This retrospective study included 119 consecutive treatment-naïve patients with biopsy confirmed primary HNSCC, who were treated at the Department of Otolaryngology, Head and Neck Surgery, Chungnam National University Hospital between 1998 and 2010. All patients underwent primary surgery including neck dissection. Tumor staging was conducted according to the criteria of the 7^th edition of the AJCC (16). Clinicopathological parameters, including histopathological and surgical studies, were obtained from the medical charts.

Follow-up data were collected by a combination of chart review and the local government office for the registration of residents. At the time of the analysis, 51 (42.9%) patients had succumbed to the disease; 45 (37.8%) of them due to the tumor. Overall, 54 patients developed recurrent disease, including 21 (17.6%) local recurrences, 29 (24.3%) subsequent regional lymph node metastases and 8 (6.7%) distant recurrences. The average follow-up time was 40.6 (range, 2–144) months.

The immunohistochemical (IHC) study of the HNSCC patients was approved by the local institutional review board. Formalin-fixed and paraffin-wax-embedded tissues from 119 HNSCC lesions - 44 in the oral cavity, 17 in the oropharynx, 48 in the larynx and 10 in the hypopharynx - were retrieved from the archives of the Department of Pathology, Chungnam National University Hospital, Korea. Sections (4 µm) of the paraffin-wax-embedded tissue blocks were used for IHC studies according to the procedures provided in the Vectastain ABC kit (Vector, Burlingame. CA, USA). All slides were scored, as described by Yuan et al (14), by at least two pathologists who were blinded to the clinical information of the patients. For scoring, representative areas of each tissue section were selected and evaluated independently. Cytoplasm and/or membrane immunoreactivity was considered to indicate podoplanin expression. Quantitative scores of 0 to 5 were assigned when 0, 1–10, 11–30, 31–50, 51–80 or 81–100% of the tumor cells were positive, respectively. The staining intensity was rated on a scale of 0 to 3: 0 = negative, 1 = weak, 2 = moderate, and 3 = strong. The raw data were then converted to a German Immunoreactive Score (IRS) by multiplying the quantity and staining intensity scores. Theoretically, the scores may range from 0 to 15. An IRS score above the median (≥7) was considered high reactivity and 0–6, low. Consensus opinions were used to assign final IRS scores in disputed cases before data analysis.

The HNSCC cell lines FaDu and SNU-1041 were purchased from the American Type Culture Collection (ATCC; Rockville, MD, USA) and the Korean Cell line Bank (KClB; Seoul, Korea) and maintained in DMEM and RPMI supplemented with 100 U/ml penicillin, 100 µg/ml streptomycin, 25 ng/ml amphotericin B and 10% fetal bovine serum (FBS) (Gibco, Carlsbad, CA, USA) at 37°C in a humidified incubator with 5% CO₂, respectively.

Podoplanin-specific siRNA and control siRNA were purchased from Bioneer (Daejeon, Korea). The targeted sequences of podoplanin were sense siRNA, 5′-GAU AAC ACG UGU GGU GAA CAA TT-3′ and antisense, 5′-UUG UUC ACC ACG UGU UAU GTT-3′. Podoplanin siRNA transfection was performed in opti-MEM media with the transfection reagent lipofectamine 2000 (Invitrogen life Technologies, Carlsbad, CA, USA) following the manufacturer’s instructions.

Human podoplanin cDNA (NM_006474.4) was PCR-amplified from a cDNA library which was purchased from Origene (Rockville, MD, USA) and cloned into the NotI/NotI site of pCMV6-Xl5 (Origene). The PDPN vector containing the podoplanin coding region or the mock vector was transfected into the FaDu or SNU-1041 cell line with lipofectamine 2000 according to the manufacturer’s instructions, respectively.

Oligonucleotides of the VEGF family genes such as VEGF-A, -B, -C and -D were purchased from Bioneer. The primer sequences for GAPDH were 5′-CCATCACCATCT TCCAGGAG-3′ (sense) and 5′-ACAGTCTTCTGGGTGGCA GT-3′ (antisense). The VEGF family-specific primers used for PCR were described in a previous study (17). Relative gene expression levels were normalized to GAPDH expression.

