SLC39A4 expression is associated with enhanced cell migration, cisplatin resistance, and poor survival in non-small cell lung cancer

The zinc transporter SLC39A4 influences epithelial cell morphology and migration in various cancers; however, its role in regulating cell invasion and chemotherapeutic resistance in human lung cancer is not yet clear. Here, integrated analysis of gene expression in non-small cell lung cancer revealed that SLC39A4 expression is significantly correlated with increased tumour size and regional lymph node spread, as well as shorter overall survival (OS) and disease-free survival (DFS). SLC39A4 silencing by lentivirus-mediated shRNA blocked human lung cancer cell epithelial-mesenchymal transition and metastasis in vitro and in vivo, respectively. Moreover, SLC39A4 knockdown enhanced cancer cell sensitivity to cisplatin-induced death by inhibiting stemness in lung cancer cells. Collectively, these data suggest that SLC39A4 may be a novel therapeutic target and predictive marker of tumour metastasis in non-small cell lung cancer.

The growing prevalence of high-throughput genomics has generated large amounts of data pertinent to cancer research 19 , leading to the formation of multiple cancer-specific gene expression databases, such as the Gene Expression Omnibus (GEO) and the Cancer Genome Atlas (TCGA) 20,21 . To date, little is known about the clinical significance and function of SLC39A4 in NSCLC pathogenesis; therefore, this study aimed to investigate its relevance in meta-analyses of data obtained from GEO and TCGA. We subsequently confirmed our results by tissue arrays, SLC39A4 functional studies in cultured cells and animal models.
Here, we showed that SLC39A4 upregulation is a prognostic indicator of NSCLC. Cell-based assays also demonstrated an important role for SLC39A4 in the diminished migratory potential, stemness, and chemoresistance of A549 lung cancer cells both in vitro and in vivo. Collectively, this research suggests that SLC39A4 may be a novel therapeutic target in NSCLC.

Results
SLC39A4 is an independent predictor of NSCLC. To study the role and function of SLC39A4 in NSCLC pathogenesis, we first assessed the levels of SLC39A4 expression in normal and lung carcinoma tissues in a meta-analysis of eight GEO datasets containing clinicopathological information (Supplementary Table S1). Notably, we found that SLC39A4A expression was significantly higher in lung cancer samples than in normal lung tissues (Z = 4.86, P < 0.00001) (Fig. 1a). Furthermore, a meta-analysis of SLC39A4 expression specificity and sensitivity supported its use as a diagnostic indicator (Fig. 1b). As shown in Fig. 1c, SLC39A4 expression is also markedly higher in stage III-IV cancers than in stage I-II counterparts (Z = 2.97, P < 0.003), indicating that SLC39A4 expression is associated with NSCLC progression.

SLC39A4 expression is negatively correlated with overall survival (OS) and disease-free survival (DFS) in NSCLC.
We queried the GEO and TCGA databases and found seven datasets that included data on various survival outcomes. A meta-analysis of these studies revealed that increased tumour SLC39A4 expression was associated with shortened OS (Z = 2.05, P < 0.04) and DFS (Z = 2.46, P < 0.01) (Supplementary Figure 1a,b). To confirm this finding, we examined the association between SLC39A4 expression and survival outcomes in lung tumour arrays by immunohistochemistry. For this, the 90 tumour samples were dichotomized into groups with high or low SLC39A4 expression based on the median staining intensity observed in the array (μ = 6.516) (Fig. 2a), and then evaluated for OS and DFS by Kaplan-Meier analysis (Fig. 2b,c). Notably, we found that high SLC39A4 expression was a significant indicator of poor OS and DFS in NSCLC (OS: HR = 2.918, 95% CI: 2.353-3.482, P = 0.0333; DFS: HR = 2.498, 95% CI: 1.924-3.072, P = 0.0180), consistent with the results of our meta-analysis. Moreover, we also determined that SLC39A4 expression was associated with increased tumour volume (P = 0.0105), regional lymph node spread (P = 0.0030), and clinical stage (P = 0.0412) ( Table 1). More importantly, regional lymph nodes (HR = 2.562, P = 0.013), clinical stage (HR = 2.118, P = 0.022) and SLC39A4 overexpression (HR = 3.067, P = 0.019) were also independent predictors for OS in the multivariate analysis ( Table 2). As shown in Fig. 2d, we also observed an increase in SLC39A4 expression corresponding with the progression of in situ tumours to metastatic lesions.

