Identification of potential oncogenes in triple-negative breast cancer based on bioinformatics analyses

Abstract

Triple-negative breast cancer (TNBC) is a subtype with high rates of metastasis, poor prognosis and limited therapeutic options. The present study aimed to identify the potential pivotal genes for prognosis and treatment in TNBC. A total of two microarray expression datasets, GSE38959 and GSE65212, were downloaded from the Gene Expression Omnibus database, and RNA-sequencing data of breast cancer from The Cancer Genome Atlas database were analyzed to screen out differentially expressed genes (DEGs) between TNBC tissues and normal tissues. The intersection of DEGs was submitted to Gene Ontology and Kyoto Encyclopedia of Genes and Genomes enrichment analyses. A protein-protein interaction (PPI) network was constructed and visualized using Cytoscape software. Furthermore, module, centrality and survival analyses were performed to identify the potential hub genes. Reverse transcription-quantitative (RT-q)PCR analysis was performed to detect the expression levels of key genes in TNBC samples, and 377 DEGs were identified. Functional analysis revealed that the DEGs were significantly involved in cell cycle process, nuclear division and the p53 signaling pathway. A PPI network was constructed with these DEGs, and 66 core genes with high centrality features in module 1 were selected. Relapse-free survival analysis confirmed that high expression levels of five genes [cyclin B1 (CCNB1), GINS complex subunit 2, non-SMC condensin I complex subunit G (NCAPG), minichromosome maintenance 4 (MCM4) and ribonucleotide reductase regulatory subunit M2 (RRM2)] were significantly associated with poor prognosis in TNBC. RT-qPCR analysis demonstrated that CCNB1, NCAPG, MCM4 and RRM2 were significantly upregulated in 25 TNBC tissues compared with adjacent normal breast tissues. Furthermore, gene set enrichment analysis revealed that CCNB1, NCAPG, MCM4 and RRM2 were closely associated with tumor proliferation. Taken together, these results suggest that CCNB1, NCAPG, MCM4 and RRM2 are associated with tumorigenesis and TNBC progression, and thus may act as promising prognostic biomarkers and therapeutic targets for TNBC.

Introduction

Breast cancer is the third most common cancer worldwide in 2016 and is considered a paramount public health issue that seriously endangers the lives of millions of women (1). Globally, 1 in 20 women develop breast cancer in their lifetime, and the incidence continues to increase (1). Triple-negative breast cancer (TNBC), whose expression levels of progesterone receptor (PR), estrogen receptor (ER) and human epidermal growth factor receptor 2 (HER2) are negative, comprises ~15% of all breast cancers, with the worst prognosis compared with other subtypes irrespective of race, age, or stage (2). TNBCs are characterized by a poor prognosis and high rates of proliferation and metastases, and occur frequently in younger patients, where tumors generally present unfavorable clinical features, such as larger size, higher histologic grade and lymph node involvement (3–5). Due to the defect of promising molecular markers, conventional chemotherapy and radiation are the primary systemic therapeutic strategies (5). Thus, it remains critical to discover novel biomarkers for therapy patients with TNBC.

With the rapid development of genomic and proteomic technologies, bioinformatics have facilitated the discovery of reliable biomarkers for diagnosis, survival and prognosis of diseases (6). Recent studies have focused on the therapeutic targets of TNBC by microarray analysis of gene expression profiles, including CCNA2, CDC20 and BUB1, which are upregulated in TNBC tissues compared with normal tissues (7,8). However, lack of direct experimental validation of the upregulated genes decreases the reliability of these conclusions.

To identify differentially expressed genes (DEGs) in TNBC tissues compared with adjacent normal breast tissues, the present study analyzed two microarray expression datasets, GSE38959 (9) and GSE65212 (10), from the Gene Expression Omnibus (GEO) database and RNA sequencing (RNA-seq) data of TNBC tissues and adjacent normal breast tissues from The Cancer Genome Atlas (TCGA) database. Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analyses were performed to determine the significant functional terms of overlapping DEGs across the three datasets. Centrality and survival analyses were performed to determine the pivotal genes with higher importance and prognostic values. Reverse transcription-quantitative (RT-q)PCR analysis was performed to detect the expression levels of the hub genes in clinical TNBC tissues and adjacent normal tissues. Gene set enrichment analysis was performed to investigate the potential biological functions associated with the hub genes.

