Research Article

The Expression and Prognostic Value of NOX4 in Gastric Cancer

Lu Qiao1,2, Ning Xie1,2, Yuru Bai1,2, Bin Qin1,2, Baicang Zou1,2, Yongquan Shi3, Yan Chen4, Jinhai Wang1,2* and Na Liu1,2*
1Department of Gastroenterology, Second Affiliated Hospital of Xi’an Jiaotong University, China
2Shaanxi Key Laboratory of Gastrointestinal Motility Disorders, Xi’an Jiaotong University, China
3Xijing Hospital of Digestive Diseases, Fourth Military Medical University, Xi'an, China
4Department of Oncology, Xijing Hospital, the Fourth Military MedicalUniversity, Xi'an, China

*Corresponding author: Jinhai Wang, Department of Gastroenterology, Second Affiliated Hospital of Xi’an Jiaotong University, Xi’5 Road, Xi’an, Shaanxi, 710004, China
Na Liu, Department of Gastroenterology, Second Affiliated Hospital of Xi’an Jiaotong University, Xi’5 Road, Xi’an, Shaanxi, 710004, China

Published: 18 Aug, 2018
Cite this article as: Qiao L, Xie N, Bai Y, Qin B, Zou B, Shi Y, et al. The Expression and Prognostic Value of NOX4 in Gastric Cancer. Clin Oncol. 2018; 3: 1505.


Background: Gastric Cancer (GC) is one of the most common malignant tumors worldwide, while modest progress in diagnosis and treatment of gastric cancer have been made.
Methods: Multiple databases including The Cancer Genome Atlas (TCGA) and Gene Expression Omnibus (GEO) were explored to identify the expression pattern and prognostic value of Nicotinamide Adenine Dinucleotide Phosphate Oxidase 4 (NOX4), the main source of reactive oxygen species, in gastric cancer.
Results: We found that NOX4 was upregulated in tumors compared with adjacent non-tumor tissues and overexpression was correlated with including tumor invasion and TNM stage. Furthermore, NOX4 level could be an independent prognostic marker for GC, unacted on the choice of therapy or HER2 expression. Besides, we investigated the potential mechanisms of NOX4 in gastric cancer.
Conclusion: Our findings proved that NOX4 may be a prognostic factor or therapeutic marker for gastric cancer.
Keywords: NOX4; Prognostic values; Gastric cancer


Gastric cancer is the third leading cause of cancer-related deaths in worldwide with an estimated 951,600 cases and 723,100 deaths per year [1,2]. Surgical treatment and adjuvant chemotherapy are the main therapeutic approaches for gastric cancer. However, due to low early detection rate, the great majority of gastric cancer patients presents with advanced-stage disease, resulting in unsatisfactory 5-year survival rates and poor quality of life. Despite tremendous efforts in cancer researches, only modest improvements in early detection and survival are seen. The identification of therapeutic targets and prognostic indicators is therefore highly desirable. NADPH oxidases 4 (NOX4), belongs to NADPH family, which consists of membrane-bound enzymes and can convert oxygen to superoxide or hydrogen peroxide [3]. Reactive Oxygen Species (ROS) stand for a class of extracellular metabolic product, which may react with cellular constituents to generate an array of oxidative lesions compromising genome stability which is critical for long-term cellular homeostasis [4]. NADPH family, a major source of ROS, has been proved to participate in occurrence and development of multiple kinds of tumors. NOX1/2/5 was reported to be strongly correlated with genes associated with cancer cell proliferation and metastasis, while NOX4 and DUO1 played an opposite role in liver cancer [5]. Depletion of NOX1 can induce apoptosis through attenuating the AKT signaling pathway in oral squamous cell carcinoma cell lines [6]. Researches have revealed that NOX4 levels are altered in pancreatic cancer, colorectal cancer, hepatocellular carcinoma, and thyroid cancer [7-10]. NOX4 can proliferate TGF- β -induced chemokinesis in pancreatic adenocarcinoma cells through a ROS/p38 MAPK cascade [11]. In addition, inhibition of NOX4 can induce apoptosis in pancreatic cancer cells via the AKT-ASK1 pathway [12]. Nevertheless, its role in gastric cancer and the underlying mechanisms are not clear. Rapid developments in next-generation sequencing and bioinformatics have provided us a tool to understand molecular characterization of cancers comprehensively and rapidly. In our study, we analyzed the expression level of NOX4 with multiple databases, including one mRNA expression data set from GEO (http://www.ncbi., mRNA sequencing (mRNA-seq) data from TCGA and Oncomine databases. Additionally, we studied the correlation between expression of NOX4 and characteristic features in gastric cancer. We identified its prognostic value with multiple databases and explored the coexpression genes to investigate the underlying mechanisms.

