCF7 (Fig. 9,B). Thus, it is evident that inhibition of breast cancer cell invasion and migration is not restricted to PAR1 and may be elicited to a greater or lesser extent by other Gi activators. Cell viability experiments were performed, and the inhibitors or activators used above (PTx, GFX, PMA, genistein, mastoparan, or LY) caused little or no growth-inhibitory or stimulatory effects (0.73–1.32-fold relative to control) over the 40–48 h time period encompassed by the invasion experiments (Table 1).

The present studies came to the surprising conclusion that activation of the PAR1 thrombin receptor strongly inhibits migration and invasion of MDAMB231 breast cancer cells. We demonstrated that this inhibition requires that the PAR1 agonists (thrombin, SFLLRN, and TFLLRN) be placed as a concentration gradient in the same direction as the chemotactic stimulus. We propose to call this phenomenon “chemoinhibition.” The MDAMB231 cells also express high levels of PAR2 and PAR4. Activation of these protease receptors did not inhibit cell migration but instead slightly stimulated invasion. This suggests that selective stimulation of each of the PARs could lead to different migratory versus static outcomes depending on the context of the cancer cells and the particular extracellular milieu.

We identified the G-protein signaling pathways that are responsible for both chemotaxis and the PAR1-dependent chemoinhibition of breast cancer cells. Many studies using hematopoietic cells have shown that cell movement is directed by polyphosphoinositide lipid signals such as PIP3, which are internally polarized to the side of the cell nearest to the chemotactic agent (28). This spatially restricted signal emanates from free βγ subunits released from Gi-coupled chemokine receptors, which, in turn, activate PI(3)Kγ to generate PIP3. The transiently produced PIP3 binds pleckstrin homology domain-containing proteins such as phosphoinositide-dependent kinase and serine/threonine protein kinase B, which then control Rac/Rho/Cdc42-dependent pathways. In accordance with the studies in inflammatory cells, we showed that both chemotaxis and inhibition of chemotaxis in breast cancer cells require Gi and PI(3)K. Activation of Gi-signaling with the α2-AR ligand epinephrine or by the wasp venom mastoparan blocked basal invasion of the MDAMB231 cells. Mastoparan was also a very potent inhibitor of migration of the poorly invasive MCF7 cells (Fig. 9). Thus, the observed chemoinhibition seen on activation of PAR1 is not restricted to PAR1 alone, but may be achieved by other Gi-coupled receptors.

We propose three possible pathways, which are not mutually exclusive, that would explain the inhibition of cell motility and invasion of the breast cancer cells by activated PAR1 (Fig. 10). The first explanation is that by activating PAR1 we have simply desensitized the chemokine receptors by receptor phosphorylation (31). We ruled out participation of PKC in the PAR1-mediated chemoinhibition of the MDAMB231 cells, because addition of the PKC inhibitor GFX did not block the inhibitory effects of thrombin on invasion (Fig. 8). If cross-desensitization of chemokine receptors is an important contributor to the PAR1-dependent chemoinhibition, it was not elicited on equally robust stimulation of the homologous PAR2 and PAR4 receptors (Fig. 2). A second possibility is that activated PAR1 steals downstream components that are necessary for proper chemokine signaling. These sequestered components could include certain Gi(βγ) isoforms or other immediate effectors in the PAR1 microenvironment that are not accessible to PAR2 or PAR4. A third possibility is that activation of PAR1 causes inappropriate or hyperstimulation of signaling pathways, which interfere with the proper cyclic pattern of signaling invoked by ligand-internalized and recycled chemokine receptors (31). Hyperstimulation could occur at essentially any level, from the immediate downstream effectors of PAR1 such as βγ→PI(3)Kγ to the more temporally remote components that regulate focal adhesion turnover and the contraction/relaxation apparatus. Hyperstimulation by PAR1 could be exacerbated by the fact that PARs are irreversibly cleaved by their cognate proteases. Because stimulation of PAR1, the α2-AR, or the addition of the direct Gi activator mastoparan all cause inhibition of MDAMB231 cell migration and invasion, we should consider the likelihood that a common mechanism through the Giβγ→PI(3)Kγ→Rho family GTPases disrupts coordinated cell motion.

