Gα₁₂-Mediated Pathway Promotes Invasiveness of Nasopharyngeal Carcinoma by Modulating Actin Cytoskeleton Reorganization

In the western world, nasopharyngeal carcinoma (NPC) is a rare head and neck malignancy; in Southeast Asia, it is highly prevalent. Many factors, including viral, genetic, dietary habit, and environment, have been implicated in the etiology of NPC (1). Undifferentiated NPC, the most common type, is characterized by overwhelming lymphocytic infiltration. The association between it and EBV distinguishes NPC from head and neck squamous cell carcinomas, which are often associated with human papillomavirus (2, 3).

NPC is more metastatic than other head and neck carcinomas. Although tumor invasion and metastasis is thought to involve deregulation of cell motility, the mechanisms underlying this dysregulation remain largely unknown. Several extracellular factors, including cytokines and growth factors, may regulate cell motility by modulating the reorganization of cytoskeleton, and many of these extracellular signals are mediated by G protein–coupled receptors (GPCR). The aberrant expression of GPCR has been associated with tumorigenesis of adrenocortical carcinoma (4), and the ectopic expression of vGPCR (a constitutively active GPCR) has been found to help tumor cells escape immune surveillance (5, 6). Moreover, GPCR signaling pathways may play an important role in tumorigenesis in various carcinomas, including non–small cell lung cancer, breast cancer, prostate cancer, melanoma, gastric cancer, diffused large B-cell lymphoma (7), and head and neck squamous cell carcinoma (8, 9). There is no report, however, of the contribution of GPCR to NPC tumorigenesis.

GPCR signaling is modulated primarily through the activation of heterotrimeric G proteins, which are composed of α and βγ subunits. The Gα subunit has four subfamilies: G_s, G_i, G_q, and G₁₂. To activate Gα₁₂, the receptor needs to be bound to ligands, such as lysophosphatidic acid, sphingosine-1-phosphate, thrombin, or thromboxane A₂ (10). The activation of Gα₁₂ triggers RhoA-dependent downstream signaling, regulating actin stress fiber formation, focal adhesion assembly, and phosphorylation of focal adhesion-associated proteins, including FAK, p130 CAS, and paxillin (11, 12). Gα₁₂ mediates divergent physiologic responses, including cytoskeleton organization, cell growth, and activation of mitogen-activated protein kinase and phospholipase D and C, by activating small G protein Rho (13, 14). Furthermore, Gα₁₂ has also been found to regulate cell migration through E-cadherin/β-catenin (14) and tight junction formation through Hsp90/Src signaling (15–17).

Gα₁₂ has been found to activate oncogenic transformation and tumor cell invasion (18–20), although it is not known whether it is also involved in epithelial-to-mesenchymal transition (EMT). EMT is a pathophysiologic process of tumor cells characterized by loss of epithelial junctional proteins and cell polarity and gain of mesenchymal markers. IQ motif containing GTPase activating protein 1 (IQGAP1), an EMT-related factor, is an important effector of Rac1 and Cdc42 for actin dynamics and cell migration (21, 22). Overexpression of IQGAP1 has been associated with decreases in E-cadherin-based cell-cell adhesion (23) and depletion of IQGAP1 with decreases in cell migration and actin meshwork (24). Together, these findings suggest that the regulation of IQGAP1 expression may be critical to the induction of EMT; therefore, it may be important to investigate the molecular mechanism that regulates IQGAP1 during the process of EMT.

In this study, our microarray studies suggest that the Gα₁₂ signaling might be dysregulated during the tumorigenesis and progression of NPC. Functionally, changes in Gα₁₂ expression in NPC cells influenced many cellular activities associated with NPC cell invasiveness and metastasis. Further microarray studies of NPC cells in which Gα₁₂ was knocked down found changes in expression of genes involved in actin cytoskeleton reorganization and EMT. IQGAP1 was found to be a downstream effector of Gα₁₂ signaling in regulating epithelial morphology. In NPC patients, we found a significant correlation between Gα₁₂ level and NPC tumorigenesis and lymph node metastasis. Together, these studies suggest that Gα₁₂ signaling plays a critical role in the tumorigenesis and metastasis of NPC.

Clinical samples. Nine NPC samples and 32 normal nasopharyngeal samples were used for microarray analysis. Clinical data on patients with NPC are reported in Supplementary Table S1. Informed consent was obtained from all participants.

