Co-immunoprecipitation analysis was carried out using a commercial kit [Immunoprecipitation kit (Protein G), Sigma-Aldrich, St. Louis, MO, USA]. Subconfluent SAS cells in a 10-cm culture dish were lysed with the lysis buffer, and the total extract was immunoprecipitated with 5 µg of anti-TPD52, -TPD53, -TPD54 or -GAPDH antibodies, or 5 µg of pre-immune mouse IgG (described above), according to the manufacturer's protocol. Subsequently, the precipitated proteins were eluted in 1X SDS sample buffer (Bio-Rad, Hercules, CA, USA) containing 5% of 2-mercaptoethanol, and co-precipitating proteins were then analyzed by western blotting as described in another subsection.

Total cellular proteins were collected with Triton X-100 lysis buffer [50 mM Tris-HCl (pH 8.0), 300 mM NaCl, 0.5% Triton X-100, 5 mM ethrylendiaminetetraacetic acid (EDTA), 1 mM sodium o-vanadate] supplemented with Complete Mini protease inhibitor tablet (Roche Diagnostics, Mannheim, Germany) 48 h after gene-transfection. Protein concentrations were measured using the Quick Start Bradford reagent (Bio-Rad) with bovine serum albumin (BSA) as a standard. For western blot analysis, 20 µg of total cellular protein was analyzed by sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) using a 4–20% gradient gel (Bio-Rad) and blotted onto a polyvinylidene difluoride membrane using iBlot 2 (Thermo Fisher Scientific, Waltham, MA, USA). After blocking of non-specific binding with 0.2% non-fat dry milk (Cell Signaling Technology, Danvers, MA, USA) in Tris-buffered saline (Takara Bio, Shiga, Japan), the membrane was incubated with the following primary antibodies: anti-TPD52 [1/1,000 dilution; rabbit monoclonal antibody, Abcam, Branford, CT, USA (ab182578)]; anti-TPD53 [1/1,000 dilution; rabbit polyclonal antibody, Proteintech, Rosemont, IL, USA (14732-1-AP)]; anti-TPD54 [1/1,000 dilution; rabbit polyclonal antibody, Proteintech (11795-1-AP)]; anti-hemagglutinin (HA) [1/1,000 dilution; rabbit polyclonal antibody, Clontech, Mountain View, CA, USA (631207)]; anti-green fluorescent protein (GFP) [1/1,000 dilution; rabbit polyclonal antibody, Santa Cruz, Dallas, TX, USA (sc-8334)]; or anti-glyceraldehyde 3-phosphate dehydrogenase (GAPDH) [1/10,000 dilution; mouse monoclonal antibody, Sigma-Aldrich (G9295)] antibody and horseradish-peroxidase-conjugated secondary antibody (for rabbit, donkey polyclonal antibody (NA934V); for mouse, sheep polyclonal antibody (NA931V) (GE Healthcare UK Ltd., Buckinghamshire, UK) as described previously (39). Subsequent washings were conducted using TBS-T. The protein bands were visualized using Amersham ECL Western Blotting Detection reagents (GE Healthcare UK Ltd.) and a ChemiDoc XRS Plus ImageLab System (Bio-Rad).

All samples were acquired from patients who underwent treatment for squamous cell carcinoma in Showa University Dental Hospital from January 2001 through March 2015. All patients provided informed consent before enrollment in the study, in accordance with the protocol approved by the Institutional Review Board at Showa University Dental Hospital (approval no. DH2015-013). Primary lesions were resected from tongue (Fig. 2A-a-c and B-a, -b) and gingiva (Fig. 2B-c). Differentiation grade was determined according to the pathological reports submitted from the Division of Pathology, Department of Oral Diagnostic Sciences, School of Dentistry, Showa University.

SAS (40), HSC2, and HSC3 cells (41) (human oral squamous cell carcinoma-derived cell lines) were grown in high-glucose Dulbecco's modified Eagle's medium (HDMEM) (Wako, Osaka, Japan) supplemented with 10% fetal bovine serum (FBS), 100 U/ml penicillin, and 100 mg/ml streptomycin (Thermo Fisher Scientific) at 37°C in an atmosphere containing 5% CO₂.

