Oral squamous carcinoma cancer cell lines SAS and OECM1 and immortalized human gingival keratinocyte S-G cells were used in the present study. The SAS cell line was established from a poorly differentiated squamous cell carcinoma of the tongue. The OECM-1 cell line was established from OSCC cells surgically excised from a Taiwanese patient. The S-G cell is an epithelial-like cell line established from an area of clinically normal adult human attached gingiva and thus is used as a normal control. The S-G and SAS cells were grown in Dulbeccos modified Eagles medium (DMEM; Gibco/Life Technologies), the OECM1 cells in RPMI-1640 medium (Gibco/Life Technologies), and HSC-3 cells were in DMEM/F-12 medium (Gibco/Life Technologies) with 10% fetal bovine serum (FBS; HyClone Laboratories, Inc., Logan, UT, USA) and antibiotics (100 U/ml penicillin and 100 ml/ml streptomycin) at 37°C in an incubator with 5% CO₂. L-glutamine (2 mM) was supplemented to all cell lines except HSC-3 and pyruvate (2 mM) was supplemented to the S-G cells. After reaching confluency, the cells were harvested and washed with phosphate-buffered saline (PBS, 10 mM Na₂HPO₄; 1.8 mM KH₂PO₄ 140 mM NaCl and 2.7 mM KCl, pH 7.4), resuspended and lysed in buffer A [50 mM Tris-HCl, pH 7.4; 150 mM NaCl; 1 mM EDTA; 1% Triton X-100; 1x complete protease inhibitor (Roche); 10 mM NaF] by sonication. The lysate was centrifuged and the supernatant was designated as the cell extract. The protein concentration was determined by the BCA protein assay (Thermo Fisher Scientific, Waltham, MA, USA) using bovine serum albumin (BSA) as the standard.

Protein samples were separated by SDS-polyacrylamide gel electrophoresis (PAGE) and then transferred to nitrocellulose membranes (Hybond-C Extra, Gelman Sciences; Amersham Biosciences, Little Chalfont, UK) as previously described (38). The membranes were blocked in 7% skimmed dry milk in TTBS (10 mM Tris-HCl, pH 7.5; 100 mM NaCl; 0.1% Tween-20) for 1 h, incubated with primary antibodies (1:1,000 for anti-PRMT1, UPSTATE and 1:1,000 for anti-methylarginine antibodies ADMA from Cell Signaling, Danvers, MA, USA) overnight, washed three times in TTBS, then incubated with secondary antibody for 1 h. Chemiluminescent detection using the VisGlow substrate for HRP (Visual Protein, Taipei City, Taiwan) was performed according to the manufacturers instructions.

Aliquots of cultured cells were mixed with trypan blue and the cell numbers were counted using a hemocytometer. The MTT assay was used to evaluate cell proliferation. Cells seeded in 6-well plates (10,000 cells/well) were incubated with the MTT solution (0.25 mg/ml; Sigma-Aldrich, St. Louis, MO, USA) for 3 h. The resulting formazan crystals were dissolved in dimethyl sulfoxide (DMSO) and measured by spectrophotometry at the absorbance of 570 nm.

Cells (5×10⁴) were seeded into wells of the Oris Cell Migration Assembly Kit-FLEX (Platypus Technologies Llc, Fitchburg, WI, USA) and migration assays were conducted in accordance with the manufacturers instructions. After attached for 10 h, cells were allowed to migrate into the clear field with the removal of the well inserts. After migration for a period of time, cells were fixed with formaldehyde, stained with crystal violet and photographed. The pre-migration and post-migration images were analyzed using ImageJ software (http://rsb.info.nih.gov/ij/).

Lentiviral particles produced in packaging vector pLKO_TRC005 with short hairpin RNA (shRNA) targeting human PRMT1 (A1 with the target sequences: 5′- GTGTTCCAGTAT CTCTGATTA-3′; B1 with the target sequences: 5′-CCGGCAGTACAAAGACTACAA-3′) and a negative control were obtained from the National RNAi Core Facility (Academia Sinica, Taipei City, Taiwan). Cells were infected in complete growth medium supplemented with polybrene (Sigma-Aldrich). After 24 h, cells were grown and selected in medium containing in 5 µg/ml puromycin (Sigma-Aldrich). The knockdown effects were examined by PRMT1 protein expression using immunoblotting.