Sodium dodecyl sulfate-polyacryl-amide gel electrophoresis (SDS-PAGE) was conducted using a Mini-Protean Tetra Cell (Bio-Rad, Hercules, CA, USA) and a 12% gel according to the manufacturer’s instructions. Proteins were transferred to a PVDF membrane and probed with primary antibodies followed by an HRP-conjugated secondary antibody. Immunolabeled proteins were detected by incubation with enhanced chemiluminescence (ECl) substrate followed by exposure of the membrane to autoradiography film.

Cells were plated on a 60-mm culture dish with 90% confluence and an injury line with a width of 2 mm was made by scraping across the cell monolayer with a yellow tip. After floating cell debris was removed by washing with PBS, cell migration was monitored under a phase-contrast microscope and photographed.

FaDu and SNU-1041 cells were cultured for 24 h after the transfection of podoplanin siRNA or PDPN vector in growth medium containing 10% FBS. The following procedures are described in a previous study (18). The cells were counted by taking photomicrographs at a magnification, ×100. Cells in five different fields of each well were counted with two wells per treatment. The mean values were obtained from three replicate experiments and were subjected to a t-test.

The SPSS software (ver. 14; SPSS, Chicago, Il, USA) was used to perform statistical analyses. Univariate analysis of podoplanin expression and clinicopathological parameters was performed using the Fisher’s exact test. Significant variables in the univariate analysis were included in a multinomial logistic regression test (multivariate analysis). The Kaplan-Meier method (assessed by log-rank test) and Cox regression model were used for univariate and multivariate overall and disease-specific survival analyses. A P-value <0.05 was considered to indicate statistical significance.

As expected, podoplanin was highly expressed in the endothelial cells of lymphatic vessels, but was not detectable in the endothelial cells of blood vessels. In histologically normal squamous epithelium adjacent to the tumors, podoplanin expression was either not detectable or extremely low in some basal cells. In primary HNSCC, podoplanin expression was generally heterogeneous and differentially increased (Fig. 1). Podoplanin transcript and translational levels were expressed differentially in the HNSCC cell lines (Fig. 2).

Immunohistochemical staining images are shown in Fig. 1. Of the 119 cases, 26 (21.8%) showed no podoplanin expression, 32 (26.9%) had weak, 37 (31.1%) moderate and 24 (20.2%) strong expression. For statistical purposes, we divided the cases into two groups; those with scores (IRS) ≤6 (the median value) were considered to have low podoplanin expression, whereas those with scores >6 were considered to have high expression. Thus, podoplanin expression was low in 58 (48.7%) cases and high in 61 (51.3%).

Relationships between the degree of podoplanin expression and the clinicopathological features of the 119 HNSCC patients are shown in Table I. There was a statistically significant correlation between high podoplanin expression and the presence of lymph node metastasis, advanced AJCC stage, and poor histological grade. In the multivariate analysis, high podoplanin expression was significantly associated with ~3- and 5-fold increases in the presence of positive lymph node metastasis and poor histological grade, respectively (P<0.05; Table II).

High podoplanin expression had a marked impact on the overall (P<0.001) and disease-specific survival rate (P<0.001; Fig. 1E and F). The 5-year overall and disease-specific survival rates were 77 and 86%, respectively, in patients with low podoplanin expression, while 37 and 44%, respectively, in those with high podoplanin expression.

Cox proportional hazards regression analysis was performed to determine whether the effect of podoplanin expression on the disease-specific survival rate was dependent on other known risk factors. Poor histological grade and high podoplanin expression were independent significant factors for worse disease-specific survival rates (P<0.05; Table III). The disease-specific death risk in HNSCC patients with high podoplanin expression was ~3-fold higher than in those with a lower podoplanin expression (Table III).

