Matrix metalloproteinase 12 is induced by heterogeneous nuclear ribonucleoprotein K and promotes migration and invasion in nasopharyngeal carcinoma

The NPC-derived cell line, TW02 [38], derived from a keratinizing squamous cell carcinoma [World Health Organization (WHO) type I], was provided by Dr. C. T. Lin (National Taiwan University, Taiwan). The NPC-derived cell line, HK1 [39], derived from a keratinizing squamous cell carcinoma (WHO type I), was provided by Dr. S. W. Tsao (Hong Kong University, SAR, China). NPC-TW02 and NPC-HK1 cells were culture in Dulbecco’s modified Eagle’s medium (DMEM, Gibco BRL, Rockville, MD) and RPMI1640 (Gibco BRL), respectively. All NPC cell lines were supplemented with 10% fetal calf serum (FCS), 100 U/ml penicillin, and 100 μg/ml streptomycin at 37°C under 5% CO₂.

RNA samples from hnRNP K-knockdown NPC-TW02 cells, control NPC-TW02 cells, nine individual NPC tissues and one pool of the corresponding adjacent non-tumor tissues, were isolated using the TRIzol reagent (Invitrogen, Carlsbad, CA). The biotinylated oligonucleotide was hybridized to the Human Genome U133 Plus 2.0 Array (Affymetrix, Santa Clara, CA) by the National Yang-Ming University Genomics Center (Taiwan), following the manufacturer’s instructions. Microarray data were analyzed using the GeneSpring software version GX 7.31 (Silicon Genetics, Redwood, CA). The ratio of the average hybridization intensity between hnRNP K-knockdown/control NPC-TW02 cells or NPC tumor/normal tissue was taken as the relative gene expression level.

RNA samples from NPC-TW02, and -HK1 cells and NPC tissues were isolated using the TRIzol reagent (Invitrogen). Reverse transcription of RNA (1 μg) was performed using oligo(dT)₂₀ primers (Invitrogen) and Moloney Murine Leukemia Virus (M-MLV) Reverse Transcriptase (Invitrogen) according to the manufacturer’s instructions. The primers used to amplify the cDNA corresponding to MMP1, MMP12, MMP13, MMP28 and GAPDH are presented in Additional file 1: Table S1. Quantitative RT-PCR was performed on a Light-Cycler (Roche Diagnostics, Mannheim, Germany), using the FastStart DNA Master SYBR Green I reagent (Roche Diagnostics). The gene expression results were normalized with regard to the expression of the GADPH. For mRNA half-life assessment, actinomycin D (5 μg/ml) was added 48 hours after cells were transfected with control or hnRNP K-targeting siRNA, and RNA was prepared at the indicated times.

The retrospective cohort comprised 82 NPC patients who had been admitted to Chang Gung Memorial Hospital (Lin-Kou, Taiwan) from 1990 to 1998. Clinical stage was defined according to the 2002 cancer staging system revised by the American Joint Committee on Cancer. The study population included 17 stage-I-II and 65 stage-III-IV patients comprising 61 men and 21 women ranging from 22 to 78 years of age (median age 44). Histological typing was done according to the WHO classification criteria, as previously described [40]. This study was reviewed and approved by the institutional review board and ethics committee of Chang Gung Memorial Hospital. Informed consent was obtained from all patients.

Immunohistochemical analyses were performed as described previously [5, 7, 37, 40], using an automatic IHC-staining device (Bond-max Automated Immunostainer; Vision Biosystems, Melbourne, Australia), according to the manufacturer’s instructions. Tissue sections were retrieved using Bond Epitope Retrieval Solution 1 (Vision BioSystems) and stained with antibodies against hnRNP K (mouse monoclonal antibody, 1:300 dilution; Santa Cruz Biotechnology, Santa Cruz, CA, USA) and MMP-12 (goat polyclonal antibody, 1:10 dilution; Santa Cruz Biotechnology). A polymer detection system (Bond Polymer Refine; Vision BioSystems) was used to reduce nonspecific staining. Tissue sections were treated with liquid DAB reagent; 3’-diaminobenzidine tetrahydrochloride was used as the chromogen, and hematoxylin was used as the counterstaining reagent. For analysis of total hnRNP K expression, specimens in which > 50% of the tumor cells displayed strong staining were defined as having ‘high-level total hnRNP K’ expression, and those in which ≤ 50% of tumor cells showed strong stained were defined as having ‘low-level total hnRNP K’ expression. For analysis of cytoplasmic hnRNP K, we used the method described previously [5], a sample was defined as ‘cytoplasmic positive’ in cases where >10% of the tumor cells exhibited cytoplasmic staining and as ‘cytoplasmic negative’ where ≤10% of cells were stained. For analysis of nuclear hnRNP K expression, specimens in which >50% of tumor cells displayed strong staining were defined as ‘high level of nuclear hnRNP K’ and those where ≤50% of tumor cells stained strongly were defined as ‘low level of nuclear hnRNP K.’ For analysis of MMP-12 expression, specimens in which > 20% of tumor cells displayed positive staining were defined as having ‘high-level’ MMP-12 expression, and those in which ≤ 20% tumor cells displayed positive staining were defined as having ‘low-level’ MMP-12 expression. MMP-12- and hnRNP K-positive tumor cells in representative microscopic fields were scored independently by two experienced pathologists.

