Critical Role of Peroxisome Proliferator-Activated Receptor γ on Anoikis and Invasion of Squamous Cell Carcinoma

Human oral squamous cell carcinoma is the most common neoplasm in oral cavity cancer and the incidence of lingual carcinoma has recently been increasing. The optimal treatment or therapy for early carcinoma of the tongue remains a controversial issue and surgical operation is still an effective therapy for squamous cell carcinoma in the field of maxillofacial surgery. However, the recovery of lost function after a large surgical excision is accompanied by many problems such as speech impediment and dysphagia (1). Local-regional control of head and neck cancer has improved in recent decades. Nevertheless, overall survival remains largely unchanged (2, 3). The major reason for this discrepancy is distant metastasis and second neoplasms (4). The incidence of clinically detected distant metastasis of the head and neck ranges from 11% to 23% and most of the metastasis are observed in the lymph nodes and lungs (5).

Peroxisome proliferator-activated receptor γ (PPARγ) that belongs to the nuclear hormone receptor family is mainly expressed in adipose tissue and plays a central role in adipocyte differentiation and insulin sensitivity (6). Recent studies have shown that, in addition to its classic role, PPARγ is implicated as a putative therapeutic target for cancer in a variety of tumors because several observations suggest that stimulation of PPARγ function may inhibit carcinogenesis and tumor cell growth (7, 8). However, the exact role of PPARγ on carcinogenesis and tumor cell growth is still controversial because of the many conflicting reports that variably provide evidence for a tumor suppressor or promoter role (9–11). Recent investigations by our colleagues have shown that PPARγ is overexpressed in hepatocellular carcinoma and that PPARγ inhibitors interfered with adhesion and caused anoikis in two hepatocellular carcinoma cell lines (12). Our investigations presented here also strongly suggest that PPARγ inhibitors might be useful in treating squamous cell carcinoma.

In this study, we examined whether PPARγ is expressed in primary oral squamous cell carcinoma and lymph node metastasis, and whether PPARγ is a potential target for tumor therapy. We therefore studied the effects of PPARγ agonists and antagonists on the viability of oral squamous cell carcinoma cell lines. The specific PPARγ antagonists T0070907, GW9662, and BADGE were able to induce apoptosis in squamous cell carcinoma cell lines by interfering with adhesion to the extracellular matrix and disrupting survival signals, and thus inducing anoikis. Furthermore, the PPARγ antagonists strongly inhibited the invasion of squamous cell carcinomas. These results imply a potentially important and novel role for the inhibition of PPARγ function via the use of specific antagonists in the treatment of oral squamous cell carcinoma and the prevention of tumor invasion and metastasis.

Chemicals. PPARγ agonists, pioglitazone and rosiglitazone, were provided by Takeda Pharmaceutical Co. (Osaka, Japan) and Glaxo SmithKline (Tokyo Japan), respectively. The PPARγ-specific antagonists T0070907, GW9662, and BADGE were purchased from Cayman Chemical (Ann Arbor, MI), Sigma (St Louis, MO), and Tokyo Kasei (Tokyo, Japan), respectively. Other drugs were reagent grade.

Tissue samples. Twenty samples of squamous cell carcinoma located in the tongue were obtained from surgical resection tissue specimens at Osaka University Hospital after informed consent was obtained. The patients, who received no preoperative therapy including chemotherapy and irradiation therapy, were randomly selected (Table 1). Thirteen patients were males and seven were females, and the age range was 40 to 76 years (58.7 ± 11.7 years). Metastatic lymph node samples were obtained from six of them. Normal tongue tissue was obtained from a patient who had enforced operation of tongue resection.

Immunohistochemical staining of peroxisome proliferator-activated receptor γ. Five-micron sections of paraffin-embedded tissues were mounted on glass slides and the expression of PPARγ within tissues was detected using a PPARγ-specific polyclonal antibody (Santa Cruz Biotechnology, Santa Cruz, CA) using standard immunohistochemical techniques on paraffin-embedded sections. The Vectastain ABC kit (Vector Laboratories, Burlingame, CA) was used with a 3,3′-diaminobenzidine substrate kit (Vector Laboratories) according to the instructions of the manufacturer.

