Correlation of Protease-Activated Receptor-1 With Differentiation Markers in Squamous Cell Carcinoma of the Head and Neck and Its Implication in Lymph Node Metastasis

Metastasis is a major reason for the poor prognosis of patients with squamous cell carcinoma of the head and neck (SCCHN), and it is the single most predictive feature of clinical outcome. Many patients with SCCHN present with lymph node metastases at the time of diagnosis, and the 5-year survival rate for patients with nodal metastases is ∼30% to 40% (1). Cancer cell metastasis is a multistep process including detachment from the primary tumor site, migration and implantation in a secondary site, survival, and rapid expansion in the new environment (2, 3). Each of these steps requires activation and deactivation of multiple specific proteins (4, 5). Modulation of expression and/or functions of these proteins may prevent tumor cells from metastasizing. Therefore, the identification of crucial proteins involved in the metastatic process as either positive or negative regulators will aid in the development of new therapeutic approaches to this disease.

Accumulating evidence suggests that thrombin and its receptors may play an important role in cancer metastasis. Thrombin, a serine protease that is generated after endothelial cell damage, is a multifunctional protein with a variety of biological functions. In addition to its functions in blood coagulation, platelet adhesion, platelet aggregation, and mitogenesis of fibroblast and smooth muscle cells, thrombin has also long been thought to play a role in tumor cell metastasis (6, 7, 8, 9, 10).

Many of the effects of thrombin are mediated through its receptors, including its major receptor, protease-activated receptor-1 (PAR-1). PAR-1, a 66-kDa single polypeptide encoded from a 3.5-kb cDNA first sequenced by Vu et al. (11) in 1991 is a G-protein-coupled receptor with seven transmembrane domains. Unlike many other cell membrane receptors, PAR-1 does not simply form a ligand-receptor complex to become active. Instead, the thrombin binds to the receptor and then cleaves the NH₂ terminus from the receptor, thereby producing an irreversibly activated form of cell surface protein that provides additional cell signaling (11).

PAR-1 expression has been detected in human melanoma (12), colon adenocarcinoma (13), pancreatic cancer (14), and SCCHN (15, 16). However, the role of PAR-1 in metastasis has not been clearly defined in current publications. Even-Ram et al. (17) and Henrikson et al. (18) demonstrated that PAR-1 was expressed at a higher level in highly metastatic breast carcinoma cell lines as compared with breast cancer cell lines with low metastatic potential. An aberrant expression and activation of PAR-1 in human colon cancer cells induced cell proliferation and motility (19). PAR-1 was also reported to be a rate-limiting factor in thrombin-enhanced experimental pulmonary metastasis from a mouse melanoma cell line (20). However, Zain et al. (21) and our group have demonstrated relatively low concentrations of thrombin-enhanced tumor cell growth, whereas higher concentrations impaired cell growth and induced apoptosis (16). Both tumor cell growth inhibition and apoptosis were PAR-1–specific, p53-independent, STAT1-dependent, and associated with up-regulation of p21^waf/cip1 and caspases (21, 22). Furthermore, Kamath et al. (23) also showed that PAR-1 could inhibit migration and invasion of breast cancer cells. The diversity and complexity of PAR-1–mediated signal transduction definitely reflects cell-type specificity. Whether PAR-1 positively or negatively correlates to metastasis is still unclear.

To study metastasis of SCCHN, we previously established highly metastatic SCCHN cell lines from a rarely metastatic SCCHN cell line through in vivo selection using a mouse model of lymph node metastasis of human tumors (24, 25). By using these cell lines, we observed that PAR-1 was significantly down-regulated in the selected highly metastatic cell lines as compared with their poorly metastatic parental cell line (24). Our current study complements this previous work by extending PAR-1 expression studies to human tumor tissues. Furthermore, we investigated the biological consequence for PAR-1 down-regulation in metastatic SCCHN lesions by correlating PAR-1 expression with differentiation markers of SCCHN.