Intriguingly, we show for the first time that both PAR2 and PAR4 can act as chemokine receptors in MDAMB231 (Fig. 3,B) and in MCF7 cells (Fig. 7). This is relevant because tumors would be expected to generate trypsin-like proteases, which can activate these two receptors. Like PAR1 (35, 36), PAR2 is a known Gi-coupler (21), and addition of PTx to the MDAMB231 cells causes a 50% reduction in Ca2+ levels in response to the PAR2 agonist SLIGKV.5 Thus, to the extent that chemotaxis requires a receptor that can activate Gi to release the appropriate βγ subunits (23), both PAR2 and PAR4 likely can also interact with Gi in both invasive and noninvasive breast cancer cells. Hence the ability to couple to Gi may be a necessary but not sufficient condition to cause chemoinhibition of a highly invasive breast cancer cell such as MDAMB231. Subtle differences in the intracellular loop structures of the PAR1 versus PAR4 and/or perhaps the magnitude and duration of signals emanating from Gi/βγ subunits or from other immediate effectors may dictate whether a particular PAR can act as either a chemotactic or chemoinhibitory receptor.

From an analysis of the gene expression patterns of the NCI-60 cancer cell panel (34), we note that unlike their epithelial/carcinoma progenitors, highly invasive and motile breast cancer cells such as MDAMB231 gain high levels of expression of Gi-couplers such as PAR1 and the α2-ARs. These receptors are expressed in platelets and other vascular cells and are essential for adhesion of the platelets to the vascular wall during blood coagulation. A compelling biological rationale for our results is that the breast cancer cells use these receptors to adhere to the endothelium and underlying matrix on entering the blood stream during the later vascular invasive and metastatic stage. Corroborating evidence for this mechanism has been shown in melanoma cells, which express PAR1 and tissue factor (20). The trypsin-activated PAR2 and PAR4 receptors are also expressed in the invasive cancer cells. Therefore, it is possible that proteases with trypsin-like specificity could be used to facilitate invasion of cancer cells by recruiting the chemotactic abilities of PAR2 and PAR4. Once the cell has penetrated into the vasculature and encounters the highly specific serum protease thrombin, then PAR1 would activate adhesion proteins for attachment at distant sites and metastasis. This scenario presents an exciting opportunity to impact the initial mobilization of breast cancer cells prior to their invasion into the vascular system. If Gi-couplers are critical for the evolution of breast cancer cells as they acquire the ability to mobilize and invade the vasculature, then therapeutics that target selected Gi-couplers, such as PAR1, represent a novel strategy to alter the biological behavior of tumor cells in the context of hormonal or other ongoing treatments.

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.

      
1

The costs of publication of this article were defrayed in part by 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.

      
2

Supported by NIH Grant R01HL57905 and by a Scholar Award from the PEW Scholars Program in the Biomedical Sciences.

            
4

The abbreviations used are: ECM, extracellular matrix; PAR, protease-activated receptor; PI(3)K, phosphoinositide-3-OH kinase; PKC, protein kinase C; PMA, phorbol-12,13-dibutyrate; MTT, 3-(4,5-dimethylthiazole-2-yl-2,5-diphenyl tetrazolium; PTx, pertussis toxin; PPACK, D-phenylalanyl-L-propyl arginine chloromethylketone; TLCK, L-1-chlor-3-(4-tosylamido)-7-amino-2-heptanon-hydrochloride; FBS, fetal bovine serum; CM, conditioned medium; GFX, GF109203X; LY, LY294002; PIP3, phosphatidyl inositol-3,4,5-triphosphate; α2-AR, α2-adrenergic receptor.

      
5

L. Kamath, and A. Kuliopulos, unpublished results.

Fig. 1.

PAR expression profiles of MDAMB231 and MCF7 cells. Cells (1 × 106) were probed with 1:400, 1:100, 1:100, and 1:30 dilutions of anti-PAR1, anti-PAR2, anti-PAR3, and anti-PAR4 antibodies (1.05, 0.2, 1.8, and 0.42 μg/ml stock solutions, respectively) and treated with FITC-labeled secondary antibody. FITC-labeled cells were analyzed by flow cytometry. Thin lines, background fluorescence (secondary antibody alone); thick lines, fluorescence shift attributable to PAR expression. Traces shown are representative of one of five independent experiments.

Fig. 1.

PAR expression profiles of MDAMB231 and MCF7 cells. Cells (1 × 106) were probed with 1:400, 1:100, 1:100, and 1:30 dilutions of anti-PAR1, anti-PAR2, anti-PAR3, and anti-PAR4 antibodies (1.05, 0.2, 1.8, and 0.42 μg/ml stock solutions, respectively) and treated with FITC-labeled secondary antibody. FITC-labeled cells were analyzed by flow cytometry. Thin lines, background fluorescence (secondary antibody alone); thick lines, fluorescence shift attributable to PAR expression. Traces shown are representative of one of five independent experiments.

Close modal
Fig. 2.

