Identification and Verification of Biomarker in Clear Cell Renal Cell Carcinoma via Bioinformatics and Neural Network Model

Liu, Bin; Xiao, Yu; Li, Hao; Zhang, Ai-li; Meng, Ling-bing; Feng, Lu; Zhao, Zhi-hong; Ni, Xiao-chen; Fan, Bo; Zhang, Xiao-yu; Zhao, Shi-bin; Liu, Yi-bo

doi:https://doi.org/10.1155/2020/6954793

BioMed Research International

On this page

Abstract Introduction Methods Results Discussion Conclusion Data Availability Disclosure Conflicts of Interest Authors’ Contributions Acknowledgments References Copyright Related Articles

Research Article | Open Access

Volume 2020 | Article ID 6954793 | https://doi.org/10.1155/2020/6954793

Identification and Verification of Biomarker in Clear Cell Renal Cell Carcinoma via Bioinformatics and Neural Network Model

Bin Liu,¹Yu Xiao,²Hao Li,³Ai-li Zhang,¹Ling-bing Meng,⁴Lu Feng,⁵Zhi-hong Zhao,¹Xiao-chen Ni,¹Bo Fan,¹Xiao-yu Zhang,¹Shi-bin Zhao,⁶and Yi-bo Liu¹

Academic Editor: Paul Harrison

Received20 Dec 2019

Revised10 May 2020

Accepted22 May 2020

Published17 Jun 2020

Abstract

Background. Clear cell renal cell carcinoma (ccRCC) is the most common subtype of kidney cancer, which represents the 9th most frequently diagnosed cancer. However, the molecular mechanism of occurrence and development of ccRCC is indistinct. Therefore, the research aims to identify the hub biomarkers of ccRCC using numerous bioinformatics tools and functional experiments. Methods. The public data was downloaded from the Gene Expression Omnibus (GEO) database, and the differently expressed genes (DEGs) between ccRCC and normal renal tissues were identified with GEO2R. Protein-protein interaction (PPI) network of the DEGs was constructed, and hub genes were screened with cytoHubba. Then, ten ccRCC tumor samples and ten normal kidney tissues were obtained to verify the expression of hub genes with the RT-qPCR. Finally, the neural network model was constructed to verify the relationship among the genes. Results. A total of 251 DEGs and ten hub genes were identified. AURKB, CCNA2, TPX2, and NCAPG were highly expressed in ccRCC compared with renal tissue. With the increasing expression of AURKB, CCNA2, TPX2, and NCAPG, the pathological stage of ccRCC increased gradually (). Patients with high expression of AURKB, CCNA2, TPX2, and NCAPG have a poor overall survival. After the verification of RT-qPCR, the expression of hub genes was same as the public data. And there were strong correlations between the AURKB, CCNA2, TPX2, and NCAPG with the verification of the neural network model. Conclusion. After the identification and verification, AURKB, CCNA2, TPX2, and NCAPG might be related to the occurrence and malignant progression of ccRCC.

1. Introduction

Worldwide, renal cell carcinoma (RCC) represents the 9th most frequently diagnosed cancer in men and the 10th in women, accounting for 5% and 3% of all oncological diagnoses, respectively, [1]. According to the most updated data provided by the World Health Organization, there are more than 140 000 RCC-related deaths yearly, with RCC ranking as the 13th most common cause of cancer death worldwide. Age and gender factors are closely related to the risk of RCC. Other potential risk factors include lifestyle, complications, drugs, and environmental factors [2]. The diagnosis and management of RCC have changed remarkably rapidly in the past decades through the unremitting efforts of generation after generation of researchers. Despite progression in cancer control and survival, locally advanced disease and distant metastases are still diagnosed in a notable proportion of patients. Nevertheless, uncertainties, controversies, and research questions remain [3]. Further advances are expected from the diagnosis, treatment, and prognosis evaluation.

RCC is a group of heterogeneous tumors with different genetic and molecular changes, clear cell renal cell carcinoma (ccRCC), papillary RCC (type 1 and type 2), and chromophobe RCC are the most common solid RCC, accounting for 85.90% of all malignant RCC [4]. Among them, ccRCC is the most common subtype of kidney cancer. Both sporadic and inherited RCC are usually associated with structural changes in the short arm of chromosome 3 [5]. In addition, the occurrence of RCC is related to multiple gene alterations, such as VHL, PBRM1, BAP1, SETD2, TCEB1, and KDM5C [3]. Although our understanding of the biology of RCC has improved, surgery is still the main treatment method of RCC. Drugs and comprehensive therapies, identification of new target pathways, and optimal sequencing and combination of existing targeted drugs are areas that are worth researching [6].

Bioinformatics tools can screen differentially expressed genes (DEGs) between diseased and normal tissues [3, 7, 8]. These DEGs are related to the pathological stage, lesion grade, and prognosis of patients. Zou et al. used microarray technology to identify the hub genes between malignant glioblastoma and normal brain tissue and obtained the important targets related to brain glioma [9]. Through a series of bioinformatics analysis, Meng et al. concluded that TPM2 may be an important biomarker for the occurrence and development of atherosclerosis [10].

Therefore, this study will use bioinformatics technology to explore the gene molecular markers of abnormal expression during the occurrence of ccRCC and discuss the related potential mechanisms. These differentially expressed genes may affect the initiation and malignant progression of ccRCC and can be used as targets for diagnosis and treatment.

2. Material and Methods

2.1. Download Public Data

The Gene Expression Omnibus (GEO) database (http://www.ncbi.nlm.nih.gov/geo) is the largest, most comprehensive, and publicly available source of gene expression data.