Cell culture. Primary nasopharyngeal epithelial cells obtained from explant cultures were cultured with defined keratinocyte serum-free medium (Invitrogen). Preconfluent primary nasopharyngeal cells were harvested during passages 3 to 10 for RNA extraction. Two NPC-derived cell lines, CNE1 (25) and NPC-TW06 (26) cells, were cultured in DMEM supplemented with 10% fetal bovine serum.

Immunofluorescence and immunohistochemistry. Normal nasopharyngeal epithelium and NPC cells were fixed and immunostained using mouse anti-human cytokeratin AE1/AE3 (DAKO), mouse anti-CD90/Thy-1 (Dianova), rabbit anti-IQGAP1 (Santa Cruz Biotechnology), mouse anti-paxillin (BD Transduction Laboratories), mouse anti-vimentin (Sigma), and mouse anti-vinculin (Sigma) as primary antibodies. Rhodamine-phalloidin was used to label F-actin. Immunohistochemical staining of tissue sections was done as described previously (27) using anti-Gα₁₂ primary antibody (1:100; sc-409; Santa Cruz Biotechnology). Breast and prostate cancer samples were used as the positive controls for Gα₁₂ staining (18, 19). The primary antibody was omitted in the negative controls. The results of immunohistochemistry were reviewed by two independent observers. Staining intensity was defined as “−,” “+,” “++,” and “+++” scoring (negative or <20%, 21-50%, 51-70%, and >71% of positive cells, respectively).

RNA extraction and cDNA microarray experiments. Total RNA was extracted with the RNeasy kit (Qiagen) and analyzed on an Agilent 2100 Bioanalyzer. A pool of total RNA from 32 normal nasopharyngeal epithelium cell strains served as reference. Labeled cDNA samples were hybridized on a customized microarray containing 46,657 cDNA clones (IMAGE consortium), representing ∼26,000 Unigene clusters. The arrays were scanned on a GenePix 4000B scanner and analyzed with the GenePix software package (Axon Instruments). The microarray data can be found at GEO (accession no. GSE14262).

Cell transfection. For small interfering RNA (siRNA) transfection, CNE1 and NPC-TW06 cells were seeded at a density of 5 × 10⁴ per well in the 24-well culture plate 24 h before transfection with either Gα₁₂- or IQGAP1-siRNA or a nontargeting control siRNA (Dharmacon) at a concentration of 90 nmol/L. For functional assays, cells were collected 1 to 3 days post-transfection. For microarray experiments, RNA was extracted from NPC-TW06 cells 48 h post-transfection. For exogenous protein expression, NPC cells were transiently transfected with one of the following expression plasmids, full-length IQGAP1 (pCMV2-Flag-FL-IQGAP1), a COOH-terminal fragment of IQGAP1 (pCMV2-Flag-IQDN4; ref. 28), wild-type Gα₁₂ (pcDNA3-Gα₁₂-WT), or a constitutively active Gα₁₂ (pcDNA3-Gα₁₂Q231L), using Fugene HD transfection reagents (Roche). The transfection efficiency was measured by cotransfection with pEGFP-C1. The Gα₁₂-WT and Gα₁₂Q231L plasmids were obtained from the Missouri S&T cDNA Resource Center.

Western blot. Equal amounts of protein were separated by SDS-PAGE and transferred to polyvinylidene fluoride membranes. Blots were probed with primary antibodies against Gα₁₂ (Santa Cruz Biotechnology), IQGAP1 (Santa Cruz Biotechnology), LAMB3 (Santa Cruz Biotechnology), E-cadherin (BD Transduction Laboratories), vinculin (Sigma), paxillin (BD Transduction Laboratories), vimentin (Sigma), and glyceraldehyde-3-phosphate dehydrogenase (Lab Frontier) followed by horseradish peroxidase–conjugated secondary antibody and developed with enhanced chemiluminescence detection reagents (Amersham Life Science).

Wound-healing assay. At 48 h post-transfection with Gα₁₂-siRNA, IQGAP1-siRNA, or a nontargeting control siRNA, we scratched confluent cell layers with plastic tips. At 0 and 24 h later, we photographed wound widths under a phase-contrast microscope. All experiments were done in triplicate.

Matrigel invasion assay. Invasion capacity was done in an invasion chamber consisting of cell culture inserts with 8 μm pore-sized PET membranes coated with 80 μg Matrigel (BD Biosciences) according to the manufacturer's instructions. CNE1 and NPC-TW06 cells transfected with Gα₁₂-siRNA or a control siRNA were harvested 48 h post-transfection. Aliquots of resuspended cells were seeded into the top chamber. After 30 h incubation at 37°C, the invaded cells were stained with 1% gentian violet. At least five distinct fields were counted for each duplicate chamber.

