Suppression of NF-κB activity by NDRG2 expression attenuates the invasive potential of highly malignant tumor cells

Abstract

Downregulation of the N-myc downstream-regulated gene 2 (NDRG2) gene is involved in the progression of aggressive forms of cancer, along with the poor prognosis of cancer patients. In the current study, we examined the effect of NDRG2 expression on the metastatic potential of HT1080 human fibrosarcoma and B16F10 murine melanoma cells in both in vitro and in vivo systems. In gelatin zymography, NDRG2 expression remarkably suppressed the matrix metalloproteinase (MMP)-9 activity and slightly inhibited MMP-2 activity of both cell lines. Tumor migration and invasion in vitro were significantly reduced by NDRG2 expression, and NDRG2 inhibited tumor cell proliferation in an anchorage-independent semisolid agar assay. Specifically, we found that NDRG2 affects invasion through suppression of nuclear factor kappa B (NF-κB) activity. In animal experiments, subcutaneously injected B16F10-NDRG2 cells showed delayed tumor growth compared with B16F10-mock cells. Furthermore, severe metastasis from primary tumor mass into the draining lymph nodes was observed after injection of B16F10-mock cells, but not with B16F10-NDRG2 cells. Pulmonary metastasis after intravenous injection of B16F10 cells was also reduced by NDRG2 expression. Intra- and peritumoral angiogenesis that is critical for the tumor growth and metastasis was clearly found in tumors after injection with B16F10-mock cells, whereas it was impaired in tumors after injection with B16F10-NDRG2 cells. Collectively, our data show that NDRG2 expression significantly suppresses tumor invasion by inhibiting MMP activities, which are regulated through the NF-κB signaling. Moreover, results from animal experiments provide evidence for the regulatory role of the NDRG2 gene in metastatic tumors.

Introduction

Tumor metastasis, one of the fundamental properties of malignant cancer cells and a major cause of cancer mortality, is a multistep process including adhesion to the extracellular matrix (ECM), proteolytic digestion of the ECMs, invasion of lymph and blood vessels and migration. Matrix metalloproteinases (MMPs), especially MMP-2 and MMP-9, are key enzymes known to degrade surrounding ECM components during cancer invasion and metastasis. Several recent studies have revealed that invasive cells have higher MMP-9 expression, and its expression correlates with vascular invasion ( 1–4 ). Levels of MMP-2 have also been found to correlate with enhanced metastasis and to associate with poor prognosis and relapse in breast cancer patients ( 5 ). Tumor angiogenesis is critical for the growth, invasion and metastasis of solid tumors. Numerous studies have shown a significant relationship between the increased intratumoral microvessel density and the risk of metastasis and/or decreased survival of patients with solid tumors. In tumor angiogenesis, MMP-9 has been implicated as an important protease component ( 6 ).

N-myc downstream-regulated gene 2 (NDRG2) is the second member of the NDRG family, which is composed of four members named NDRG1, NDRG2, NDRG3 and NDRG4. Human NDRG2 shows 57% identity to NDRG1 and NDRG3 and 65% identity to NDRG4. In addition, there is 92% identity between human NDRG2 and mouse NDRG2. NDRG family members exhibit distinct expression patterns during development and adult life. While NDRG1 is a widely expressed gene, NDRG2 and NDRG3 are mainly expressed in the brain, heart, skeletal muscle and kidney. NDRG4 expression is strictly restricted to the brain and heart ( 7 ). NDRG1, which is the best-characterized gene from the NDRG family, has been reported to play roles in terminal differentiation, stress responses, atherosclerosis, hormone responses, hypoxia and tumor suppression ( 8–15 ). Expression of NDRG1 is inhibited by N-myc ( 16 ), and the protein has been shown to be at a lower level in transformed cells, whereas it is upregulated in growth-arrested differentiating cells ( 17 ). Our group reported that NDRG2 is expressed during the differentiation of dendritic cells, and the expression is differentially regulated by maturation-inducing stimuli such as lipopolysaccharide and CD40 ( 18 ). In addition, NDRG2 is found in brain lesions of Alzheimer's disease-affected patients and is thought to be associated with the progression of Alzheimer's disease ( 19 ). In recent studies, it is reported that NDRG2 expression is downregulated in a variety of carcinomas, including liver cancer, pancreatic cancer, thyroid cancer and meningioma, suggesting a possible role for NDRG2 in tumor suppression ( 20–24 ). Promoter methylation, mutation and genomic deletion are involved in decreased NDRG2 expression in cancer cell lines ( 25 ). It has been demonstrated that NDRG2 inhibits tumor cell proliferation and increases Fas-, p53- or hypoxia-mediated apoptosis and its expression is correlated with patient survival and prognosis ( 26–29 ). However, the role of NDRG2 in regulating metastasis remains unclear. Therefore, using two metastatic tumor cell lines, HT1080 human fibrosarcoma and B16F10 mouse melanoma cells, we examined the effect of NDRG2 overexpression on metastatic potential and further probed the underlying molecular mechanisms of any NDRG2 effects.

Materials and methods

Mice and cell cultures

HT1080 human fibrosarcoma cells were maintained in Dulbecco's modified Eagle's medium (DMEM; Gibco/Invitrogen, Carlsbad, CA) supplemented with 10% (vol/vol) heat-inactivated fetal bovine serum (FBS; Gibco/Invitrogen) and 100 U/ml of penicillin/100 μg/ml of streptomycin (Gibco/Invitrogen) at 37°C in a humidified 5% CO ₂ incubator. B16F10 murine melanoma cells, which are metastatic in the lungs of C57BL/6J mice, were also grown at 37°C under 5% CO ₂ in DMEM supplemented with 10% FBS and 100 U/ml of penicillin/100 μg/ml of streptomycin (Gibco/Invitrogen). For animal experiments, specific pathogen-free female C57BL/6J mice were purchased from Daehan BioLink (Umsong, Republic of Korea) and maintained in our animal facility for at least 2 weeks before use. The mice were housed under specific pathogen-free conditions in a barrier facility with 12-h light–dark cycles. Experimental procedures were performed at 6–8 weeks of age.