The coding regions of human tpd52 and -53 cDNAs were amplified by an RT-PCR technique using single-stranded cDNA reverse-transcribed from total RNA of SAS cells that was used as a template. All of the primer sequences used are shown in Table I. Final concentrations of the primers were 10 µM, respectively. PCR cycles were 95°C for 3 min; then, 94°C for 30 sec, 60°C for 30 sec, 72°C for 1 min (×40 cycles); and followed by 72°C for 5 min. The amplified products were subcloned into a pGEM T-Easy vector (Promega, Madison, WI, USA) using a TA-cloning technique. For construction of expression vectors of hemagglutinin (HA)-tagged human tpd52 and -53, cDNA of the coding region was amplified by PCR employing a sense-BglII-adaptor primer and an antisense KpnI-adaptor primer. The amplicons were double-digested with BglII and KpnI and inserted into the corresponding site of the pCMV-HA vector (Clontech Laboratories) as described previously (39). The resulting expression vectors were confirmed by sequencing using an ABI PRISM 310 Genetic Analyzer (Thermo Fisher Scientific) and the BigDye Terminator v3.1 Cycle Sequencing kit (Thermo Fisher Scientific). The expression vector for human tpd54 was that used in a previously described experiment (40). siRNA against human tpd52 and -54 was purchased from Sigma-Aldrich and Thermo Fisher Scientific, respectively. Control siRNA was purchased from Thermo Fisher Scientific. Expression vectors and siRNAs were transfected with Lipofectamine 2000 (Thermo Fisher Scientific) according to the manufacturer's protocol.

tpd52 and tpd54 overexpressing (OE) stable clones were obtained by using GFP Fusion TOPO TA Expression kits (Thermo Fisher Scientific). Single-stranded cDNAs were reverse-transcribed from total RNA of SAS cells that was used as a template. The amplified products were subcloned into a pcDNA3.1/NT-GFP-TOPO vector using TOPO cloning technology (Thermo Fisher Scientific). pcDNA3.1/NT-GFP (Thermo Fisher Scientific) was used as a control vector for OE experiments. TPD54 knocked-down (KD) stable clones were obtained by using BLOCK-iT Pol miR RNAi Expression Vector kits (Thermo Fisher Scientific). A double-stranded oligo targeting tpd54 was prepared by annealing two single-stranded oligos. The amplified products were subcloned into a pcDNA6.2-GW/Em-GFP-miR vector using Gateway cloning technology. pcDNA6.2-GW/Em-GFP-miR-neg (Thermo Fisher Scientific) was used as a negative control miRNA vector. All of the primer sequences used are shown in Table II. The resulting expression vectors were verified by sequencing and were transfected with Lipofectamine 2000 (Thermo Fisher Scientific). To obtain tpd52 and tpd54 OE stable clones, transfected SAS cells were cultivated in the presence of 0.5 mg/ml of Geneticin (Thermo Fisher Scientific). tpd54 knock-down stable clones were obtained by cultivation in the presence of 50 µg/ml of Blasticidin S HCl (Thermo Fisher Scientific), according to the manufacturer's protocol.

One thousand cells were seeded into a 96-well tissue culture plate in triplicate. After 48 h, cell growth was assayed using the tetrazolium salt 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assay, as described previously (42). The experiment was performed in triplicate.

The anchorage-independent growth assay was carried out by using a commercial kit (CytoSelect 96-Well In Vitro Tumor Sensitivity assay, Cell Biolabs, San Diego, CA, USA) according to the manufacturer's protocol. In total, 2500 transfected cells were grown in soft agar. After a week, the cells were examined under a microscope (Eclipse TS100/TS100-F, Nikon, Tokyo, Japan), photographed with a digital CCD camera (DS-Fil, Nikon), and formed colonies were counted (colonies/field). A colony was defined as a cell cluster that was >50 µm in diameter. Thereafter, cell growth was assayed using the MTT assay. The experiment was performed in triplicate.

