Src-homology 2 domain-containing tyrosine phosphatase 2 promotes oral cancer invasion and metastasis

Twenty-one pairs of primary oral cancer and histologically normal oral mucosa adjacent to the tumors were obtained after surgical resection at Chi-Mei Medical Center, Liouying, Tainan, Taiwan, and stored at -80°C until use. All of the human tissue specimens in this study were processed and used with prior approval from the Chi-Mei Medical Center Institutional Review Board and the National Health Research Institute Institutional Review Board (IRB1000202-R2). Samples containing > 70% tumor cells were selected after microscopic examination of representative tissue sections from each tumor.

Immunohistochemistry (IHC) was performed to evaluate SHP2 expression in paraffin-embedded oral squamous cell carcinoma specimens. The slides were stained with a SHP2 antibody (1:200, GeneTex Inc., Irvine, CA, USA) by using an automatic slide stainer BenchMark XT (Ventana Medical Systems), and counterstained with Harris hematoxylin. Two independent pathologies evaluated the slides under a light microscope. Immunoreactivity was classified by estimating the percentage (P) of tumor cells exhibiting characteristic staining (from an undetectable level, 0%, to homogeneous staining, 100%) and by estimating the intensity (I) of staining (1, weak staining; 2, moderate staining; and 3, strong staining). Results were scored by multiplying the percentage of positive cells by the intensity, (i.e. quick score Q = P × I; maximum = 300) [21].

Real-time reverse transcription-polymerase chain reaction (RT-PCR) analysis of SHP1, SHP2, Snail, Twist1 and GAPDH was conducted using SYBR-Green Master Mix (Roche Applied Science, Basel, Switzerland) according to the manufacturer's instructions. PCR amplifications were performed using an ABI7900 thermal cycler by applying the following amplification conditions: 50°C for 2 min, 95°C for 10 min, for 40 cycles at 95°C for 15 s (denaturation step), and 60°C for 1 min (annealing/extension steps). GAPDH was amplified as an internal control. All of the experiments were performed in duplicate. Relative expression of the target genes (SHP1, SHP2, Snail, and Twist1) to the control gene (GAPDH) was calculated using the ΔC_T method: relative expression = 2^-ΔC _T, where ΔC_T = C_{T (Target)} - C_{T (GAPDH)}[22]. The oligonucleotide primers for human SHP1, SHP2, Snail, Twist1, and GAPDH are listed as follows: SHP2, forward: 5’-TCGTATAAATGCTGCTGAAAT-3’, reverse: 5’- TCCTGTTGTTGTAGTGTCT-3’; SHP1, forward: 5’-GCAGTACAAGTTCATCTA-3’, reverse: 5’-CAGGTTCTCATACACATC-3’; Snail, forward: 5’-ACGAGGTGTGACTAACTATG-3’, reverse: 5’-GACAAGTGACAGCCATTAC-3’; Twist1, forward: 5’- TGATAGAAGTCTGAACAGTTGT-3’, reverse: 5’-GCACGACCTCTTGAGAAT-3’; GAPDH, forward:5’-ACACCCACTCCTCCACCTTT-3’, reverse: 5’- AGCCAAATTCGTTGTCATACC-3’.

HSC3 cells (JCRB, JCRB0623) were provided by Dr. Lu-Hai Wang, Institute of Molecular and Genomic Medicine, National Health Research Institute, Taiwan. The HSC3 cells were cultured in Dulbecco’s modified Eagle’s medium supplemented with 100 μL/mL of fetal bovine serum [23].

The highly invasive HSC3 cell line was established using the Falcon Cell Culture Inserts with a Matrigel coating (BD Biosciences, CA, USA). Briefly, cells (5 × 10⁴) were harvested, re-suspended in a serum-free medium with 0.1-% bovine serum albumin (BSA) (Sigma-Aldrich, Inc., St. Louis, MO, USA), and then plated in a transwell chamber. The chamber was incubated for 18 h with a complete culture medium added to the lower chamber. After 18 h of incubation, cells migrating to the lower surface of the filter were collected [23]. This in vitro selection protocol was used in selecting cells from 4 to 8 cycles to derive the highly invasive sub-lines, HSC3-Inv4 and HSC3-Inv8; in these terms, the number following “Inv” denotes the number of cycles of selection. After invasion selection, the lines were tested for their migratory and invasive ability by performing a Boyden chamber migration/invasion assay [24].