Using an Institutional Review Board-approved consent for tissue acquisition, specimens for this study were obtained from surgical specimens from patients with SCCHN diagnosed at the University of Pittsburgh. The selection criteria of the available formalin-fixed and paraffin-embedded tissue blocks included three groups: primary tumors with positive lymph nodes (Tu-1) and their paired lymph node metastases (Met-1) and primary tumors with negative lymph nodes (Tu-2). The specimens were collected randomly in two separate batches due to their availability. The sample number and distribution were the following: n = 15 (Tu-1), n = 15 (Met-1), and n = 20 (Tu-2) in the first batch and n = 13 (Tu-1), n = 13 (Met-1), and n = 21 (Tu-2) in the second batch, respectively. The clinical information of the second batch was obtained from the surgical pathology files in the University of Pittsburgh after the regulations of the Health Insurance Portability and Accountability Act. The relevant clinical characteristics including original primary tumor site, pathology stage, and histologic grade are listed in Table 1.

The SCCHN cell lines Tu686 and 686LN are paired cell lines established from a primary tumor (Tu) in base of tongue and a lymph node metastasis (LN) from the same patient, respectively (26). They were kindly provided by Dr. Peter G. Sacks (New York University College of Dentistry, New York, NY). 686LN-M3a2, 686LN-M3a3, 686LN-M3b2, and 686LN-M3b3 are highly metastatic cell lines generated by in vivo selection from 686LN that has low metastatic potential in the lymph node of nude mouse as described previously (24). SCCHN cell lines Tu212 and 886LN were established from a primary hypopharynx tumor and cervical lymph node metastases from a laryngeal primary tumor, respectively, from the University of Texas M.D. Anderson Cancer Center (Houston, TX) and provided by Dr. Gary L. Clayman (M.D. Anderson Cancer Center). Additional SCCHN cell lines UPCI-4A, -37A, -37B, and -6B were established from larynx (trans glottis), tonsil, and larynx (epiglottis), respectively, from the University of Pittsburgh Cancer Institute (Pittsburgh, PA), in which UPCI-37A and -4A were from primary tumors, whereas UPCI-37B and -6B were from metastases. They were gifts from Dr. Theresa Whiteside (University of Pittsburgh Cancer Institute). The cell lines were maintained as monolayer cultures in DMEM/F12 medium (1:1) supplemented with 10% fetal bovine serum at 37°C in a humidified atmosphere with 5% CO_2.

PAR-1 expression in tissue specimens was determined using a mouse monoclonal antihuman PAR-1 antibody, SPAN12 (Immunotech-Coulter Corp., Miami, FL). A peptide (NATLDPRSFLLR) recognized by SPAN12 served as a blocking peptide to confirm specificity of the immunohistochemical analysis. This peptide covers thrombin cleavage site from amino acid residues 35 to 46. Therefore, SPAN12 binds to intact PAR-1. Mouse IgG at 1:100 dilutions was used as a negative control. Immunohistochemical analysis on formalin-fixed, paraffin-embedded human specimens was performed according to a modified procedure. In brief, after deparaffinization with xylene and rehydration with EtOH, endogenous peroxidase activity was blocked by incubating the slides in 3% hydrogen peroxide with methanol for 15 minutes. To retrieve the antigens, the tissue slides were heated in a microwave oven in 100 mmol/L of sodium citrate buffer (pH 6.0) for 10 minutes and then allowed to remain at room temperature for 20 minutes. After being washed in PBS, the slides were incubated with 2.5% normal horse serum (Vector Laboratories, Burlingame, CA) to decrease the background signal. Next, the slides were incubated with a 1:50 dilution of anti-PAR-1 primary antibody overnight at 4°C, left at room temperature for 20 minutes, and washed with PBS. Then the slides were incubated with a biotinylated secondary antibody for 20 minutes at room temperature and with biotin-avidin peroxidase conjugate (ABC kit, Vector Laboratories) for 15 minutes at room temperature. The substrate was then added (0.1% 3,3′-diaminobenzidine solution, Sigma Chemical Co., St. Louis, MO, in PBS with 0.01% hydrogen peroxide). Finally, the slides were counterstained with hematoxylin for 50 seconds (Vector Laboratories) and then observed by light microscopy.