Materials and methods

Data source

A total of two microarray expression datasets (GSE38959 and GSE65212) were downloaded from the GEO database (https://www.ncbi.nlm.nih.gov/geo). The GSE38959 dataset had 30 TNBC tissues and 13 adjacent normal breast tissues, while the GSE65212 dataset had 41 TNBC tissues and 11 adjacent normal breast tissues. All samples included both TNBC tissues and normal breast tissues, and each microarray contained >40 samples. In addition, gene expression profiles together with corresponding clinical data of 1,109 breast cancer tissues and 113 adjacent normal tissues were obtained from TCGA database (http://gdac.broadinstitute.org/runs/analyses_2016_01_28/data/BRCA). Following filtration via immunohistochemistry (IHC) information of ER/PR/HER2 in clinical data of TCGA-BRCA, 88 TNBC tissues and 6 normal tissues, with detailed clinical information and without history of neoadjuvant chemotherapy, were enrolled in the present study. The aforementioned information was freely available online.

Tissue samples

A total of 25 TNBC tissues and matched adjacent normal tissues were collected from patients diagnosed with TNBC via biopsy IHC staining at the First Affiliated Hospital of Chongqing Medical University (Chongqing, China) between January 2019 and October 2019. The age of patients ranged from 37–75 years (median age, 50 years). The extracted normal tissues were 3 cm away from the tumor border, all tissue samples were snap-frozen in liquid nitrogen after surgery, and subsequently stored in liquid nitrogen until subsequent experimentation. Patients' initial treatment was surgery without receiving prior treatment with radiation and/or chemotherapy. The present study was approved by the Ethics Committee of the First Affiliated Hospital of Chongqing Medical University (Chongqing, China, approval no. 2020-124) and written informed consent was provided by all patients prior to the study start.

Identification of DEGs

The gene expression datasets from the GEO database (GSE38959 and GSE65212) were analyzed using GEO2R (http://www.ncbi.nlm.nih.gov/geo/geo2r), an online tool that can compare gene expression levels between two sample groups (11), to identify DEGs between TNBC tissues and adjacent normal tissues. The RNA-seq level 3 normalized data from TCGA database was performed using R package of edgeR (v3.28.1, http://bioinf.wehi.edu.au/edgeR). Genes with |log2 fold change|>1.5 and P<0.05 were differentially expressed. Venny 2.1 (http://bioinfogp.cnb.csic.es/tools/venny), a Venn diagram web tool, was used to identify the overlapping DEGs across the three datasets.

Functional enrichment analysis of DEGs

GO functions were analyzed based on overlapped genes, whose functions were classified into biological process (BP), molecular function (MF) and cellular component (CC) terms. The Search Tool for the Retrieval of Interacting Genes (STRING) database (version 11.0; http://string-db.org) was used to export results of GO enrichment analysis (12). The Database for Annotation, Visualization and Integrated Discovery (DAVID) (version 6.8, http://david.ncifcrf.gov) (13) was used to perform KEGG pathway enrichment analysis of the upregulated and downregulated DEGs, respectively. P<0.05 was considered to indicate statistically significant GO terms and KEGG pathways. Gene count thresholds for the GO terms and KEGG pathways were set to ≥20 and ≥4, respectively.

Protein-protein interaction (PPI) network construction

To assess the potential associations among DEGs, the STRING database (12) was used to construct a PPI network. The results were visualized using Cytoscape software v3.7.1 (14). A combined score of >0.4 was considered to indicate a statistically significant result.

Centrality analysis based on the PPI network

Based on the PPI network, two significant topological parameters, degree and betweenness centrality, were used to identify potential pivotal genes in this network. The two centrality scores of each node were exported using NetworkAnalyzer (v2.7) in Cytoscape software (15). The Venn diagram was applied to demonstrate the intersections of top 50% DEGs sorted by the degree value and the betweenness value.

Module analysis of the PPI network

MCODE Cytoscape plugin (v1.6.1) was applied to screen the modules considered essential parts of the network (16). For each significant module, the default criteria were as follows: Degree cut-off, 2; node score cut-off, 0.2; k-core, 2 and max depth, 100. The genes in the 1st ranked module with high degree and betweenness values were selected as candidate genes for further analysis.

Survival analysis

To assess the clinical outcome, the candidate genes were subjected to the Kaplan-Meier plotter (http://kmplot.com/analysis), which assessed the effects of 22,277 genes on breast cancer prognosis, using microarray data (17). A total of 255 patients with TNBC were selected from 3,955 patients in the Kaplan-Meier plotter breast cancer database by restricting the IHC negative status of ER, PR and HER2. In the present study, relapse-free survival (RFS) curves were drawn and exported using the online survival analysis tool, Kaplan-Meier plotter. According to the median expression of each gene, the cohorts were divided into two groups, high expression group (127 patients) and low expression group (128 patients). Log-rank P<0.05 was considered to indicate a statistically significant difference.