Functional activity of PARs in breast cancer cells. Intracellular calcium fluxes in response to PAR agonists were measured on a Perkin-Elmer LS50B Luminescence Spectrofluorometer. Cells were lifted with PBS/1.5 mm EDTA, thoroughly washed, resuspended in KRB buffer (118 mM NaCl, 1 mM CaCl2, 1 mM KH2PO4, 1.2 mM MgSO4·7H2O, 1 mg/mL BSA, 0.72 mg/mL dextrose, 20 mM Hepes, pH 7.2) (12) at 106 cells/ml, and labeled with 2.5 μm fura2/AM at 37°C for 30 min with gentle shaking. Ca2+ fluorescence experiments were performed at 25°C with emission recorded at 510 nm with dual excitation at 340 and 380 nm as described (12). The individual traces shown are representative of four independent experiments.

Fig. 2.

Functional activity of PARs in breast cancer cells. Intracellular calcium fluxes in response to PAR agonists were measured on a Perkin-Elmer LS50B Luminescence Spectrofluorometer. Cells were lifted with PBS/1.5 mm EDTA, thoroughly washed, resuspended in KRB buffer (118 mM NaCl, 1 mM CaCl2, 1 mM KH2PO4, 1.2 mM MgSO4·7H2O, 1 mg/mL BSA, 0.72 mg/mL dextrose, 20 mM Hepes, pH 7.2) (12) at 106 cells/ml, and labeled with 2.5 μm fura2/AM at 37°C for 30 min with gentle shaking. Ca2+ fluorescence experiments were performed at 25°C with emission recorded at 510 nm with dual excitation at 340 and 380 nm as described (12). The individual traces shown are representative of four independent experiments.

Close modal
Fig. 3.

Activation of PAR1 but not PAR2 or PAR4 inhibits basal invasion of breast cancer cells through Matrigel. MDAMB231 cells (25,000) were seeded in the upper chamber of a transwell apparatus equipped with a 6.5-mm polycarbonate filter (8 μm pore) layered with 40 μg of Matrigel. Lower wells contained 600 μl of NIH3T3-CM with or without PAR agonist. Data are means (n = 4) of experiments repeated 2–5 times; bars, ± SE. Basal invasion (2,600–3,100 cells) toward CM alone was normalized to 100%. A, statistically significant (P < 0.0001) inhibition of cell invasion occurred at ≥10 nm thrombin with IC50 of 5 ± 2 nm. B, statistically significant (P < 0.018) inhibition of cell invasion occurred at SFLLRN concentrations ≥0.1 μm with IC50 of 200 ± 100 nm but no statistically significant differences in invasion were seen for the treatment with 0.1–100 μm SLIGKV (P ≥ 0.19). The only statistically significant difference from untreated control in the GYPGKF data was at 100 μm peptide (P = 0.02). ▪, percentage invasion of MDAMB231 cells when 100 nm thrombin (A) or 500 μm SFLLRN (B) was added to top well in 200 μl 0.1% BSA with CM alone added to the bottom well. No statistically significant (P ≥ 0.18) differences were observed between the samples treated with agonist in the top well as compared with untreated controls.

Fig. 3.

Activation of PAR1 but not PAR2 or PAR4 inhibits basal invasion of breast cancer cells through Matrigel. MDAMB231 cells (25,000) were seeded in the upper chamber of a transwell apparatus equipped with a 6.5-mm polycarbonate filter (8 μm pore) layered with 40 μg of Matrigel. Lower wells contained 600 μl of NIH3T3-CM with or without PAR agonist. Data are means (n = 4) of experiments repeated 2–5 times; bars, ± SE. Basal invasion (2,600–3,100 cells) toward CM alone was normalized to 100%. A, statistically significant (P < 0.0001) inhibition of cell invasion occurred at ≥10 nm thrombin with IC50 of 5 ± 2 nm. B, statistically significant (P < 0.018) inhibition of cell invasion occurred at SFLLRN concentrations ≥0.1 μm with IC50 of 200 ± 100 nm but no statistically significant differences in invasion were seen for the treatment with 0.1–100 μm SLIGKV (P ≥ 0.19). The only statistically significant difference from untreated control in the GYPGKF data was at 100 μm peptide (P = 0.02). ▪, percentage invasion of MDAMB231 cells when 100 nm thrombin (A) or 500 μm SFLLRN (B) was added to top well in 200 μl 0.1% BSA with CM alone added to the bottom well. No statistically significant (P ≥ 0.18) differences were observed between the samples treated with agonist in the top well as compared with untreated controls.

Close modal
Fig. 4.