Antibodies and chemicals

Anti-phospho-IκB kinase (IKK) α/β (Ser176/180), anti-IκB-α and anti-phospho-IκB-α (Ser32) antibodies were purchased from Cell Signaling Technology (Danvers, MA). Antibodies against nuclear factor kappa B (NF-κB) p65 and platelet-endothelial cell adhesion molecule-1 (PECAM-1) were obtained from Santa Cruz Biotechnology (Santa Cruz, CA). Anti-MMP-2 and anti-MMP-9 antibodies were obtained from Epitomics (Burlingame, CA). Anti-NDRG2 monoclonal antibody was generated by our laboratory ( 26 ).

Overexpression of the NDRG2 gene in HT1080 fibrosarcoma and B16F10 melanoma cells

Both human and murine NDRG2 complementary DNA were cloned into the pCMV-Taq2B vector. HT1080 cells and B16F10 cells were transfected with the plasmid pCMV-Taq2B/human NDRG2 and pCMV-Taq2B/murine NDRG2, respectively, using a WelFect-EX™ PLUS Transfection Reagent (WelGENE, Daegu, Republic of Korea). Stable clones were selected using complete medium containing 1 mg/ml neomycin (G418, Geniticin, Gibco/Invitrogen), and NDRG2 expression was confirmed by reverse transcription–polymerase chain reaction (PCR) and western blotting.

RNA extraction and reverse transcription–PCR (RT–PCR)

Total RNA was extracted from the cells using TRIzol reagent (Invitrogen, Carlsbad, CA) following the manufacturer's instructions and reverse transcribed to complementary DNA using M-MLV reverse transcriptase (Promega, Madison, WI) and oligo(dT). Complementary DNA aliquots corresponding to 5 μg RNA were analyzed by semiquantitative PCR. PCR products were electrophoresed on 1% agarose gels containing ethidium bromide.

In vitro cell growth and soft agar colony formation assays

For in vitro cell growth analysis, cells were seeded into a six-well plate (5 × 10 ⁴ cells per well). After incubation for 3 days, the cells were harvested and counted using the trypan blue exclusion test. For determination of anchorage-independent cell growth, 1 × 10 ⁴ cells were suspended in 3 ml of the medium containing 0.3% agar and 10% FBS and they were applied onto the pre-solidified 0.6% agar (3 ml) in 60 mm culture dishes. Triplicate plates were prepared for each cell line. After 2 weeks of incubation, colonies on soft agar were observed under a phase-contrast microscope and photographed.

Gelatin zymography

Enzymatic activities of MMP-2 and MMP-9 were analyzed by gelatin zymography. Serum-free culture medium was collected after incubation for 24–48 h. In the case of the B16F10 melanoma cells, aliquots of the supernatants were concentrated using a Centricon (Amicon, Beverly, MA), and equivalent volumes of the culture supernatant fractions were electrophoresed on a 10% sodium dodecyl sulfate–polyacrylamide gel containing 0.1–0.2% gelatin. Gels were washed twice with washing buffer (50 mM Tris–HCl, pH 7.5, 100 mM NaCl, 2.5% Triton X-100). Gels were then treated with incubation buffer (50 mM Tris–HCl, pH 7.5, 150 mM NaCl, 10 mM CaCl ₂ , 0.02% NaN ₃ , 1 μM ZnCl ₂ ) at 37°C for 18–36 h, stained (0.05% Coomassie blue, 10% isopropanol, 10% acetic acid) and destained (10% isopropanol, 10% acetic acid). MMPs were detected as transparent bands on the blue background of Coomassie blue-stained slab gels.

In vitro tumor cell migration and invasion assay

The in vitro migration and invasion assay was performed using 24-well transwell unit with polycarbonate filters that have a diameter of 6.5 mm and a pore size of 8.0 μm (Corning Costar, Cambridge, MA). For the invasion assay, after filling the lower part of the transwell with DMEM medium plus 5% FBS as a chemoattractant, the transwell was coated with 20 μl of a 1:2 mixture of Matrigel:DMEM (Matrigel; BD Biosciences, Bedford, MA), which was assembled as the intervening invasive barrier on a 24-well plate. Cells (5 × 104) were suspended in serum-free DMEM medium, added to the upper part of the transwell and incubated for 12–20 h at 37°C. Cells attached to the upper surface of the filter were completely removed by wiping with a cotton swab, and filters were stained with a 0.2% crystal violet/20% methanol (wt/vol) solution. Alternatively, the cell number was determined using calcein-acetomethylester (AM, Molecular Probes®, Eugene, OR), which can be cleaved by an intracellular esterase to produce calcein, a strong fluorescent compound. Fluorescent intensity was measured with a fluorescence microplate reader set (PerkinElmer Victor™, Fremont, CA). Background fluorescence was corrected by subtracting the value derived from the non-cell control. The migration assay was performed using a transwell culture system without any ECM coating.

Transient transfection and luciferase reporter assay

In order to perform the luciferase reporter assay, semi-confluent cells grown in 12-well plates were transiently transfected with a luciferase reporter plasmid and the pCMV/β-Gal plasmid using Lipofectamine 2000 (Invitrogen, San Diego, CA). pCMV/β-Gal was used to control the variability in transfection efficiency. After incubation for 18 h, cells were harvested in passive lysis buffer, and the luciferase activities were measured on a VICTOR ³ 1420 multilabel counter with the luminescence microplate reader set (PerkinElmer Victor™) using the Luciferase Assay System according to the manufacturer's instructions (Promega). The β-galactosidase activities were measured using o -nitrophenyl β-galactopyranoside as a substrate. Luciferase activity was calculated relative to the β-galactosidase activity.