RT-qPCR

Total RNA was extracted from TNBC tissues and adjacent normal tissues using TRIzol^® reagent (Takara Bio, Inc.), according to the manufacturer's instructions. Briefly, TRIzol^® reagent was added to each tissue sample and homogenized. Subsequently, chloroform was used to separate the components and isopropanol was added to precipitate RNA. The eluted RNA precipitation was assessed using NanoDrop 2000 (Thermo Fisher Scientific, Inc.). Total RNA was reverse transcribed into cDNA using the PrimeScript RT reagent kit (Takara Bio, Inc., cat. no. RR047A). Temperature protocol for RT was as follows: 37°C for 15 min, 85°C for 5 sec and 4°C for 15 min. qPCR was subsequently performed using SYBR Premix Ex Taq™ (Takara Bio, Inc.) to determine the amplification of mRNAs in the CFX96 Real Time system (Bio-Rad Laboratories, Inc.). The primer sequences used for qPCR are listed in Table I. Relative expression levels were calculated using the 2^−ΔΔCq method (18) and normalized to the internal reference gene GAPDH. All experiments were performed in triplicate.

Table I.

Primer sequences used for quantitative PCR.

Primer	Sequence (5′-3′)
GAPDH-F	GTCTTCCTGGGCAAGCAGTA
GAPDH-R	CTGGACAGAAACCCCACTTC
CCNB1-F	AACTTTCGCCTGAGCCTATTTT
CCNB1-R	TTGGTCTGACTGCTTGCTCTT
GINS2-F	AGGGTCTCGTTCTGTCATCC
GINS2-R	TCTTTTGGTCCCAGTCTTCC
NCAPG-F	TTTGTATTGGTGTGCCCTTT
NCAPG-R	AGCCAGCAGTTTTTTTCTTC
MCM4-F	CTCATCCACAACCGCTCC
MCM4-R	TTCACTCTGTCCCCAGGC
RRM2-F	CTCCAAGGACATTCAGCAC
RRM2-R	GGAAGCCATAGAAACAGCG

F, forward; R, reverse; CCNB1, cyclin B1; GINS2, GINS complex subunit 2; NCAPG, non-SMC condensin I complex subunit G; MCM4, minichromosome maintenance 4; RRM2, ribonucleotide reductase regulatory subunit M2.

Gene set enrichment analysis (GSEA)

GSEA was performed on TCGA RNA-seq data using R package ‘clusterprofiler’ (19). Based on Spearman's correlation coefficients between the expression levels of the hub genes and other genes in TNBC samples of TCGA cohort, GSEA was implemented on a set of 50 hallmark signatures (20). Gene signatures with adjusted P<0.05 were significantly enriched. The reference gene set ‘h.all.v7.0.symbols.gmt.txt’ was downloaded from the Molecular Signatures Database (MSigDB, http://broadinstitute.org/gsea/msigdb/index.jsp).

Statistical analysis

Statistical analysis was performed using the ggpubr package (v 0.4.0, http://CRAN.R-project.org/package=ggpubr) in R version 3.6.2 (https://www.R-project.org). Paired Student's t-test was used to compare differences between two groups. P<0.05 was considered to indicate a statistically significant difference.

Results

Filtration of DEGs

A total of 1,800, 2,347 and 2,244 DEGs were obtained from the GSE38959 and GSE65212 datasets and TCGA TNBC cohort, respectively. A total of 377 overlapping DEGs were identified between TNBC tissue samples and non-tumor breast tissue samples via Venn analysis. Among these genes, 260 genes were upregulated (Fig. 1A) and 117 genes were downregulated (Fig. 1B).

Figure 1.

Common DEGs among the GSE65212, GSE38959 and TCGA datasets, and GO and KEGG functional enrichment analyses of 377 DEGs. (A) Upregulated and (B) downregulated DEGs. (C) Significantly enriched GO terms of DEGs, including top 10 cellular components, molecular functions and biological processes, according to the -log₁₀ (P-value). (D) Significantly enriched KEGG pathway terms for upregulated (left) and downregulated (right) DEGs. The size of each node represents the gene number in the corresponding pathway, whereas the color change from blue to red or from green to salmon indicates the P-values from the big to the small for the corresponding pathway. DEGs, differentially expressed genes; TCGA, The Cancer Genome Atlas; GO, gene ontology; KEGG, Kyoto Encyclopedia of Genes and Genomes.