Inhibitors of thrombin and serine proteases or a PAR-1 blocking antibody do not prevent high basal invasion of MDAMB231 cells. Matrigel invasion assays were conducted as in Fig. 3. Lower wells contained 600 μl of NIH3T3-CM with or without thrombin and inhibitors. Basal invasion was normalized to 100% (2200 ± 300 cells). Experiments were conducted twice using two separate isolates of MDAMB231 cells and show the mean for each condition (n = 4) pooling data from both experiments; bars, ± SE. The only statistically significant difference in invasion from untreated control was with 10 nm thrombin alone (P = 0.025).

Fig. 4.

Inhibitors of thrombin and serine proteases or a PAR-1 blocking antibody do not prevent high basal invasion of MDAMB231 cells. Matrigel invasion assays were conducted as in Fig. 3. Lower wells contained 600 μl of NIH3T3-CM with or without thrombin and inhibitors. Basal invasion was normalized to 100% (2200 ± 300 cells). Experiments were conducted twice using two separate isolates of MDAMB231 cells and show the mean for each condition (n = 4) pooling data from both experiments; bars, ± SE. The only statistically significant difference in invasion from untreated control was with 10 nm thrombin alone (P = 0.025).

Close modal
Fig. 5.

Thrombin and SFLLRN inhibit chemotaxis of MDAMB231 cells when applied as a concentration gradient in the direction of cell migration. Migration of MDAMB231 cells (50,000/upper well) in response to the indicated concentrations of thrombin or SFLLRN placed in upper and/or lower well was determined using a transwell apparatus. Lower wells contained 600 μl of NIH3T3-CM as the chemoattractant. Data are number of cells (in thousands) that migrated to the underside of the membrane in 5 h at 37°C. Each data point gives the mean (n = 4); bars, ± SE. The experiment was performed 4 times with essentially identical results.

Fig. 5.

Thrombin and SFLLRN inhibit chemotaxis of MDAMB231 cells when applied as a concentration gradient in the direction of cell migration. Migration of MDAMB231 cells (50,000/upper well) in response to the indicated concentrations of thrombin or SFLLRN placed in upper and/or lower well was determined using a transwell apparatus. Lower wells contained 600 μl of NIH3T3-CM as the chemoattractant. Data are number of cells (in thousands) that migrated to the underside of the membrane in 5 h at 37°C. Each data point gives the mean (n = 4); bars, ± SE. The experiment was performed 4 times with essentially identical results.

Close modal
Fig. 6.

MDAMB231 cancer cells migrate toward CM by a chemotactic but not by a chemokinetic mechanism. Migration of MDAMB231 cells (50,000/upper well) was determined using a transwell apparatus. Upper and lower wells contained various dilutions of NIH3T3 CM. Data shown (n = 4) are number of cells (in thousands) that migrated to the underside of the membrane in 5 h at 37°C and are representative of three independent experiments; bars, mean ± SE.

Fig. 6.

MDAMB231 cancer cells migrate toward CM by a chemotactic but not by a chemokinetic mechanism. Migration of MDAMB231 cells (50,000/upper well) was determined using a transwell apparatus. Upper and lower wells contained various dilutions of NIH3T3 CM. Data shown (n = 4) are number of cells (in thousands) that migrated to the underside of the membrane in 5 h at 37°C and are representative of three independent experiments; bars, mean ± SE.

Close modal
Fig. 7.

The poorly invasive MCF7 breast cancer cell line exhibits chemotaxis toward PAR2 and PAR4 agonists. Migration of MCF7 cells (50,000/upper well) was measured in a transwell apparatus as in Fig. 5. Each data point gives the mean (n = 4); bars, ± SE. The experiment was performed 3 times with similar results. A, statistically significant (P = 0.05) increase in cell invasion occurred at 100 nm thrombin and near statistical significance (P = 0.08) at 10 nm thrombin. Addition of TFLLRN in B caused no statistically significant differences relative to untreated control. The only statistically significant increase from untreated control in the SLIGKV data (C) was at 1 μm peptide (P = 0.04). D, addition of 1 mm GYPGKF caused a near statistically significant (P = 0.06) increase in cell migration as compared with untreated control.

Fig. 7.

The poorly invasive MCF7 breast cancer cell line exhibits chemotaxis toward PAR2 and PAR4 agonists. Migration of MCF7 cells (50,000/upper well) was measured in a transwell apparatus as in Fig. 5. Each data point gives the mean (n = 4); bars, ± SE. The experiment was performed 3 times with similar results. A, statistically significant (P = 0.05) increase in cell invasion occurred at 100 nm thrombin and near statistical significance (P = 0.08) at 10 nm thrombin. Addition of TFLLRN in B caused no statistically significant differences relative to untreated control. The only statistically significant increase from untreated control in the SLIGKV data (C) was at 1 μm peptide (P = 0.04). D, addition of 1 mm GYPGKF caused a near statistically significant (P = 0.06) increase in cell migration as compared with untreated control.