Mouse monoclonal antibody, K 1 + 10 + 11, against a complex of human keratins-1, -10, and -11 (Novus Biological Inc., Littleton, CO), was also used for immunohistochemical analysis in the same specimens as used for PAR-1 analysis. The procedure was the same as described above except for the use of a 1:100 dilution of the antikeratin antibody with an incubation time of 1 hour at room temperature.

The intensity of immunohistochemical staining was measured using a numerical scale (0 = no expression, 1⁺ = weak expression, 2⁺ = moderate expression, and 3⁺ = strong expression) and quantified as Weight Index [WI = % Positive Stain (> 0) in Tumor × Intensity Score].

A 3.5-kb par-1 cDNA fragment was obtained from Dr. Shaun R. Coughlin (University of California, San Francisco, CA). The fragment with appropriate modifications was inserted into the pcDNA3 vector in both sense and antisense orientations and was transfected into the 686LN cell line according to a standard Lipofectamine protocol (Invitrogen Life Technologies, Inc., Carlsbad, CA). Stable transfectants were selected with the antibiotic G418 (400 μg/mL). PAR-1 expressions in the stable transfectants were confirmed by both Northern blots and immunoblots.

Northern blotting was performed using the method described previously (24). In brief, mRNA was prepared using the Oligotex mRNA Purification kit (Qiagen, Valencia, CA) from par-1 cDNA-transfected cell lines. Two micrograms of mRNA from each of the samples was used for Northern blot analysis with a 1.0-kb par-1 cDNA fragment as probe. The 1.5-kb glyceraldehyde-3-phosphate dehydrogenase (G3PDH) cDNA fragment was used as an internal control probe. The probes were labeled with ³²P-dCTP. The hybridization signal was quantified by image analysis using an α Imager 2000 system (α Innotech Corporation, San Leandro, CA). The expression level of PAR-1 mRNA for each sample was normalized with G3PDH mRNA expression.

Cells were washed twice in PBS before lysing (20 mmol/L Tris base, 137 mmol/L NaCl, 10% glycerol, 1% NP40, 0.5% Triton X-100, 1 mmol/L phenylmethylsulfonyl fluoride, and 10 μg/mL leupeptin and aproptin) at 4°C for 30 minutes. The lysate was centrifuged at 13,000 rpm at 4°C for 20 minutes. Fifty micrograms of total protein for each sample were separated by 8% to 16% gradient SDS-PAGE (Bio Whittaker Mol. Appl., Inc., Rockland, ME), transferred onto a polyvinylidene difluoride membrane (Millipore Co., Bedford, MA), then probed with corresponding antibodies. Primary antibodies for immunoblotting were mouse antihuman PAR-1 (SPAN 12; 1:200 dilution, Immunotech-Coulter), mouse antihuman involucrin (1:500 dilution, Sigma-Aldrich), mouse antihuman phospho-extracellular signal-regulated kinase (ERK)1/2 (1:500 dilution, Santa Cruz Biotechnology Inc, Santa Cruz, CA), rabbit antihuman ERK1/2 (1:500 dilution, Santa Cruz Biotechnology), and mouse antihuman β-actin (1:5000 dilution, Sigma-Aldrich) as an internal control. Bound antibody binding was detected using the SuperSignal West Pico Chemiluminecsence system (Pierce, Inc., Rockford, IL).