Functional enrichment analysis of DEGs

The functions of the 377 filtrated DEGs were assessed via GO function and KEGG enrichment analyses (Table SI). GO analysis demonstrated that the DEGs were associated with ‘mitotic cell cycle process’, ‘mitotic nuclear division’, ‘DNA conformation change’, ‘chromosome segregation’, ‘centromeric region’, ‘condensed chromosome’ and ‘binding of protein, ATP and microtubule’ (Fig. 1C). KEGG analysis for the upregulated DEGs demonstrated that the genes were markedly enriched in the ‘p53 signaling pathway’, ‘cell cycle’, ‘DNA replication’, ‘alcoholism’, ‘extracellular matrix (ECM)-receptor interaction’ and ‘progesterone-mediated oocyte maturation’. For the downregulated DEGs, the most enriched pathways were ‘axon guidance’, ‘cGMP-PKG signaling pathway’ and ‘aldosterone-regulated sodium reabsorption’ (Fig. 1D).

Construction of the PPI network

To determine the interactions of the 377 identified DEGs, a PPI network was constructed, which comprised of 335 nodes and 6,026 edges (Fig. S1). These DEGs were regarded as potential crucial genes in TNBC pathogenesis.

Centrality analysis of the PPI network

Centrality analysis of the PPI network was performed based on two significant parameters, degree and betweenness centrality. Degree centrality refers to the sum of edges connected to other vertexes, which symbolizes importance of each node in the network. While betweenness centrality refers to the sum of times each vertex is included in all-pairs shortest paths, indicating the intermediate influence of each node (21). The results demonstrated that the degree and betweenness values displayed power-law distributions (Fig. 2A and B). Subsequently, the top 50% of each parameter was chosen for further investigations, and 111 DEGs were obtained based on the results of the Venn analysis (Fig. 2C).

Figure 2.

Centrality analyses of differentially expressed genes based on triple-negative breast cancer related protein-protein interaction network. The density distribution of (A) degree centrality and (B) betweenness centrality. (C) Venn diagram of hub genes common to top 50% degree values and top 50% betweenness values.

Modules analysis of the PPI network

Modules analysis was performed using Cytoscape software, and the module with the highest score, module 1 (Fig. 3), was further screened from the PPI network. And the results demonstrated that module 1 contained 96 nodes and 4,064 edges, with a 85.56 MCODE score. Among the 96 nodes in module 1, 66 nodes had high degree and betweenness values, suggesting that these nodes may act as potential key genes with essential physiological or pathological regulatory functions. Thus, these 66 nodes were selected as candidate genes for further analyses.

Figure 3.

Most significant module of triple-negative breast cancer related PPI network. Module 1 contained 96 nodes and 4,064 edges. The color depth and size of nodes refers to the degree value in the PPI network. PPI, protein-protein interaction.

Survival analysis to identity the hub genes

RFS analysis in the Kaplan-Meier plotter platform was performed to determine the prognostic value of the 66 potential candidate genes. The results demonstrated that upregulated CCNB1, GINS complex subunit 2 (GINS2), NCAPG, MCM4 and RRM2 expression levels were associated with unfavorable RFS of patients with TNBC (log-rank P<0.05; Fig. 4). Further details of the five hub genes are presented in Table II.

Figure 4.

RFS analysis of hub genes in patients with TNBC using Kaplan-Meier plotter. High expression levels of (A) CCNB1, (B) GINS2, (C) NCAPG, (D) MCM4, (E) RRM2 were significantly associated with unfavorable prognosis of patients with TNBC. RFS, relapse-free survival; TNBC, triple-negative breast cancer; CCNB1, cyclin B1; GINS2, GINS complex subunit 2; NCAPG, non-SMC condensin I complex subunit G; MCM4, minichromosome maintenance 4; RRM2, ribonucleotide reductase regulatory subunit M2; HR, hazard ratio.

Table II.

Information on the five hub genes from the protein-protein interaction network.

Gene symbol	Gene description	Expression in TNBC	Degree value	Betweenness value	P-value
CCNB1	Cyclin B1	Up	132	4,591.6	0.0078
GINS2	GINS complex subunit 2	Up	97	712.8	0.0120
NCAPG	Non-SMC condensin I complex subunit G	Up	113	199.9	0.0120
MCM4	Minichromosome maintenance 4	Up	103	204.0	0.0260
RRM2	Ribonucleotide reductase regulatory subunit M2	Up	115	1,317.9	0.0330

Validation of hub genes in clinical samples via RT-qPCR analysis

RT-qPCR analysis was performed to detect the expression levels of the five hub genes in 25 clinical TNBC tissues and adjacent normal breast tissues. The results demonstrated that the expression levels of CCNB1, NCAPG, MCM4 and RRM2 were elevated in TNBC tissues compared with adjacent normal tissues (Fig. 5A-D). These experimental results were in accordance with the bioinformatics-predicted results. However, no significant difference in GINS2 expression was observed between the TNBC tissues and adjacent normal tissues (Fig. 5E).

