N-myc downstream-regulated gene 2 expression is associated with glucose transport and correlated with prognosis in breast carcinoma

This study was approved by the Ethics Committee of the Fourth Military Medical University. All patients from whom we obtained the 30 pairs of breast carcinoma and adjacent normal breast tissue specimens, as well as the 269 breast carcinoma sample study cohort, provided their full consent to participate in the study at the Xijing Hospital of the Fourth Military Medical University (Xi’an, China). NDRG2 and GLUT1 expression were detected in all specimens. Tissue specimens were examined separately by two pathologists under double-blinded conditions without prior knowledge of the clinical status of the specimens.

Immunohistochemistry (IHC) was performed using the avidin-biotin-peroxidase complex method on all breast carcinoma samples. All sections were deparaffinised in xylenes and dehydrated through a gradient concentration of alcohol before endogenous peroxidase activity was blocked using 0.5% H₂O₂ in methanol for 10 minutes. After nonspecific binding was blocked, the slides were incubated with NDRG2 antibody (1:200; Abnova, Taipei, Taiwan) or GLUT1 antibody (1:200; Santa Cruz Biotechnology, Santa Cruz, CA, USA) in phosphate-buffered saline (PBS) at 4°C overnight in a humidified container. Biotinylated goat anti-rabbit immunoglobulin G (IgG) (1:400; Sigma-Aldrich, St Louis, MO, USA) was incubated with the sections for 1 hour at room temperature and detected using a streptavidin-peroxidase complex. The brown colour indicative of peroxidase activity was developed by incubation with 0.1% 3,3′-diaminobenzidine (Sigma-Aldrich) in PBS with 0.05% H₂O₂ for 5 minutes at room temperature. The appropriate positive and negative controls were included in each run of IHC.

An immunoreactivity score system based on the proportion and intensity of positively stained cancer cells was applied. The two extensional standards taken were as follows: (1) the number of positively stained cells ≤5%, scored 0; 6% to 25%, scored 1; 26% to 50%, scored 2; 51% to 75%, scored 3; and >75%, scored 4; and (2) the intensity of stain colourless, scored 0; pallideflavens, scored 1; yellow, scored 2; and brown, scored 3. Extensional standards (1) and (2) were multiplied, and the staining grade was stratified as absent (score 0), weak (score 1 to 4), moderate (score 5 to 8) or strong (score 9 to 12). Specimens were rescored if the difference of scores from the two pathologists was greater than 3. Tumours with moderate or strong immunostaining were classified as having high expression, and tumours with absent or weak immunostaining were classified as having low expression.

T-47D and SK-BR-3 cells were obtained from the American Type Culture Collection (Manassas, VA, USA) and cultured in a humidified incubator under 5% CO₂ in Dulbecco’s modified Eagle’s medium (DMEM) (Invitrogen/Life Technologies, Carlsbad, CA, USA) supplemented with 10% foetal bovine serum (FBS) and 2 mM L-glutamine. NDRG2 antibody was purchased from Abnova. GLUT1, hemagglutinin (HA), actin, flag and tubulin antibodies were obtained from Santa Cruz Biotechnology.

The cells (1 × 10⁶ cells/well) were seeded into six-well plates and transfected with the following constructs using Lipofectamine 2000 (Invitrogen/Life Sciences) according to the manufacturer’s instructions as follows: NDRG2 expression plasmid (pCMV-flag-NDRG2), GLUT1 expression plasmid (pCMV-eGFP-GLUT1), HA-ubiquitin or small interfering RNA (siRNA) that targeted NDRG2. The target sequences of NDRG2 siRNA and control siRNA are given in Additional file 1: Table S1.

Cells were seeded into six-well plates at a density of 1 × 10⁶ cells/well and incubated to reach approximately 80% confluence. After the medium was removed, adenovirus expressing NDRG2 (Ad-NDRG2) or the negative control adenovirus expressing LacZ (Ad-LacZ) was added to serum-free DMEM, incubated for 2 hours, replaced with fresh DMEM supplemented with 10% FBS and incubated for another 48 hours. Recombinant adenoviruses carrying NDRG2 or LacZ were purchased from Benyuan Zhengyang Gene Technology Company (Beijing, China).