Figure 1

Another alt text

Figure 1
The expression pattern and prognostic value from TCGA and GEO:
(a) The levels of NOX4 in GC from TCGA and GSE79973, the red dots: tumor tissues, the black dots: normal tissues, ****: P value <0.0001; ***: P value <0.001.
(b) Overall survival by NOX in GC, the red line: NOX4 high expression group, the blue line: NOX4 low expression group.

Materials and Methods

Gene expression and characteristics of TCGA dataset
The gene expression quantification including 57288 genes from 374 tumor tissues and 32 adjacent normal tissues of gastric cancer was obtained from TCGA. Their clinical information dataset was downloaded from the TCGA data portal (March 2017). The R package ‘edgeR’ was used to analyzed mRNA-seq data to discover the Differentially Expressed Genes (DEGs). DEGs were selected significantly with the criterion of combined adjusted P<0.001 and logFC >1.5. After excluding 8 cases without clinical survival data, we finally got 367 cases, including 133 females and 234 males. There were 157 patients who were younger than 65 years old, while the other 210 were older than 65. Among them, 49 patients were diagnosed TNM stage I, 110 were stage II, 152 were stage III, 40 were stage IV, and the remaining 16 patients were not clear. The median follow-up period was 440 days (ranging from 0 to 3720 days).
Gene expression data in GEO
The microarray data from GEO (accession number GSE 79973) were explored. There were 10 pairs of GC tissue and adjacent nontumor mucosa in GSE79973. The R package ‘limma’ was used to identify DEGs. DEGs were selected significantly with the criterion of combined adjusted P<0.001 and logFC >1.5.
Oncomine database analysis
Oncomine (, an online microarray database, was utilized to examine the mRNA expression difference of NOX4 between tumor and normal tissues GC. Cancer type, fold change, t-test value, p-value and sample sizes were obtained from studies that showed statistically significant differences.
Kaplan-Meier Plotter database analysis
The KM Plotter (, which is capable to assess the effect of 54,675 genes on survival using 1065 gastric cancer patients, was applied to evaluate the prognostic values of NOX4 in GC. Patients were split into high and low expression group by the median values of mRNA expression. Then the desired probe ID was separately entered into the database. After that, survival analyses were carried out to achieve Kaplan-Meier plots. P-Value <0.01 was considered to indicate a statistically significant result. Survival outcome, HRs, 95% CIs and p-values were summarized from the KM plotter webpage; some representative plots were also displayed.
Linked Omics database analysis
We made use of Linked Omics database (http://www.linkedomics. org) to explore TCGA to discover all the associated genes with NOX4. We select mRNA expression data from TCGA, and Pearson test was utilized. Gene symbols, Pearson correlation coefficient and P value were obtained from the LinkFinder. The association figures were downloaded from LinkFinder.
Functional enrichment analysis
We performed functional enrichment analysis for associated genes through the Database for Annotation, Visualization and Integrated Discovery (DAVID) to uncover biological processes involved in NOX4. Gene Ontology (GO) terms and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways with p-value were considered as significantly enriched. 4 of the maximum hits were presented.
Statistical analyses
The correlation between NOX4 level and characteristic features was analyzed with χ2 test. OS (Overall survival) was measured from the date of the initial operation until the date of death or last followup. Cancer patients were divided into high and low expression group by the median values of mRNA expression and survival curves were created by R 3.3.3. The Kaplan-Meier method and log-rank test were applied to assess OS. Univariate Cox regression proportional hazards analysis was used to assess clinicopathologic characteristics significantly related to OS with HRs and 95% CIs. Multivariate Cox proportional hazards analysis was performed to determine whether the gene expression is an independent prognostic factor. A two-sided p value of <0.05 was considered statistically significant.