Cell culture and treatment with peroxisome proliferator-activated receptor γ agonists and antagonists. We used four human oral squamous cell carcinoma cell lines (SCCTF, SCCKN, SAS, and CA9-22). SCCTF, SCCKN, and SAS are established tongue squamous cell carcinoma cell lines (13), and CA9-22 is a gingival squamous cell carcinoma cell line (14, 15). SCCTF, SCCKN, and SAS were maintained in DMEM containing 10% fetal bovine serum (FBS); CA-9-22 was maintained in DMEM containing 0.6% glutamine and 10% FBS at 37°C under 0.5% CO₂. For cell growth and adhesion experiments, cells were trypsinized and the PPARγ agonists (pioglitazone and rosiglitazone) and antagonists (BADGE, GW9662, and T0070907) were added at the time of replating.

Cell survival assay. Squamous cell carcinoma cells were treated with PPARγ agonists and antagonists dissolved in DMSO for 24 and 48 hours in culture medium. 3-[4,5-Dimethylthiazol-2-yl]-2, 5-diphenyltetrazolium bromide (MTT) solution (Sigma) was then added to the well with gentle pipetting. After 4 hours, 20% SDS was added to the solution, and the cells were incubated for 5 hours at room temperature. After treatment, the absorbance was determined at a wavelength of 595 nm. The inhibition of cell growth was compared with vehicle (DMSO) control and the percentage of differences was determined.

Apoptosis analysis. Apoptosis was analyzed by two different methods. First, chromatin staining with Hoechst 33342 was done. Briefly, the SCCKN treated with antagonists were cultured on Laboratory-Tek Chamber Slides for 24 hours. The cells were fixed with 4% paraformaldehyde in PBS for 10 minutes and then rinsed. Chromatin staining was done with Hoechst 33342 (Sigma) to detect nuclear condensation.

We also did annexin V and propidium iodide exclusion double staining (Roche Annexin V-FLUOS Staining Kit, Roche Applied Science, Penzberg, Germany). SCCKN were treated either with the vehicle alone (DMSO) or with 30 or 60 μmol/L of T0070907 dissolved in DMSO for 48 hours. Then, cells were washed with PBS and resuspended in an incubation buffer. Annexin V-FITC and propidium iodide were added at a concentration of 1 × 10⁶ cells/ mL. Control cells stained with annexin V or propidium iodide alone were used to compensate for flow cytometric analysis. Samples and controls were incubated for 15 minutes at room temperature in the dark and FACScan analysis was done by FACScanG3 equipped with Cell Quest software. Annexin V–negative, propidium iodide–negative cells were defined as live cells; annexin V–positive, propidium iodide–negative cells were defined as early apoptotic cells; and annexin V and propidium iodide double-positive cells were defined as late apoptotic and necrotic cells.

Western blot analysis. Adherent or suspended cells were washed in PBS, and cell extracts were prepared by lysing cells in lysis buffer. Protein concentrations were measured using Bio-Rad Protein Assay Reagent (Bio-Rad, Richmond, CA) following the suggested procedure of the manufacturer. Ten micrograms of protein were separated by 10% SDS-PAGE. After electrophoresis, the proteins were transferred to nitrocellulose membrane (Millipore, Bedford, MA), blocked for 3 hours in TBS with 5% skim milk at room temperature, and reacted with primary polyclonal antibody overnight. After three washings, the membranes were incubated for 1 hour at room temperature with secondary antibody, and immune complexes were visualized by using the enhanced chemiluminescence detection kit (Amersham, London, United Kingdom) following the suggested procedure of the manufacturer.

Adhesion assay. SCCKN was pretreated with DMSO and T0070907 (30 μmol/L) for 9 hours, then cells were plated into fibronectin-coated plates (Becton Dickinson, Bedford, MA) with DMEM without FBS. After 0.5 to 6 hours, adherent cells were counted from six fields that were randomly selected. The percentage of adherent cells was expressed as per dish in each group.

Caspase inhibition. For experiments with caspase inhibition, 200 μmol/L Z-VAD-FMK (BD Biosciences, Sun Jose, CA) was directly added to the cells when the cells were replated 90 minutes before the addition of various concentrations of T0070907. Control cells had matched concentrations of vehicle (DMSO) added in place of the inhibitor and/or Z-VAD-FMK. After 48 hours of incubation, inhibition of cell growth was measured by MTT assay and expressed as a percentage of inhibition.