Close modal
Fig. 8.

Effect of signal transduction activators and inhibitors on invasion of MDAMB231 cells. MDAMB231 cells (25,000/upper well) were seeded in the upper well. Lower wells contained 600 μl of NIH3T3-CM as the chemoattractant with or without agonist/inhibitor. Basal invasion (1900 ± 50 cells) was normalized to 100%. Inhibitors and/or agonists used in the experiment were: CM alone; CM + 100 nm thrombin; CM + 100 ng/ml PTx; CM + 100 ng/ml PTx + 100 nm thrombin (thr); CM + 10 μm PMA (PKC activator); CM + 10 μm PMA + 100 nm thr; CM + 10 μm GFX (PKC inhibitor); CM + 10 μm GFX + 100 nm thr; CM + 10 μm genistein (tyrosine kinase inhibitor); CM + 10 μm genistein + 100 nm thr; and CM + 10 μm LY [LY; PI(3)K inhibitor]; and CM + 10 μm LY + 100 nm thr. Data shown are percentage of cells (n = 4) that invaded the Matrigel in 40 h at 37°C and are representative of four independent experiments; bars, mean ± SE. The data with 10 μm GFX in the absence of thrombin gave near statistically significant (P = 0.07) inhibition of cell invasion. All of the other conditions caused statistically significant (P ≤ 0.0002) decreases in cell invasion relative to untreated (no thrombin) control.

Fig. 8.

Effect of signal transduction activators and inhibitors on invasion of MDAMB231 cells. MDAMB231 cells (25,000/upper well) were seeded in the upper well. Lower wells contained 600 μl of NIH3T3-CM as the chemoattractant with or without agonist/inhibitor. Basal invasion (1900 ± 50 cells) was normalized to 100%. Inhibitors and/or agonists used in the experiment were: CM alone; CM + 100 nm thrombin; CM + 100 ng/ml PTx; CM + 100 ng/ml PTx + 100 nm thrombin (thr); CM + 10 μm PMA (PKC activator); CM + 10 μm PMA + 100 nm thr; CM + 10 μm GFX (PKC inhibitor); CM + 10 μm GFX + 100 nm thr; CM + 10 μm genistein (tyrosine kinase inhibitor); CM + 10 μm genistein + 100 nm thr; and CM + 10 μm LY [LY; PI(3)K inhibitor]; and CM + 10 μm LY + 100 nm thr. Data shown are percentage of cells (n = 4) that invaded the Matrigel in 40 h at 37°C and are representative of four independent experiments; bars, mean ± SE. The data with 10 μm GFX in the absence of thrombin gave near statistically significant (P = 0.07) inhibition of cell invasion. All of the other conditions caused statistically significant (P ≤ 0.0002) decreases in cell invasion relative to untreated (no thrombin) control.

Close modal
Fig. 9.

Activation of Gi signaling by epinephrine or mastoparan inhibits invasion of MDAMB231 and migration of MCF7 cells. Cells (50,000) were seeded in the upper well. Data shown are percentage of cells (n = 4) that invaded the Matrigel in 40 h at 37°C (MDAMB231) or those that migrated to the underside of the membrane in 5 h in the absence of Matrigel (MCF7); bars, mean ± SE. Basal invasion (1,900 ± 100 MDAMB231 cells for ≤100 nm and 5,400 ± 700 MDAMB231 cells for ≥100 nm mastoparan 5280 and epinephrine) and migration (2,300 ± 300 MCF7 cells) was normalized to 100%. Data shown are selected from one of four independent experiments. Statistically significant (P ≤ 0.02) inhibition of MDAMB231 invasion was seen at ≥1 μm mastoparan with IC50 of 0.4 ± 0.2 μm, and ≥10 μm epinephrine with IC50 of 1.5 ± 0.8 μm. Statistically significant (P < 0.02) inhibition of MCF7 migration was seen at ≥10 nm mastoparan with IC50 of 6 ± 3 nm.

Fig. 9.