Both cells and breast tissues were lysed in radioimmunoprecipitation assay buffer (0.05 M Tris-HCl (pH 7.4), 0.15 M NaCl, 0.25% deoxycholic acid, 1% Nonidet P-40, 1 mM ethylenediaminetetraacetic acid (EDTA), 1 mM phenylmethylsulfonyl fluoride, 10 mg/ml aprotinin and 10 mg/ml leupeptin). Protein concentrations were measured using a bicinchoninic acid protein assay (Pierce Biotechnology, Rockford, IL, USA). Proteins were resolved by SDS-PAGE and transferred to Hybond enhanced chemiluminescence nitrocellulose membranes (Amersham Biosciences, Piscataway, NJ, USA). The blots were probed with the different primary antibodies and species-matched secondary antibodies. The bands were detected using enhanced chemiluminescence (Pierce Biotechnology) or the Odyssey Imaging System (LI-COR Biosciences, Lincoln, NE, USA).

The RNA extracted from cells with TRIzol reagent (Invitrogen/Life Technologies, Carlsbad, CA, USA) was converted to combinational cDNA with the RevertAid First Strand cDNA Synthesis Kit (Fermentas/Thermo Scientific; Pittsburgh, PA, USA). Real-time PCR analysis was performed using the Prism 7500 Real-Time PCR System (Applied Biosystems, Foster City, CA, USA) and the SYBR Premix Ex Taq II (Tli RNase H Plus) kit (TaKaRa Bio, Shiga, Japan) according to the manufacturer’s instructions. The relative gene expression levels were calculated using the 2^-ΔΔCt method, in which Ct represented the threshold cycle and β-actin was used as a reference gene. The primer sequence is given in Additional file 1: Table S1.

Cell growth following transfection was evaluated by 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assay. Cells were seeded into a 96-well plate (1 × 10⁴ cells per well) and incubated for 24 hours. The cells were then incubated with 0.5 mg/ml MTT (Sigma-Aldrich). Four hours later, the medium was replaced with 100 μl of dimethyl sulfoxide (Sigma-Aldrich) and vortexed for 10 minutes. Absorbance was then recorded at 490 nm using an Easy Reader 340 AT plate reader (SLT-Lab Instruments, Salzburg, Austria). Relative values of optical density were calculated as a percentage of the control. All experiments were performed three times independently.

Prior to being harvested, adherent cultures of control and NDRG2 adenovirus- or siRNA-treated cells in DMEM containing 25 mM glucose were washed twice with cold PBS and then lysed with ion-free H₂O for 5 minutes on ice. The glucose content was measured with a D-glucose measurement kit (GAHK-20; Sigma-Aldrich) according to the manufacturer’s protocol.

Cells were fixed in a freshly prepared solution of 4% paraformaldehyde, rinsed and permeabilised with 0.1% Triton X-100 in PBS. Permeabilised cells were then incubated with horse serum in PBS to block nonspecific binding. After being washed with PBS, the cells were incubated overnight at 4°C with mouse anti-NDRG2 antibody (diluted 1:150), rabbit anti-GLUT1 antibody (diluted 1:150) and fluorescein isothiocyanate (FITC)-conjugated anti-mouse antibody (diluted 1:400; Sigma-Aldrich) or cyanine 3 (Cy3)-conjugated anti-rabbit antibody (diluted 1:400; Sigma-Aldrich). The isotype mouse and rabbit IgGs were used as negative controls. Dual-colour detection was performed using a laser confocal microscope after treatment with 4′,6-diamidino-2-phenylindole (DAPI) to label nuclear DNA.