NPC cells treated with hnRNP K-targeting siRNA were cultured in serum-free medium for 48 h, and the conditioned medium was harvested and concentrated 20-fold using an Amicon Ultra-4 centrifugal filter (Millipore). The protein concentration was quantified using the Bradford reagent and protein was mixed with non-reducing sample buffer. The protein mixture was heated at 37°C for 30 min and separated by electrophoresis on an SDS-polyacrylamide gel containing 1 mg/ml α-casein. The gel was washed twice with 2.5% Triton X-100 for 30 min at room temperature (RT), and incubated in developing buffer (50 mM Tris–HCl, pH 7.6, 15 mM NaCl, 10 mM CaCl₂ and 0.02% NaN3) for 15 min at RT with gentle agitation. The gel was then transferred to fresh developing buffer and incubated at 37°C for 48 h, and then incubated in fixing buffer for 15 min at RT with gentle agitation. The gel was stained with 0.125% Coomassie blue at RT for 1 hr and destained with fixing buffer; the solution was changed every 15 min until caseinolytic bands were visible. The caseinolytic band found at 54 kDa was subjected to zymographic measurement of MMP12 activity.

The promoter sequences of human MMP12 were obtained from the UCSC genome browser. Using human genomic DNA isolated from normal peripheral blood mononuclear cells (PBMCs) as the template, we obtained the MMP12 promoter −2000 (−2000 to +97) fragment by PCR using the following primers: forward, 5'-TCCCC CGGGA CATAG AAAAA TTATC TAGTC CTACG TGTA-3' and reverse, 5'-CCGCT CGAGT GTAAA CTTCT AAACG GATCA ATTCA GTTT-3' (including SmaI and XhoI sites, respectively). The resulting PCR product was ligated into the SmaI and XhoI sites of the pGL3-basic vector (Promega, Madison, WI). To generate 5' serial deletions of the MMP 12 promoter, fragments −1500 (−1500 to +97), −999 (−999 to +97), −499 (−499 to +97), −399 (−399 to +97), −299 (−299 to +97), −199 (−199 to +97), −99 (−99 to +97), −42 (−42 to +97), −32 (−32 to +97) or +2 (+2 to +97) were amplified from pGL3-MMP12 -2000 and ligated into the SmaI/XhoI-treated pGL3-basic vector.

NPC-TW02 cells in 24-well plates were co-transfected with 0.4 ng of pRL-TK (Promega) and 0.8 μg of pGL3-basic vector with or without MMP12 promoter fragments, using Lipofectamine (Invitrogen) according to the manufacturer’s instructions. After 24 hours, Firefly and Renilla luciferase activities were measured using the Dual-Glo Luciferase Assay System (Promega) according to the manufacturer's instructions. Firefly luciferase activities were normalized to Renilla activities. Each bar represents an average of at least three independent experiments, and the error bars show standard deviations calculated using Microsoft Office Excel.