After cells were transfected with the podoplanin siRNA for 24 h, total protein was extracted, and western blot analysis was performed. There were various degrees of podoplanin protein suppression, depending on the cell types and there was no change in the cellular morphology of cells that were treated with podoplanin siRNA (data not shown). As shown in Fig. 3A, control siRNA-expressing FaDu or SNU-1041 cells repaired the wounded area rapidly at 24 h. Meanwhile, treatment with podoplanin siRNA in FaDu and SNU-1041 cells inhibited wound repair significantly at 24 h, respectively. In the Transwell invasion assay, the percentages of the invasive cells in the podoplanin siRNA-expressing FaDu and SNU-1041 were decreased to 72 and 80.7% of that in the control siRNA-expressing FaDu and SNU-1041 cells, respectively (Fig. 3B). In addition, inhibition of podoplanin led to suppression of the VEGF-C transcript level (Fig. 3C) and protein expression (Fig. 3D) in the FaDu and SNU-1041 cells, respectively. Other VEGF family such as VEGF-A, -B and -D were not altered in the podoplanin siRNA-expressing cells when compared to levels in the control cells (Fig. 3C).

As shown in Fig. 4A, mock vector-expressing FaDu or SNU-1041 cells slightly repaired the wounded area at 24 h. However, PDPN vector-overexpressing FaDu and SNU-1041 cells induced wound repair rapidly at 24 h, respectively. In a Transwell invasion assay, the percentages of invasive cells in the podoplanin-expressing FaDu and SNU-1041 were increased to 148.3 and 113.9% of that in the mock vector-expressing FaDu and SNU-1041 cells, respectively (Fig. 4B). Furthermore, overexpression of podoplanin led to an increase in the VEGF-C transcript level (Fig. 4C) and protein expression (Fig. 4D) in the FaDu and SNU-1041 cells, respectively. Other VEGF family such as VEGF-A, -B and -D were not altered in the podoplanin-overexpressing cells compared to that in the mock cells (Fig. 4C). These results suggest that podoplanin regulates cell wound healing activity and invasiveness through interaction of VEGF-C in HNSCC cells.

To determine whether podoplanin modulates HNSCC cell metastasis, we examined cell wound healing migration and a Transwell invasion assay in VEGF-C-suppressed cells transfected with VEGF-C siRNA. We confirmed that VEGF-C protein expression was suppressed in the VEGF-C siRNA-expressing FaDu and SNU-1041 cells compared with levels in the control siRNA-expressing FaDu and SNU-1041 cells, respectively (Fig. 5A). Control siRNA-expressing FaDu and SNU-1041 cells repaired the wound area rapidly at 24 h, while the VEGF-C siRNA-expressing FaDu and SNU-1041 cells suppressed the wound repair significantly at 24 h, respectively (Fig. 5A). In addition, the percentage of invasive cells in the VEGF-C siRNA-expressing FaDu and SNU-1041 were decreased to 71.9 and 75.4% of that in the control siRNA-expressing FaDu and SNU-1041 cells, respectively (Fig. 5B). Additionally, inhibition of VEGF-C led to suppression of podoplanin expression in the FaDu (Fig. 5C) and SNU-1041 (Fig. 5D) cells, respectively.

Next, to determine a correlation between podoplanin and VEGF-C, we examined the cell wound healing migration and invasion assays in both podoplanin overexpression vector and VEGF-C siRNA-cotransfected FaDu and SNU-1041 cells, respectively. The podoplanin-overexpressing FaDu and SNU-1041 cells rapidly repaired the wound area compared with this rate in the mock cells, respectively (Fig. 6A and C). In contrast, both podoplanin vector and VEGF-C siRNA-coexpressing FaDu or SNU-1041 cells inhibited wound repair compared with the podoplanin vector-expressing FaDu and SNU-1041 cells (Fig. 6A and C). In addition, the percentages of invasive cells in the podoplanin vector-expressing FaDu and SNU-1041 were increased to 163.9 and 184% of that in the mock vector-expressing FaDu and SNU-1041 cells (Fig. 6B and D). In the cell wound healing assay, the percentages of invasive cells in both the podoplanin vector and VEGF-C siRNA-coexpressing FaDu and SNU-1041 cells were significantly decreased to 48.8 and 20.7% of these percentage in the podoplanin vector-expressing FaDu and SNU-1041 cells, respectively (Fig. 6B and D). These results indicated that podoplanin regulates metastasis via interaction of VEGF-C in HNSCC cells.