RNA interference approach. The 19-nucleotide small interfering RNA (siRNA) for PPARγ was designed according to the method of Kelly et al. (16). The sequence of the sense strand of PPARγ siRNA was 5′-GCCCTTCACTACTGTTGAC-3′. For the transfection, PPARγ siRNA or control GL2 siRNA (5′-fluorescein–labeled luciferase GL2 siRNA duplex, Dharmacon, Inc., Tokyo, Japan) solution was added to DMEM medium containing Lipofectamine 2000 (Invitrogen, Inc., Tokyo, Japan) and allowed to incubate for 20 minutes at room temperature to create the transfection mixture. SCCKN was trypsinized and resuspended in DMEM without FBS, and the cells were separated approximately 1 × 10⁵ cells for each dish. The transfection mixture was then added to the cells with the final concentration of siRNA at 10, 20 and 40 nmol/L. Twenty-four hours after the start of transfection, the medium was changed to DMEM containing 10% FBS. The MTT assay was done 96 hours after the incubation.

Confocal microscopy. SCCKN and SCCTF cells were cultured on a Lab-Tek II chamber slide (Nalge Nunc International, Naperville, IL). Cells were treated with the PPARγ-specific antagonist T0070907 at a concentration of 30 and 60 μmol/L for 24 hours. Vehicle was treated with DMSO. The cells were then fixed with 4% paraformaldehyde in PBS for 15 minutes, permeabilized in PBS containing 0.1% Triton X-100 for 15 minutes, and then rinsed with PBS. The fixed cells were incubated with anti–F-actin antibody conjugated with Alexa Fluor 594 phalloidin (Molecular Probes, Eugene, OR) for 16 hours at 4°C and rinsed with PBS. Nuclear staining was done by incubation with Syto Green Fluorescent Nucleic Acid Stains (24 dye, Ex 490 nm, Em 515 nm; Molecular Probes). A confocal laser scanning microscope (LSM410, Carl Zeiss, Tokyo, Japan) was used to visualize the structure of F-actin.

Invasion assay. To quantify the invasion, the membrane invasion assay was carried out in Matrigel-coated invasion chambers (Becton Dickinson Labware). SCCKN was detached by trypsin-EDTA and resuspended in DMEM with 1% FBS containing PPARγ antagonist (T0070907 60 μmol/L) and plated in the upper chamber. DMEM with 10% FBS and antagonist was also added in the lower chamber. Following 48 hours of incubation, the cells in the upper chamber and on the Matrigel were mechanically removed with a cotton swab. The cells adherent to the outer surface of the lower side of the membrane were fixed with methanol and stained with hematoxylin. The invaded cells were examined, counted, and photographed by microscopy. Six fields were randomly selected and counted per filter in each group.

Flow cytometric analysis for CD151 expression. SCCKN were treated either with vehicle or with T0070907 (60 μmol/L) for 48 hours. Then, cells were incubated with anti-CD151 antibody (PharMingen, San Diego, CA). The cells were stained with Alexa-conjugated goat anti-mouse immunoglobulin G (Molecular Probes) and the fluorescence intensity was analyzed using a flow cytometer.

Oral squamous cell carcinoma cell lines and tumor cells in tissues express peroxisome proliferator-activated receptor γ protein. Squamous cell carcinoma primary tissues and lymph node metastases were stained using an anti-PPARγ-specific antibody, and a similar pattern of PPARγ expression was observed in all of the specimens (Table 1). Tumor cells in primary tissue were positively stained for PPARγ (Fig. 1A, left and middle). On the other hand, PPARγ staining was not observed in normal tongue tissue (data not shown). Interestingly, tumor cells in the metastatic lymph node were also stained for PPARγ (Fig. 1A, right). Western blot analysis showed the expression of PPARγ in all squamous cell carcinoma cell lines (Fig. 1B), and high expression levels of PPARγ protein were observed in SCCKN, SCCTF, and SAS. The expression of PPARγ was observed both in human squamous cell carcinoma tissues and squamous cell carcinoma cell lines, but not in normal human lingual tissue, indicating the principal role of PPARγ in squamous cell carcinomas.