The sense and antisense par-1 cDNA-transfected 686LN cells were seeded in 6-well plates at a concentration of 5 to 8 × 10⁵ cells per well and serum-starved for 24 hours. MAPK activity was then induced by incubation of the cells with 10 μmol/L of thrombin receptor peptide (TRP, Sigma-Aldrich) for 5, 10, and 60 minutes. The expression of phospho-ERK1/2 and total ERK1/2 in whole cell lysates were detected by immunoblotting.

The sense par-1 cDNA and pcDNA3 vector-transfected 686LN cells were seeded into 60-mm dishes (1 × 10⁶ cells per dish). After fasting for 24 hours, two groups of these cells were pretreated with U0126, a mitogen-activated protein/extracellular signal-regulated kinase kinase (MEK) inhibitor (Promega) or LY294002, a phosphatidylinositol 3′-kinase (PI3K) inhibitor (Calbiochem, La Jolla, CA) for 1 hour at concentrations of 20 μmol/L and 50 μmol/L, respectively, and then incubated with 10 μmol/L TRP for another 48 hours. Whole cell lysates were used for immunoblotting analyses of phospho-ERK1/2, ERK, and involucrin expression.

The expression, or the weight index, for each biomarker was calculated as the product of the intensity times the percentage of positive staining in a given sample. In cases where multiple metastatic nodes were collected on a given individual, expression values were averaged over a given individual to yield a single expression value for metastatic samples within the given subject (and thus satisfy the statistical assumption of independent observations). Separate summary statistics were calculated within each of the following groups: Met-1, Tu-1, and Tu-2. Because expression values followed highly skewed distributions, all of the significance levels were calculated using nonparametric statistics (which assessed significance based on the ranks of the data, rather than assuming normality). Significance levels for comparisons of matched samples (i.e., Met-1 versus Tu-1) were calculated using the Wilcoxon signed-rank test, which is the nonparametric analog to the paired t test. For all of the other group comparisons, P values were calculated using either the Wilcoxon rank-sum test (nonparametric unpaired t test) or the Kruskal-Wallis test (nonparametric ANOVA as an overall test of significance between more than two groups). Correlations were calculated using Spearman’s rank correlation.

Distributions of PAR-1 and Keratin expression, by tumor groups, were described using box plots, which are graphical representations of the median, 25th and 75th percentiles, and range of the data. The bottom and top of the box correspond to the 25th and 75th percentiles of the distribution. The median is represented by a horizontal line inside the box. The horizontal bracket and dotted lines extending outside the box represent the entire range of the data.

Expression of PAR-1 protein in primary tumors and lymph node metastases of SCCHN was studied by immunohistochemical analysis. The tissue samples included lymph node-positive primary tumors (Tu-1), their lymph node metastases (Met-1), and lymph node-negative primary tumors (Tu-2). Incubation of PAR-1–specific antibody SPAN12 in the presence of the SPAN12 blocking peptide abolished PAR-1 immunostaining in tissue samples with strong expression of PAR-1, conforming specificity of PAR-1 immunostaining by antibody SPAN12 (Fig. 1,A). The staining pattern of PAR-1 was consistent between the two batches of specimens that were collected separately, so the two were combined for analysis. PAR-1 was expressed in 33 of 41 lymph node-negative primary tumors (80%) and 19 of 28 node-positive primary tumors (68%). The metastatic tumors stained less intensely for PAR-1 than the paired primary tumors (13 of 28, 46%). PAR-1 expression was mainly located at the membrane and in the cytoplasm of the tumor cells, as well as in the endothelial cells (Fig. 1,B). The primary tumors demonstrated a strong staining for PAR-1 protein, whereas the lymph node metastases showed only weak staining of PAR-1. Interestingly, the expression of PAR-1 was often confined to well-differentiated and keratinized squamous cell carcinoma components of the primary tumors, suggesting a possible correlation between PAR-1 expression and squamous cell differentiation. PAR-1 expression was also observed in the spinous and granular cell layers of normal epithelium. Quantitative analysis of PAR-1 immunostaining showed that expression PAR-1 was significantly down-regulated in lymph node metastases as compared with the node negative primary tumors (Median WI = 10.0 for Met-1 versus Median WI = 60.0 for Tu-2, P = 0.01; Fig. 2 A).