On 20 December, 2019, we set key words “(clear cell renal cell carcinoma) AND (normal kidney)” to detect the datasets, using a filter of “expression profiling by array.” The inclusion criteria includes a diagnosis of clear cell renal cell carcinoma (data from papillary renal cell carcinoma diagnoses were excluded), the dataset including the gene expression profile of normal kidney (datasets which were composed of only tumor data were excluded), a sample number of more than forty per dataset (samples of less than forty were excluded), data from Homo sapiens (data from other species were excluded), and a series entry type, expression profiling by array (data using methylation profiling only by array were excluded).

Therefore, GSE105288 (GPL10558, Illumina HumanHT-12 V4.0 expression beadchip) and GSE66272 (GPL570(HG-U133_Plus_2) Affymetrix Human Genome U133 Plus 2.0 Array) were obtained from the GEO database. A total of 44 samples, including 35 ccRCC tissues and 9 normal renal tissues, were selected from GSE105288. A total of 53 samples, including 26 ccRCC tissues and 27 normal renal tissues, were selected from GSE66272.

2.2. Differentially Expressed Genes (DEGs) between Normal and PCRC

GEO2R (http://www.ncbi.nlm.nih.gov/geo/geo2r) could import data of the GEO database into the R language and perform differential analysis, essentially through the following two R packages, including limma packages and GEOquery. Therefore, through the GEO2R tool, DEGs were identified between the normal and ccRCC groups. The values < 0.001 was defined as significant. The gene symbols were necessary. SangerBox (https://shengxin.ren), one open tool, was used to draw volcano maps. Venn diagrams were delineated using FunRich software (http://www.funrich.org/), which would visualize common DEGs shared between GSE105288 and GSE66272.

2.3. GO and KEGG Analysis

One online tool, DAVID (https://david.ncifcrf.gov/home.jsp) (version 6.8, Maryland, America), was applied to carry out the functional annotation for DEGs. Gene Ontology (GO) [11] generally performs enrichment analysis of genomes. And there are mainly cellular components (CC), biological processes (BP), and molecular functions (MF) in the GO analysis. Kyoto Encyclopedia of Genes and Genomes (KEGG) (https://www.kegg.jp/) [12] is a comprehensive database of genomic, chemical, and systemic functional information. Therefore, DAVID was used to make the analysis of GO and KEGG. The Biological Networks Gene Oncology tool (BiNGO) (version 3.0.3) was used to analyze and visualize the DEGs’ cellular component, biological process, and molecular function [13].

2.4. Protein-Protein Interaction (PPI) Network

The common DEGs, shared between GSE105288 and GSE66272, were converted into differently expressed proteins. The STRING (Search Tool for the Retrieval of Interacting Genes) online database (http://string-db.org) could construct the PPI network, which was visualized by Cytoscape (version 2.8) [14].

2.5. Significant Module and Hub Genes

Molecular Complex Detection tool (MCODE) (version 1.5.1) [15], an open plug-in of Cytoscape, was performed to identify tested most significant module from the PPI network, and the criteria was that the maximum , MCODE 5, , , and . Then, cytoHubba [16], a free plug-in of Cytoscape, was applied to authorize the hub genes, when the .

2.6. Expression Analysis of Hub Genes

The clustering analysis of expression level of hub genes was performed using heat maps based on the GSE105288 and GSE66272. Also, the expression profiles of hub genes in the ccRCC and normal groups were analyzed using Gene Expression Profiling Interactive Analysis (GEPIA, http://gepia.cancer-pku.cn/) [17].

2.7. Effect of Hub Gene Expression for Pathological Stage and Overall Survival

The effect of hub gene expression for pathological stage and overall survival was analyzed by the GEPIA. Finally, the correlation and linear regression analyses between AURKB, CCNA2, TPX2, and NCAPG were performed. And the receiver operator characteristic (ROC) curve analysis was performed to test the sensitivity and specificity of the hub gene expression for the diagnosis of ccRCC. The SPSS software (version 21.0; IBM; New York; America) was used to conduct all the statistical analysis. A value < 0.05 was defined as statistically significant.

2.8. RT-qPCR Assay

A total of 10 ccRCC participates were recruited. After surgery, 10 ccRCC tumor samples from ccRCC patients and 10 adjacent normal kidney tissues samples were obtained. The research conformed to the Declaration of Helsinki and was authorized by the Human Ethics and Research Ethics Committees of the Fourth Hospital of Hebei Medical University. The informed consents were obtained from all participates.

Total RNA was extracted from 10 ccRCC tumor samples and 10 adjacent normal kidney tissue samples by the RNAiso Plus (TRIzol) kit (Thermo Fisher, Massachusetts, America and reverse transcribed to cDNA. RT-qPCR was performed using a Light Cycler® 4800 System with specific primers for genes. Table 1 presents the primer sequences used in the experiments. The RQ values (2^−ΔΔCt, where Ct is the threshold cycle) of each sample were calculated and are presented as fold change in the gene expression relative to the control group. GAPDH was used as an endogenous control.

2.9. The Confirmation Using The Cancer Genome Atlas (TCGA) Data

The gene expression dataset of ccRCC in the TCGA was downloaded using the University of California Santa Cruz (UCSC) Xena (https://xena.ucsc.edu/welcome-to-ucsc-xena/). There were a total of 944 samples including 537 ccRCC samples and 407 normal renal samples. The IlluminaHiSeq was selected as gene expression RNAseq in the research. In addition, the gene expression levels of VEGFA, AURKB, CCNA2, MCM2, MCM7, SMC4, TPX2, SLC2A1, MCM5, and NCAPG between ccRCC and normal renal samples were compared using the one-way ANOVA.