The expression pattern of NOX4 in multiple datasets
We utilized one mRNA microarray from GEO and TCGA data to explore the expression of NOX4 in GC. As shown in Figure 1A, NOX4 was significantly up-regulated (logFC=1.618) in 374 tumor tissues than the 32 normal tissues from TCGA (P=1.75E-10). In 10 pairs of tumor and adjacent normal tissues of GSE79973, NOX4 has raised 2.08 times (P=2.18E - 05).In addition, to address the mRNA expression differences of NOX4 between tumor and adjacent non-cancerous tissues in GC, we performed an analysis using the Oncomine database. In four gastric datasets, NOX4 were proved to be overexpressed in tumor than normal tissues. Cho’s data revealed that the level of NOX4 was 2.522 times higher in diffuse gastric adenocarcinoma than in normal gastric tissues, while study from DErrico showed that NOX4 was up-regulated in gastric intestinal type adenocarcinoma, suggesting that in different pathological types of GC, NOX4 was highly expressed. All the statistically significant results were summarized in Supplementary Table 1.
The correlation between NOX4 level and characteristic features
NOX4 has been demonstrated to be up-regulated in GC. We further proceeded to find the association between the expression of NOX4 and clinicopathologic data. We downloaded clinical data from TCGA and discovered that NOX4 level was related to tumor invasion (P=0. 0.00329) and TNM stage (P=0.004948) statistically. The deeper tumor invaded, the higher NOX4 expressed. Meanwhile, GC patients who were diagnosed more advanced-stage; the higher NOX4 was upregulated. The expression pattern of NOX4 was not associated with patient’s age, gender, tumor metastasis and lymph node invasion (Table 1).
The prognostic value of NOX4 in GC
We next investigated whether NOX4 is associated with the prognosis of GC patients from TCGA. Higher NOX4 implied worse Overall Survival (OS) (P=0.01741, HR=1.48 (1.07-2.06)), clearly demonstrating that high NOX4 expression was associated with a shorter survival rate. Survival curve was presented in Figure 1B. Furthermore, univariate and multivariate Cox regression analyses were performed to confirm the possibility that NOX4 could be useful as an independent risk factor for poor prognosis in GC from TCGA. First, we performed univariate analysis and recognized that age, TNM stage and NOX4 expression were correlated with short overall survival time (Table 2). Thus, multivariate analyses were utilized to discover whether high NOX4 expression could imply a poor prognosis independently. As shown in Table 3, NOX4 could be an independent prognostic factor (P=0.01392), independent on age and TNM stage. Besides, we measured the overall survival time in different sub-groups according to treatment and HER2 expression by using Kaplan-Meier Plotter database. 631 GC patients were followed up and the overall survival rate was shown in Figure 2A. Among all these patients, 380 patients chose to have an operation only, while 34 had 5-FU adjuvant therapy. HER2 was upregulated in 153 patients. Meanwhile, HER2 was negative in 195 patients. It was found that the OS was shorter in GC patients with higher NOX4 levels, no matter which therapy methods were chosen or whether HER2 was positive (Figure 2B-2E).
The associated genes with NOX4 in GC
To uncover the underlying mechanisms of NOX4 in GC, we explored the LinkedOmics database to identify all the genes related to NOX4. There were 434 positively correlated significant genes (correlation coefficient >0.5, P value <0.05), while coefficients of association of negatively correlated genes were under 0.5. All the detailed data were shown in Supplementary Table 2. As (Figure 3A) shown, genes with maximum correlation were labeled on the heat map. Moreover, we performed functional enrichment analysis to determine which Biological Processes (BP) and pathways related genes involved in to uncover the possible mechanisms. We detected that associated genes of NOX4 were most likely linked to cell adhesion and extracellular matrix organization (Figure 3B), which were responsible for epithelial-mesenchymal transition and tumor metastasis. Top 4 of maximum hits of Kyoto Encyclopedia of Genes and Genomes (KEGG) was presented in Figure 3C, and the associated genes were annotated to take part in kinds of classical tumor development and metastasis pathways, for example, the PI3K-Akt signaling pathways, focal adhesion, pathways in cancer and ECM-receptor interaction.

Figure 2

Another alt text

Figure 2
Survival analyses of NOX4 in gastric cancer from KM plotter:
(a): Overall survival analysis of NOX4 in gastric cancer patients.
(b,c): Prognosis analysis of NOX4 in patients with surgery alone and 5-FU adjuvant therapy.
(d): Prognosis analyses of NOX4 in HER2 negatively expressed patients.
(e): Prognosis analyses of NOX4 in HER2 positively expressed patients.