Wilcoxon rank-sum test was used to correlate PAR-1 expression (WI) in primary tumors (Tu-1 and Tu-2) with relevant clinical information available for the second batch of the samples. Significant difference in PAR-1 expression was observed between well-differentiated tumors (WD) and poorly differentiated tumors (PD; Table 1). Although there was no significant difference in PAR-1 expression between node-positive and node-negative tumors, poorly differentiated node positive tumors (PD and N+) illustrated significantly different PAR-1 level from well-differentiated node negative tumors (WD and N−). PAR-1 expression was not specifically located in any tumor site in these samples and was not correlated to tumor stage and age of the patients (data not shown). PAR-1 expression in moderate differentiated tumors did not show significant difference as compared with well-differentiated tumors or poorly differentiated tumors, because moderate differentiated tumors are usually an intermediate status between well-differentiated tumors and poorly differentiated tumors (data not shown).

To determine a possible correlation between PAR-1 expression and epithelial differentiation, the same specimens used for PAR-1 stain were studied for expression of common differentiation markers, keratin-1, -10, and -11 (K1–10–11) by immunohistochemical analysis (Fig. 1,B). As we expected, K1–10–11 were highly expressed in well-differentiated tumor nests. Similar to PAR-1, the expression level of K1–10–11 was significantly lower in lymph node metastasis than the node-negative primary tumors (Median WI = 15 versus Median WI = 135, P = 0.005; Fig. 2,B). Statistical analysis showed that the WI for K1–10–11 was positively correlated to the WI of PAR-1 within each tumor groups (Met-1: r = 0.81, P = 0.002; Tu-1: r = 0.72, P = 0.008; and Tu-2: r = 0.71, P = 0.001; Fig. 3 A–C).

Consistent with observations in the lymph node-metastatic mouse model and the human specimens, down-regulation of PAR-1 expression in metastatic SCCHN was also observed in human cell lines. Among nine SCCHN cell lines examined, four of them were established from primary tumors and another five were originally from the lymph node metastases. As shown in Fig. 4A, cell lines from the metastases showed only a trace amount of PAR-1 expression, whereas cell lines from the primary tumors demonstrated higher levels of PAR-1 expression than the metastatic cell lines.

The localization of PAR-1 protein in well-differentiated and keratinized SCCHN components prompted us to examine whether PAR-1 expression might correlate with epithelial differentiation. To test this hypothesis, we initially examined the expression of the differentiation marker, involucrin, in the highly metastatic cell lines 686LN-M3a3 and 686LN-M3b3 established from the lymph node metastasis mouse model. These cell lines expressed lower PAR-1 than their poorly metastatic parental cell line 686LN (24). We found that the expression of involucrin was significantly down-regulated in highly metastatic cell lines as compared with 686LN (Fig. 4,B). The expression pattern was positively correlated with PAR-1 expression in these cell lines. A similar correlation pattern between PAR-1 and involucrin was also observed in the SCCHN cell lines, with an exception in UPCI-4A cells (Fig. 4 A).

To additionally study this correlation, cell line 686LN was stably transfected with sense and antisense par-1 cDNA. PAR-1 expression was higher in the sense par-1 transfectants and lower in the antisense par-1 transfectants than that in the vector-transfected cell control by both Northern (Fig. 4,C) and Western blot analyses (Fig. 4,D). The results also showed that the expression level of involucrin was significantly up-regulated in the sense par-1 transfectants and down-regulated in the antisense par-1 transfectants as compared with the vector-transfected cell controls, suggesting that PAR-1 expression facilitates expression of involucrin (Fig. 4 D).