Furthermore, the effect of the gene expression of VEGFA, AURKB, CCNA2, MCM2, MCM7, SMC4, TPX2, SLC2A1, MCM5, and NCAPG on overall survival was analyzed by using the TCGA data.

2.10. The Construction of Neural Network Model

The training group was randomly divided into the calibration data and training data according to the proportion of 3 : 7. There were 6 samples in the calibration data, and 20 samples in the training data. We used MATLAB (version 8.3) to accomplish the normalization processing of variable values, network simulation, network training, and network initialization. The number of input neurons in the input layer is the same as the number of input variables, and the number is two. The hidden layer is designed as 1 layer, and the output layer is also designed for 1 layer. One output variable is the intima-media thickness. When training to 2000 steps after repeated training, the falling gradient is 0, and the training speed is uniform [10]. At the same time, the training , and the (relativity) value reached 0.9906.

3. Results

3.1. DEGs between Normal Kidney and ccRCC Samples

There are plenty of DEGs on all chromosomes between the ccRCC and normal samples (Figure 1(a)). One volcano plot presents the DEGs in the GSE105288 (Figure 1(b)), and another volcano plot presents the DEGs in the GSE66272 (Figure 1(c)). The Venn diagram manifested that a total of 251 DEGs exist in the two datasets (GSE105288 and GSE66272) simultaneously (Figure 2(a)).

(a)

(b)

(c)

(a)

(b)

3.2. Construction of the PPI Network

After construction of the PPI network for the common DEGs, there are 189 nodes and 406 edges in the PPI network (Figure 2(b)).

3.3. The Functional Enrichment Analysis of DEGs via GO and KEGG

GO analysis manifested that variations in DEGs related with biological processes (BP) were significantly enriched in canonical glycolysis, glycolytic process, peptidyl-proline hydroxylation to 4-hydroxy-L-proline, angiogenesis, cell proliferation, fructose metabolic process, cell division, DNA replication initiation, regulation of insulin secretion, mitotic nuclear division, regulation of actin cytoskeleton organization, carbohydrate phosphorylation, glycine catabolic process, glycine decarboxylation via glycine cleavage system, cellular response to hypoxia, and so on (Figure 3(a)). The variations in DEGs related with cellular components (CC) were significantly enriched in the basolateral plasma membrane, endoplasmic reticulum lumen, membrane, extracellular exosome, melanosome, cytosol, MCM complex, and so on (Figure 3(b)). The variations in the DEGs related with molecular functions (MF) were significantly enriched in protein binding, procollagen-proline 4-dioxygenase activity, identical protein binding, anion transmembrane transporter activity, apolipoprotein binding, L-ascorbic acid binding, actin binding, and so on (Figure 3(c)). The KEGG pathway enrichment analysis showed that the top pathways related with DEGs were fructose and mannose metabolism, carbon metabolism, collecting duct acid secretion, DNA replication, HIF-1 signaling pathway, and so on (Figure 3(d)).

(a)

(b)

(c)

(d)

(e)

The BP analysis for DEGs is presented in Figure 3(e) via the BiNGO software (Figure 3(e)). The CC analysis for DEGs is presented in Figure 4(a) via the BiNGO software (Figure 4(a)). The MF analysis for DEGs is presented in Figure 4(b) via the BiNGO software (Figure 4(b)).

(a)

(b)

3.4. Significant Module Network and Identification of Hub Genes

A significant module was screened from the PPI network, and one module network consisted of 14 nodes and 84 edges (Figure 5(a)). Another module network consisted of 15 nodes and 32 edges (Figure 5(b)). And ten hub genes were identified, including VEGFA, AURKB, CCNA2, MCM2, MCM7, SMC4, TPX2, SLC2A1, MCM5, and NCAPG (Figure 5(c)).

(a)

(b)

(c)

(d)

3.5. Difference of Expression of Hub Genes between ccRCC and Normal Kidney Samples

Hierarchical clustering allowed for simple differentiation of ccRCC tissues from normal colorectal tissues via the expression levels of hub genes in the GSE105288 and GSE66272 datasets. One heat map showed that the expressions of all the hub genes were higher in the ccRCC samples than the normal samples in the GSE105288 (Figure 6(a)). Another heat map also showed that the expressions of all the hub genes were higher in the ccRCC samples than the normal samples in the GSE66272 (Figure 6(b)). Through the GEPIA analysis, the expressions of hub genes in the ccRCC patients were higher than the normal individuals (Figure 7(a)).

(a)

(b)

(a)

(b)

3.6. Association between Hub Gene Expression, Pathological Stage, and Overall Survival

The results of GEPIA manifested that the expressions of AURKB, CCNA2, TPX2, and NCAPG were significantly positively related with pathological stage (), while the expressions of VEGFA, MCM2, MCM7, SMC4, SLC2A1, and MCM5 were not (Figure 7(b)). The results showed that the expression level of VEGFA was not related with the overall survival of ccRCC patients (, Figure 8(a)). The overall survival analysis showed that ccRCC patients with high expression levels of AURKB (Figure 8(b)) and CCNA2 (Figure 8(c)) had poorer overall survival times than those with low expression levels (). The expression level of MCM2 was not related with the overall survival of ccRCC patients (, Figure 8(d)). The expression level of MCM7 was not related with the overall survival of ccRCC patients (, Figure 8(e)). The expression level of SMC4 was not related with the overall survival of ccRCC patients (, Figure 8(f)). The overall survival analysis showed that ccRCC patients with high expression levels of TPX2 had poorer overall survival times than those with low expression levels (, Figure 9(a)). The expression levels of SLC2A1 (Figure 9(b)) and MCM5 (Figure 9(c)) were not related with the overall survival of ccRCC patients (). The overall survival analysis showed that ccRCC patients with high expression levels of NCAPG had poorer overall survival times than those with low expression levels (, Figure 9(d)).