Chemotaxis, haptotaxis, and invasion assays were carried out using a commercially available Boyden chamber kit (Chemotaxicell; Kurabo, Osaka, Japan). For the chemotaxis assay, the cells were starved in FBS-free medium for 24 h and then 10,000 of the transfected cells were seeded into a chamber inserted in a 24-well culture plate containing 10% FBS as a chemotactic agent. After 48 h, the cells were fixed with 10% formalin (Wako) and stained using crystal violet (43). Migrated cells were photographed, and the numbers of migrated cells were counted (cells/field). For the haptotaxis assay, the membranes were pre-coated with type I collagen (Nitta Gelatin Inc., Osaka, Japan), according to the manufacturer's protocol. Approximately 100,000 transfected cells were seeded into a chamber inserted in a 24-well culture plate. After 48 h, the cells were fixed and stained using crystal violet. Migrated cells were photographed, and were then counted, as described for the chemotaxis assay. The wound healing assay was carried out as described previously (44). The transfected cells (100,000) were seeded into a 24-well culture plate. Once the cells were confluent, a wound was made by scratching (0.9 mm width). After wound closure was first observed, the cells were fixed, stained using crystal violet and photographed. The wound width was then measured. For the invasion assay, the cells were starved in FBS-free medium for 24 h. Similar to the chemotaxis assay, 10,000 of the transfected cells were seeded into a chamber, which was pre-coated with type I collagen (Nitta Gelatin Inc.), and was then inserted in a 24-well culture plate. After 48 h, the cells were fixed and stained using crystal violet. Migrated cells were photographed and were counted in a manner similar to the chemotaxis and haptotaxis assays. Each experiment was performed in triplicate.

This study was approved by the Institutional Animal Care and Use Committee (approval no. 15024) and was carried out according to the Showa University Guidelines for Animal Experiments. Four-week-old female balb/c nu/nu mice (n=3 for each experimental group) were purchased from Clea Japan, Inc. (Tokyo, Japan) and were maintained under pathogen-free conditions. Approximately 1.0×10⁷ cells in 100 µl of phosphate-buffered saline (PBS) were injected subcutaneously into a unilateral flank. Tumor-bearing mice were treated with 1 nmol of siRNAs (Koken, Tokyo, Japan) in 100 µl of AteloGene (Koken), which were injected into subcutaneous spaces around tumor sections once a week for 7 weeks from day 7. The AteloGene and siRNA mixture was generated according to the manufacturer's protocol. At the end of the experiment, mice were sacrificed by CO₂ asphyxiation, and resected tumor, liver, and lung were fixed in formalin (Wako) and stained with hematoxylin and eosin (H&E) (Sakura Finetek Japan, Tokyo, Japan), as described previously (45). Tumor volume was determined by direct measurement, and was calculated using the formula π/6 × (large diameter) × (small diameter)² (45). The Kaplan-Meier method was used for survival analysis (46).

Resected specimens were fixed with 10% formalin, embedded in paraffin, stained with H&E, and then immunohistochemically stained for TPD52 family proteins as described previously (46). Antigen retrieval was carried out with citrate-phosphate buffer (0.01 M, pH 6.0) at 121°C for 20 min. Endogenous peroxidases were blocked by incubating the sample with 10% H₂O₂ (Wako) for 10 min. Proteins were blocked using a commercial kit (Dako, Carpinteria, CA, USA) according to the manufacturer's protocol. The sections were incubated at 4°C overnight with primary antibodies (anti-TPD52 antibody, 1:50 dilution, rabbit polyclonal antibody; Biorbyt, Cambridgeshire, UK (orb100564); anti-TPD53 antibody, 1:200 dilution, rabbit polyclonal antibody, Proteintech (14732-1-AP); anti-TPD54 antibody, 1:200 dilution, rabbit polyclonal antibody, Proteintech (11795-1-AP); anti-cytokeratin 10/13, 1/200 dilution, mouse monoclonal antibody, Santa Cruz (DE-K13); control pre-immune IgG, 1/200 dilution, mouse IgG, Santa Cruz (sc-2762). The next day, the sections were incubated with secondary antibodies (EnVision⁺ system-HRP labelled polymer anti-rabbit/anti-mouse; Dako). Finally, sections were reacted with a 3, 3′-diaminobenzidine (DAB) peroxidase substrate kit (Dako) for color development and were examined under a microscope.