Probes corresponding to the potential binding elements within the MMP12 promoter were generated by PCR using the appropriate biotinylated primers, as follows: −42/+97 (−42 to +97) forward, 5'-GGGAT GATAT CAACT ATGAG TCACT CATAG G-3' and reverse, 5'-TGTAA ACTTC TAAAC GGATC AATTC AG-3'; and +2/+97 (+2 to +97) forward, 5'-AGAAC CCGGA CTAAG GGC-3' and reverse, 5'-TGTAA ACTTC TAAAC GGATC AATTC AG-3'. The biotinylated probes were conjugated with M-280 Streptavidin Dynabeads (Invitrogen) in binding buffer (10 mM Tris–HCl, pH 7.5, 50 mM KCl, 1 mM MgCl₂, 1 mM EDTA, pH 8.0, 1 mM Na₃VO₄, 5 mM DTT, 5% glycerol, 0.3% NP-40) for 40 min at room temperature. NPC-TW02 cells were extracted using the Compartmental Protein Extraction Reagent (Millipore), and nuclear fractions (50 μg) were incubated with unconjugated Dynabeads in the presence of 25 μg/ml poly (dI:dC) for 20 min at RT. The unbound fraction was incubated with 250 μg of Dynabeads bound to 50 pmol of immobilized probe for 1 h at RT. The Dynabead-bound complexes were collected using a Dynal MPC-S magnetic particle concentrator (Dynal Biotech, Oslo, Norway) and washed with binding buffer. The DNA-bound proteins were eluted in SDS sample buffer and assayed by Western blotting.

ChIP assays were performed using a Magna ChIP™ Kit (Millipore) according to the manufacturer’s protocol, with modifications. Briefly, NPC-TW02 cells (1×10⁷) were cross-linked by treatment with 1% formaldehyde for 10 min on ice, and the reactions were quenched with glycine (0.125 M) for 10 min on ice. Fixed cells were washed twice with ice-cold PBS and lysed for 15 min on ice with the provided cell lysis buffer and protease inhibitors. The samples were then centrifuged at 800 x g for 5 min at 4°C, the supernatants were removed, and the pellets were resuspend with the provided nuclear lysis buffer and protease inhibitors. Chromatin was sheared by sonication (eight times for 10 seconds each at 30% output) on ice and centrifuged at 10,000 × g for 10 min at 4°C. The supernatant was collected and diluted 10-fold with ChIP dilution buffer (0.01% SDS, 1.1% Triton X-100, 1.2 mM EDTA, 16.7 mM Tris–HCl, pH 8.1, 167 mM NaCl) containing protease inhibitors. The diluted samples were incubated overnight at 4°C with 4 μg of an anti-hnRNP K antibody (Invitrogen) and magnetic protein A/G beads (Millipore). Mouse IgG (Millipore) was used as a control antibody. The immunocomplexes were collected using a Dynal MPC-S magnetic particle concentrator (Dynal Biotech) and washed once each in low-salt buffer, high-salt buffer, LiCl buffer, and Tris-EDTA buffer. The samples were resuspended in ChIP elution buffer (200 mM NaCl, 1% SDS, 20 mM Tris–HCl, pH 7.4) containing 100 μg/ml proteinase K, incubated for 2 h at 62°C, and then incubated for 10 min at 95°C. The DNA fragments were further purified using a QIAquick PCR Purification Kit (QIAGEN, Hilden, Germany), and quantitative PCR was performed using primers against the potential hnRNP K-binding elements in the MMP12 promoter.

The negative control shRNA (lacZ) and two shRNAs targeting different sequences of human MMP12 (KD1, clone ID TRCN0000050209; and KD2, clone ID TRCN0000372998) in the pLKO.1-puro vector backbone were purchased from the National RNAi Core Facility of Academia Sinica (Taipei, Taiwan). For lentiviral production, 293 T cells were seeded at 4 × 10⁵/well in 6 well-plates and transfected with 1.8 μg pCMV-Δ8.91, 0.2 μg pMD.G and 2 μg lentiviral vector. Six hours after transfection, the culture medium was change to DMEM supplemented with 1% FCS. Supernatants were collected at 24 and 48 h after transfection, pooled, filtered through a 0.22-μm filter, and frozen at −80°C until use. For lentiviral transduction, NPC-TW02 cells were seeded at 2 × 10⁵/well in 6-well plates and infected with lentivirus in the presence of 8 μg/mL of polybrene. The transduced cells were selected with 1 μg/ml puromycin for 2–3 weeks.

Equal numbers of MMP12-knockdown cell clones were dispensed to 6-well plates, and total cell numbers were counted on days 1, 2, 3 and 4 after plating. The results are presented as the mean ± SD from four independent experiments.