(a)

(b)

(c)

(d)

(e)

(f)

(a)

(b)

(c)

(d)

3.7. The Interaction Analysis among the Hub Genes

Through the Pearson correlation test, heat maps manifested that there were strong correlations among hub genes in the GSE105288 (Figure 10(a)) and GSE66272 (Figure 10(b)) datasets. The correlation between AURKB, CCNA2, TPX2, and NCAPG was strong (Figures 10(c)–10(h)).

(a)

(b)

(c)

(d)

(e)

(f)

(g)

(h)

3.8. ROC Analysis

To identify accurate thresholds for hub genes to predict ccRCC, we constructed ROC. The expression of all hub genes was associated with a diagnosis of ccRCC. The ROC curve of AURKB in the GSE105288 was shown in Figure 11(a). The ROC curve of CCNA2 in the GSE105288 was shown in Figure 11(b). The ROC curve of TPX2 in the GSE105288 was shown in Figure 11(c). The ROC curve of NCAPG in the GSE105288 was shown in Figure 11(d). The ROC curve of AURKB in the GSE66272 was shown in Figure 11(e). The ROC curve of CCNA2 in the GSE66272 was shown in Figure 11(f). The ROC curve of TPX2 in the GSE66272 was shown in Figure 11(g). The ROC curve of NCAPG in the GSE66272 was shown in Figure 11(h).

(a)

(b)

(c)

(d)

(e)

(f)

(g)

(h)

3.9. Results of RT-qPCR Analysis

As presented in Figure 12, the relative expression levels of VEGFA, AURKB, CCNA2, MCM2, MCM7, SMC4, TPX2, SLC2A1, MCM5, and NCAPG were significantly higher in the ccRCC samples, compared with the normal kidney tissues groups. The result demonstrated that VEGFA, AURKB, CCNA2, MCM2, MCM7, SMC4, TPX2, SLC2A1, MCM5, and NCAPG might be considered biomarkers for ccRCC.

(a)

(b)

(c)

(d)

(e)

(f)

(g)

(h)

(i)

(j)

3.10. The Verification by TCGA

According to the above expression analysis, VEGFA, AURKB, CCNA2, MCM2, MCM7, SMC4, TPX2, SLC2A1, MCM5, and NCAPG were markedly upregulated in ccRCC tumor samples compared with the normal renal samples. After confirmation using the TCGA data, these gene expression levels in the ccRCC samples were also significantly higher than the normal renal samples (, Figure 13).

The results showed that the expression level of VEGFA (), MCM2 (), MCM7 (), SMC4 (), SLC2A1 (), and MCM5 () was not related with the overall survival of ccRCC patients. The overall survival analysis showed that ccRCC patients with high expression levels of AURKB (), CCNA2 (), TPX2 (), and NCAPG () had poorer overall survival times than those with low expression levels (Figure 14).

3.11. The Neural Network Prediction Model between AURKB, CCNA2, TPX2, and NCAPG

The mean squared error is <0.05 (Figure 15(a)). The relativity of training is 0.9906. The relativity of validation is 0.99768. The relativity of test is 0.93812. And the relativity of all procedure is 0.97977 (Figure 15(b)). Through verifying the predicted value of the data against the actual value, we found that there are only small differences in the comparison chart of training results (Figure 15(c)) and error analysis diagram (Figure 15(d)). Based on the above result, we could speculate that there were strong correlations between AURKB, CCNA2, TPX2, and NCAPG.

(a)

(b)

(c)

(d)

(e)

(f)

Through the cubic spline interpolation algorithm, we find the high-risk warning indicator of TPX2: and . The three-dimensional stereogram could present the warning range well (Figure 15(e)). The plane graph is also shown (Figure 15(f)).

4. Discussion

RCC is a common disease in the urinary system. According to the statistics of the World Health Organization in 2018, its incidence is second only to prostate cancer and bladder cancer and is increasing year by year [1]. Although many genes are considered potential therapeutic targets and prognostic predictors of RCC, the molecular mechanism of the occurrence and development of RCC remains controversial.

With the continuous progress of science, microarray technology, as a special data mining method, is very influential at present. This revolutionary technology transforms traditional molecular research from a situation that relies on personal experience and subjective guesses to a more objective science [18–20].

In this paper, bioinformatics tools are used to mine the targeted biomarkers of ccRCC. The results showed that AURKB, CCNA2, TPX2, and NCAPG were highly expressed in ccRCC compared with renal tissue. With the increasing expression of AURKB, CCNA2, TPX2, and NCAPG, the pathological stage of ccRCC increased gradually. Compared with the individuals with low expression of AURKB, CCNA2, TPX2, and NCAPG, patients with high expression of AURKB, CCNA2, TPX2, and NCAPG have a poor overall survival.