Protein preparation

For the preparation of whole-cell lysates, cells were lysed on ice in M-PER Mammalian Protein Extraction Reagent (Pierce Biotechnology, Rockford, IL) for 30 min. Supernatant fractions were recovered by centrifugation (14 000 g × 20 min, 4°C), and the protein concentration was determined using the BCA (bicinchoninic acid) assay. Nuclear and cytoplasmic extracts were prepared using NE-PER Nuclear and Cytoplasmic Extraction Reagents (Pierce Biotechnology), according to the manufacturer's instructions. Concentration of proteins from the cultured media was performed by trichloroacetic acid precipitation. Proteins after immunoblotting were visualized with an enhanced chemiluminescence system (Amersham Pharmacia Biotech) using a LAS 3000 imaging system (FUJIFILM Corporation, Tokyo, Japan). The band intensity was analyzed using TINA 2.0 software.

Matrigel plug assay

The cells grown at subconfluence were harvested, washed twice with phosphate-buffered saline (PBS) and resuspended in serum-free medium at 1.5 × 10 ⁷ cells/ml. Aliquots of the cells (0.2 ml) were mixed with 0.4 ml of growth factor-reduced Matrigel, and the mixture was subcutaneously (s.c.) injected into the flanks of mice. The mice were euthanized at day 10 after implantation, and the Matrigel plugs were carefully removed. The excised plugs were placed in cold PBS at 4°C for 15 h to be liquefied. After centrifugation at 14 000 r.p.m. for 5 min, hemoglobin levels in the supernatant were quantified with Drabkin's reagent (Sigma–Aldrich, St Louis, MO). In addition, vascular endothelial growth factor (VEGF)-α level in cultured media was determined with the Murine VEGF ELISA Development kit (Peprotech, Rocky Hill, NJ).

In vivo tumorigenicity and metastasis assay

For the in vivo tumorigenicity assays, 6- to 8-week-old female C57BL/6J mice were shaved on the ventral side and challenged s.c. with 2 × 10 ⁵ cells per 0.2 ml of PBS of B16F10 melanoma cells. Tumor diameters were measured every other day with a vernier caliper, and their volumes were calculated by the formula: volume (mm ³ ) = width ² (mm ² ) × length (mm) × 0.52. Tumor-bearing animals were killed when the tumors displayed severe ulceration. Peritumoral vascularization was observed at the time of tumor excision. To examine lymph node metastasis, the draining lymph nodes, especially inguinal lymph nodes, were investigated for melanotic B16F10 cells at day 20 after s.c. injection of tumor cells. Experimental metastatic potential of the B16F10 cells was measured by the lung colonization assay. In brief, the cells were injected into the tail vein of mice at a density of 2 × 10 ⁵ cells per 0.2 ml of PBS. Twenty days later, the mice were killed, lungs were fixed with Bouin's solution (Sigma, St. Louis, MO) and metastatic colonies on the lung surface were counted macroscopically.

Downregulation of NDRG2 gene expression by NDRG2 siRNA transfection

NDRG2 siRNA (sc-40758) was purchased from Santa Cruz Biotechnology. Fluorescein-conjugated control siRNA (sc-36869) was used as negative control and indicator of transfection efficiency. Transfection of siRNA was performed following the manufacturer's instruction. NDRG2-silenced SNU-620 cells (human gastric cancer cell line, Korean Cell Line Bank, Seoul, Korea) were obtained by transfection with pSUPER-retroviral vector encoding NDRG2 siRNA using retrovirus and by selection with puromycin ( 26 ).

Histology

Lung tissues were fixed in 10% neutral-buffered formalin solution for 6 h. After washing in fresh PBS, fixed tissues were dehydrated, cleared and embedded in paraffin (PARAPLAST, McCormick Scientific, St. Louis, MO). Sections (6 μm) were prepared on microscope slides, deparaffinized and stained with hematoxylin and eosin as routine procedures. Immunohistochemical staining against PECAM-1 (CD31) in tumor masses was performed as routine procedures.

Statistical analysis

Results are presented as means ± SDs. All experiments were repeated at least three times and data were analyzed for statistical significance using the Student's t -test. P values of <0.05 were considered significant.

Results

NDRG2 overexpression inhibits anchorage-independent growth of HT1080 human fibrosarcoma cells

To elucidate the effects of NDRG2 overexpression on the phenotypic changes of malignant cancer, we initially established stable clones of the HT1080 human fibrosarcoma cells by transfecting with a pCMV-Taq2B (mock) or pCMV-Taq2B-NDRG2 plasmid. After transfection, clones were selected and expanded based on their resistance to G418. As shown in Figure 1A , levels of NDRG2 messenger RNA (mRNA) in HT1080-NDRG2-transfected cells were significantly higher than that in the HT1080-mock-transfected cells. Strong expression of the NDRG2 protein in HT1080-NDRG2 cells was also confirmed by western blot analysis using a monoclonal antibody against NDRG2. HT1080-NDRG2 cells displayed no significant differences in cell morphology and cell proliferation ( Figure 1B ) when compared with HT1080-mock cells. It has been reported that the ability of cells to form colonies in a semisolid medium is generally considered a marker of anchorage independence, which is positively correlated with metastatic potential ( 30 ). As shown in Figure 1C , HT1080-mock cells displayed rapid proliferation compared with HT1080-NDRG2 cells and formed sizable colonies in the soft agar. However, the HT1080-NDRG2 cells showed a remarkably decreased ability to form soft agar colonies including a lower number of sizable colonies and a reduced colony size. The inhibition of colony formation was clearly not due to the cell growth rate, as determined in Figure 1B .

Fig. 1.