For inoculation into nude mice, SK-BR-3 or MDA-MB-231 cells were washed with PBS, digested with trypsin and resuspended in serum-free DMEM. After centrifugation (800 rpm), the cell pellets were resuspended in DMEM. The cell suspension (1 × 10⁶ cells in a 100-ml volume of PBS) was injected subcutaneously into the hind legs of 4-week-old female BALB/c athymic (nu/nu) mice (SLAC Laboratory Animal Company, Shanghai, China). When the tumours reached a volume of approximately 200 mm³, the mice were arbitrarily assigned to different groups (n = 6 each) to receive intratumoural injections of 0.5, 1 or 2 × 10⁹ plaque-forming units (PFU) of Ad-NDRG2, 2 × 10⁹ Ad-LacZ or PBS. Intratumoural injections were repeated every 3 days for a total of 21 days. Tumours were measured (perpendicular diameters) every 3 days, and their volumes were calculated. On day 21, the mice were killed and their tumours were removed for analysis. Tumour volumes were calculated based on caliper measurements of the length and width of the lesions using the following formula: 0.5 × length × (width²). The tumour growth curve was then derived from these data.

All of the experimental procedures were conducted in accordance with the Detailed Rules for the Administration of Animal Experiments for Medical Research Purposes issued by the Ministry of Health of China and received ethical approval by the Animal Experiment Administration Committee of the Fourth Military Medical University (Xi’an, China). All efforts were made to minimise the animals’ suffering and reduce the number of animals used.

In vitro experiments were performed three times, and each experiment was performed in triplicate. Data from all quantitative assays are expressed as the means ± SD and were analysed statistically using one-way analysis of variance, independent samples t-test or Student’s t-test. In the clinical specimens study, the associations between NDRG2 expression and categorical variables were analysed using the χ² test or Fisher’s exact test as appropriate. Correlations between NDRG2 expression and the expression of other molecules were analysed by using the Spearman correlation test. Kaplan–Meier analysis was used to evaluate disease-free survival and overall survival. P < 0.05 was considered statistically significant.

To assess the significance of NDRG2 protein expression in the development and progression of breast cancer, we compared the histopathological characteristics of 269 breast cancer samples with available NDRG2 protein status. The correlations between NDRG2 expression and different clinical histopathological factors are presented in Table 1. Significant correlations were found between low NDRG2 expression and advanced TNM staging (P < 0.0001), high proliferation index (Ki67 status; P = 0.006), positive human epidermal growth factor receptor 2 (HER2) status (P = 0.010) and poor histological differentiation (P < 0.0001). However, high NDRG2 expression was correlated with positive oestrogen receptor (ER) status (P = 0.010). Correlation coefficients are presented in Table 2. Furthermore, Kaplan–Meier analysis was used to evaluate the disease-free survival and overall survival of patients with breast cancer and NDRG2 protein expression. The results show that patients with high NDRG2 expression in breast tumour tissues had better disease-free survival than those with low NDRG2 expression (P = 0.0066 by logrank test) (Figure 1A). Breast cancer patients with low NDRG2 expression had a higher risk of relapse than those with high NDRG2 expression. A statistically significant association between long overall survival and high NDRG2 protein levels was found in breast cancer patients. Breast cancer patients with high NDRG2 expression had longer overall survival than patients with low NDRG2 expression P = 0.0007 by logrank test) (Figure 1B).

Tumour cells depend on glucose as a major substrate for energy production, and many tumour suppressor genes play important roles in the regulation of glucose metabolism [27]. GLUT1 is broadly expressed in the body tissues and is also involved in glucose uptake in the basic state, especially so for tumour glucose metabolism [18]. In our present study, we first investigated this relationship between NDRG2 and GLUT1 expression by IHC and immunoblotting in 30 pairs of breast carcinoma and adjacent normal breast tissue specimens. Positive staining of anti-NDRG2 was found predominantly in the cytoplasm of both normal breast cells and breast cancer cells, but weaker staining for NDRG2 was observed in the cancer specimens compared with normal tissues. Conversely, strong positive staining of GLUT1 was observed in the cytoplasm and cytosolic membrane of tumour cells, and weaker staining was found in noncancerous breast tissue (Figure 1C and Additional file 2: Figures S1 and S2). Furthermore, immunoblotting data revealed that NDRG2 protein expression in cancer samples was much lower than that in the paired adjacent normal tissue. However, GLUT1 expression was higher in cancer tissue than in the paired adjacent normal tissue (Figure 1D and Additional file 2: Figure S3). Indeed, in normal tissue and cancer tissue, the means of immunoreactivity score of NDRG2 IHC staining were 6.03 and 1.97 (Figure 1E), respectively, and the means of immunoreactivity score of GLUT1, IHC Staining were 3.13 and 6.53 (Figure 1F), respectively. Similarly, we identified the degree of correlation between NDRG2 expression and GLUT1 expression in 269 breast cancer specimens. The Spearman correlation analysis indicated that NDRG2 protein expression was inversely correlated with GLUT1 expression (Pearson correlation coefficient r _s = -0.179; P = 0.003) (Table 3). The data imply that NDRG2 expression might be involved in the development and progression of breast cancer and that NDRG2 expression is inversely correlated with GLUT1 expression in breast carcinoma.