SLC39A4 silencing inhibits lung cancer cell metastasis in vitro.
EMT is as a key step in the metastatic initiation of cancer cells 22 . Because tumour metastasis is significantly detrimental to the survival of patients with lung cancer, we further investigated the effects of SLC39A4 expression on lung cancer cell migration. A549 cells were infected with lentivirus expressing SLC39A4 shRNA or control lentivirus and potent knockdown was confirmed by qPCR and western blot analysis for mRNA and protein expression, respectively (P < 0.01; Fig. 3a,b). We found that SLC39A4 knockdown inhibits the Zn 2+ concentration of cells in vivo (Fig. 3c). Wound-healing, transwell, and matrigel invasion assays revealed that SLC39A4 knockdown significantly inhibited A549 cell migration ( Fig. 3d-i). Consistently, immunofluorescence and western blot analysis showed a marked upregulation of the epithelial marker E-cadherin accompanied by a concomitant downregulation of the mesenchymal markers FSP-1 and N-cadherin in response to SLC39A4 silencing (Fig. 3j,k). The same experimental results were obtained from SPC-A-1 cells (Supplementary Figure 2). In addition, the function of SLC39A4 in EMT was investigated in normal lung epithelial cell line BEAS-2B; we found that SLC39A4 knockdown reduced the Zn 2+ concentration and inhibited the expression of EMT makers (Supplementary Figure 3). This pattern was also observed in the GEO meta-analysis, where SLC39A4 levels were negatively correlated with E-cadherin expression (Z = 7.60, P < 0.00001; Supplementary Figure 4a), but positively associated with FSP-1 (Z = 6.56, P < 0.00001) and N-cadherin (Z = 6.59, P < 0.00001; Supplementary Figure 4b SLC39A4 silencing promotes sensitivity to cisplatin in human NSCLC cells. Our   shown that the EMT conveys resistance to radiotherapy or chemotherapy 12 , we speculated that SLC39A4 silencing could affect the chemotherapeutic response by regulating stemness. SLC39A4 expression was significantly correlated with the expression of the lung cancer stem cell biomarkers CD44 and CD133 23 (Z = 7.11, P < 0.00001; Fig. 5a,b). This finding was further confirmed by immunofluorescence and western blot analysis of CD44 and CD133 expression in shSLC39A4 and Ctl A549 cells (Fig. 5c,d).
Cancer stem cells are associated with chemoresistance 24 . Cisplatin is the first-line drug for lung cancer therapy; however, resistance is highly prevalent in patients and significantly limits survival. Thus, we further explored the functional significance of SLC39A4 expression in cisplatin resistance. Notably, the CCK-8 assay and colony formation assays revealed that shSLC39A4 sensitized A549 cells to cisplatin-induced cell death (Fig. 5e-j), suggesting that SLC39A4 enhances cellular resistance to cisplatin therapy. Similar findings were also observed in a subcutaneous tumour model with cisplatin treatment, where shSLC39A4 A549 tumours caused less body weight loss (Fig. 5h) and were less proliferative and resistant to cisplatin treatment (Fig. 5i,j).