Activation of Gi signaling by epinephrine or mastoparan inhibits invasion of MDAMB231 and migration of MCF7 cells. Cells (50,000) were seeded in the upper well. Data shown are percentage of cells (n = 4) that invaded the Matrigel in 40 h at 37°C (MDAMB231) or those that migrated to the underside of the membrane in 5 h in the absence of Matrigel (MCF7); bars, mean ± SE. Basal invasion (1,900 ± 100 MDAMB231 cells for ≤100 nm and 5,400 ± 700 MDAMB231 cells for ≥100 nm mastoparan 5280 and epinephrine) and migration (2,300 ± 300 MCF7 cells) was normalized to 100%. Data shown are selected from one of four independent experiments. Statistically significant (P ≤ 0.02) inhibition of MDAMB231 invasion was seen at ≥1 μm mastoparan with IC50 of 0.4 ± 0.2 μm, and ≥10 μm epinephrine with IC50 of 1.5 ± 0.8 μm. Statistically significant (P < 0.02) inhibition of MCF7 migration was seen at ≥10 nm mastoparan with IC50 of 6 ± 3 nm.

Close modal
Fig. 10.

Proposed mechanisms for PAR1-dependent inhibition of breast cancer cell migration and invasion. Akt, serine/threonine protein kinase B; PDK, phosphoinositide-dependent kinase; GEF, guanine nucleotide exchange factor; PH, pleckstrin homology domain.

Fig. 10.

Proposed mechanisms for PAR1-dependent inhibition of breast cancer cell migration and invasion. Akt, serine/threonine protein kinase B; PDK, phosphoinositide-dependent kinase; GEF, guanine nucleotide exchange factor; PH, pleckstrin homology domain.

Close modal
Table 1

Effects of PAR1 agonists and signaling pathway activators or inhibitors on viability of MDAMB231 cells

ReagentMTT-cell viabilitya (%)
Control (RPMI 1640 + 10% FBS) 100 ± 7 
Vincristine (10 μm–50 μm)b <0.3c 
Thrombin  
 0.1 nm 90 ± 6 
 1 nm 92 ± 2 
 10 nm 96 ± 6 
 100 nm 90 ± 5 
SFLLRN  
 0.1 μm 93 ± 3 
 1 μm 92 ± 1 
 10 μm 97 ± 3 
 100 μm 98 ± 1 
100 ng/ml PTx 132 ± 3d 
100 ng/ml PTx + 100 nm thrombin 119 ± 1d 
10 μm PMA 84 ± 1c 
10 μm GFX 91 ± 5 
10 μm genistein 86 ± 1c 
10 μm mastoparan 5280 99 ± 6 
10 μm LY 73 ± 3d 
ReagentMTT-cell viabilitya (%)
Control (RPMI 1640 + 10% FBS) 100 ± 7 
Vincristine (10 μm–50 μm)b <0.3c 
Thrombin  
 0.1 nm 90 ± 6 
 1 nm 92 ± 2 
 10 nm 96 ± 6 
 100 nm 90 ± 5 
SFLLRN  
 0.1 μm 93 ± 3 
 1 μm 92 ± 1 
 10 μm 97 ± 3 
 100 μm 98 ± 1 
100 ng/ml PTx 132 ± 3d 
100 ng/ml PTx + 100 nm thrombin 119 ± 1d 
10 μm PMA 84 ± 1c 
10 μm GFX 91 ± 5 
10 μm genistein 86 ± 1c 
10 μm mastoparan 5280 99 ± 6 
10 μm LY 73 ± 3d 
a

Cells (10,000/well) were cultured in 96-well microtiter plates in the presence of the listed reagents for 40–48 h at 37°C. Cell viability was assessed by the MTT mitochondrial activity assay (22) with the optical density measured at 570 and 670 nm. The data show the mean ± SE of quadruplicate samples.

b

Vincristine was added as a positive control for apoptosis in all experiments.

c

P < 0.0001 (two-tailed t test).

d

P ≤ 0.003 (two-tailed t test).

We thank Jaya Goel for helpful suggestions to standardize the invasion assay experiments, Vimla Band for the use of her microscopy equipment, Lidija Covic and John Erban for many thoughtful discussions, and members of the Kuliopulos lab for critical review of the manuscript.