NDRG2 overexpression inhibits anchorage-independent growth of HT1080 human fibrosarcoma cells. ( A ) Expression levels of NDRG2 mRNA and protein in HT1080-mock and -NDRG2-transfected cells were examined by RT–PCR (upper lanes) and western blot (lower lanes) analysis, respectively. ( B ) Cell morphology was observed with phase-contrast microscopy and cell proliferation was determined with an 3-(4,5-dimethylthiazole-2-yl)-2,5-biphenyl tetrazolium bromide assay. ( C ) Cells (1 × 10 ⁴ ) were suspended in 3 ml of the medium containing 0.3% agar and 10% FBS and were applied onto the pre-solidified 0.6% agar (3 ml) in 60 mm culture dishes. After 2 weeks of incubation, colonies on soft agar were observed under a phase-contrast microscope and the diameter of representative 20 colonies was measured; * P < 0.01 (values are compared with HT1080-mock cells).

Effects of NDRG2 overexpression on the migration and invasion of HT1080 fibrosarcoma cells in an in vitro assay

To further evaluate the effect of NDRG2 overexpression on metastatic activity, we performed in vitro transwell migration assays and Matrigel-coated transwell invasion assays with HT1080-mock and HT1080-NDRG2-transfected cells. Representative micrographs were taken from the lower surface of the transwell filter, and cells that migrated or invaded were marked by using calcein-AM. As shown in Figure 2A , the migratory activity of the HT1080-NDRG2 cells was significantly lower than that of the HT1080-mock cells. In addition, the invasive potential, which is determined by the cells’ ability to invade a Matrigel barrier, was also considerably suppressed in NDRG2-transfected cells ( Figure 2B ). We next investigated the effect of NDRG2 overexpression on cell invasion in response to phorbol-12-myristate-13-acetate (PMA) stimulation. Treatment with 5 nM PMA led to a remarkable increase of invasiveness in the HT1080-mock cells, but any change in invasion of the HT1080-NDRG2 cells was insignificant, even after PMA treatment ( Figure 2C ). Quantitation of invading cells using calcein-AM revealed that NDRG2 overexpression in the HT1080 cells inhibited invasive activity by ∼88% in a resting state and 85% in a PMA-stimulated condition compared with HT1080-mock cells ( Figure 2D ).

Fig. 2.

Effects of NDRG2 overexpression on the migration and invasion of HT1080 fibrosarcoma cells in an in vitro assay. ( A ) A fixed number of cells were plated onto the upper part of the transwell chamber. The media compartment of the lower chamber contained 10% FBS. After incubation for 20 h, cells migrating to the lower surface of the membrane were stained with a 0.2% crystal violet/20% methanol (vol/vol) solution and were observed with phase-contrast microscopy. Migrating cells were also examined using calcein-AM as a substrate for intracellular esterase. The fluorescent intensity of free calcein was measured with a fluorescence microplate reader set (PerkinElmer Victor™). ( B ) A fixed number of cells were plated onto the upper part of the Matrigel-coated transwell chamber. After incubation for 20 h, invasive cells were characterized as described in Figure 2A. ( C and D ) The effect of NDRG2 overexpression on cell invasion in response to PMA stimulation was also investigated. * P < 0.01 (values are compared with HT1080-mock cells).

NDRG2 overexpression in HT1080 fibrosarcoma cells reduces PMA-induced MMPs expression and proteolytic activity

In order to elucidate the inhibitory mechanism of NDRG2 overexpression on migration and invasion, we examined the level of both mRNA and protein expression of MMPs in these cells. In a resting state, HT1080-mock cells expressed various MMPs at a high level, including MMP-1, -2, -9 and membrane type 1-MMP. In addition, PMA stimulation of HT1080-mock cells remarkably increased MMP expression in a dose-dependent manner (data not shown). However, as shown in Figure 3A , mRNA expression of MMP-9 expression in HT1080-NDRG2 cells was significantly lower than that in the HT1080-mock cells, and this observation was true in both a resting state and a PMA-stimulated condition. Since MMP-2 and MMP-9 play major roles in facilitating cancer metastasis by degrading various ECMs, we next examined whether NDRG2 overexpression in HT1080 cells modulates the secretion of MMP-2 and MMP-9 by western blot and gelatin zymography. As shown in Figure 3B , secretion of active MMP-2 was detected only in PMA-stimulated HT1080-mock cells, and secretion of MMP-9 was significantly suppressed in HT1080-NDRG2 cells in both a resting state and a PMA-stimulated condition. In addition, gelatin zymographic analyses revealed that HT1080-mock cells constitutively secreted MMP-2 and -9. PMA stimulation increased secretion of MMP-9 and partially converted latent pro-MMP-2 to its active form. However, in HT1080-NDRG2 cells, MMP-9 secretion was markedly reduced in a resting state and under PMA stimulation. MMP-2 activation by PMA treatment was slightly inhibited in HT1080-NDRG2 cells ( Figure 3C ). Similar to mRNA expression and gelatin zymography, assays for promoter activity of MMP-2 and -9 using a luciferase assay system revealed that NDRG2 overexpression markedly reduced the activities of the MMP-9 and also reduced the activities of MMP-2, especially under PMA stimulation ( Figure 3D ).

Fig. 3.

NDRG2 overexpression in HT1080 fibrosarcoma cells reduces PMA-induced MMP expression, proteolytic activities and NF-κB activation. ( A ) mRNA expression of MMPs in a resting condition or PMA-simulated condition was measured by RT–PCR. Cells were treated with PMA for 24 h at concentrations of 2.5 and 5 nM. ( B ) Cells were incubated in serum-free media for 24–48 h, with or without 5 nM of PMA, and CM was collected. Concentrated CM by trichloroacetic acid precipitation was subjected to western blot for MMP-2 and -9. ( C ) CM was subjected to gelatin zymography for detecting MMP-2 and -9 activities. ( D ) Cells were transiently transfected with a luciferase reporter construct using Lipofectamine 2000 reagent. After incubation for 4 h, 5 nM of PMA was added. After 18 h of additional incubation, cells were harvested in passive lysis buffer, and the luciferase activities were measured by a luminometer using the luciferase assay system. Changes in luciferase activity, with respect to the mock control, were calculated. ( E ) NF-κB and AP-1 luciferase activities after PMA or TNF-α stimulation were measured using the luciferase assay system. * P < 0.01 (values are compared with HT1080-mock cells). ( F ) HT1080 wild-type cells were transfected with dominant negative forms of dnIKKα and dnIKKβ and then exposed to PMA (5 nM) for 24 h. CM was subjected to gelatin zymography. ( G and H ) HT1080-mock and -NDRG2 cells were stimulated with 5 nM PMA for the indicated time, and cell lysates were subjected to western blot analysis. To examine the nuclear translocation of the NF-κB p65 subunit, cell lysates were fractionated into cytosolic (C) and nuclear (N) compartments. Results are from a single analysis, representative of two independent experiments.