To determine the possible signaling pathway involved in involucrin induction by PAR-1, we initially examined activation of the MAPK pathway by the PAR-1–specific activator TRP in sense and antisense par-1 cDNA-transfected 686LN cells. PAR-1 has been suggested to induce MAPK activity in several cell types (27, 28). MAPK was activated by TRP within 5 minutes, and the peak lasted at least 30 to 60 minutes (Fig. 5 A). Moreover, the sense par-1 transfectant showed higher MAPK activity than the antisense transfectant, suggesting that a certain PAR-1 level is essential for MAPK activation in 686LN cells.

To determine whether involucrin expression is regulated by PAR-1 through MAPK signaling, a sense par-1 cDNA transfectant was pretreated with the MEK inhibitor U0126 and the PI3K inhibitor LY294002 to block phosphorylation of ERK1/2 and Akt before TRP challenge. Western blot analysis of involucrin was then done to detect changes in protein levels. U0126 at 20 μmol/L significantly blocked phosphorylation of ERK1/2 and led to down-regulation of involucrin in par-1 cDNA-transfected 686LN (Fig. 5 B). A similar result was achieved with U0126 at 5 μmol/L (data not shown), whereas LY294002 up to 50 μmol/L did not show any effect on involucrin expression (data not shown), suggesting that PAR-1 induces involucrin expression at least partially through the MAPK pathway but not the PI3K signaling pathway.

PAR-1, as a G-protein-coupled receptor, has been suggested to involve multiple signal transduction pathways. Studies of a PAR-1-negative mouse fibroblast cell line established from a par-1 knockout mouse demonstrated that PAR-1 is necessary and sufficient for activation of the MAPK signaling pathway (28). This result is consistent with findings that thrombin can activate two major MAPKs, ERK2 and c-Jun NH₂-terminal kinase 1 (29). A possible explanation for MAPK activation by PAR-1 might be similar to other G-protein-coupled receptors that transactivate epidermal growth factor receptor and then induce phosphorylation of MAPK (30, 31), but there has been no direct evidence to date showing that PAR-1 can transactivate epidermal growth factor receptor. On the other hand, it has been reported that G-protein-coupled receptor signal transduction usually involves activating phospholipase C, generating inositol 1,4,5, -triphosphate, increasing cellular Ca⁺², and inducing Ca⁺²-dependent protein kinase C (PKC-α; refs. 32, 33, 34). Recently, Wang et al. (35) proposed a model showing that PAR-1–mediated stimulation of astrocyte proliferation involved the PKC-α, phospholipase C, and PI3K signaling pathways leading to MAPK activation. This model was supported by observations in endothelial cells (36).

Although MAPK activation induced by PKC-α is usually involved in cell proliferation, PKC-α and phospholipase C can also induce differentiation of keratinocytes, the cells that form SCCHN. Activation of PKC-α and phospholipase C induces expression of profilaggrin, loricrin, and involucrin, the basic elements that form the cornified envelope in a normal stratified epithelial layer (37). PKC-α has also been found to induce G₁ arrest, removing cells from the cell cycle (38). Because PKC-α is one of the downstream signal transducers for PAR-1 (35), it is reasonable to speculate that PAR-1 may also contribute to epithelial cell differentiation.

Our observations from both human tissue specimens and cell lines clearly showed that PAR-1 expression is significantly down-regulated in lymph node metastases of SCCHN. Although PAR-1 is expressed in SCCHN, strong staining for PAR-1 was found to be associated with well-differentiated tumor nests and correlated with expression of keratinocyte differentiation markers K1–10–11. Statistical analysis showed that there was significant difference in PAR-1 expression between poorly differentiated and well-differentiated primary tumors. Furthermore, PAR-1 expression in poorly differentiated node-positive tumors (PD and N+) was significantly different from that in well-differentiated node-negative tumors (WD and N−). To explore the effect of PAR-1 expression on differentiation in metastasis, we examined involucrin as a keratinocyte differentiation marker and initially compared its expression in SCCHN cell lines established from primary tumors with those from the respective metastases. We also compared highly metastatic SCCHN cell lines established through in vivo selection with their poorly metastatic parental cell lines. All of these comparisons showed a positive correlation between the expression of PAR-1 and involucrin, which was also down-regulated in metastatic SCCHN cell lines. The sense and antisense par-1 cDNA-transfected SCCHN cell lines provided additional confirmation of this correlation. We found that the expression level of involucrin was significantly up-regulated in the sense par-1 transfectants and down-regulated in the antisense par-1 transfectants as compared with the vector-transfected controls. Our results indicated that a certain level of PAR-1 is essential for involucrin expression, suggesting that PAR-1 may contribute to epithelial differentiation.