NADPH oxidases family, generating ROS by transporting electrons from cytosolic NADPH across biological membranes to molecular oxygen, has been reported to have positive and negative impacts on cellular health and tumor [4]. NOX4, a member of NADPH oxidases family, have been revealed to have different effects in various types of cancers. Silence of NOX4, through pyruvate kinase-M2 isoform, can sensitize human renal carcinoma cells to drug - induced cell death [3]. Paradoxically, increased NOX4 expression was proved to be associated with genes that inhibit hepatocellular carcinoma progression. Our study focused on the levels of NOX4 in GC and its correlation with clinicopathologic data. Besides, we discussed the predicted molecular mechanisms of NOX4 in GC. We explored TCGA and 1microarray (GSE79973) from GEO and discovered that mRNA expression of NOX4 was upregulated in tumor than normal tissues, agreeing with outcomes from Oncomine. Additionally, Oncomine datasets proved to us that in different histological types of GC, NOX4 levels were all over expressed. Multiple databases certified that NOX4 levels were raised significantly in cancerous tissues, implying its function in development of stomach adenocarcinoma. Furthermore, we detected that NOX4 overexpression was co relative with tumor invasion and TNM stage statistically. Fundamentally, patients with higher NOX4 expression had shorter survival time, compared with those with lower levels. Univariate and multivariate regression analyses illustrated that NOX4 could be an unfavorable prognostic biomarker for GC, independent of TNM stage and age. We also analyzed other datasets from KM Plotter to demonstrate that the prognostic value of NOX4 has no relationship with HER2 levels or treatment options, prompting its extensively potential clinical utility. Lee, J developed and validated a prognostic algorithm for GC with a microarray gene expression profiling and NOX4 was a member of gene signature in stage II GC patients [5]. We reported that NOX4 had a significant prognostic role in GC, no matter which stage GC patients were diagnosed with. In general, our findings suggested that NOX4 might be a cancer - promoting gene and have a crucial place in GC occurrence and progression. To elucidate the underlying biological functions of NOX4 in GC, we summarized all the linked genes with NOX4 and found that the positively related genes were enriched in cell adhesion and metastasis. Accumulating evidence has proved that NOX4 can alter cell adhesion in various types of tumors. Silence of NOX4 with siRNA can inhibit GC cell lines MGC-803 and BGC-823 invasion [13], revealing that NOX4 can enhance GC metastasis in vivo and identifying with our results. Treatment of papillary thyroid cancer cells with NOX4-targeted siRNA down regulates induced thyroid iodide-metabolizing gene repression, inhibiting cancer cell invasion [14]. NOX4 inhibition leads to loss of cell mobility and down-regulation of epithelial mesenchymal transformation in hepatic and neuroblastoma tumors [15]. We discovered that the most likely pathway NOX4 involved in was PI3K-Akt signaling pathway with bioinformatics prediction approaches. Evidence revealed that in Non-Small Cell Lung Cancer (NSCLC), there was a mutual positive regulation between NOX4 and PI3K/Akt signaling. Overexpression of NOX4 could enhance cell proliferation and invasion, and produce larger tumors, shorter survival time, and more lung metastasis in nude mice than control cells in NSCLC [16]. The product of NOX4, superoxide ROS, was reported to be key molecule in Akt-VEGF pathway in melanoma [17]. Moreover, inhibition of NOX4 can induce pancreatic cancer cells apoptosis through the Akt pathway [18,19,12]. Taken together, our predictions are in accord with previous researches, implying that NOX4 may proliferate gastric cancer cell growth and metastasis through PI3K-Akt signaling pathway. There are few studies focusing on the role of NOX4 in GC. One determined depletion of NOX4 could inhibit GC cell invasion in vivo, and another recognized NOX4 was a member of GC prognostic gene signature. Our research took advantage of multiple datasets with detailed and complete clinical features to identify biological functions of NOX4. However, there is need to explore pathogenesis and mechanisms of NOX4 in GC comprehensively. We further plan to collect clinical GC cases to build our cohort to validate its prognostic significance and design experiments in vivo and in vitro to certify our bioinformatics findings.

Figure 3

Another alt text

Figure 3
Analysis of the associated genes with NOX4 in gastric cancer:
(a) The heatmap ofNOX4 positively associated genes. Functional enrichment analysis of the associated genes of NOX4 in GC.
(b) Biological progress enrichment of NOX4 associated genes.
(c) KEGG pathway enrichment of NOX4 associated genes.


In conclusion, this study uses high throughput data to clarify expression pattern of NOX4 in GC and demonstrate that NOX4 mRNA expression is an unfavorable prognostic biomarker in gastric cancer patients. It sheds a new light on better understanding the complex and crucial role of NOX4 in gastric cancer and uncovers a new target for GC diagnosis and therapy.