1
Grand R. J., Turnell A. S., Grabham P. W. Cellular consequences of thrombin-receptor activation.
Biochem. J.
,
313
:
353
-368,  
1996
.
2
Huang Y-Q., Li J-J., Karpatkin S. Thrombin inhibits tumor cell growth in association with UP-regulation of p21waf/cip1 and caspases via a p53-independent, STST1-dependent pathway.
J. Biol. Chem.
,
275
:
6462
-6468,  
2000
.
3
Covic L., Gresser A. L., Kuliopulos A. Biphasic kinetics of activation and signaling for PAR1 and PAR4 thrombin receptors in platelets.
Biochemistry
,
39
:
5458
-5467,  
2000
.
4
Fischer E. G., Wolfram R., Mueller B. M. Tissue factor-initiated thrombin generation activates the signaling thrombin receptor on malignant melanoma cells.
Cancer Res.
,
55
:
1629
-1632,  
1995
.
5
Nierodzik M. L., Bain R. M., Liu L. X., Shivji M., Takeshita K., Karpatkin S. Presence of the seven transmembrane thrombin receptor on human tumour cells: effect of activation on tumour adhesion to platelets and tumor tyrosine phosphorylation.
Br. J. Haematol.
,
92
:
452
-457,  
1996
.
6
Vu T-K. H., Hung D. T., Wheaton V. I., Coughlin S. R. Molecular cloning of a functional thrombin receptor reveals a novel proteolytic mechanism of receptor action.
Cell
,
64
:
1057
-1068,  
1991
.
7
Nystedt S., Emilsson K., Wahlestedt C., Sundelin J. Molecular cloning of a potential proteinase activated receptor.
Proc. Natl. Acad. Sci. USA
,
91
:
9208
-9212,  
1994
.
8
Ishihara H., Connolly A. J., Zeng D., Kahn M. L., Zheng Y. W., Timmons C., Tram T., Coughlin S. R. Protease-activated receptor 3 is a second thrombin receptor in humans.
Nature (Lond.)
,
386
:
502
-506,  
1997
.
9
Xu W-F., Andersen H., Whitmore T. E., Presnell S. R., Yee D. P., Ching A., Gilbert T., Davie E. W., Foster D. C. Cloning and characterization of human protease-activated receptor 4.
Proc. Natl. Acad. Sci. USA
,
95
:
6642
-6646,  
1998
.
10
Kahn M. L., Zheng Y-W., Huang W., Bigornia V., Zheng D., Moff S., Farese R. V., Tam C., Coughlin S. R. A dual thrombin receptor system for platelet activation.
Nature (Lond.)
,
394
:
690
-694,  
1998
.
11
Vassallo R. R., Kieber-Emmons T., Cichowski K., Brass L. R. Structure-function relationships in the activation of platelet thrombin receptors by receptor-derived peptides.
J. Biol. Chem.
,
267
:
6081
-6085,  
1992
.
12
Kuliopulos A., Covic L., Seeley S. K., Sheridan P. J., Helin J., Costello C. E. Plasmin desensitization of the PAR1 thrombin receptor: kinetics. sites of truncation, and implications for thrombolytic therapy.
Biochemistry
,
38
:
4572
-4585,  
1999
.
13
Even-Ram S., Uziely B., Cohen P., Grisaru-Granovsky S., Maoz M., Ginzburg Y., Reich R., Vlodavsky I., Bar-Shavit R. Thrombin receptor overexpression in malignant and physiological invasion processes.
Nat. Med.
,
4
:
909
-914,  
1998
.
14
Henrikson K. P., Salazar S. L., Fenton J. W., Pentecost B. T. Role of thrombin receptor in breast cancer invasiveness.
Br. J. Cancer
,
79
:
401
-406,  
1999
.
15
Luboshits G., Shina S., Kaplan O., Engelberg S., Nass D., Lifshitz-Mercer B., Chaitchik S., Keydar I., Ben-Baruch A. Elevated expression of the CC chemokine regulated on activation, normal T cell expressed and secreted (RANTES) in advanced breast carcinoma.
Cancer Res.
,
59
:
4681
-4687,  
1999
.
16
Youngs S. J., Ali S. A., Taub D. D., Rees R. C. Chemokines induce migrational responses in human breast carcinoma cell lines.
Int. J. Cancer
,
71
:
257
-266,  
1997
.
17
Kleiner D. E., Stetler-Stevenson W. G. Matrix metalloproteinases and metastasis.
Cancer Chemother. Pharmacol.
,
43
:
S42
-S51,  
1999
.
18
Horwitz A. R., Parsons J. T. Cell migration-Movin’ on.
Science (Wash. DC)
,
286
:
1102
-1103,  
1999
.
19
Nierodzik M. L., Kajumo F., Karpatkin S. Effect of thrombin treatment of tumor cells on adhesion of tumor cells to platelets in vitro and metastasis in vivo.