Involucrin, loricrin, and transglutaminase-I are considered as precursors for cornified envelope (39, 40). They are cross-linked into the insoluble cornified envelope in the calcium-sensitive and membrane-bound forms in both mouse and human keratinocytes (41, 42). Keratin-1 and -10 are usually considered as early differentiation markers, whereas involucrin, loricrin, transglutaminase-I, and profilaggrin are late markers of kerantinocyte differentiation (37). Expression of involucrin and other differentiation markers can be induced through the PKC-α signal transduction pathway. G-protein-coupled receptor signal transduction usually involves activation of PKC-α (43). PKC-α, in turn, induces and/or activates members of the AP-1 transcription factor family directly through activation of the MAPK pathway. The MAPK pathway then regulates the expression of genes encoding the proteins involved in cornified envelope formation such as involucrin, transglutaminase-I, and loricrin (37).

Because expression level of PKC-α was barely detectable in SCCHN cell line 686LN (data not shown), we directly examined the effect of MAPK signaling on expression of involucrin using its sense and antisense par-1 stable transfectants. The PAR-1–specific inducer, TRP, was used to induce MAPK activity in these cells. The sense par-1 transfectants showed much higher MAPK activity than the antisense transfectants. However, U0126, a MEK inhibitor, blocked MAPK activation and, consequently, inhibited involucrin expression, indicating that PAR-1–induced involucrin expression is regulated through the MAPK pathway. We realize that the concentration of U0126 used in these experiments could completely blocked phosphorylation of ERK1/2, but only partial reduction of involucrin was achieved, suggesting that MAPK is not the only pathway that regulates involucrin expression.

No defined relationship between differentiation and metastasis of SCCHN has been described. However, an inverse correlation between the rate of metastasis and the degree of differentiation has been reported in SCCHN (44, 45). This series of experiments explored the relationship between expression of SCCHN differentiation markers and metastasis. Although pathologically defined differentiation of SCCHN may not be identical to differentiation of keratinocytes, our data clearly illustrated that differentiation markers K1–10–11 were significantly down-regulated in lymph node metastases. Also, involucrin was down-regulated in metastatic cell lines selected from the lymph node-metastatic mouse model, implicating that expression of proteins related to keratinocyte differentiation is not favorable for SCCHN metastasis. On the other hand, alteration of these differentiation markers changes the structure of the cellular skeleton, possibly affecting motility and survival of the SCCHN cells in the circulatory system and lymph nodes. Whether PAR-1 or PAR-1–induced alteration of differentiation markers actively inhibits lymph node metastasis of SCCHN or just serves as a marker for metastasis of SCCHN will require more investigation. Regardless the exact role PAR-1 plays in SCCHN metastasis, our observations suggest that expression of PAR-1 may be a prognosis or a target for prevention of this disease.

Because PAR-1 is known to be involved in multiple signaling pathways, the exact contribution of this G-protein-coupled receptor in cancer progression and metastasis has not been elucidated. Perhaps these multiple signaling transduction pathways stimulated upon PAR-1 activation account for the multiple biological functions or dual functions of PAR-1. These different functions may affect cancer progression and metastasis at different stages and may vary with the specific type of cancer. The exact function of PAR-1 in cancer progression should be defined in different types of cancer.