Aurora kinase B (AURKB) is a serine/threonine kinase that participates in the regulation of chromosome arrangement and segregation by binding to microtubules [21]. Numerous studies have found that the overexpression of AURKB exists in a variety of cancer cell lines [22–24]. Sorrentino et al. found that AURKB is highly expressed in thyroid carcinoma, and its expression level is related to malignant degree. The block of AURKB expression or by using an inhibitor of Aurora kinase activity significantly reduced the growth of thyroid carcinoma cells [23]. Katayama et al. have similar findings in colorectal cancer [25], Smith et al. in lung cancer [22], and Chieffi et al. in prostate cancer [24]. Abnormal mitotic regulation can induce the production of aneuploid cells and act as a driving role in the process of malignant progression, while serine/theronine protein kinases of the Aurora family genes play a critical role in the regulation of key cell cycle processes. The abnormal expression of AURKB can produce malignant and invasive aneuploid cells. This further indicates that AURKB is related to tumorigenesis [26, 27]. With the discovery of abnormal expression of AURKB in cancer cells, researchers realized that it may become a new target for cancer treatments. At present, many AURKB inhibitors have been developed, including AZD1152, AT9283, VX-680/MK-0457, PHA-680632, AMG-900, PHA-739358, and CYC-116, and some of them have entered clinical trials [28].

The proteins encoded by CyclinA2 (CCNA2) belong to a highly conserved cyclin family, which promotes cell transformation by binding and activating cyclin-dependent kinases (CDKs) through G1/S and G2/M [29]. Previous studies have found that the overexpression of CyclinA2 occurs in lung cancer [30, 31], breast cancer [32, 33], colorectal cancer [34], and other tumors and related to poor prognosis of cancer patients. Aaltomaa et al. found that CyclinA2 was expressed in the cytoplasm of RCC but not in the normal tissue near the tumor, and the overexpression of CyclinA2 was related to the survival time of patients with RCC, suggesting that it may be a prognostic indicator of RCC [35]. The increase of the CyclinA2 expression is related to the uncontrolled and accelerated cell cycle, which leads to gene amplification and chromosome ectopia. Gopinathan et al. found that knockout CyclinA2 in mice can inhibit tumorigenesis [36]. Liang et al. found that the increased expression of sclerostin domain-containing protein1 (SOSTDC1) can inhibit CyclinA2, while SOSTDC1 can inhibit tumor growth [37]. CyclinA2 can not only be used as a predictor of prognosis and survival in patients with RCC but also has great potential in cancer treatment.

TPX2 microtubule nucleation factor (TPX2) encodes a microtubule-associated protein that activates cell cycle kinase called Aurora A and regulates mitotic spindles. The overexpression of TPX2 is related to the genesis of different cancers and is closely related to chromosome instability. The uncontrolled expression of TPX2 may eventually become the driving force of cancer development by inducing aneuploidy [38]. Zhang et al. found that compared with human bronchial epithelial cells (16HBE), TPX is overexpressed in malignant transformed 16HBE cells(16HBE-C) through anti-benzo[a]pyrene-trans-7,8-dihydrodiol-9,10-epoxide, in which TPX2 RNA interference (RNAi) can lead to S-phase arrest, inhibit cell proliferation, and induce cell apoptosis. TPX2 is tyrosine phosphorylated in malignant transformed 16HBE-C, and this phosphorylation may be involved in the malignant proliferation of cancer cells [39]. Ma et al. found that the level of the TPX2 protein in normal bronchial epithelium and alveoli was very low, while the level of TPX2 protein increased gradually in squamous metaplasia, dysplasia, and carcinoma in situ and invasive tumor. The immunohistochemical labeling index of TPX2 was related to the degree of differentiation, stage, and lymph node metastasis of lung squamous cell carcinoma, and the overexpression of TPX2 was significantly correlated with the decrease of 5-year survival rate [40]. Similar results were found in a variety of cancers, such as colorectal cancer [41], cervical cancer [42], and prostate cancer [43]. The expression of TPX2 in RCC was significantly higher than that in normal renal tissue, and it was related to tumor size, histological grade, tumor stage, and poor prognosis [44–47]. This may be due to the significant upregulation of TPX2 in RCC tissues, thus increasing the proliferation and invasive ability of renal cancer cells. From this point of view, TPX2 can not only become a target for RCC treatment but also play a role as an independent prognostic factor of RCC.

Non-SMC condensin I complex subunit G (NCAPG) coding a condensed protein complex subunit is responsible for chromosome condensation and stabilization during mitosis and meiosis [48]. In recent years, there are more and more studies on the abnormal expression of NCAPG in prostate cancer [49], lung cancer [50], breast cancer [51], and other cancers. In the study of Liu et al., NCAPG was found to be overexpressed in hepatocellular carcinoma compared with the adjacent normal tissue, and high levels of NCAPG expression were found to significantly correlate with recurrence, the time of recurrence, metastasis, differentiation, and TNM stage. The knockdown of NCAPG expression also inhibited tumor cell migration and the cell invasive capacity in vitro [52]. Through genome-wide functional knockout screen, Wang et al. believe that NCAPG is a necessary clinical-related target for the growth of hepatocellular carcinoma cells [53]. Ai et al. found that microRNA-181c (miR-181c) inhibits cancer by downregulating the expression of NCAPG, affecting the infiltration, migration, proliferation, and apoptosis of hepatoma cells [54]. In the study of Arai et al., microRNA-99a-3p downregulated the expression of NCAPG, thereby inhibiting cancer cell invasion in castration-resistant prostate cancer [49]. In conclusion, NCAPG represents a promising novel target and a prognostic biomarker for clinical management.

However, this study also has some shortcomings. Although the core genes screened in this study may play an important role in the occurrence of ccRCC, more clinical samples and patient prognosis information are still needed for verification.