Discussion
Lung cancer is the leading cause of cancer-related death in men and women 25 , with more than one million new cases diagnosed each year. Despite recent advances in treatment modalities, including surgery, chemotherapy, and radiotherapy, cancer metastasis remains the major determinant of poor outcome, with an overall 5-year survival rate of 15% 26 . Thus, it is important to discover novel lung cancer metastasis-related genes and their functional mechanisms to encourage the development of additional cancer therapeutics.
The combined use of tumour-related gene expression data and clinical specimens has proven to be a valuable method for identification of potential drug targets [27][28][29] . Studies have indicated that zinc transport and homeostasis play important roles in cancer progression [30][31][32] ; however, their functional significance in lung cancer is not clear. In this study, a meta-analysis of data extracted from the GEO and TCGA databases revealed the clinical significance of SLC39A4 in lung cancer. Specifically, SLC39A4 expression was markedly higher in NSCLC tissues than in normal controls and was a prognostic indicator of increased metastatic stage, as well as poor OS and DFS. Collectively, these data suggest that SLC39A4 may play an important role in lung cancer pathophysiology.
The EMT plays a significant role during tumour invasion and metastasis 33 , and is characterized by the acquisition of mesenchymal markers such as vimentin or FSP-1, and loss of epithelial cell adhesion molecules, such as E-cadherin 22,34,35 . To further study the effect of SLC39A4 on the aggressive cellular characteristics of NSCLC, we infected A549 cells with lentivirus expressing shSLC39A4 or empty vector control. Notably, knockdown of SLC39A4 induced a more epithelial-like phenotype with increased E-cadherin and decreased FSP-1 expression, suggesting that SLC39A4 promotes EMT in A549 cells. The clinical relevance of these results was confirmed with in vitro migration assays and a mouse model of tumour metastasis, indicating that SLC39A4 plays an important role in NSCLC cell migration.
Many studies have demonstrated that the EMT facilitates resistance to radiotherapy and chemotherapy 36,37 . Cisplatin (cis-diamminedichloro-platinum II) is the primary chemotherapeutic agent used in lung cancer therapy, particularly NSCLC 38 . Because acquired resistance is a common occurrence in NSCLC patients 39 , insights into the molecular mechanisms underlying cisplatin resistance is necessary for the development of novel therapeutic strategies. Consistently, SLC39A4 knockdown cells displayed heightened sensitivity to cisplatin-induced cell death when compared to control cells.
In conclusion, the present study demonstrated that SLC39A4 is overexpressed in NSCLC and correlates with increased staging and diminished patient survival. Moreover, silencing of SLC39A4 induced an epithelial-like phenotype, decreased cancer stem cell marker expression, and increased cisplatin sensitivity. Thus, these findings suggest that SLC39A4 may serve as a prognostic biomarker and putative therapeutic target to enhance chemosensitivity in NSCLC.