Quantitative real-time reverse transcription-PCR. Quantitative real-time reverse transcription-PCR (QRT-PCR) was done in a LightCycler PCR system (Roche) with the FastStart DNA Master SYBR Green I Kit (Roche Applied Science). Refer to Supplementary Table S2 for the primers used. The expression level was corrected relative to the level of MAP4, the expression level of which remained unchanged across all microarray experiments. Relative quantification was calculated using RelQuant software (Roche).

Gα₁₂ signaling activated in NPC tumorigenesis. Primary epithelial populations isolated from nasopharyngeal biopsies were used for microarray analysis. We compared the differential gene expression profiles in 9 primary NPC cell strains and 5 NPC-derived cell lines against a pool of 32 normal primary normal nasopharyngeal epithelium cell strains. By selecting significantly (P ≤ 0.05) enriched gene sets, we identified the gene ontology terms and classified them by k-means clustering to identify the cellular processes commonly altered during tumorigenesis or specifically altered in advanced NPC cells (data not shown). A large number of genes involved in Gα₁₂ signaling were up-regulated in NPC (Fig. 1). These genes included Gα₁₂, ARHGEF12, RhoA, SLC9A1, ROCK1, PFN1, and JNK. This up-regulation suggests that Gα₁₂ signaling might be activated in the tumorigenesis of NPC.

Gα₁₂ depletion inhibited cell migration and invasion and reversed fibroblastic changes in NPC cells. To test whether the increased Gα₁₂ was critical to increased NPC aggressiveness, we used siRNA to knock down endogenous Gα₁₂ and assessed the effects on cell mobility. Compared with the siRNA controls, Gα₁₂ mRNA level in both CNE1 and NPC-TW06 cells were reduced by >80% by Gα₁₂-siRNA at 24 h post-transfection (Fig. 2A). Wound-healing assay showed that rate of wound healing was markedly less in Gα₁₂-depleted NPC cells than in siRNA controls. At 24 h post-scratching, the Gα₁₂-depleted cells had a wound gap that was ∼89% that of open gap measured at 0 h. Mock control cells and control siRNA cells had a gap of 33% and 39%, respectively (Fig. 2B). Matrigel invasion assay was done to evaluate the invasion ability of the NPC cells. The Gα₁₂-siRNA-transfected NPC cells were less invasive than siRNA controls at 30 h after plating (Fig. 2C,, left). The difference between their invasiveness was significant (P < 0.0001; Fig. 2C , right). These results suggest that Gα₁₂ may function to regulate the invasive ability of NPC cells.

A spindle-like fibroblastic morphology generally correlates with increased cell mobility in many types of epithelial cancers. To find out whether depleting Gα₁₂ in CNE1 and NPC-TW06 cells could reverse fibroblastic changes in epithelial cell morphology, we examined their morphologies and actin cytoskeletons. At 48 h post-transfection of siRNA, the morphology of Gα₁₂-depleted cells, which were flattened and multipolar and had a larger cell-cell contact area, was different from that of either the mock cells or the cells transfected with a control siRNA, which were small and round bipolar with a spindle-like appearance (Supplementary Fig. S1A). Moreover, the spike-like F-actin protrusion at the bipolar ends of mock control cells or cells transfected with control siRNA became less common or even disappeared in Gα₁₂-depleted cells (Supplementary Fig. S1B, compare inserts in middle and right). Together, these results suggest that Gα₁₂ plays a role in the invasive/metastatic potential of NPC cells.

Reduction of Gα₁₂ down-regulated key actin cytoskeleton regulators. We performed microarray analysis to identify transcripts regulated by the depletion of Gα₁₂ in NPC-TW06 cells. Most differentially regulated genes involved in actin polymerization and stabilization, focal adhesion assembly, and actomyosin-mediated contractility were down-regulated in the Gα₁₂-depleted cells (Fig. 3A).