NDRG2 overexpression significantly inhibits PMA-induced NF-κB activation

Previous reports have demonstrated that the expression of ECM-degrading proteases is regulated by several transcription factors including NF-κB and activator protein-1 (AP-1), and any pathway in tumor cells leading to activation of these transcription factors may contribute to increased MMP transcription and enhanced tumor invasion. In order to examine whether the inhibitory effect of NDRG2 overexpression on MMP expression and invasion was linked to NF-κB and/or AP-1 activities, NF-κB and AP-1 responsive reporter vectors were transiently transfected into HT1080-mock and HT1080-NDRG2 cells. As shown in Figure 3E , NDRG2 overexpression in HT1080 cells reduced NF-κB promoter activity by ∼20%. In addition, treatment with PMA led to a pronounced increase of NF-κB activity in HT1080-mock cells (∼4-fold), but NDRG2 overexpression strongly suppressed NF-κB activity, even after PMA stimulation. In contrast, neither PMA stimulation nor NDRG2 overexpression influenced AP-1 promoter activity in HT1080 cells. We confirmed suppression of NF-κB activation by NDRG2 overexpression under tumor necrosis factor (TNF)-α stimulation condition. TNF-α stimulation could increase AP-1 activity in HT1080-mock cells, but in this condition, AP-1 activity was not significantly suppressed by NDRG2 overexpression. In gelatin zymography, NDRG2 overexpression considerably inhibited the TNF-α-mediated secretion of MMP-9 and active MMP-2 (data not shown). Since NF-κB is critically involved in the metastasis via MMPs activation, we first examined the effect of transient expression of dominant negative 1κB kinase, dnIKKα or dnIKKβ on PMA-induced gelatinolytic MMP-2 and MMP-9 secretion in HT1080 cells. As shown in Figure 3F , PMA-induced MMP-9 secretion was significantly decreased by transiently expressing dnIKKα or dnIKKβ, which was similar to the observation made in HT1080-NDRG2 cells. Meanwhile, PMA-induced active MMP-2 secretion was not inhibited by dnIKKα or dnIKKβ. The dissociation of the inhibitory subunit IκB, which is mediated by its phosphorylation and subsequent degradation, and the translocation of the active p65 subunit to the nucleus are critical steps in NF-κB activation. The key regulatory step in this pathway involves activation of a high-molecular weight IKK complex. As shown in Figure 3G , PMA stimulation in HT1080-mock cells remarkably induced a phosphorylation of IKKα/β, and NDRG2 overexpression could significantly inhibit the phosphorylation of IKKα/β under PMA stimulation. Next, we examined the levels of IκBα and phospho-IκBα in these cells, along with the nuclear translocation of p65. Treatment of HT1080-mock cells with PMA immediately increased the level of IκBα phosphorylation, along with IκBα degradation. In addition, the p65 subunit rapidly translocated to the nucleus after treatment with PMA ( Figure 3H ). However, in HT1080-NDRG2 cells, the increase of IκBα phosphorylation and the nuclear translocation of p65 following PMA treatment were insignificant, compared with HT1080-mock cells.

NDRG2 overexpression in murine B16F10 melanoma cells also suppresses metastatic potential

To demonstrate the role of the NDRG2 gene as a candidate tumor suppressor in vivo , we utilized murine B16F10 melanoma cells, which are highly metastatic in syngeneic C57BL/6J mice. First, we established B16F10-mock and -NDRG2 stable clones, as described in the Materials and Methods. We subsequently investigated the effect of strong NDRG2 expression on the in vitro metastatic potential of these cells. As shown in Figure 4A and B , B16F10-NDRG2-transfected cells showed a lower level of MMP-9 expression, compared with B16F10-mock-transfected cells. In addition, in B16F10-mock cells, gelatinolytic MMP-9 secretion was significantly increased by TNF-α stimulation. Secretion of MMP-9 in B16F10-NDRG2 cells was insignificant even after TNF-α stimulation ( Figure 4C ). MMP-2 secretion was also reduced by NDRG2 overexpression in a resting state and TNF-α-stimulating condition. As demonstrated in the HT1080 cell system, NDRG2 expression in the B16F10 cells also completely prevented colony formation in soft agar ( Figure 4D ). Along with decreased MMP secretion, the migratory and invasive capabilities of B16F10-NDRG2 cells were remarkably suppressed compared with B16F10-mock cells ( Figure 4E ). TNF-α-induced MMP-9 secretion in B16F10 cells was significantly decreased by transiently expressing dnIKKα or dnIKKβ (data not shown). This result reinforces that suppression of NF-κB is critical for the metastasis inhibition via MMP-9 activation. We then examined the effect of NDRG2 expression on TNF-α-induced NF-κB promoter activity. As expected, treatment with TNF-α increased NF-κB activity in B16F10-mock cells (∼4-fold), but NDRG2 expression prevented an increase of NF-κB activity, even after TNF-α treatment ( Figure 4F ). However, AP-1 activity in B16F10 cells was not affected by NDRG2 overexpression. Furthermore, NDRG2 overexpression could significantly inhibit TNF-α-mediated increase of IKKα/β phosphorylation ( Figure 4G ). These results indicate again that NDRG2 expression inhibits MMP-9 secretion via suppression of NF-κB activation.