  1. Auer S, Rinnerthaler M, Bischof J, Streubel MK, Breitenbach-Koller H, Geisberger R, et al. The human NADPH oxidase, Nox4, regulates cytoskeletal organization in two cancer cell lines, HepG2 and SH-SY5Y. Front Oncol. 2017;7:111.
  2. Azouzi N, Cailloux J, Cazarin JM, Knauf JA, Cracchiolo J, Al Ghuzlan A, et al. NADPH oxidase NOX4 is a critical mediator of BRAFV600E - induced down regulation of the sodium/iodide symporter in papillary thyroid carcinomas. Antioxid Redox Signal. 2017;26(15):864-77.
  3. Edderkaoui M, Nitsche C, Zheng L, Pandol SJ, Gukovsky I, Gukovskaya AS. NADPH oxidase activation in pancreatic cancer cells is mediated through Akt-dependent up-regulation of p22phox. J Biol Chem. 2013;288(51):36259.
  4. Eun HS, Cho SY, Joo JS, Kang SH, Moon HS, Lee ES, et al. Gene expression of NOX family members and their clinical significance in hepatocellular carcinoma. Sci Rep. 2017;7:11060.
  5. Gao X, Sun J, Huang C, Hu X, Jiang N, Lu C. RNAi-mediated silencing of NOX4 inhibited the invasion of gastric cancer cells through JAK2/STAT3 signaling. Am J Transl Res. 2017;9(10):4440-9.
  6. Govindarajan B, Sligh JE, Vincent BJ, Li M, Canter JA, Nickoloff BJ, et al. Overexpression of Akt converts radial growth melanoma to vertical growth melanoma. J Clin Invest. 2007;117(3):719-29.
  7. Ito K, Ota A, Ono T, Nakaoka T, Wahiduzzaman M, Karnan S, et al. Inhibition of Nox1 induces apoptosis by attenuating the AKT signaling pathway in oral squamous cell carcinoma cell lines. Oncol Rep. 2016;36(5):2991-8.
  8. Lee J, Sohn I, Do IG, Kim KM, Park SH, Park JO, et al. Nanostring-based multigene assay to predict recurrence for gastric cancer patients after surgery. PLoS One. 2014;9(3):e90133.
  9. Lin XL, Yang L, Fu SW, Lin WF, Gao YJ, Chen HY, et al. Overexpression of NOX4 predicts poor prognosis and promotes tumor progression in human colorectal cancer. Oncotarget. 2017;8(20):33586-600.
  10. Liu Z, Tu K, Wang Y, Yao B, Li Q, Wang L, et al. Hypoxia accelerates aggressiveness of hepatocellular carcinoma cells involving oxidative stress, epithelial - mesenchymal transition and non - canonical hedgehog signaling. Cell Physiol Biochem. 2017;44(5):1856-68.
  11. Nitsche C, Edderkaoui M, Moore RM, Eibl G, Kasahara N, Treger J, et al. The phosphatase PHLPP1 regulates Akt2, promotes pancreatic cancer cell death, and inhibits tumor formation. Gastroenterology. 2012;142(2):377-87.
  12. Shanmugasundaram K, Nayak BK, Friedrichs WE, Kaushik D, Rodriguez R, Block K. NOX4 functions as a mitochondrial energetic sensor coupling cancer metabolic reprogramming to drug resistance. Nat Commun. 2017;8:997.
  13. Thomasz L, Oglio R, Salvarredi L, Perona M, Rossich L, Copelli S, et al. Regulation of NADPH oxidase NOX4 by delta iodolactone (IL-δ) in thyroid cancer cells. Mol Cell Endocrinol. 2018;470:115-26.
  14. Torre LA, Bray F, Siegel RL, Ferlay J, Lortet-Tieulent J, Jemal A. Global cancer statistics, 2012. CA Cancer J Clin. 2015;65(2):87-108
  15. Torre LA, Siegel RL, Ward EM, Jemal A. Global cancer incidence and mortality rates and trends--An update. Cancer Epidemiol Biomarkers Prev. 2016;25(1):16-27.
  16. Vaquero EC, Edderkaoui M, Pandol SJ, Gukovsky I, Gukovskaya AS. Reactive oxygen species produced by NAD(P)H oxidase inhibit apoptosis in pancreatic cancer cells. J Biol Chem. 2004;279(33):34643-54.
  17. Weyemi U, Redon CE, Parekh PR, Dupuy C, Bonner WM. NADPH Oxidases NOXs and DUOXs as putative targets for cancer therapy. Anticancer Agents Med Chem. 2013;13(3):502-14.
  18. Witte D, Bartscht T, Kaufmann R, Pries R, Settmacher U, Lehnert H, et al. TGF - β1 - induced cell migration in pancreatic carcinoma cells is RAC1 and NOX4 - dependent and requires RAC1 and NOX4 - dependent activation of p38 MAPK. Oncol Rep. 2017;38(6):3693-701.
  19. Zhang C, Lan T, Hou J, Li J, Fang R, Yang Z, et al. NOX4 promotes nonsmall cell lung cancer cell proliferation and metastasis through positive feedback regulation of PI3K/Akt signaling. Oncotarget. 2014;5(12):4392-405.