Paraffin embedded specimens (Pathological blocks) were from patients with head and neck cancers of different stages. The patients recruited in the Department of Otolaryngology, Chung-Shan Medical University Hospital (Taichung, Taiwan) between 2001 and 2010 were de-linked from their identification information and were randomly numbered. The present study was approved by the Institutional Research Board of Chung-Shan Medical University Hospital. IHC was conducted with the ultraView Universal DAB Detection kit (Roche) using the Ventana BenchMark XT automated IHC/ISH slide staining system (Roche) at the short CC antigen retrieval condition. Anti-PRMT1 antibodies from Upstate Biotechnologies were diluted (1:200) and incubated at 37°C for 16 min.

We first examined protein arginine methylation in three oral squamous carcinoma cancer cell lines (SAS, OECM-1 and HSC-3). Immortalized S-G cells established from clinically normal gingiva was used as a normal control. In comparison, cell extracts prepared from other human cancer cell lines such as HeLa (cervical cancer cell line) and MCF-7 (breast cancer cell line) cells that have been investigated for PRMTs or arginine methylation (39–41) were included. The major type I protein arginine methyltransferase PRMT1 and ADMA-containing proteins were detected using specific antibodies. Western blot analyses showed that the expression patterns of PRMT1 in oral cancer cells were similar to that of HeLa and MCF-7 cells with the major signals of about the same molecular masses (Fig. 1A). The expression levels of PRMT1 was lowest in HSC-3.

The levels of ADMA-containing proteins were high in OECM-1 and SAS, and lower in S-G cells (Fig. 1B). HSC-3 only showed few weak ADMA-containing signals. To analyze the effects of arginine methylation in the oral cancer cells, we treated the cells with an indirect methyltransferase inhibitor AdOx that has been widely used in protein methylation studies. AdOx treatment clearly decreased ADMA signals in oral cancer cell line SAS and OECM-1 that express high level of PRMT1, the major type I PRMT catalyzing the formation of ADMA (Fig. 1B; 50 mM AdOx). Decreased ADMA signals following the methylation inhibition by AdOx also confirmed that the signals detected by the ADMA-specific antibodies are from methylation modification. However, methylarginine levels reduced less significantly for HSC-3 and S-G cells.

As a general transmethylation inhibitor, AdOx has been shown to block proliferation of different cancer cells (33,36,37). To examine whether AdOx can affect growth of oral cancer cells, we first determined the growth curves of the four cell lines. SAS and OECM-1 proliferated at a higher growth rate compared to the S-G cells (Fig. 2A). On the other hand, HSC-3 cells that expressed low level of PRMT1 and ADMA-containing proteins, proliferated at a slower rate similar to the S-G cells. We then determined the effects of AdOx treatment on the proliferation of these cells. Addition of AdOx at the concentration that could effectively reduce the ADMA level (50 mM) also blocked the proliferation of SAS and OECM-1 cells significantly (Fig. 2B). In contrast, cell growth of the S-G cells was only slightly inhibited by the same AdOx treatment. Limited growth inhibition of the HSC-3 cells by AdOx treatment was observed (Fig. 2B).

AdOx treatment has also been demonstrated to suppress tumor cell invasion and migration (35). We thus evaluated the effects of AdOx on migration of the oral cancer cells. Analyses of the migration activity of the oral cancer cells showed migration activity of SAS and HSC-3 but limited migration ability of OECM-1 (data not shown). We thus used SAS and HSC-3 cells to study the effects of AdOx on migration. It is apparent that HSC-3 migrated faster than SAS (Fig. 3A, 16 h). Treatment with 50 mM of AdOx reduced the migration of SAS and HSC-3 cells (Fig. 3A). To analyze the effects of AdOx treatment on the migration activity of the cells, migration area of the AdOx-treated cells was compared with that of the untreated ones. After the treatment, the area covered by the migrated SAS cells was only about 30% of the area covered by the migrated untreated cells. The migration activity of HSC-3 cells was slightly suppressed with the migration area close to 70% of that of untreated cells (Fig. 3B).