Fig. 4.

NDRG2 overexpression in B16F10 murine melanoma cells also decreases their metastatic potential and suppresses NF-κB activation. ( A ) mRNA expression of MMP-2 and MMP-9 was measured by RT–PCR. ( B ) Concentrated CM was subjected to western blot analysis. ( C ) Gelatinolytic activities were analyzed by gelatin zymography using CM. TNF-α (10 ng/ml) was used to treat the cells for 24 h. ( D ) To examine anchorage-independent cell growth, a soft agar colony formation assay was performed. ( E ) A fixed number of cells were plated onto the upper part of the transwell chamber. After incubation for 20 h, cells migrating to the lower surface of the membrane were stained and observed by phase-contrast microscopy. Migrating cells were also determined using calcein-AM. A fixed number of cells were plated onto the upper part of the Matrigel-coated transwell chamber. After incubation for 20 h, invasive cells were characterized. ( F ) NF-κB and AP-1 luciferase activity, after treating cells with TNF-α (10 ng/ml) for 18 h, was measured using the luciferase assay system. * P < 0.01 (values are compared with B16F10-mock cells). ( G ) After treating cells with TNF-α (10 ng/ml) for 15 min, cell lysates were subjected to western blot analysis.

NDRG2 overexpression inhibits in vivo tumor growth, pulmonary metastasis and secondary metastasis into lymph nodes

To clarify the role of NDRG2 expression on tumor metastasis in vivo , we next examined the inhibitory effect of NDRG2 expression on the ability of B16F10 cells to colonize the lung after intravenous injection. B16F10-mock cells injected into the tail vein of C57BL/6J mice metastasized to the lung and formed a considerable number of black colonies that were visible from the surface ( Figure 5A ). However, B16F10-NDRG2 cells formed a reduced number of colonies in the lung, suggesting that NDRG2 expression can inhibit the metastatic potential of B16F10 cells in vivo . Microscopic evaluation of the histology sections showed that >80% of the lung area in B16F10-mock-injected mice was covered with the melanoma colonies. In B16F10-NDRG2-injected mice, only 10–20% of the lung area was occupied by melanoma colonies ( Figure 5B ). As shown in HT1080 cell culture, B16F10-NDRG2 cells also had a similar growth rate in culture to B16F10-mock cells (data not shown). These findings indicate that reduced pulmonary metastasis of B16F10-NDRG2 cells was not primarily due to the rate of cell proliferation. However, it cannot be excluded that the cell proliferation of B16F10-mock and -NDRG2 cells may be differentially influenced by in vivo environmental conditions. Interestingly, when inoculated into syngeneic C57BL/6J, B16F10-mock cells developed a palpating tumor mass more rapidly and formed larger primary tumors at the injection site, compared with B16F10-NDRG2 cells ( Figure 5C ). Furthermore, at day 20 after the s.c. injection, we observed the metastasized melanotic cells in the inguinal lymph nodes. At this time point, the average tumor volume in the mice ( n = 3) inoculated with B16F10-mock and B16F10-NDRG2 cells was 27.5 ± 4.4 and 16.2 ± 0.3 (×10 ³ mm ³ ), respectively. The average tumor weight for the mock- and NDRG2-transfected cells was 6.5 ± 0.3 and 4.2 ± 0.03 g, respectively (data not shown). As shown in Figure 5D , lymph node metastases, especially to the side of the injection (right), were observed in B16F10-mock-injected mice, whereas lymph node metastasis in B16F10-NDRG2-injected mice was remarkably suppressed. When tumor cells were injected in the center of abdomen, metastasized melanotic cells were discovered in both sides of the lymph nodes ( Figure 5E ). However, compared with B16F10-mock cells, metastasis of B16F10-NDRG2 cells into the lymph nodes from the established primary tumor mass was almost completely suppressed. Moreover, considerable pulmonary metastases after s.c. injection were only detected in B16F10-mock-inoculated mice ( Figure 5F ). These results provide strong evidence that NDRG2 overexpression in B16F10 cells can reduce in vivo metastasis from a primary tumor mass to draining lymph nodes and distant organs.

Fig. 5.

NDRG2 overexpression in B16F10 melanoma cells reduces pulmonary metastases and in vivo tumorigenesis. ( A ) Experimental metastatic potential was measured by a lung colonization assay. ( B ) Representative histological photographs of lung tissue sections stained with hematoxylin and eosin (×200). NL, normal lung tissue; T, metastatic tumor lesion. ( C ) For the in vivo tumorigenicity assay, 6- to 8-week-old female C57BL/6J mice were shaved on the ventral side and challenged s.c. with 2 × 10 ⁵ cells per 0.2 ml of PBS. Tumor diameters were measured every other day with a vernier caliper. ( D ) To examine lymph node metastasis, melanotic B16F10 cells were characterized in draining lymph nodes at day 20 after s.c. injection of tumor cells on the right buttocks. The pictures of lymph nodes are representative of inguinal lymph nodes (each group, n = 3). ( E ) Cells were s.c. injected in the center of the abdominal region. After 20 days, inguinal lymph nodes on both sides were observed for melanotic tumor cells. ( F ) Lung colonization of tumor cells in the mice showing severe lymph node metastasis was observed macroscopically.