To further understand how Gα₁₂ modulated actin reorganization, we extracted 95 differentially expressed genes involved in actin cytoskeleton signaling (Supplementary Table S3) from mock control cells and then compared them with those transcripts in Gα₁₂-depleted cells. Using k-means clustering algorithms, we identified genes that have distinct expression patterns (Fig. 3B). Many of the actin-related differentially expressed genes were found to have opposite expression patterns in cells treated with Gα₁₂-siRNA compared with mock cells (Fig. 3B , bottom). Twenty-nine genes were suppressed and only 7 genes were up-regulated by the reduction of Gα₁₂. To validate our microarray results, we performed QRT-PCR on 10 selected genes, Gα₁₂, IQGAP1, IRS1, ARHGEF12, ARPC5, v-MYC, WASL, FYN, JAK1, and MAP4, in Gα₁₂-depleted NPC cells and in respective control cells. QRT-PCR results corresponded well to those of microarray analysis (data no shown).

IQGAP1 expression was regulated by Gα₁₂ and depletion of IQGAP1 impaired migration of NPC cells. Because increases in the motility and invasiveness of cancer cells are indicative of EMT and are closely dependent on active remodeling of actin cytoskeleton, we identified 7 EMT-related genes (including IQGAP1, ARHGEF12, JAK1, IRS1, ARPC5, v-MYC, and WASL) from the above-mentioned actin-related differentially expressed genes. The QRT-PCR analysis confirmed that the Gα₁₂-siRNA drastically down-regulated their expressions in NPC-TW06 cells. A known Gα₁₂-regulated gene, RhoA, was used as an informative positive control for the assay (Fig. 4A). In contrast, overexpression of Gα₁₂-WT or Gα₁₂Q231L in NPC-TW06 cells increased the expression of IQGAP1 (Fig. 4B). Furthermore, the immunostaining of IQGAP1 in cells overexpressing Gα₁₂-WT or Gα₁₂Q231L also showed stronger signal intensities in the cell membrane and bipolar regions than those found in control cells (data not shown).

Because depletion of Gα₁₂ resulted in a reduction in mesenchymal-like cell formation and IQGAP1 has been previously implicated in the induction of EMT (22), we wanted to investigate whether depletion of Gα₁₂ in NPC cells would influence the expression of EMT markers. Result of Western blot analysis showed that the EMT markers, including LAMB3, vimentin, and paxillin, were down-regulated by depletion of Gα₁₂ and up-regulated by overexpression of a Gα₁₂-WT or Gα₁₂Q231L in NPC-TW06 cells. However, the epithelial markers, including E-cadherin and vinculin, were not noticeably affected by Gα₁₂-siRNA (Fig. 4C), suggesting that the reduction of mesenchymal-like cell formation does not necessarily up-regulate the expression of epithelial cell markers. To test whether IQGAP1 also contributed to maintenance of EMT in NPC cells, we depleted IQGAP1 by siRNA and found a similar effect to that seen by Gα₁₂-siRNA. The protein levels of LAMB3, vimentin, and paxillin were down-regulated by IQGAP1-siRNA and up-regulated by overexpression of IQGAP1 (Fig. 4C). In addition, the epithelial cell markers, the protein levels of which were found to be unaffected by the Gα₁₂ expression level, were markedly up-regulated by IQGAP1-siRNA (Fig. 4C , right). Together, these results suggest that Gα₁₂ and IQGAP1 may act together to regulate EMT in NPC cells.

To further understand the role of IQGAP1 in the invasive potential of NPC cells, we depleted endogenous IQGAP1 in NPC-TW06 cells. As was observed in Gα₁₂-depleted cells (Fig. 2B), the wound-healing assay clearly showed that the migration ability was markedly decreased by the IQGAP1-siRNA at 24 h post-wounding compared with control cells (Fig. 4D). Collectively, these biochemistry, morphology, and functional assays indicate that IQGAP1 plays a critical role in invasive potential of NPC cells.

IQGAP1 is a key downstream effector of Gα₁₂ for the conversion from epithelial to fibroblastoid morphology in cultured NPC cells. Because both IQGAP1 and vinculin are required for the assembly of adherens junctions, which are important for cell migration (29), we wanted to analyze whether suppression of Gα₁₂ would alter the expressions or subcellular localizations of IQGAP1 and vinculin. The immunocytochemical results showed that Gα₁₂-depleted cells had less IQGAP1 and vinculin immunofluorescence than the control cells. The depletion of IQGAP1 also reduced vinculin immunofluorescence (Fig. 5A,, bottom right). Particularly, we found marked reductions in vinculin in the cell-matrix junctions and IQGAP1 in the cell-cell contacts in cells treated with the siRNA for Gα₁₂ or IQGAP1 (Fig. 5A). In addition, the immunofluorescent signals of vimentin and paxillin (two known mesenchymal markers) in Gα₁₂- or IQGAP1-siRNA-treated cells were found dramatically decreased (Supplementary Fig. S2A). Furthermore, using phase-contrast microscopy, we observed cell shape changes in IQGAP1-depleted cells. Their rounded, bipolar, spindle-shaped, fibroblastoid appearance changed to a more epithelial-like, large, flat, spread-out appearance, findings that were similar to those we observed when knocking down Gα₁₂ with siRNA (Fig. 5B , top). These results suggest that the cell-cell contacts, cell-matrix junctions, and mesenchymal morphology are coregulated by Gα₁₂ and IQGAP1 in NPC cells.