Cell viability was measured using the 3-(4, 5-dimethylthiazol-2-yl)-2, 5-diphenyl-2H- tetrazolium bromide (MTT) colorimetric assay. The HSC3 cells were plated at 10³ cells/well in a 96-well plate (100 μL/well) and incubated for 24 h. After 24 h, the culture medium was removed, and 200 μL of a fresh medium containing 20 μL of MTT (5 mg/mL; Sigma-Aldrich Japan, Tokyo, Japan) was added to each well. The cells were incubated at 37°C for 4 h. After 4 h, the liquid was discarded and DMSO (200 μL/well) was added, after which the samples were mounted on a micromixer for 15 min to make dissolve the blue granules in the samples thoroughly. The culture plate was then placed on the microplate reader, and optical density (OD) was measured at 570 nm [23].

Total RNA was isolated from normal human oral keratinocytes (HOK cells) by using the Trizol reagent (Life Technologies, New York, NY, USA). Two microgram aliquots were reverse-transcribed using SuperScript II reverse transcriptase (Life Technologies) and the oligo dT primer according to the manufacturer’s instructions [22]. The human SHP2 coding region (GeneBank: NM_002834) was amplified by performing PCR using the forward primer 5’-GGATCCATGACATCGCGGAGATGGTTT-3’, which introduced a BamHI site, and the reverse primer 5’- GAATTCTTCATCTGAAACTTTTCTGCTG-3’, which introduced an EcoRI site, under the following conditions: denaturing for 30 s at 94°C, annealing for 30 s at 62°C and elongation for 1 min at 72°C for 35 cycles. The full-length of SHP2 was subcloned into the constitutive mammalian expression vector pCMV Tag 2B vector (Stratagene, La Jolla, CA, USA). The SHP2C459S (SHP2C/S) mutant was generated using the QuikChange Lighting Site-Directed Mutagenesis kit (Agilent Technologies, Inc., Wilmington, USA). The HSC3 cells were transfected with the pCMV Tag 2B-SHP2 wild type (WT) or the SHP2C/S mutant and empty vector by using a lipofectamine reagent (Life Technologies), according to the manufacturer’s protocol, and then subjected to invasion, metastasis assays and western blot analysis. The pEGFP-SHP2 WT and C/S mutant were engineered by inserting a coding region into the SalI and BamHI sites of pEGFP vector (Stratagene). The HSC3 cells were transfected with the pEGFP-SHP2 WT or the SHP2 C/S mutant and empty vector, and harvested for use in the immunoprecipitation assay.

The HSC3 cells were transfected at 50% confluence with SHP2 siRNA or a scrambled control (Invitrogen Stealth™ RNAi Negative Control LOGC, Life Technologies), Lipofetamine RNAimax (Life Technologies) and Optimen I (Life Technologies) according to the manufacturer's instructions [24]. The RNAi sequences for human SHP2 are listed as follows: SHP2#1, sense: 5’-UAA AUCGGUACUGUGCUUCUGUCUG-3’, antisense: 5’-CAGACAGAAGCACAG ACCGAUUUA-3’; SHP2#2, sense: 5’-AAUAUUUGUAUAUUCGUGCCCUUU C-3’, antisense: 5’- GAA AGG GCACGAAUAUACAAAUAUU-3’. The target sequence for si-RNA is within the SHP2 coding region.