In vivo tumor angiogenesis is inhibited by NDRG2 overexpression

Angiogenesis is the formation of new blood vessels from pre-existing ones, and it has been shown to be essential for growth, invasion and metastasis of solid tumors. Vascular proliferation and tumor angiogenesis are regulated by angiogenic factors produced by tumor cells. Since the growth of B16F10-NDRG2 cells was retarded and the tumor masses from B16F10-NDRG2 cells were smaller than those of B16F10-mock cells at the same time point, we investigated peritumoral vascular formation when the tumor volume was similar between B16F10-mock and -NDRG2 cells ( Figure 6A , day 12 versus day 18 in mock versus NDRG2 cells). Regardless of tumor volume, vessel density around B16F10-mock tumors was higher than that around B16F10-NDRG2 tumors. To measure the intratumoral vessel density, we examined PECAM-1 (CD31) levels in each tumor mass. PECAM-1 mRNA levels in B16F10-mock tumor masses were increased in proportion to tumor volume, whereas the levels were very low in B16F10-NDRG2 tumor masses ( Figure 6A ). Additionally, to visualize microvessels in the tumor masses, immunohistochemical staining with an anti-PECAM-1 antibody was performed. As shown in Figure 6B , the numbers of intact capillaries in the B16F10-mock tumor masses were remarkably higher than in the B16F10-NDRG2 tumor masses. To further analyze the anti-angiogenic effect of NDRG2, we examined angiogenesis using a Matrigel plug assay, in which hemoglobin levels are measured as a relative index of angiogenesis. In this study, growth factor-reduced Matrigel was utilized because there was little hemoglobin in the Matrigel plug alone. As shown in Figure 6C , the hemoglobin levels in the plugs with B16F10-mock and -NDRG2 cells were 2.17 and 0.69 mg/g, respectively. In addition, the level of VEGF-α after incubation under normoxic or hypoxic condition was examined. As shown in Figure 6D , VEGF-α mRNA expression as well as VEGF-α secretion from B16F10-NDRG2 cells were remarkably lower than those of control cells. These results indicate that NDRG2 overexpression was associated with decreased angiogenesis, and this correlation may contribute to the inhibition of tumor growth and to the reduced metastases in lymph nodes and distant organs.

Fig. 6.

NDRG2 overexpression in B16F10 melanoma cells inhibits in vivo tumor angiogenesis. ( A ) Peritumoral vascularization was observed at day 12 or at day 18 after s.c. injection of B16F10-mock or -NDRG2 cells, respectively. Tumor volume of both mice was similar at the time of observation. After RNA extraction from a B16F10-mock and -NDRG2 tumor mass (at day 18 and 23 after s.c. injection), PECAM-1 mRNA expression was examined by RT–PCR. ( B ) The vascular density in solid tumors derived from s.c. injection of B1F10-mock or -NDRG2 cells was determined using immunohistochemical staining with an anti-CD31/PECAM-1 antibody. Tissues were counterstained with hematoxylin. Brownish staining indicates endothelial cells (×400). ( C ) B16F10-mock or -NDRG2 cells (3 × 10 ⁶ cells per 0.2 ml of serum-free medium) with 0.4 ml of growth factor-reduced Matrigel were s.c. injected into the flank of C57BL/6J mice. After 7 days, Matrigel plugs were removed from the mice, and hemoglobin levels of Matrigel plugs were quantified with Drabkin's reagent. Three independent experiments ( n = 2) were performed, and photographs and hemoglobin levels of representative Matrigel plugs were shown. ( D ) VEGF-α mRNA and protein secretion in the normoxia and hypoxia condition (10% CO ₂ and 1% O ₂ ) were measured by RT–PCR and enzyme-linked immunosorbent assay, respectively. * P < 0.01 (values are compared with B16F10-mock cells).

siRNA-mediated knockdown of NDRG2 rescues NF-κB activity and increases migration as well as invasion of tumor cells

To confirm the inhibitory effect of NDRG2 expression on metastasis and NF-κB activation, we examined whether knockdown of NDRG2 could rescue the migratory and invasive potential of tumor cells. As shown in Figure 7A , siRNA-mediated NDRG2 knockdown in B16F10-NDRG2 cells significantly increased migration and invasion, compared with control siRNA-transfected B16F10-NDRG2 cells. Moreover, NDRG2-silenced cells considerably increased NF-κB activity, especially under TNF-α stimulation ( Figure 7B ). In the experiments using NDRG2-high SNU-620 cells, retrovirus-mediated NDRG2 silencing could also increase gelatinolytic MMP-9 secretion, anchorage-independent cell proliferation and Matrigel invasion ( Figure 7C and D ). Collectively, these results, along with those from in vitro cell migration and invasion assays using HT1080 and B16F10 cells, support the notion that NDRG2 may be a candidate suppressor of tumor metastasis.

Fig. 7.

siRNA-mediated knockdown of NDRG2-rescued NF-κB activity and increased migration as well as invasion. ( A and B ) The decrease of NDRG2 mRNA levels in B16F10-NDRG2 cells transfected with NDRG2-siRNA was confirmed by RT–PCR. Cell migration and invasion were analyzed after siRNA-mediated gene silencing of NDRG2. NF-κB luciferase activity was also examined. * P < 0.01 (values are compared with control siRNA-transfected cells). ( C ) Increase in MMP-9 activity was observed in NDRG2-silenced SNU-620 cells. ( D ) NDRG2-silenced SNU-620 cells rescued the ability to proliferate in semisolid agar (upper) and to invade Matrigel barrier (lower).

Discussion

In this study, we demonstrated that NDRG2 expression is significantly downregulated in metastatic tumor cell lines and upregulation of its expression diminishes the metastatic capacity of cells in vitro and in animal tumor models. In particular, the present study is the first effort to specifically elucidate how NDRG2 expression inhibits migration, invasion and angiogenesis of tumor cells. NDRG2 expression in these cells was accompanied by a decrease in MMP activities via inactivation of NF-κB. First, our study confirmed that NDRG2 expression is almost completely downregulated in HT1080 human fibrosarcoma and B16F10 murine melanoma cell lines, which are both known to retain potent metastatic capacity. Previous studies reported that loss of NDRG2 expression was caused by methylation of promoter CpG sites ( 25 , 31 ). In fact, treatment with 5-azacytidine, a potent inhibitor of DNA methyltransferase, remarkably increased NDRG2 expression in both cell lines (data not shown). However, the levels of NDRG1 in both cell lines were considerable and were not altered by 5-azacytidine treatment or NDRG2 overexpression (data not shown).