All values are expressed as means ± standard deviation of triplicate data sets. The statistical significance of differences between groups was analyzed using a paired Student's t-test. A p-value of <0.05 was considered statistically significant.

Since previous studies (2,4) showed that TPD52 family proteins form homo-and hetero-complexes in various cells, we investigated whether such molecular interactions were observed in SAS cells by a co-immunoprecipitation (co-IP) assay (Fig. 1). In western blotting of the total cellular protein (input), the blotted bands of TPD54, TPD53, and TPD52 gave strong, weak, and faint signals, respectively, in agreement with our previous study (39). Of more importance, TPD52, -53, and -54 were found to co-immunoprecipitate with each other. Co-IP with control pre-immune mouse IgG indicated the specificity of the binding, despite the presence of a very faint pseudo-positive signal that was probably due to insufficient washing of the blots.

We recently showed that TPD54 is highly expressed in oral squamous cell carcinomas (38). However, the expression of TPD52 and -53 in OSCC, and the differences in TPD52 family protein expression in highly and poorly differentiated tumors are still unclear. We therefore analyzed the detailed expression patterns of TPD52 family proteins in OSCC using immunohistochemical staining (Fig. 2). Three cases each of highly and poorly differentiated tumors were randomly chosen from tongue (Fig. 2A-a-c and B-a, -b) and gingiva (Fig. 2B-c). Differentiation grade was determined according to the pathological reports. Immunohistochemical staining showed that cytokeratin 10/13 [a marker of SCC cells (47)] was highly expressed in all of these OSCCs. A cancer 'pearl' structure was observed in some of the highly differentiated tissues (arrows in Fig. 2 indicate a typical cancer pearl structure), but was not observed in poorly differentiated tissue. Additionally, staining with pre-immune IgG (negative control) indicated the specificity of TPD52 family immune staining. In agreement with our previous study (38), TPD54 was highly expressed in the cancer region and the surrounding connective tissue, regardless of the tumor differentiation level. In contrast, the expression of TPD52 in either highly or poorly differentiated OSCC was lower than that of TPD54, and TPD52 was barely expressed in normal tissue. TPD53 was moderately expressed in the cancer region.