SHP2 activity was analyzed using a Human Active SHP-2 kit (R&D Systems Inc., Minneapolis, MN, USA). Briefly, cells were lysed in a lysis buffer ([50 mM HEPES, 0.1 mM EGTA, 0.1 mM EDTA, 120 mM NaCl, 0.5-% Nonidet-P40 [NP-40], pH 7.5 supplemented with fresh protease-inhibitor-mixture tablets (Roche Applied Science). The SHP2 proteins were then immunoprecipitated using active SHP2 immunoprecipitation beads (R&D Systems Inc.), and washed 3 times in the lysis buffer and 4 times in a phosphatase assay buffer (10 mM HEPES, 0.1 mM EGTA, 0.1 mM EDTA, 0.5-% BSA, 1 mM dithiothreitol [DTT], pH 7.5). The phosphatase reaction was initiated by incubating the immunocomplexes for 30 min at 37°C in the presence of tyrosine phosphatase substrate I, DADEY (PO3) LIPQQG, according to the manufacturer's instructions. Phosphatase activity was determined using a microplate reader (SpectraMax 190 Absorbance Microplate Reader; Molecular Devices) at 620 nm.

The HSC3 cells were lysed in a RIPA buffer (50 mM Tris–HCl, pH 7.8; 150 mM NaCl; 5 mM EDTA; 5 μL/mL of Triton X-100; 5 μL/mL of NP-40; 1 μL/mL of sodium deoxycholate) and subjected to western blot analysis with the indicated antibodies. The bands were detected and revealed by applying enhanced chemiluminescence (ECL) using ECL western blotting detection reagents and exposed to X-ray film (GE Healthcare, Little Chalfont, Buckinghamshire, UK). Western blot images were captured using an AlphaImager Mini System (Alpha Innotech, Corp., San Leangro, CA, USA) [22]. Detailed antibodies and reagents were described in the Additional file 1.

The HSC3 cells were transfected with the pEGFP-SHP2 or the C/S mutant and treated with a lysis buffer (50 mM KP [pH 7.5], 100 mM KCl, 1 mM MgCl₂, 10-% glycerol, 0.2-% NP-40, 1 mM EGTA, 1 mM NaF, 1 mM sodium pyrophosphate) supplemented with 1 mM DTT, 0.1 mM PMSF, 1 mM sodium orthovanadate and protease inhibitor cocktail tablets (Roche Applied Science). Cell lysates were mixed with an antiserum against Flag, GFP and the immunocomplexes were collected on protein A/G-Sepharose beads (Amersham Pharmacia Biotec) [25]. Western blotting of proteins was performed as described previously.

The migration and invasion of oral cancer cells were assessed using Falcon Cell Culture Inserts with or without a Matrigel coating (BD Biosciences, CA, USA). Briefly, cells (5 × 10⁴) were harvested, re-suspended in a serum-free medium with 0.1-% BSA (Sigma-Aldrich, Inc., St. Louis, MO, USA), and then plated in a transwell chamber. The chamber was incubated for 18 h with a complete culture medium added to the lower chamber. Cells migrating to the lower chamber were stained with crystal violet. Photomicrographs of 3–5 regions were captured from duplicated chambers and the numbers of cells were counted [26].

The HSC3 cells grown on glass coverslips were fixed with 4-% paraformaldehyde for 10 min, permeabilized with 0.5-% Triton X-100 for 10 min, and blocked with 10-% BSA for 1 h. The cells were then incubated with a rabbit anti-E-cadherin antibody (1:200) for 1 h, before being incubated with FITC-conjugated anti-rabbit immunoglobulin (1:200; Life Technologies) for 30 min. Fluorescence images were captured using a Leica TCS SP5 confocal microscope [27].