The migration and invasion of NPC cells were evaluated using Transwell inserts (Corning Inc., Corning, NY, USA) and Biocoat Matrigel invasion chambers (BD Biosciences, Bedford, MA), respectively. For cell migration assays, the cells were washed twice with serum-free medium and resuspended in serum-free medium, and 1.8 × 10⁵ cells in 0.1 ml were added to the upper chamber of the apparatus. The lower chamber contained 0.6 ml medium with 10% FBS. For cell invasion assays, the same procedures were used, except that 2.5 × 10⁵ cells were resuspended in 0.5 ml of serum-free medium, and added to the upper chamber of the apparatus, while the lower chamber contained 0.75 ml medium with 10% FBS. After 24 h at 37°C, the migrated and invading cells were fixed and stained for 20 min with 0.25% crystal violet, 10% formaldehyde and 80% methanol, and the filters were washed five times with ddH₂O to remove non-adherent cells. Ten to fifteen random fields (× 100 magnification) were captured for each membrane. The migrated or invading cells were counted and averages were calculated; results were obtained from three independent experiments. The relative fold-change in the number of migrated or invasive cells is shown, with the results from control cells given as 1.0. The effect of MMP12-specific inhibitor PF-356231 (Enzo Life Science, Farmingdale, NY) [41, 42] on the migration of NPC cells was determined after culturing for 24 h in the presence of indicated concentrations of inhibitor or DMSO (as vehicle control). The invasive activities of NPC cells were determined after 24 h (NPC-TW02) or 36 h (NPC-HK1) of treatment with inhibitor.

All statistical analyses were performed using the SPSS 13.0 statistical software package. The relationship between MMP12 expression and hnRNP K expression was evaluated using the Pearson Chi-Square test. In vitro data were analyzed with the Student’s t test. Differences were considered significant at a level of P < 0.05.

We previously showed that hnRNP K contributes to the metastasis of NPC cells in part by regulating downstream genes [7]. Since the MMP family proteins are well known to be involved in tumor metastasis, we tested if they could be regulated through hnRNP K. We used Affymetrix cDNA microarrays to compare the expression profiles of MMP family genes in NPC-TW02 cells transiently transfected with hnRNP K-targeting siRNA versus those transfected with negative control siRNA, and in NPC tissue samples and adjacent normal tissues. The 7 out of 23 MMP genes showed reduced expression (1.5-fold decrease) in hnRNP K-knockdown cells (Additional file 2: Table S2), while 11 out of 23 were elevated (1.5-fold increase) in NPC tissues (Additional file 3: Table S3). Among these differentially expressed genes, MMP1, MMP12, MMP13 and MMP28 were consistently reduced in hnRNP K-knockdown cells (Figure 1A) but elevated in tumor cells (Figure 1C). We further confirmed our microarray results using quantitative RT-PCR, and found that the mRNA levels of MMP1, MMP12, MMP13 and MMP28 were significantly reduced (to 0.77-, 0.28-, 0.46- and 0.09-fold, respectively; P < 0.05) in hnRNP K-knockdown cells compared with control siRNA-treated NPC-TW02 cells (Figure 1B). On the other hand, the mRNA levels of MMP1 and MMP12 were significantly elevated (14.2- and 56.3-fold, respectively; P < 0.05) in nine matched-pairs of NPC tumor and adjacent normal tissues. NPC tumor samples compared with adjacent normal tissues, whereas the mRNA levels of MMP13 and MMP28 were not significantly different between the tumor and adjacent normal tissues (Figure 1D). As MMP12 has not previously been examined in the context of NPC, it was chosen for further study.

The epithelial-stromal cell cross contamination is known to be one of problems in the analysis of RNA/protein expression from solid tumor. Therefore, 82 NPC biopsy specimens were subjected to immunohistochemical (IHC) analysis and the differential expression of MMP12 and hnRNP K between the tumor and normal epithelial tissues were investigated. Patient characteristics and clinical features are summarized in Table 1. In general, our IHC data demonstrated that the NPC tumor cells expressed higher levels of MMP12 compared to adjacent normal cells. As shown in Figure 2A-C, consecutive tissue slides of the same set of specimens were used to evaluate the protein expression levels of MMP12 and hnRNP K. We further analyzed whether the expression level of MMP12 correlated with the subcellular localization (cytoplasmic or nuclear) of hnRNP K in NPC cells. We assessed the association between MMP12 expression (high or low) and (1) the total hnRNP K expression (high or low), or (2) the nuclear hnRNP K expression (high or low), or (3) the cytoplasmic hnRNP K expression (positive or negative). The statistical analysis was summarized in Table 2. Statistical analyses revealed that high-level MMP12 expression was significantly correlated with high-level of total hnRNP K (P = 0.026; Table 2) and nuclear hnRNP K (P = 0.042; Table 2), rather than cytoplasmic hnRNP K (P = 0.168; Table 2). These results suggest that nuclear hnRNP K was positively correlated with MMP12 in NPC tumor cells.