Peroxisome proliferator-activated receptor γ antagonists inhibit cell growth of squamous cell carcinomas. According to the data of PPARγ expression in squamous cell carcinomas, we hypothesized that inhibition of PPARγ function might affect cell growth. To investigate that hypothesis, we initially used SCCKN and SCCTF cell lines for the cell growth experiment because these cell lines express high level of PPARγ proteins compared with SAS and CA9-22. Various PPARγ antagonists, such as T0070907, GW9662 and BADGE, inhibited cell growth of both SCCTF and SCCKN in a concentration-dependent manner (Fig. 1C). In contrast, the cell growth rate of squamous cell carcinomas was not altered when cells were treated with the PPARγ agonist pioglitazone (Fig. 1C, ▪). The same results were also observed when cells were treated with another PPARγ specific agonist, rosiglitazone (data not shown).

There are many in vitro reports indicating that PPARγ agonists induce inhibition of tumor cell growth and apoptosis (17–19). Therefore, we examined the effect of high concentration of pioglitazone (100 μmol/L) and observed that the percentage of inhibition of cell growth was only 13.7%. In contrast, 30 μmol/L of PPARγ antagonists showed squamous cell carcinoma growth inhibitions of 60.7% (BADGE), 45.5% (T0070907), and 30.8% (GW9662), respectively. A significant inhibition of cell growth by antagonists was observed at the concentrations 10 to 30 μmol/L. The inhibition of cell growth by antagonists seemed to be time dependent (data not shown).

Peroxisome proliferator-activated receptor γ antagonists cause apoptosis of squamous cell carcinomas. Because the MTT assay indicated that treatment with PPARγ antagonists inhibited the growth of squamous cell carcinomas, we investigated whether PPARγ antagonists induce apoptosis of squamous cell carcinomas. The SCCKN treated with T0070907 showed nuclear condensation by chromatin staining with Hoechst 33342, which was considered to be a specific morphologic change associated with apoptosis (Fig. 1D). The same results were also observed when squamous cell carcinomas were treated with other antagonists such as BADGE or GW9662 (data not shown). The result of annexin V-FITC staining also showed that the PPARγ-specific antagonists dramatically induced apoptosis (Fig. 1E). These results strongly indicate that the inhibition of squamous cell carcinoma growth by PPARγ antagonists is due to apoptosis.

Peroxisome proliferator-activated receptor γ antagonists inhibit cell adhesion. Although treatment of squamous cell carcinoma cells with PPARγ-specific antagonists leads to apoptosis, visible differences were not observed until 9 hours after treatment. However, cells started floating 9 to 12 hours after the start of the treatment, and most cells were floating and dead after 24 hours. After 48 hours of treatment, the number of adherent living cells significantly decreased; in contrast, the number of floating dead cells increased (Fig. 2A and B). Consequently, the total number of cells decreased. These results suggest that PPARγ antagonists caused the inhibition of cell adhesion and resulted in the inhibition of cell growth (i.e., anoikis, apoptosis resulting from loss of cell-matrix interactions; ref. 20).

We next investigate the effect of PPARγ antagonists on cell-extracellular matrix adhesion. The result of adhesion assay is shown in Fig. 2C; as can be seen, control cells adhered strongly to fibronectin-coated plates, and this adhesion was significantly inhibited by pretreatment with PPARγ antagonists. These results indicate that PPARγ antagonists clearly inhibit cell-extracellular matrix adhesion, indicating anoikis.

To investigate whether the inhibitory effect of PPARγ antagonists on the ability of cell adhesion is PPARγ specific or not, we applied a PPARγ-specific agonist, pioglitazone, to improve the decreased adhesion induced by PPARγ antagonists. As shown in Fig. 2D and E, PPARγ agonist dramatically improved the decreased adhesion induced by PPARγ antagonists. These results clearly indicate that the inhibition of cell adhesion by PPARγ antagonists is mediated by blockade of PPARγ-specific pathways.