Male CB17/SCID mice (aged 4–6 weeks; 20–25 g) were obtained from BioLASCO Taiwan Co., Ltd and maintained under specific pathogen-free conditions. All experiments were approved by the Animal Care and Use Committee at the National Health Research Institutes, Taiwan (NHRI-IACUC-101117-A). HSC3 cells (1 × 10⁵) were suspended in 100 μM phosphate-buffered saline and injected into the tail vein of mice (4 in each group), before being received control si-RNA (Invitrogen Stealth™ RNAi Negative Control) or SHP2 siRNA (10 μL/g body weight) mixed with the Invivofectamine transfection reagent (Life Technologies) through tail vein injection (100 μL) every 7 d for the next 5 wks. The mice were sacrificed 5 weeks after the injection of HSC3 cells [28‐30]. The entire lung was removed, fixed, embedded in paraffin and then sectioned for hematoxylin and eosin (H&E) staining. Tissue images were captured using a Zeiss Mirax Scan 150 microscope (Carl-Zeiss, Oberkochen, Germany). SHP2 siRNA, sense: 5’-UAA AUCGGUACUGUGCUUCUGUCUG-3’, antisense: 5’-CAGACAGAAGCACAG ACCGAUUUA-3’.

The cytoplasmic and nuclear protein fractions of HSC3 cells were extracted using a NE-PER* Nuclear and Cytoplasmic Extraction Kit (Thermo Fisher Scientific, Yokohama, Japan) according to the manufacturer's instructions [31]. Briefly, cells were harvested in cytosol fractionation buffer supplemented with fresh phosSTOP Phosphatase and Protease Inhibitor Cocktail Tablets (Roche Applied Science) and incubated on ice for 10 min before being centrifuged at 16 000 × g for 10 min. The precipitated pellet was solubilized with a nuclear fractionation buffer and then centrifuged at 16000 × g for 10 min.

A MMP-2 ELISA Kit (EMD Millipore, Inc., Darmstadt, Germany) was used to detect MMP-2 secretion. Briefly, conditioned medium were collected and subjected to an immobilized capture antibody specific for MMP-2. After unbound material was washed away, a synthetic substrate was added to measure absorbance using a spectrophotometric plate reader according to the manufacturer's instructions.

All data were analyzed using the Student’s t test and are presented as the mean ± SD. Difference were considered to be statistically significant at *P < 0.05.

To assess the potential role of SHP2 in oral tumorigenesis, we evaluated SHP2 expression in human oral tumors, and paired and histologically normal oral mucosa adjacent to the tumors. We subjected 2 type tissue samples to IHC staining for SHP2 and observed a significantly higher SHP2 in tumor cells than in histologically normal oral mucosa adjacent to the tumors (Figure 1A). Real-time quantitative RT-PCR analysis supported these results and indicated significantly higher levels of the SHP2 transcript in tumor tissue than in histologically normal oral mucosa adjacent to the tumors (Figure 1B).

To investigate the biological functions of SHP2 in oral tumorigenesis, we isolated highly invasive clones from oral cancer cells by using an in vitro invasion assay. We used 4–8 cycles of HSC3 cells, which have modest migratory and invasive ability among oral cancer cell lines (data not shown), to derive the highly invasive clones, HSC3-Inv4 and HSC3-Inv8. The growth of these clones was the same as that of the parental cells (Figure 1C), but the number of HSC3-Inv4 cells that migrated through the filter was significantly higher than the number of parental cells that migrated through the filter (Figure 1D). We observed significantly upregulated SHP2 expressions in the HSC3-Inv4 and HSC3-Inv8 clones in comparison with the parental cells (Figure 1E). We observed no significant difference in the levels of the SHP1 transcript in the clones and parental cells (Additional file 2: Figure S1). SHP1 is a high homolog of SHP2. Therefore, these results suggested that SHP2 may exclusively be responsible for the migration and invasion of oral cancer cells.

To determine whether SHP2 is involved in regulating oral cancer migration and invasion, we knocked down SHP2 by using specific si-RNA. As expected, when we downregulated SHP2 expression, the oral cancer cells exhibited markedly reduced migratory and invasive ability (Figure 2A). We observed similar effects on the invasive ability of the HSC3-Inv4 and HSC3-Inv8 cells (Figure 2B). Collectively, our results indicated that SHP2 plays a crucial role in migration and invasion in oral cancer cells.Considering the crucial role of SHP2 activity in various cellular functions, we then investigated whether SHP2 activity is required for migration and invasion of oral cancer cells. We generated a flag-tagged SHP2 WT or phosphatase-dead SHP2 C459S mutant in HSC3 cells. When we analyzed the cell migration or invasion, we observed that the SHP2 mutant abrogated cell migration and invasion elicited by the SHP2 WT (Figure 2C). Overall, these data indicated that the catalytic activity of SHP2 is required for the migration and invasion of oral cancer cells.