To gain insight into the potential role of hnRNP K in regulating MMP12 expression, we tested MMP12 expression in hnRNP K-knockdown and control cells of two NPC cell lines (NPC-TW02 and -HK1). As shown in Figure 3A, the level of MMP12 mRNA was reduced significantly in hnRNP K siRNA-treated NPC cells compared with control siRNA-treated cells. To assess whether the effect of hnRNP K knockdown on MMP-12 mRNA was correlated with changes in the protein and/or enzymatic levels, we performed Western blot and zymographic analyses. Conditioned media were obtained from the above-described NPC cells and analyzed by immunoblotting and casein zymography. Western blot analysis and casein zymography showed that hnRNP K knockdown significantly reduced the protein expression and activity levels of MMP12 (54 kDa), respectively (Figure 3A and B). Thus, hnRNP K knockdown inhibited the mRNA expression, protein expression and enzymatic activity of MMP12.

We further clarified the mechanism(s) underlying the hnRNP K-mediated regulation of MMP12 expression. To discriminate between transcriptional activation and post-transcriptional regulation, we analyzed the effect of hnRNP K knockdown on MMP12 promoter activity and mRNA stability. As shown in Figure 4A and B, NPC-TW02 cells were treated with siRNA followed by transfection of constructs containing 5' serial deletions of the MMP12 promoter, and reporter activity was examined 24 h later. Our results revealed that knockdown of hnRNP K significantly inhibited the activity of MMP12 promoter constructs containing the deletion from −2000 to −42 bp of the transcription start site. (P < 0.05). There had no effect on MMP12 promoter (−32 to +97) while cells treated with hnRNP K siRNA compared with control group (Figure 4B). Moreover, the MMP12 promoter construct spanning −32 to +97 showed substantially less activity compared with that spanning −42 to +97 (61.3% vs. 100%, respectively) (Figure 4B). These results collectively suggest that the MMP12 promoter region covering −42 to −33 may be the potential hnRNP K response region. To further verify the binding of hnRNP K to the MMP12 promoter, we performed in vitro DNA pull-down assays with probes spanning −42 to +97 and +2 to +97 of the MMP12 promoter. As shown in Figure 4C, hnRNP K specifically bound to probe (−42 to +97) but not probe (+2 to +97), suggesting that the −42 to +1 region is indispensable for hnRNP K binding. To further support our contention that hnRNP K can interact with the endogenous MMP12 promoter, we performed a chromatin-immunoprecipitation analysis. As shown in Figure 4D, hnRNP K specifically immunoprecipitated with the MMP12 promoter. Together, these results indicated that the hnRNP K-responsive region is the sequence of −42 to −33 bp upstream of the MMP12 transcription start site.

In addition, we examined the effect of hnRNP K knockdown on MMP12 mRNA stability. Treatment of NPC-TW02 cells with actinomycin D to block de novo RNA synthesis, and used quantitative RT-PCR to examine MMP12 mRNA levels at 2, 4, 8, 12 and 16 h post-treatment. The half-life of the MMP12 mRNA was 31.07 h in hnRNP K-knockdown cells and 38.17 h in control cells, which was not significantly different (Additional file 4: Figure S1). Taken together, our findings indicate that the hnRNP K-mediated changes in MMP12 gene expression arise via promoter inhibition, not mRNA destabilization.

To examine the biological function of MMP12 in NPC cells, we established two MMP12-knockdown cell lines using lentiviral transduction of two different MMP12-targeting shRNA sequences. As shown in Figure 5A, the MMP12 protein and mRNA levels were reduced in the two MMP12-knockdown cell lines compared to control cells transduced with a control shRNA targeting LacZ. Importantly, cell migration (Figure 5B) and invasion (Figure 5C) were significantly and dose-dependently reduced in the MMP12-knockdown cells compared to controls (P < 0.05). However, the reduction of migration and invasion in MMP12-knockdown cells were not due to the difference in cell growth (Figure 5D) between MMP12-knockdown and control cells. We further investigated the effect of the treatment of PF-356231, a specific inhibitor of MMP12 on the migration and invasion of NPC cells. As compared to untreated control, PF-356231 treatment significantly and dose-dependently reduced the migration (Figure 6A) and invasion (Figure 6B) in NPC-TW02 cells (P < 0.05). Similar results were observed in NPC-HK1 cells (Figure 6C and D). Taken together, these results indicate that hnRNP K-mediated MMP12 expression enhances the migration and invasion of NPC cells. In addition, MMP12-mediated cell migration and invasion can be inhibited by PF-356231 treatment.