To further investigate the clinical involvement of PRMT1 in HNC, we analyzed PRMT1 expression in oral cancer specimens by immunohistochemistry. Six patients were included in the present study as described in Table I. We included tissue sections from patients of the late (IVA) or early (I) oral cavity SCC stages and the subside from tongue or buccal mucosa. Representative IHC stains of all six patients are shown in Fig. 4A. To better illustrate the PRMT1 expression in tumor and normal cells, representative regions containing tumor or normal cells from patients 4, 5 and 6 are shown in Fig. 4B with higher magnification. It is clear that PRMT1 stained weakly in the normal cells but its expression were stronger in the poorly differentiated tumor cells. The majority of the intense PRMT1 signals were concentrated in the nucleus. As illustrated in patients 4–6, the IHC stain of PRMT1 revealed strong nuclear staining in cancer cells (upper panels) and weak nuclear staining in non-neoplastic squamous cells (lower panels). Even though the samples were limited in the present study, the expression level of PRMT1 in the specimens appeared to have strong correlation with the tumors. The PRMT1 level did not appear to be correlated with the cancer stage, subside or sex.

We also analyzed the head and neck squamous carcinoma database from The Cancer Genome Atlas (TCGA; 273 samples) by SurvExpress (http://bioinformatica.mty.itesm.mx/SurvExpress) (42). In SurvExpress, all datasets from TCGA were obtained at the gene level (level 3) which were the data of RNA-Seq (42). It is apparent that the high-risk group express higher level of PRMT1 than the low-risk group (Fig. 4C). Similar results can be obtained from another database (TCGA Head and Neck squamous cell carcinoma June 2016, 502 samples). Furthermore, extended survival of the PRMT1 low expression group is statistically significant (P<0.05) compared with the high expression ones in either the stage II (Fig. 4D) or grade II samples (data not shown).

As arginine methylation stands for the majority of the AdOx-blocked protein methylation (38) and PRMT1 is the predominant type I PRMT, AdOx might affect mostly through PRMT1 for decreased migration activity. Furthermore, PRMT1 expression was strongest in the poorly differentiated cells in the clinical samples. We thus suspect that PRMT1 is critical for the proliferation and migration/invasion activity of HNC. To specifically analyze the function of PRMT1 in oral cancer cells, we knocked down PRMT1 gene in SAS cells with shRNA by lentivirus (shPRMT1A1, B1 or combined) infection. The protein level of PRMT1 decreased almost completely in cells that stably express either or both of the two different PRMT1 shRNA (PRMT1-KD-1, PRMT1-KD-2 or PRMT1-KD) compared to that in the non-infected or control shRNA infected SAS cells (SAS and Control; Fig. 5A). As PRMT1 is the major type I PRMT, the level of proteins containing asymmetric dimethylarginine (ADMA) also decreased significantly after the PRMT1 knockdown (data not shown). We then examined the growth of SAS cells with PRMT1 knockdown. The PRMT1 KD cells apparently grew slower than the non-infected or control shRNA infected cells. Treatment of the cells with AdOx further reduced the growth of all three cells with or without PRMT1 KD (Fig. 5B).

We also determined the migration abilities of the SAS cells. The migration of the SAS cells with PRMT1 knockdown showed significantly decreased migration compared with the cells with control shRNA or SAS cells (Fig. 5C and D). E-cadherin is a calcium-dependent adhesion molecule expressed on normal epithelium (43). The elevated expression of E-cadherin has been reported to suppress cell migration (44). We further examined the expression of E-cadherin in PRMT1 knockdown SAS cells and the results showed that PRMT1 knockdown upregulated the epithelial marker E-cadherin (Fig. 5A).