As shown in Figure 3A, we evaluated the changes in EMT-associated E-cadherin and vimentin in highly invasive oral cancer cells. Our results indicated that the majority of the parental HSC3 cells were polygonal in shape (Figure 3A, left upper panel); whereas, the HSC3-Inv4 cells were rather spindle shaped (Figure 3A, right upper panel), with downregulated of E-cadherin protein and upregulated of vimentin protein (Figure 3B). When we evaluated the levels of the transcripts of EMT regulators Snail/Twist1, we observed significant upregulation of Snail/Twist1 mRNA expression levels in the highly invasive clones generated from the HSC3 cells (Figure 3C).We then tested the medium from the highly invasive clones to evaluate the secretion of MMP-2. As shown in Figure 3D, increased MMP-2 secretion from oral cancer cells significantly correlated with increased cell invasion. While we analyzed the medium from SHP2-depleted cells, we observed significantly reduced MMP-2 (Figure 3E). Collectively, these results suggested that SHP2 exerts its function in several critical stages that contribute to the acquirement of invasiveness during oral cancer metastasis.

To identify the potential biochemical pathways that rely on SHP2 activity, we analyzed total tyrosine phosphorylation in SHP2 WT- and C459S mutant-expressing cells. As shown in Additional file 3: Figure S2, we observed increased protein phosphorylation in mutant-expressing cells, particularly those migrating around 40–50 kD on the gel, compared with SHP2 WT-expressing cells. We thus hypothesized that p44/42 (ERK1/2) signaling might trigger nuclear events because the phosphorylation of ERK1/2 leads to its translocation to the nucleus, which is required for the induction of several cellular responses. By immunoprecipitating exogenously expressed EGFP-tagged SHP2 and immunoblotting using anti-ERK1/2 as a probe, we identified an association between ERK1/2 and SHP2 in cells expressing SHP2 WT and mutant (Figure 4A). We observed markedly increased ERK1/2 phosphorylation in phosphatase-dead cells (Figure 4A), indicating that SHP2 catalytic activity plays a major role in the regulation of ERK1/2 activity, but is not required for the assembly of the ERK1/2/SHP2 complex.Considering the hypothesis that increased ERK1/2 phosphorylation leads to its accumulation in the nucleus (Figure 4B), we then investigated whether Snail and Twist1 are possible downstream effectors of ERK1/2 signaling. In the presence of a selective ERK1/2 inhibitor, FR180204, we observed a dose-dependent reduction at the transcript levels of Snail/Twist1 in oral cancer cells (Figure 4C). However, in the absence of SHP2 expression, we observed increased transcript levels of Snail/Twist1 (Figure 4D), as well as increased ERK1/2 phosphorylation (Figure 4E). These results supported that SHP2 modulates Snail/Twist1 at a transcript level by negatively regulating ERK1/2 activity.

We evaluated the effects of SHP2 attention on the metastasis of oral cancer cells toward the lung to establish the potential for developing SHP2 as a target for human oral cancer treatment. As shown in Figure 5, we analyzed the lungs of mice with HSC3 xenografts and SHP2 si-RNA administered through tail vein injection by using H&E staining. Analysis of lung tissue sections indicated that HSC3 tumors with SHP2 knockdown exhibited an approximate 70% reduction in metastatic capacity, compared with those with control si-RNA (Figure 5, lower panel). Overall, the result supported that SHP2 inhibits the migration, invasion, and metastasis of oral cancer cells, and indicated that SHP2 is a potential target for oral cancer treatment.