5. Conclusion

To sum up, 251 differentially expressed genes and 10 hub genes (especially AURKB, CCNA2, TPX2, and NCAPG) were screened from ccRCC and normal renal tissues by microarray technology, which could be used as diagnostic and therapeutic biomarkers for ccRCC. AURKB, CCNA2, TPX2, and NCAPG which might be related to the occurrence and malignant progression of ccRCC.

Data Availability

The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.

Disclosure

The authors ensure that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.

Conflicts of Interest

The authors declare that they have no conflicts of interests.

Authors’ Contributions

Ai-li Zhang and Ling-bing Meng contributed equally to this work.

Acknowledgments

We gratefully thank the Department of Urinary Surgery in The Fourth Hospital of Hebei Medical University for the technical assistance.

References

R. L. Siegel, K. D. Miller, and A. Jemal, “Cancer statistics, 2018,” CA: a Cancer Journal for Clinicians, vol. 68, no. 1, pp. 7–30, 2018.
View at: Publisher Site | Google Scholar
U. Capitanio, K. Bensalah, A. Bex et al., “Epidemiology of renal cell carcinoma,” European Urology, vol. 75, no. 1, pp. 74–84, 2019.
View at: Publisher Site | Google Scholar
U. Capitanio and F. Montorsi, “Renal cancer,” The Lancet, vol. 387, no. 10021, pp. 894–906, 2016.
View at: Publisher Site | Google Scholar
B. Shuch, A. Amin, A. J. Armstrong et al., “Understanding pathologic variants of renal cell carcinoma: distilling therapeutic opportunities from biologic complexity,” European Urology, vol. 67, no. 1, pp. 85–97, 2015.
View at: Publisher Site | Google Scholar
R. Srinivasan, C. J. Ricketts, C. Sourbier, and W. M. Linehan, “New strategies in renal cell carcinoma: targeting the genetic and metabolic basis of disease,” Clinical Cancer Research, vol. 21, no. 1, pp. 10–17, 2015.
View at: Publisher Site | Google Scholar
P. Singh, N. Agarwal, and S. K. Pal, “Sequencing systemic therapies for metastatic kidney cancer,” Current Treatment Options in Oncology, vol. 16, article 3, 2015.
View at: Publisher Site | Google Scholar
H. M. Huang, X. Jiang, M. L. Hao et al., “Identification of biomarkers in macrophages of atherosclerosis by microarray analysis,” Lipids in Health and Disease, vol. 18, no. 1, article 107, 2019.
View at: Publisher Site | Google Scholar
Y. Qiu, L. B. Meng, C. Y. Di et al., “Exploration of the differentially expressed long noncoding RNAs and genes of morphine tolerance via bioinformatic analysis,” Journal of Computational Biology, vol. 26, no. 12, pp. 1379–1393, 2019.
View at: Publisher Site | Google Scholar
Y. F. Zou, L. B. Meng, Z. K. He et al., “Screening and authentication of molecular markers in malignant glioblastoma based on gene expression profiles,” Oncology Letters, vol. 18, pp. 4593–4604, 2019.
View at: Publisher Site | Google Scholar
L. B. Meng, M. J. Shan, Y. Qiu et al., “TPM2 as a potential predictive biomarker for atherosclerosis,” Aging, vol. 11, no. 17, pp. 6960–6982, 2019.
View at: Publisher Site | Google Scholar
M. Ashburner, C. A. Ball, J. A. Blake et al., “Gene ontology: tool for the unification of biology,” Nature Genetics, vol. 25, pp. 25–29, 2000.
View at: Publisher Site | Google Scholar
M. Tanabe and M. Kanehisa, “Using the KEGG database resource,” Curr Protoc Bioinformatics, vol. 11, no. 1, 2005.
View at: Publisher Site | Google Scholar
S. Maere, K. Heymans, and M. Kuiper, “BiNGO: a Cytoscape plugin to assess overrepresentation of gene ontology categories in biological networks,” Bioinformatics, vol. 21, no. 16, pp. 3448-3449, 2005.
View at: Publisher Site | Google Scholar
M. E. Smoot, K. Ono, J. Ruscheinski, P. L. Wang, and T. Ideker, “Cytoscape 2.8: new features for data integration and network visualization,” Bioinformatics, vol. 27, no. 3, pp. 431-432, 2011.
View at: Publisher Site | Google Scholar
G. D. Bader and C. W. Hogue, “An automated method for finding molecular complexes in large protein interaction networks,” BMC Bioinformatics, vol. 4, no. 1, p. 2, 2003.
View at: Publisher Site | Google Scholar
C. H. Chin, S. H. Chen, H. H. Wu, C. W. Ho, M. T. Ko, and C. Y. Lin, “cytoHubba: identifying hub objects and sub-networks from complex interactome,” BMC Systems Biology, vol. 8, article S11, 2014.
View at: Publisher Site | Google Scholar
Z. Tang, C. Li, B. Kang, G. Gao, C. Li, and Z. Zhang, “GEPIA: a web server for cancer and normal gene expression profiling and interactive analyses,” Nucleic Acids Research, vol. 45, no. W1, pp. W98–W102, 2017.
View at: Publisher Site | Google Scholar
T. Flieder, M. Weiser, T. Eller et al., “Interference of DOACs in different DRVVT assays for diagnosis of lupus anticoagulants,” Thrombosis Research, vol. 165, pp. 101–106, 2018.
View at: Publisher Site | Google Scholar