Effect of knockdown of peroxisome proliferator-activated receptor γ by small interfering RNA. To investigate whether the inhibitory effect of PPARγ antagonists on cell adhesion and growth of squamous cell carcinomas is PPARγ pathway specific or not, we used the siRNA approach for PPARγ. PPARγ-specific siRNA (PPARγ siRNA) effectively decreased PPARγ protein level in squamous cell carcinomas (data not shown) and showed concentration-dependent inhibition of cell growth and adhesion of SCCKN (Fig. 3A and B). These results clearly indicate that the inhibition of cell adhesion and growth by PPARγ antagonists is mediated by blockade of PPARγ pathway.

Investigation of caspase dependency of anoikis. We also confirmed whether the lack of adhesion is the result or cause of apoptosis. The general caspase inhibitor, Z-VAD-FMK, did not affect the inhibition of attachment induced by PPARγ antagonists; however, it clearly attenuated the inhibition of cell growth by PPARγ antagonists (Fig. 3C). These data strongly provide the evidence that detachment is a cause, rather than a result, of apoptosis induced by PPARγ antagonists.

Peroxisome proliferator-activated receptor γ antagonists destroy cell skeletal structure. The results of anoikis lead us to the hypothesis that adhesion/signaling complex formation in squamous cell carcinomas might be inhibited by the blockade of PPARγ pathway. We therefore investigated the effect of PPARγ antagonists on cytoskeletal and signaling molecules. For visualization of the cell skeleton, we did fluorescent detection of F-actin by confocal laser microscopy. In the vehicle control group, the actin filaments were clearly observed and the cell shapes remained normal (Fig. 4A). On the other hand, in cells treated with antagonists (T0070907, 30 μmol/L), actin filaments were irregularly disrupted. At a high concentration of antagonist (60 μmol/L), the shapes of the cells were altered to round shapes. These results suggest that PPARγ antagonists cause the alteration of the cell skeletal structures and consequently inhibit the ability of cells to attach to extracellular matrix.

Peroxisome proliferator-activated receptor γ antagonists inhibit the phosphorylation of focal adhesion kinase. We next investigated the mechanisms of anoikis induced by PPARγ antagonists. FAK, a 125 kDa nonreceptor tyrosine kinase, is an important regulator of cell survival, invasion, migration, and cell cycle progression (21, 22). The results of the Western blotting analysis showed decreased expression of phosphorylated focal adhesion kinase (p-FAK) at 12 to 24 hours after treatment with antagonists (Fig. 4B). Consequently, the decreased expression of phosphorylated mitogen-activated protein (MAP)/Erk kinase (p-MEK) and Erk (p-Erk) was also observed 12 and 24 hours after treatment with antagonists, indicating the involvement of the MAP kinase (MAPK) pathway.

We then investigated the effects of PPARγ antagonists on already firmly adherent squamous cell carcinomas, whether PPARγ antagonists were able to induce the squamous cell carcinoma detachment. PPARγ antagonists were administered to squamous cell carcinomas when cells reached 50% confluency. The detachment of squamous cell carcinomas was observed when cells were treated with PPARγ antagonists (data not shown). The decrease in p-FAK and consequent decreases in p-MEK and p-Erk were also observed 12 and 24 hours after treatment (data not shown).

Inhibition of integrin α5 expression by peroxisome proliferator-activated receptor γ antagonists. PPARγ antagonists clearly inhibited cell-extracellular matrix adhesion, resulting in the inhibition of FAK signaling pathways. We therefore investigated the potential mechanism(s) involved in the inhibition of cell-extracellular matrix adhesion. We did Gene Chip analysis to detect altered genes and observed down-regulation of several adhesion molecule expressions on squamous cell carcinomas treated with PPARγ antagonists. Among them, we took notice of integrin α5 because integrins α5 and β1 are well-known major counter adhesion molecules for fibronectin (23, 24). As shown in Fig. 4C, the expression of integrin α5 was time-dependently increased with respect to cell adhesion. In contrast, the increase in the expression of integrin α5 was dramatically inhibited by treatment with PPARγ antagonist. The results were also confirmed by flow cytometric analysis (Fig. 4D). No marked difference in the expression of integrin β1 was observed. These results indicate that PPARγ antagonists inhibit the expression of integrin α, resulting in the inhibition of cell-extracellular matrix adhesion.