Next, we sought to explore the regulatory mechanism of NDRG2 in tumour glucose metabolism. To investigate the effect of NDRG2 on breast cancer cell proliferation in high-glucose (25 mM) or low-glucose (5.5 mM) medium, we designed the following assays. First, the basic NDRG2 expression in five breast cancer cell lines was examined by immunoblotting. The results showed that the expression of NDRG2 was relatively high in T-47D cells and relatively low in SK-BR-3 cells (Figure 2A). In the following studies, we chose T-47D and SK-BR-3 cells as experimental models. Next, the adenovirus carrying NDRG2 (Ad-NDRG2) or the siRNA targeting NDRG2 (NDRG2 siRNA) was applied to upregulate or knock down, respectively, the level of NDRG2. After infection by Ad-NDRG2 in SK-BR-3 or transfection by NDRG2 siRNA in T-47D cells, the expression of NDRG2 was successfully increased or decreased (Figure 2B). Moreover, these effects were more prominent with increasing concentrations of Ad-NDRG2 or NDRG2 siRNA (Figures 3A to 3D). The results of the MTT assays showed that in both high- and low-glucose medium, the overexpression of NDRG2 inhibited SK-BR-3 cells proliferation (Figure 2C) and silencing NDRG2 promoted T-47D cell proliferation (Figure 2D). In high-glucose medium, however, the effects of NDRG2 on cell proliferation were more prominent.

Given that the antiproliferative effects of NDRG2 on breast cancer cells were more prominent in high-glucose medium, we investigated whether NDRG2 would affect the capacity of these cells to take in glucose. First, the effect of the glucose culture environment on NDRG2 expression was investigated. Immunoblot analysis revealed that, after exposure to different concentrations of glucose medium, the transcription and translation levels of NDRG2 increased with increasing glucose levels (Figures 2E and 2F). In addition, we checked the intracellular glucose content to assess the extent to which NDRG2 contributed to cell glucose uptake. Glucose uptake was decreased with increasing Ad-NDRG2 treatment in SK-BR-3 cells (Figure 2G). T-47D cells were transfected with NDRG2 siRNA at different concentrations, and glucose uptake was significantly augmented with increasing NDRG2 siRNA administration (Figure 2H). These data suggest that a high-glucose microenvironment could promote NDRG2 expression and that NDRG2 could inhibit glucose intake in breast cancer cells.