One of our previous studies has demonstrated that an NDRG2-silenced gastric cancer cell line demonstrates slightly increased proliferation and resistance against cisplatin and Fas-mediated cell death ( 26 ). In case studies of gastric cancer patients, it has been shown that NDRG2 expression may be involved in tumor progression and the overall survival of the patients. Furthermore, loss of NDRG2 expression in hepatocellular carcinomas is significantly correlated with aggressive clinicopathologic features of hepatocellular carcinomas patients, and NDRG2 expression contributes to the suppression of liver cancer metastasis ( 31 ). Metastasis is a characteristic of highly malignant cancers with poor clinical outcome, and excess degradation of surrounding ECM is one of the hallmarks of tumor invasion and metastasis. Among ECM-degrading enzymes, MMPs are the most important metastasis-promoting genes because MMPs affect basic, necessary steps in the metastatic cascade, including angiogenesis, proliferation and apoptosis. As shown in our current study, upregulation of NDRG2 could significantly inhibit MMP-9 expression, in addition to MMP-2 and -9 activities under PMA- or TNF-α-stimulated conditions. Decreased levels of MMPs may attenuate metastatic progression, including anchorage-independent growth, migration, invasion and angiogenesis. Constitutive expression of NF-κB has been shown in cancer cell lines as well as tissue sample of cancer patients, and activation of NF-κB can lead to tumor cell proliferation, invasion, angiogenesis and metastasis. The aberrant IKK activity, shorter half-life of IκBα and TNF-α production are some of the reasons that have been postulated for constitutive NF-κB activation. We demonstrated in this study that NDRG2 overexpression reduces phospho-IKKα/β expression following suppression of subsequent IκBα degradation and p65 nuclear translocation. In particular, NDRG2 expression appeared to specifically affect tumor growth in vivo but not cell growth in vitro , suggesting a critical role for NDRG2 in modulating the tumor/stroma microenvironment. A potential function of NDRG2 as a metastatic suppressor may be due to altering angiogenesis in tumor stroma. Histological examination in fact revealed that dead cell regions are sporadically observed in B16F10-NDRG2-injected tumor masses (data not shown). As shown in Figure 6B , scanty vessel formation with NDRG2 expression may be insufficient for providing enough nutrients and oxygen to allow the tumor to grow unrestrained and to metastasize to lymph nodes or distant organs. Since MMP-9 plays a critical role in the angiogenic switch during carcinogenesis, inhibition of MMP-9 by NDRG2 expression is properly correlated with decreased angiogenesis. In addition, to generate blood vessels, release of angiogenesis-inducing signaling molecules from tumor cells is very important. In a Matrigel plug assay injected with a Matrigel-conditioned medium (CM) mixture, hemoglobin contents in the plugs mixed with B16F10-NDRG2-derived CM were slightly lower than that in the plugs mixed with B16F10-mock-derived CM. A more precise study is required to understand the underlying mechanism of how expression of such angiogenic factors could be modulated by NDRG2. Our results showed that the level of VEGF-α in the CM obtained from B16F10-NDRG2 cells was remarkably lower than that of control cells. Furthermore, NDRG2 expression induced a downregulation of the transcriptional activity of hypoxia-inducible factor 1α, which plays a key role in VEGF transcription (data not shown). We are currently investigating activators or inhibitors of angiogenesis that can be modulated by NDRG2 expression and following up on related regulatory mechanisms. It has recently been reported that NDRG2 expression is markedly increased after exposure to hypoxic conditions and similar stresses, and its expression is closely related to hypoxia-induced apoptosis ( 27 ). The findings support our observation that B16F10-NDRG2 cells that are exposed to hypoxic conditions during tumor growth in vivo are more susceptible to apoptosis than control cells, leading to growth retardation and metastasis inhibition. It is intriguing to speculate that NDRG2-negative and -positive tumor cells could differentially respond to environmental stimuli and adjust the host immune system, consequently influencing tumor progression and the prognosis of patients. Getting a global view of the biological function of NDRG2 in cancer and the immune system can therefore provide a unique and powerful tool to develop clinically relevant molecular biomarkers that have applications to basic research, diagnosis and therapy of cancer. Collectively, NDRG2 could thus affect not only metastasis/invasion but also tumor growth in vivo , suggesting that NDRG2 could be a putative metastasis suppressor gene.

Funding

Science Research Center (SRC) program of Korea Science and Engineering Foundation (Research Center for Women's Diseases, R11-2005-017-03001); Korea Research Foundation (KRF-2007-355-C00048).

Abbreviations

Abbreviations
 
• AM

  acetomethylester

 
• AP-1

  activator protein-1

 
• CM

  conditioned medium

 
• DMEM

  Dulbecco's modified Eagle's medium

 
• ECM

  extracellular matrix

 
• FBS

  fetal bovine serum

 
• IKK

  IκB kinase

 
• MMP

  matrix metalloproteinase

 
• mRNA

  messenger RNA

 
• NDRG2

  N-myc downstream-regulated gene 2

 
• NF-κB

  nuclear factor kappa B

 
• PBS

  phosphate-buffered saline

 
• PCR

  polymerase chain reaction

 
• PECAM-1

  platelet-endothelial cell adhesion molecule-1

 
• PMA

  phorbol-12-myristate-13-acetate

 
• s.c.

  subcutaneously

 
• siRNA

  small interfering RNA

 
• TNF

  tumor necrosis factor

 
• VEGF

  vascular endothelial growth factor

We thank Dr Jang-Soo Chun (Gwangju Institute of Science and Technology, Gwangju, Korea) for providing dnIKKα and dnIKKβ. We also thank Dr Geon-Kook Lee (National Cancer Center, Goyang, Korea) for doing histochemical analysis of tumor tissues. Full-length human MMP-2 promotor-luciferase construct was kindly provided by Dr Etty Benveniste (University of Alabama at Birmingham, Birmingham, AL).

Conflict of Interest Statement: None declared.