Peroxisome proliferator-activated receptor γ antagonist inhibits invasion of squamous cell carcinomas via down-regulation of CD151. Invasion and metastasis are closely related to cell adhesion and these are critical events in oral squamous cell carcinomas. We therefore investigated the effect of PPARγ antagonists on the invasion of squamous cell carcinomas. Treatment with the PPARγ antagonist T0070907 dramatically inhibited the invasion of squamous cell carcinomas (Fig. 5A and B). The same results were also observed on squamous cell carcinomas treated with other PPARγ antagonists, GW9662 and BADGE (data not shown). These results strongly indicate the effectiveness of PPARγ antagonists on the prevention of squamous cell carcinoma invasion.

To investigate the potential mechanisms, we did Gene Chip analysis to detect altered genes and observed marked down-regulation of CD151, cdc42, and Rho family expressions in squamous cell carcinomas treated with PPARγ antagonists (data not shown). CD151 and cdc42 are well-known important players in the invasion and metastasis of tumors (25). We therefore confirmed the effect of PPARγ antagonist on the down-regulation of CD151 expression by flow cytometric analysis. PPARγ antagonists strongly down-regulated the expression of CD151 (Fig. 5C), indicating one potential mechanism in the inhibition of invasion.

Although there are many reports indicating the relationship between PPARγ and carcinogenesis, the biological role of PPARγ on carcinogenesis and tumor cell growth is still unclear due to the contradictory results that have been reported (9–11). In the present study, we clearly showed that PPARγ antagonists, but not agonists, inhibited cell growth of cultured oral squamous cell carcinomas established from human tongue carcinomas. In contrast, several reports have indicated that PPARγ is expressed in tumors and that PPARγ agonists induce the inhibition of cell growth and apoptosis of adenocarcinomas and squamous cell carcinomas (17–19, 26–28). What is the reason for the discrepancies between our data and the data found in other reports? The possible reasons may be attributed to the fact that comparatively high concentrations of agonists (for instance 30-100 μmol/L) were used in the other reports, and the fact that in most cases, troglitazone, a low specificity ligand, was used. Furthermore, the previous investigators did not study the effect of PPARγ antagonists on tumor cell growth, as they only used PPARγ agonists. In our experimental conditions, PPARγ antagonists, but not agonists, strongly inhibited cell growth of squamous cell carcinomas. In addition, we used the siRNA approach to confirm the inhibitory effect of PPARγ antagonists on cell adhesion, and growth was mediated by the blockade of PPARγ pathway. PPARγ siRNA clearly inhibited the cell adhesion and growth of SCCKN. These results clearly indicate that the inhibition of cell adhesion and growth by PPARγ antagonists is mediated by the blockade of PPARγ pathway, and that only the decrease in PPARγ protein itself induces the inhibition of cell growth. Similar effects were also observed in hepatocellular carcinoma and HT-29 cell lines by our colleagues (12), which also led us to consider these effects in squamous cell carcinomas treated with PPARγ antagonists on a broader applicable phenomenon in epithelial cancers.

The inhibition of cell growth of squamous cell carcinomas by PPARγ antagonists is due to apoptosis, as evidenced by the data for nuclear condensation by chromatin staining with Hoechst 33342 and annexin V-FITC staining through fluorescence-activated cell sorting analysis. Mechanisms may exist by which apoptosis is induced by PPARγ antagonists. Furthermore, the inhibition of cell growth by antagonists was time dependent; visible differences were not observed until 9 hours after the treatment. However, cells started floating 9 to 12 hours after the start of the treatment, and most cells were floating and dead after 24 hours. Furthermore, caspase inhibitors did not affect the inhibition of attachment induced by PPARγ antagonists; however, it clearly attenuated the inhibition of cell growth by PPARγ antagonists. These data strongly support the notion that detachment is a cause, rather than a result, of apoptosis induced by PPARγ antagonists. These results suggest that PPARγ antagonists cause the inhibition of cell adhesion resulting in the inhibition of cell growth (i.e., anoikis).