The transport of glucose across the plasma membrane is the first rate-limiting step for glucose metabolism and is mediated via GLUTs. GLUTs play critical roles in glucose uptake, especially in tumour cells. To investigate whether GLUTs are regulated by NDRG2, GLUT1/2/3/4 protein levels were measured in SK-BR-3 cells infected with Ad-NDRG2 or T-47D cells transfected with NDRG2 siRNA in a high-glucose medium. Immunoblot analysis revealed that GLUT1 protein levels were decreased by increases in Ad-NDRG2-mediated NDRG2 overexpression, but that GLUT1 protein levels were increased with NDRG2 siRNA treatment (Figures 3A and 3B). There were no changes in the intensity of anti-GLUT2 and anti-GLUT4 bands from treated cells compared with the untreated controls. GLUT3 was nearly undetectable in both the SK-BR-3 and T-47D cells (Figures 3A and 3B). However, the data obtained from real-time PCR showed that GLUT1 transcriptional levels were not significantly changed with NDRG2 up- or downregulation (Figures 3E and 3F). Because the amount of GLUT1 protein can be modified by changes in the rate of synthesis or degradation, we hypothesised that NDRG2 may exert an effect on the degradation of GLUT1. To test this hypothesis, we added the ubiquitin-proteasome inhibitor MG-132 with the Ad-NDRG2-treated SK-BR-3 cells. As shown in Figure 3G, the addition of MG-132 rescued GLUT1 protein levels, indicating that the decrease in GLUT1 in response to NDRG2 overexpression was due to proteasome-dependent degradation. GLUT1 protein is synthesised and degraded in the cytosol, whereas GLUT1 protein enacts its biological function primarily when it is transported to the cell membrane. So, we separated the membrane and cytosolic content of cells to detect GLUT1 protein distribution changes. As shown in Figure 3H, the amount of GLUT1 in the membrane and cytosolic fractions was substantially decreased in NDRG2-overexpressed SK-BR-3 cells. In addition, the immunoprecipitation data revealed that ectopic NDRG2 expression led to an increase in GLUT1 ubiquitination (Figure 3I and Additional file 2: Figure S4). These results suggest that NDRG2 is able to promote GLUT1 ubiquitination and degradation and can lead to decreased GLUT1 protein level.

Our observation that NDRG2 regulated GLUT1 protein stability prompted us to examine the interaction between NDRG2 and GLUT1 and their subcellular distribution. Confocal microscopy was applied to observe the subcellular localisation of NDRG2 and GLUT1 in SK-BR-3 cells. We found that a portion of endogenous NDRG2 and endogenous GLUT1 was colocalised in the cytoplasmic region of SK-BR-3 cells (Figure 4A). The colocalisation suggests that NDRG2 and GLUT1 may physically interact with each other. To establish whether NDRG2 physically associates with GLUT1, coimmunoprecipitation was performed using lysates prepared from SK-BR-3 cells. A 41-kDa protein that corresponds to NDRG2 was precipitated by anti-GLUT1 antibody and probed by anti-NDRG2 antibody, and a 55-kDa protein corresponding to GLUT1 was precipitated by the anti-NDRG2 antibody and probed by anti-GLUT1 antibody (Figure 4B). We also demonstrated that exogenous GLUT1 coimmunoprecipitated with exogenous NDRG2 in HEK293 cells that were cotransfected with pCMV-flag-NDRG2 and pCMV-eGFP-GLUT1 plasmids and vice versa (Additional file 2: Figure S5). Collectively, the results of these experiments indicated that NDRG2 bound to GLUT1.

We found that NDRG2 could induce GLUT1 protein degradation to decrease glucose uptake in breast cancer cell lines, but whether this regulatory mechanism also functions in the tumour microenvironment must be studied in vivo. We injected Ad-NDRG2 at the concentrations of 0.5, 1 and 2 × 10⁹ PFU or 2 × 10⁹ PFU Ad-LacZ every 3 days into preestablished human SK-BR-3 breast tumours (approximately 200 mm³) grown in nude mice. As shown in Figure 5A, the Ad-NDRG2 group that received injections at 2 × 10⁹ PFU achieved a sustained and significant arrest of tumour growth (68% decrease in mean tumour volume on day 21 compared with Ad-LacZ group). The mice were killed at 21 days after the first intratumoural injection, and the tumours were removed for analysis of the glucose uptake and protein levels of NDRG2 and GLUT1. We found that the glucose uptake of tumour cells was inhibited significantly with increased Ad-NDRG2 compared with the Ad-LacZ group (Figure 5B). Consistent with the results of in vitro experiments (Figure 3A), increased NDRG2 was correlated with decreased GLUT1 in the xenograft tumours (Figure 5D and Additional file 2: Figure S6). In addition, obvious positive NDRG2 staining was detected by IHC in tumours excised from mice in the 2 × 10⁹ PFU Ad-NDRG2 group. However, GLUT1 staining was weaker in the Ad-NDRG2 group compared with the Ad-LacZ group (Figure 5C). We repeated the xenograft tumour experiments with MDA-MB-231 cells, a highly metastatic cell line, and the results (Additional file 2: Figure S7) were consistent with those of SK-BR-3 cells.