Because depletion of Gα₁₂ decreased IQGAP1 expression and depletion of IQGAP1 resulted in phenotypes similar to that found in the Gα₁₂-depleted cells, we speculated that IQGAP1 might function as a downstream effector of Gα₁₂ signaling. To test this hypothesis, we depleted Gα₁₂ but simultaneously overexpressed IQGAP1-WT or COOH-terminal IQGAP1 (IQDN4; ref. 28) to determine whether this would reverse the morphologic changes induced by depletion of Gα₁₂. Indeed, overexpression of either IQGAP1 proteins reversed, at least in part, Gα₁₂-siRNA-induced morphologic changes (Fig. 5B , bottom) and expression level of mesenchymal markers (Supplementary Fig. S2B). Together, these results support a model in which IQGAP1 functions downstream of Gα₁₂ in NPC cells.

Immunohistochemical assays of Gα₁₂ expression in NPC tumorigenesis. To validate the role of Gα₁₂ in the pathogenesis of NPC, we conducted immunohistochemical analysis of Gα₁₂ on 50 nasopharyngeal biopsies. Gα₁₂ immunoreactivity was predominantly negative in the normal nasopharyngeal epithelium (12 of 13), 1-positive to 2-positive in dysplasic lesions of varying severity (mild, moderate, and severe), and strongly positive in NPC (Fig. 6A). A significant association was observed between expression of Gα₁₂ and NPC (P < 0.01). Of the 31 NPC biopsies, 58% (18 of 31) were 3-positive, 35.5% (11 of 31) were 2-positive, and 6.5% (2 of 31) were 1-positive for immunoreactivity in tumor masses (Fig. 6B). Further, expression of Gα₁₂ was much higher in NPC than in adjacent basal layer epithelium (Supplementary Fig. S3). To further investigate if Gα₁₂ contributed to the invasive ability of NPC, we used quantitative QRT-PCR assay to compare the transcript levels of Gα₁₂ in NPC biopsy samples taken from patients who had neck lymph node metastasis with those who did not. We found a significant correlation between lymph node metastasis and expression of Gα₁₂ (P < 0.05; Fig. 6C). Using microarray, biochemical, cell biology, and histopathologic approaches, we have identified and characterized the Gα₁₂ signaling in tumorigenesis and metastasis of NPC.

Although it has been suggested that some EBV-encoded genes, such as latent membrane protein 1 and 2, are important for cell motility and invasiveness of NPC cells (30, 31), the molecular mechanisms underlying the EMT and metastasis of NPC remain largely unknown. In this study, we have uncovered a Gα₁₂-mediated pathway that plays an essential role in the tumorigenesis and progression of NPC. Specifically, we found that siRNA-induced depletion of Gα₁₂ resulted in decreased migration, invasiveness, and reappearance of the enlarged, flat morphology of epithelial cells. Global transcriptional analysis revealed the loss of Gα₁₂ function to be accompanied with a reorganization of the actin cytoskeleton. For example, key actin dynamic regulators involved in Cdc42-dependent and actomyosin-related cell contraction signaling pathways were repressed, supporting the hypothesis that Gα₁₂ signaling may regulate motility and invasiveness through the modulation of actin cytoskeleton reorganization in NPC cells.