G. Hussain, J. Wang, A. Rasul et al., “Role of cholesterol and sphingolipids in brain development and neurological diseases,” Lipids in Health and Disease, vol. 18, no. 1, p. 26, 2019.
View at: Publisher Site | Google Scholar
W. Wang, W. Lou, B. Ding et al., “A novel mRNA-miRNA-lncRNA competing endogenous RNA triple sub-network associated with prognosis of pancreatic cancer,” Aging, vol. 11, no. 9, pp. 2610–2627, 2019.
View at: Publisher Site | Google Scholar
F. L. Chan, B. Vinod, K. Novy et al., “Aurora kinase B, a novel regulator of TERF1 binding and telomeric integrity,” Nucleic Acids Research, vol. 45, no. 21, pp. 12340–12353, 2017.
View at: Publisher Site | Google Scholar
S. L. Smith, N. L. Bowers, D. C. Betticher et al., “Overexpression of aurora B kinase (AURKB) in primary non-small cell lung carcinoma is frequent, generally driven from one allele, and correlates with the level of genetic instability,” British Journal of Cancer, vol. 93, no. 6, pp. 719–729, 2005.
View at: Publisher Site | Google Scholar
R. Sorrentino, S. Libertini, P. L. Pallante et al., “Aurora B overexpression associates with the thyroid carcinoma undifferentiated phenotype and is required for thyroid carcinoma cell proliferation,” The Journal of Clinical Endocrinology and Metabolism, vol. 90, no. 2, pp. 928–935, 2005.
View at: Publisher Site | Google Scholar
P. Chieffi, L. Cozzolino, A. Kisslinger et al., “Aurora B expression directly correlates with prostate cancer malignancy and influence prostate cell proliferation,” Prostate, vol. 66, no. 3, pp. 326–333, 2006.
View at: Publisher Site | Google Scholar
H. Katayama, T. Ota, F. Jisaki et al., “Mitotic kinase expression and colorectal cancer progression,” Journal of the National Cancer Institute, vol. 91, no. 13, pp. 1160–1162, 1999.
View at: Publisher Site | Google Scholar
M. Tatsuka, H. Katayama, T. Ota et al., “Multinuclearity and increased ploidy caused by overexpression of the aurora- and Ipl1-like midbody-associated protein mitotic kinase in human cancer cells,” Cancer Research, vol. 58, no. 21, pp. 4811–4816, 1998.
View at: Google Scholar
T. Ota, S. Suto, H. Katayama et al., “Increased mitotic phosphorylation of histone H3 attributable to AIM-1/Aurora-B overexpression contributes to chromosome number instability,” Cancer Research, vol. 62, no. 18, pp. 5168–5177, 2002.
View at: Google Scholar
A. Tang, K. Gao, L. Chu, R. Zhang, J. Yang, and J. Zheng, “Aurora kinases: novel therapy targets in cancers,” Oncotarget, vol. 8, no. 14, pp. 23937–23954, 2017.
View at: Publisher Site | Google Scholar
D. Gong, J. R. Pomerening, J. W. Myers et al., “Cyclin A2 regulates nuclear-envelope breakdown and the nuclear accumulation of cyclin B1,” Current Biology, vol. 17, no. 1, pp. 85–91, 2007.
View at: Publisher Site | Google Scholar
M. Volm, R. Koomägi, J. Mattern, and G. Stammler, “Cyclin A is associated with an unfavourable outcome in patients with non- small-cell lung carcinomas,” British Journal of Cancer, vol. 75, no. 12, pp. 1774–1778, 1997.
View at: Publisher Site | Google Scholar
Y. Dobashi, M. Shoji, S. X. Jiang, M. Kobayashi, Y. Kawakubo, and T. Kameya, “Active cyclin A-CDK2 complex, a possible critical factor for cell proliferation in human primary lung carcinomas,” The American Journal of Pathology, vol. 153, no. 3, pp. 963–972, 1998.
View at: Publisher Site | Google Scholar
I. R. Bukholm, G. Bukholm, and J. M. Nesland, “Over-expression of cyclin A is highly associated with early relapse and reduced survival in patients with primary breast carcinomas,” International Journal of Cancer, vol. 93, no. 2, pp. 283–287, 2001.
View at: Publisher Site | Google Scholar
R. Michalides, H. van Tinteren, A. Balkenende et al., “Cyclin A is a prognostic indicator in early stage breast cancer with and without tamoxifen treatment,” British Journal of Cancer, vol. 86, no. 3, pp. 402–408, 2002.
View at: Publisher Site | Google Scholar
J. Q. Li, H. Miki, F. Wu et al., “Cyclin a correlates with carcinogenesis and metastasis, and p27^kip1 correlates with lymphatic invasion, in colorectal neoplasms,” Human Pathology, vol. 33, no. 10, pp. 1006–1015, 2002.
View at: Publisher Site | Google Scholar
S. Aaltomaa, P. Lipponen, M. Ala-Opas, M. Eskelinen, K. Syrjänen, and V. M. Kosma, “Expression of cyclins A and D and p21^(waf1/cip1) proteins in renal cell cancer and their relation to clinicopathological variables and patient survival,” British Journal of Cancer, vol. 80, no. 12, pp. 2001–2007, 1999.
View at: Publisher Site | Google Scholar
L. Gopinathan, S. L. Tan, V. C. Padmakumar, V. Coppola, L. Tessarollo, and P. Kaldis, “Loss of Cdk2 and cyclin A2 impairs cell proliferation and tumorigenesis,” Cancer Research, vol. 74, no. 14, pp. 3870–3879, 2014.
View at: Publisher Site | Google Scholar