Cancer Res.
,
52
:
3267
-3272,  
1992
.
20
Nierodzik M. L., Chen K., Takeshita K., Li J. J., Huang Y. Q., Feng X. S., D’Andrea M. R., Andrade-Gordon P., Karpatkin S. Protease-activated receptor 1 (PAR-1) is required and rate-limiting for thrombin-enhanced experimental pulmonary metastasis.
Blood
,
92
:
3694
-3700,  
1998
.
21
Miyata S., Koshikawa N., Yasumitsu H., Miyazaki K. Trypsin stimulates integrin α5β1-dependent adhesion to fibronectin and proliferation of human gastric carcinoma cells through activation of proteinase-activated receptor-2.
J. Biol. Chem.
,
275
:
4592
-4598,  
2000
.
22
Albini A., Iwamoto Y., Kleinman H. K., Martin G. R., Aaronson S. A., Kozlowski J. M., McEwan R. N. A rapid in vitro assay for quantitating the invasive potential of tumor cells.
Cancer Res.
,
47
:
3239
-3245,  
1987
.
23
Neptune E. R., Bourne H. R. Receptors induce chemotaxis by releasing the βγ subunit of Gi, not by activating Gq or Gs.
Proc. Natl. Acad. Sci. USA
,
94
:
14489
-14494,  
1997
.
24
2nd ed. Clarke R. Leonessa F. Brunner W. N. Thompson E. W. eds. .
In Vitro Models. Diseases of Breast
,
:
335
-354, Lippincott Williams & Wilkins Philadelphia, PA  
2000
.
25
Kleinman H. K., McGarvey M. L., Liotta L. A., Robey P. G., Tryggvason K., Martin G. R. Isolation and characterization of type IV procollagen, laminin, and heparin sulfate proteoglycan from the EHS sarcoma.
Biochemistry
,
21
:
6188
-6193,  
1982
.
26
Zain J., Huang Y., Feng X. S., Nierodzik M. L., Li J-J., Karpatkin S. Concentration-dependent dual effect of thrombin on impaired growth/apoptosis or mitogenesis in tumor cells.
Blood
,
95
:
3133
-3138,  
2000
.
27
Arai H., Tsou C., Charo I. F. Chemotaxis in a lymphocyte cell line transfected with C-C chemokine receptor 2B: evidence that directed migration is mediated by βγ dimers released by activation of Gi-coupled receptors.
Proc. Natl. Acad. Sci. USA
,
94
:
14495
-14499,  
1997
.
28
Servant G., Weiner O. D., Herzmark P., Balla T., Sedat J. W., Bourne H. R. Polarization of chemoattractant receptor signaling during neutrophil chemotaxis.
Science (Wash. DC)
,
287
:
1037
-1040,  
2000
.
29
Li Z., Jiang H., Xie W., Zhang Z., Smrcka A. V., Wu D. Roles of PLC-β2 and-β3 and the PI3Kγ in chemoattractant-mediated signal transduction.
Science (Wash. DC)
,
287
:
1046
-1049,  
2000
.
30
Ali H., Tomhave E. D., Richardson R. M., Haribabu B., Snyderman R. Thrombin primes responsiveness of selective chemoattractant receptors at a site distal to G protein activation.
J. Biol. Chem.
,
271
:
3200
-3206,  
1996
.
31
Ali H., Richardson R. M., Haribabu B., Snyderman R. Chemoattractant receptor cross-desensitization.
J. Biol. Chem.
,
274
:
6027
-6030,  
1999
.
32
Shao Z-M., Wu J., Shen Z-Z., Barsky S. H. Genistein exerts multiple suppressive effects on human breast carcinoma cells.
Cancer Res.
,
58
:
4851
-4857,  
1998
.
33
Keularts I. M., Gorp R. M. V., Feijge M. A., Vuist W. M., Heemskerk J. W. α(2A)-adrenergic receptor stimulation potentiates calcium release in platelets by modulating cAMP levels.
J. Biol. Chem.
,
275
:
1763
-1772,  
2000
.
34
Ross D. T., Scherf U., Eisen M. B., Perou C. M., Rees C., Spellman P., Iyer V., Jeffrey S. S., Van de Rijn M., Waltham M., Pergamenschikov A., Lee J. C. F., Lashkari D., Shalon D., Myers T. G., Weinstein J. N., Botstein D., Brown P. O. Systematic variation in gene expression patterns in human cancer cell lines.
Nat. Genet.
,
24
:
227
-235,  
2000
.
35
Barr A. J., Brass L. F., Manning D. R. Reconstitution of receptors and GTP-binding regulatory proteins (G proteins) in Sf9 cells.
J. Biol. Chem.
,
272
:
2223
-2229,  
1997
.
36
Swift S. M., Sheridan P., Covic L., Kuliopulos A. PAR1 thrombin receptor-G protein interactions: separation of binding and coupling determinants in the Gα subunit.
J. Biol. Chem.
,
275
:
2627
-2635,  
2000
.