This pathway was found to possibly play a role in the regulation of EMT in NPC cells. In tumorigenesis, EMT has been specifically linked to tumor invasiveness and metastasis, making treatment difficult (32). To transform into motile fibroblastoid cells, epithelial cells need to go through changes in cell-cell and cell-matrix adhesion contacts, extracellular matrix synthesis, and actin cytoskeleton reorganization. Because Gα₁₂ was shown by this study to be a key determinant of aggressive behavior and fibroblastoid morphology in NPC cells, we investigated whether knocking down Gα₁₂ would reduce the expression of EMT-related genes. QRT-PCR analysis revealed that IQGAP1 and several other EMT-related genes were down-regulated by the depletion of Gα₁₂ in NPC cells. Because previous genome-wide expression analyses of different malignancies, including breast, lung, gastric, and colon cancers, have implicated IQGAP1 dysfunction in tumorigenesis and tumor progression (33–37), we wanted to determine whether depletion of IQGAP1 would contribute to inhibition of cell motility. Similar to what we observed in NPC cells lacking Gα₁₂, depletion of IQGAP1 disrupted adherens junctions and transformed the spindle, fibroblastoid-like cells into flattened and enlarged multipolar ones. We also found a decrease in migration in IQGAP1-depleted cells. Interestingly, although several studies reported different signaling partnerships with Gα₁₂ to be associated with cell-cell adhesion and cell migration (14), none have mentioned the role that IQGAP1 may play in this process. Our epistasis analysis suggested that IQGAP1 may be a potential downstream effector of Gα₁₂ signaling in NPC cells. Having established the link between Gα₁₂ and IQGAP1, we set out to investigate whether Gα₁₂ played an important role in EMT. Both depleted and overexpressed Gα₁₂ clearly altered protein levels or subcellular localization of EMT-related markers, suggesting that the Gα₁₂ signaling is essential for EMT. Our immunochemical studies of 50 nasopharyngeal biopsies confirmed that Gα₁₂ signaling played an important role in tumorigenesis and metastasis of NPC.

IQGAP1 is known to bind several specific cell-cell adhesion-related proteins, including calmodulin, actin, E-cadherin, and Cdc42, to modulate cell-cell adhesion, cell polarity, and cell migration (21). Our data showed that the depletion of either Gα₁₂ or IQGAP1 disrupted cell-cell or cell-matrix junctions in NPC cells, suggesting both proteins might be involved in the same signaling cascade. Future studies may want to focus on how IQGAP1 participates in the Gα₁₂-mediated tumorigenesis of NPC. The role of IQGAP1 in canonical Gα₁₂ signaling in vivo also needs further study. To do this, it will be necessary to measure expression level of IQGAP1 and its correlation with Gα₁₂ during the tumorigenesis and metastasis of NPC in clinical samples. Moreover, because the cross-talk between GPCR and EGFR signaling has been shown to be important for the growth and invasion of head and neck squamous cell carcinoma (8), it will be of interest to investigate whether IQGAP1 is also modulated by another GPCR or receptor of other families in NPC.

A further issue related to the association between Gα₁₂ and NPC pathogenesis is how Gα₁₂ expression is triggered. The role of EBV in NPC pathogenesis may provide clues regarding the connection between the induction of Gα₁₂ signaling and NPC pathogenesis. Nearly all undifferentiated NPC cases harboring EBV are severely infiltrated with lymphocytes. The activated immune cells and adjacent tumor cells can express lysophosphatidic acid to further activate the Gα₁₂-RhoA signaling pathway, contributing to the invasive behavior of cancer cells (38). Moreover, the EBV genome encodes a v-cytokine (vIL-10) and a vGPCR (ORF BILF1; refs. 5, 39), both thought to promote lymphocyte infiltration and chemokine secretion (40). Therefore, EBV might trigger some extracellular stimuli, such as lysophosphatidic acid, to potentiate Gα₁₂ signaling. There is some basis for this hypothesis found in a recent study showing that EBV infection of Hodgkin's lymphoma cells induces autotaxin causing lysophosphatidic acid to be generated (41). On the other hand, although the involvement of Gα₁₂/IQGAP1 in other non-EBV-induced head and neck tumors remains unexplored, it may be involved in other head and neck cancers because the E-cadherin/β-catenin pathway, which is regulated by Gα₁₂ signaling, has been shown to be deregulated in head and neck squamous cell carcinoma (42).

This study is the first to report an association between Gα₁₂-mediated pathway and invasiveness of NPC cells. This association offers new clues for further understanding of the molecular mechanisms and may serve as a basis for the development of targeted molecular cancer therapies.

No potential conflicts of interest were disclosed.

Grant support: National Health Research Institutes.

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.

We thank National Health Research Institutes Optical Biology Core for microscopy, Dr. Stephen P. Goff for IQGAP1 expression plasmids, Drs. Lu-Hai Wang and Jen-Yang Chen for comments on the article, James Steed and C.Y. Foundation for English editing assistance, and Hsin Yi Hsieh for assistance with the Western blotting.