W. Liang, H. Guan, X. He et al., “Down-regulation of SOSTDC1 promotes thyroid cancer cell proliferation via regulating cyclin A2 and cyclin E2,” Oncotarget, vol. 6, no. 31, pp. 31780–31791, 2015.
View at: Publisher Site | Google Scholar
C. Aguirre-Portolés, A. W. Bird, A. Hyman, M. Cañamero, I. Pérez de Castro, and M. Malumbres, “Tpx2 controls spindle integrity, genome stability, and tumor development,” Cancer Research, vol. 72, no. 6, pp. 1518–1528, 2012.
View at: Publisher Site | Google Scholar
L. Zhang, H. Huang, L. Deng et al., “TPX2 in malignantly transformed human bronchial epithelial cells by anti- benzo[a]pyrene-7,8-diol-9,10-epoxide,” Toxicology, vol. 252, no. 1-3, pp. 49–55, 2008.
View at: Publisher Site | Google Scholar
Y. Ma, D. Lin, W. Sun et al., “Expression of targeting protein for xklp2 associated with both malignant transformation of respiratory epithelium and progression of squamous cell lung cancer,” Clinical Cancer Research, vol. 12, no. 4, pp. 1121–1127, 2006.
View at: Publisher Site | Google Scholar
A. H. Sillars-Hardebol, B. Carvalho, M. Tijssen et al., “TPX2 and AURKA promote 20q amplicon-driven colorectal adenoma to carcinoma progression,” Gut, vol. 61, no. 11, pp. 1568–1575, 2012.
View at: Publisher Site | Google Scholar
H. Chang, J. Wang, Y. Tian, J. Xu, X. Gou, and J. Cheng, “The TPX2 gene is a promising diagnostic and therapeutic target for cervical cancer,” Oncology Reports, vol. 27, no. 5, pp. 1353–1359, 2012.
View at: Publisher Site | Google Scholar
P. Vainio, J. P. Mpindi, P. Kohonen et al., “High-throughput transcriptomic and RNAi analysis identifies AIM1, ERGIC1, TMED3 and TPX2 as potential drug targets in prostate cancer,” PLoS One, vol. 7, no. 6, article e39801, 2012.
View at: Publisher Site | Google Scholar
Q. I. Chen, B. Cao, N. Nan et al., “TPX2 in human clear cell renal carcinoma: expression, function and prognostic significance,” Oncology Letters, vol. 11, no. 5, pp. 3515–3521, 2016.
View at: Publisher Site | Google Scholar
Z. A. Glaser, H. D. Love, S. Guo et al., “TPX2 as a prognostic indicator and potential therapeutic target in clear cell renal cell carcinoma,” Urologic Oncology, vol. 35, no. 5, pp. 286–293, 2017.
View at: Publisher Site | Google Scholar
Y. Gu, L. Lu, L. Wu, H. Chen, W. Zhu, and Y. He, “Identification of prognostic genes in kidney renal clear cell carcinoma by RNA-seq data analysis,” Molecular Medicine Reports, vol. 15, no. 4, pp. 1661–1667, 2017.
View at: Publisher Site | Google Scholar
W. Wei, Y. Lv, Z. Gan, Y. Zhang, X. Han, and Z. Xu, “Identification of key genes involved in the metastasis of clear cell renal cell carcinoma,” Oncology Letters, vol. 17, no. 5, pp. 4321–4328, 2019.
View at: Publisher Site | Google Scholar
A. Eberlein, A. Takasuga, K. Setoguchi et al., “Dissection of genetic factors modulating fetal growth in cattle indicates a substantial role of the non-SMC condensin I complex, subunit G (NCAPG) gene,” Genetics, vol. 183, no. 3, pp. 951–964, 2009.
View at: Publisher Site | Google Scholar
T. Arai, A. Okato, Y. Yamada et al., “Regulation of NCAPG by miR-99a-3p (passenger strand) inhibits cancer cell aggressiveness and is involved in CRPC,” Cancer Medicine, vol. 7, no. 5, pp. 1988–2002, 2018.
View at: Publisher Site | Google Scholar
S. Li, Y. Xuan, B. Gao et al., “Identification of an eight-gene prognostic signature for lung adenocarcinoma,” Cancer Management and Research, vol. 2018, pp. 3383–3392, 2018.
View at: Publisher Site | Google Scholar
J. Chen, X. Qian, Y. He, X. Han, and Y. Pan, “Novel key genes in triple-negative breast cancer identified by weighted gene co-expression network analysis,” Journal of Cellular Biochemistry, vol. 120, no. 10, pp. 16900–16912, 2019.
View at: Publisher Site | Google Scholar
W. Liu, B. Liang, H. Liu et al., “Overexpression of non-SMC condensin I complex subunit G serves as a promising prognostic marker and therapeutic target for hepatocellular carcinoma,” International Journal of Molecular Medicine, vol. 40, no. 3, pp. 731–738, 2017.
View at: Publisher Site | Google Scholar
Y. Wang, B. Gao, P. Y. Tan et al., “Genome-wide CRISPR knockout screens identify NCAPG as an essential oncogene for hepatocellular carcinoma tumor growth,” The FASEB Journal, vol. 33, no. 8, pp. 8759–8770, 2019.
View at: Publisher Site | Google Scholar
J. Ai, C. Gong, J. Wu et al., “MicroRNA-181c suppresses growth and metastasis of hepatocellular carcinoma by modulating NCAPG,” Cancer Management and Research, vol. 2019, pp. 3455–3467, 2019.
View at: Publisher Site | Google Scholar

Copyright

Copyright © 2020 Bin Liu et al. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

PDF Download Citation

Download other formats

Order printed copies

Views

730

Downloads

1019

Citations