Anoikis was first documented in normal epithelial cells and endothelial cells, and it helps maintain a dynamic balance between cell turnover and survival (20). Malignant tumor cells are often resistant to anoikis, enabling them to survive after detachment from the extracellular matrix and colonize a secondary site. In our present study, the inhibition of FAK phosphorylation in squamous cell carcinomas treated with PPARγ antagonists was observed. FAK, a 125 kDa nonreceptor tyrosine kinase, is an important regulator of cell survival, invasion, migration, and cell cycle progression (21, 22). The overexpression of FAK was observed in a number of human malignant cells, with the degree of overexpression correlating with greater aggressiveness (29). FAK is functionally important in transducing intracellular messages associated with growth factor signaling and cell-extracellular matrix interactions that are tightly related to the formation of cell adhesion complex (30). The intracellular messages link p-FAK at Tyr925 to signaling pathways that modify the cytoskeleton and activate MAPK cascades. In our present study, inhibition of MEK and Erk phosphorylation after the inhibition of FAK-phosphorylation by PPARγ antagonists was observed. Although the detailed mechanisms are unclear on how PPARγ antagonists inhibit MAPK pathway, our results indicate that the mechanisms of the inhibition of cell growth by PPARγ antagonists are in part due to the FAK-MAPK pathway. Other possible mechanisms are due to the direct inhibition of MAPK pathway by PPARγ antagonists because it has been reported that PPARγ agonists, such as ciglitazone and troglitazone, directly activate MAPK via a PPARγ-independent pathway (31). PPARγ antagonists may directly inhibit the MAPK pathway.

Anchorage-independent growth and the ability to avoid detachment-induced apoptosis (anoikis) are hallmarks of transformed epithelial cells. Malignant cells are able to resist anoikis to varying degrees, and this property has been proposed to contribute to tumorigenesis and metastasis (32). Our data from the present study clearly indicate that PPARγ antagonists induce apoptosis in squamous cell carcinoma cell lines via induction of anoikis involved in the inhibition of p-FAK. Although the detailed mechanisms are unknown on how PPARγ antagonists inhibit the FAK signaling pathways, we observed the inhibition of integrin α5 expression by treatment with PPARγ antagonists. Integrin α5 is well known as the counter receptor for fibronectin and the decrease in integrin α5 expression leads to the inhibition of cell-extracellular matrix adhesion (23, 24). In fact, in our experimental conditions, we confirmed the decrease in integrin α5 expression by PPARγ antagonist treatment, followed by inhibition of cell-extracellular matrix adhesion. Cell-extracellular matrix adhesion is required for the continuous activation of FAK signaling pathway. Therefore, the decrease in integrin α5 expression by PPARγ antagonists might be one of the potential mechanisms to inhibit the cell-extracellular matrix adhesion and resulting inhibition of FAK signaling pathways. Further investigations might be necessary to clarify the detailed mechanisms.

We also showed the strong inhibition of squamous cell carcinoma invasion that occurs during treatment with PPARγ antagonists. We observed the decrease in the expression of CD151, which is well known to be an important molecule for invasion and metastasis of tumors (25). It is also reported that CD151 associates with the integrin family and modulates cell adhesion (33, 34). Our data indicate the effectiveness of PPARγ antagonists in the prevention of squamous cell carcinoma invasion via the down-regulation of cell surface adhesion molecules. Together with those data, the treatment of patients with PPARγ antagonists might thus prevent the local invasion of squamous cell carcinomas and attachment of metastatic cells to establish new sites such as lymph nodes. Therefore, the inhibition of PPARγ by specific antagonists may be a novel and potent therapy for preventing not only the invasion of squamous cell carcinomas but also their metastasis into lymph nodes.

In summary, we have identified a novel and potentially important role for PPARγ antagonists in the treatment of human oral squamous cell carcinoma, where PPARγ is noticeably overexpressed. The effects on cell adhesion and subsequent survival in squamous cell carcinoma cell lines may indicate a role in preventing further tumor growth and survival away from the extracellular matrix. Further experiments are required to delineate the precise contributions and mechanisms of PPARγ inhibition in terms of tumor transformation, growth, survival, and metastasis.

We thank Drs. Lawrence J. Saubermann and Katherine L. Schaefer, Boston Medical Center, for their useful suggestions.