Materials and Methods
Public data analysis. Gene expression data were obtained from the Gene Expression Omnibus (GEO) and The Cancer Genome Atlas (TCGA) databases. The meta-analysis included eight datasets encompassing 942 lung cancer patients (Supplementary Table S1), whereas the prognostic analysis included seven datasets with 1623 total patients (Supplementary Table S2). Raw CEL files were downloaded from the GEO database (http://www. ncbi. nlm.nih.gov/geo/), and background correction and data extraction was performed in R software (R version 3.3.0). The meta-analysis was conducted in Review Manager (RevMan Version 5.3, Copenhagen, Denmark), using a random-effects model because the expression data were acquired by different means. Results are presented in forest plots. Cochran χ2 and I 2 analyses were performed to assess heterogeneity among the studies involved.
Reagents. The anti-SLC39A4 antibody was obtained from Abcam (Cambridge, UK). All other antibodies were obtained from Proteintech Group, Inc. (Wuhan, China). Secondary antibodies were purchased from Santa Cruz Biotechnology (Dallas, TX, USA). All other kits and reagents were purchased from the Beyotime Institute of Biotechnology (Shanghai, China). Tissue arrays were from Outdo Biotech Co., Ltd. (Shanghai, China). SLC39A4 immunohistochemistry. Tissue arrays were dewaxed and antigens retrieved using high pressure. Endogenous peroxidases were blocked with 3% hydrogen peroxide for 10 min. After immersion in normal goat serum for 30 min, tissues were incubated with the primary antibody at 4 °C overnight, washed with phosphate-buffered saline (PBS), and then incubated with a biotin-conjugated secondary antibody for 30 min at 37 °C. After washing, the sections were incubated with horseradish peroxidase (HRP) complex for 30 min at 37 °C and visualized using diaminobenzidine (DAB). All immunohistochemical images were obtained under an Olympus BX51 microscope equipped with a 20× , a 40× , or 100× objective lens (Olympus, Tokyo, Japan) and a DP 50 camera (Olympus). Images were processed using DPC controller software (Olympus).
Immunofluorescence. Cultured cells were fixed with 4% paraformaldehyde, washed twice with PBS, and then blocked with PBS containing 10% normal goat serum. Cells were then stained with anti-E-cadherin, anti-vimentin, or anti-FSP1 polyclonal antibody for 30 min at 37 °C, washed twice with PBS, stained with Cy3 (red) or Alexa Flour 488 (green)-conjugated secondary antibody for 30 min at 37 °C, and then washed twice again before imaging. All immunofluorescence images were obtained with an Olympus BX51 microscope equipped with a 20 × or 40 × objective lens (Olympus) and a DP 50 camera (Olympus). Images were processed using DPC controller software (Olympus).
Transwell assays. Cells were cultured in 10-cm plates, and fresh medium was added 18 h before each assay.
Cells were trypsinized, washed twice, resuspended in serum-free medium, and then counted using a haemocytometer. RPMI 1640 with 10% serum was added to the lower wells of the transwell chamber apparatus, and 10,000 cells in 200 μL of serum-free media were added to each upper well. The loaded chamber was incubated for 24 h at 37 °C, at which time the cells on the upper membrane surface were removed by scraping to leave only those that had migrated through the membrane. The transwell membranes were then fixed in methanol, stained with 0.1% crystal violet, and air-dried. The number of cells in each field was quantified and presented as an average from five fields of triplicate wells for each test condition.
Western blotting. Protein samples were resolved by SDS-PAGE on 12% gels and transferred to nitrocellulose membranes, which were then blocked for 1 h at room temperature in Tris-buffered saline (TBS) containing 0.1% Tween 20 and 5% fat-free milk. Primary antibody incubations were performed for 18 h at 4 °C. HRP-conjugated secondary antibody incubations were performed at room temperature for 1 h and visualized with enhanced chemiluminescence (SuperSignal; Pierce, Rockford, IL) or ECL Plus (Amersham Pharmacia Biotech, Buckinghamshire, United Kingdom) substrates according to the manufacturers' instructions.
Cell viability assays. Cell viability was assessed by colony formation assays and cell counting kit (CCK)-8 assay. Briefly, cells were plated at 500 cells per well in a 6-well plate (Corning, Corning, NY, USA) after being treated with different concentrations of cisplatin (0, 1.25, 2.5, or 5 μg/mL). Cells were cultured for 10 days with medium changes every 3 days. Colonies were washed with PBS, fixed in methanol, and then stained with crystal violet. The CCK-8 assay was completed according to the manufacturers' instructions.
Mouse xenograft model. All xenograft experiments were performed in accordance with the guidelines of the Laboratory Animal Ethical Committee at Chengdu Medical College. All experimental protocols were approved by the Laboratory Animal Ethical Committee at Chengdu Medical College (Approval date: 2017.07.20; Approval number: CYYFYEC2016003). An A549 lung metastasis model was established to investigate the effect of SLC39A4 on cell migration and invasion in vivo. Female nude mice (6-8 weeks old, 20-22 g) were purchased from the Experimental Animal Center of Sichuan University (Chengdu, Sichuan Province, China). Briefly, ten nude mice were intravenously inoculated with 5 × 10 5 A549 cells expressing shSLC39A4 or the empty vector control and euthanized 30 days later to count the pulmonary metastatic nodules. The lungs were fixed in Bouin's solution after removal. Formalin-fixed, paraffin-embedded sections of each lung tissue sample were stained routinely with H&E. A subcutaneous A549 model was established to investigate the effect of SLC39A4 on cisplatin treatment and EMT in vivo. Ten nude mice were subcutaneously inoculated with 5 × 10 5 A549 cells expressing shSLC39A4 or an empty vector control. The tumour samples were fixed in Bouin's solution after removal. Formalin-fixed, paraffin-embedded sections of each sample were stained routinely with EMT biomarkers (E-cadherin, N-cadherin, and FSP-1). Histopathological examination and immunofluorescence imaging were completed under an Olympus microscope. Statistical analysis. Each experiment was performed at least three times independently. The association of SLC39A4 expression with clinicopathological parameters was analysed by paired t-test or one-way ANOVA in GraphPad Prism 5 (GraphPad, San Diego, CA, USA). Statistical significance was defined as P < 0.05.