The higher expression level of TPD54 compared to TPD52 and -53 suggested that TPD54 might play a more important role in OSCCs. Furthermore, our previous study (39) showed that TPD54 affects cell attachment to the ECM and cell migration of OSCC cells. Based on those findings, we hypothesized that TPD54 might be a key protein in OSCC cells, and might negatively affect tumor progression mediated by TPD52 and -53. To further assess this possibility, tpd52, -53, and -54 were overexpressed in SAS cells. Additionally, in this experiment, HA-tagged tpd52 or -53 expression vectors were transfected into SAS cell lines that stably expressed GFP-tpd54 or GFP-miRNA against-tpd54. Subsequent western blot analysis (Fig. 3) showed that protein expression resulting from transient overexpression of tpd52 or -53 and stable overexpression or knock-down of tpd54 was observed as expected. Expression marker tags (HA and GFP) and internal control (GAPDH) proteins were also detected. Since expression profiles of control miRNA cell lines were similar to those of control cell lines (Fig. 3), protein expression in the control miRNA cell lines in further experiments are not shown to simplify the experiments. These cells were then assayed using an MTT assay of the monolayer cultured cells (Fig. 4A), a soft agar colony formation assay (Fig. 4B and C) and an MTT assay of the colonies formed in soft agar (Fig. 4D). Significant differences in cell growth between the control and several of the differently transfected cells were observed in the MTT assay of monolayer cultures (^*p<0.05 versus control cells) (Fig. 4A), but the differences were smaller than those between the cells in colony formation assays (Fig. 4B–D). Colony formation of the tpd54-overexpressing cells was greatly limited in an anchorage-independent manner versus control cells, regardless of the co-expression of tpd52 or tpd53 (Fig. 4B). Conversely, knock-down of tpd54 enhanced the number of colonies formed (Fig. 4C). MTT assay of the cells in this colony forming experiment showed no differences in cell growth (Fig. 4D), indicating that cell viability was not an important factor even in an anchorage-dependent culture. These findings suggested that TPD54 might be capable of attenuating the tumorigenicity of other TPD52 family proteins, and that TPD54 and other TPD52 family proteins might have opposite effects on primary tumor formation. Next, it was examined whether these effects on cell growth are reproduced in other OSCC cell lines, or not. The same experiments were carried out in HSC2 (Fig. 5) and HSC3 (Fig. 6) cells. HSC2 and HSC3 cells also showed significant differences between transfected cells versus control cells in an anchorage-independent manner in colony formation (^*p<0.05 versus control cells) as well as in cell growth in a monolayer culture. However, the effects observed in HSC2 and HSC3 cells were slightly different than those observed in SAS cells. Therefore, SAS cells, in which OE and KD of TPD52 proteins were the most effective were used as representative OSCC cells in the following experiments.

Our previous report (39) showed that TPD54 plays a crucial role in the cell migration and invasion of OSCC cells. To further assess the effects of TPD52, -53, and -54 on the cell migration of OSCC cells, we performed chemotaxis, haptotaxis and wound healing assays of the transfected cells (Fig. 7). tpd54 overexpression reduced cell migration versus control cells in a chemotaxis assay when expressed either alone or in combination with overexpression of tpd52 or -53 (Fig. 7A). However, little effects on haptotaxis were observed following overexpression or knock-down of tpd54 alone or combined with overexpression of tpd52 or -53 (Fig. 7B). Overexpression of tpd52 or -53 alone did not affect either chemotaxis or haptotaxis. tpd54-overexpressing cells (tpd54⁺, 53⁻, 52⁻) showed decreased wound closure compared to control in a wound healing assay (^*p<0.05, Fig. 7C). No differences versus control were observed for the other experimental groups (Fig. 7C). We additionally investigated the effect of the different cell transfection combinations on cell invasion (Fig. 7D). However, cell invasion was barely modulated by any of these combinations. These results indicated that TPD54 might regulate cell migration through the modulation of molecular events such as integrin signaling.

Since our previous in vitro study suggested that TPD54 has negative effects on cell proliferation and migration in OSCC cells, we further examined the ability of TPD54 to modulate tumor growth and/or metastasis in an in vivo study. The scheme of this study is outlined in Fig. 8A. In these experiments, one nanomole of siRNAs in Atelogene was injected once a week into each mouse (n=3 per group) with a xenografted tumor of OSCC cells, in order to obtain sustainable knock-down of tpd52 or -54. A significant difference in body weight between the tpd54 overexpressing group and control was seen (^*p<0.05) (Fig. 8B) although no significant tumor suppression was observed in the tpd54 overexpressing group (Fig. 8C and E). However, contrary to our expectations, knock-down of tpd52 in tpd54 overexpressing cells showed little effect on tumor growth. No significant differences in survival rate between the different groups of mice, as calculated by Kaplan-Meier's method, were observed (Fig. 8D). At the endpoint, specimens of the original tumor, and of lung and liver tissue were H&E stained for confirmation of metastasis (Fig. 9). However, there were few histological differences between the groups. Metastasis was not observed in any group by microscopic examination. These results indicated that TPD54 might function as a negative regulator of tumor growth, suggesting that interactions between TPD52 family protein members might be involved in the growth of OSCC.