Figure 5.

Validation of the gene expression levels of CCNB1, RRM2, NCAPG, MCM4, GINS2 between TNBC tissues and adjacent normal breast tissues via reverse transcription-quantitative PCR analysis. (A-D) CCNB1, RRM2, NCAPG and MCM4 expression levels were significantly upregulated in TNBC tissues compared with adjacent normal breast tissues. (E) No significant difference in GINS2 expression was observed between the TNBC tissues and adjacent normal breast tissues. ***P<0.001, ****P<0.0001. TNBC, triple-negative breast cancer; ns, no significance; CCNB1, cyclin B1; GINS2, GINS complex subunit 2; NCAPG, non-SMC condensin I complex subunit G; MCM4, minichromosome maintenance 4; RRM2, ribonucleotide reductase regulatory subunit M2.

Four hub oncogenes significantly associated with tumor proliferation

To further investigate the potential biological functions associated with hub genes, GSEA was performed on mRNA expression data of TCGA TNBC samples. The results demonstrated prominent enrichments of hallmark proliferation gene sets for genes associated with high expression levels of the four hub oncogenes (CCNB1, NCAPG, MCM4 and RRM2), including ‘E2F_TARGETS’, ‘G2M_CHECKPOINT’, ‘MYC_TARGETS_V1’ and ‘MYC_TARGETS_V2’ (Fig. 6A-D). Taken together, these results suggest that the four identified hub genes are significantly associated with cell proliferative processes.

Figure 6.

Gene set enrichment analysis of the four key oncogenes in The Cancer Genome Atlas Triple-negative breast cancer samples. (A-D) Top three gene signatures (sorted by normalized enrichment score) significantly enriched for high expression levels of each hub gene. CCNB1, cyclin B1; NCAPG, non-SMC condensin I complex subunit G; MCM4, minichromosome maintenance 4; RRM2, ribonucleotide reductase regulatory subunit M2.

Discussion

In 2016, breast cancer was the third most common cancer worldwide and the leading cause of cancer-associated mortality (535,000 deaths) in women (1). TNBC is a unique subtype of breast cancer characterized by poor prognosis and limited effective treatments (2). Due to the absence of targeted therapies, conventional chemotherapy and radiation are the primary systemic therapeutic strategies (5). Recently, the rapid development of next generation sequencing in GEO and integrated multi-omics measurements in TCGA database has revealed significant molecular heterogeneity of breast cancer (22). Thus, bioinformatics analyses are performed to identify specific molecular targets for TNBC.

In the initial stages of the study, three microarrays were assessed (GSE65212, GSE38959 and GSE76250). However, GSE76250 was excluded due to the difference in its design from the other two datasets. Based on the GSE65212 and GSE38959 datasets from the GEO database and a breast cancer cohort from TCGA database, 377 DEGs between TNBC tissues and adjacent normal human breast tissues were screened, including 260 upregulated genes and 117 downregulated genes. GO and KEGG functional enrichment analyses demonstrated that the most enriched GO terms of the DEGs were ‘mitotic cell cycle process’, ‘chromosome segregation’ and ‘mitotic nuclear division’. KEGG pathways, such as the ‘p53 signaling pathway’, ‘progesterone-mediated oocyte maturation’, ‘DNA replication’, ‘alcoholism’ and ‘ECM-receptor interaction’ were predominantly associated with the upregulated genes, while a few pathways, such as ‘axon guidance’, ‘cGMP-PKG signaling pathway’ were enriched in the downregulated genes.

It is well-known that defects in cell cycle regulation, such as sustaining proliferation and unlimited replication, are fundamental characteristics of cancer pathogenesis (23), and some newly discovered TNBC-associated small molecule inhibitors have been demonstrated to induce cell cycle arrest (5). Similarly, chromosome segregation with nuclear division in M phase and DNA replication in S phase are essential processes during mitotic cell division (24). In tumorigenesis, driven by oncogene activation, DNA replication stress and its adverse impact on chromosome segregation are associated with genome instability (25). In addition, the p53 pathway is a classic signaling pathway involved in the occurrence and progression of cancer, which plays essential roles in tumor suppression, regulating cell migration and invasion (26). The frequency of TP53 gene mutation in basal-like breast tumors/TNBCs is ~80% (22), and based on molecular mechanisms of the p53 pathway, a few chemicals indicate potential therapeutic intervention in breast cancer (27,28). In addition, alcohol has a deleterious effect on women by increasing the risk of breast cancer (29), and in vitro experiments have demonstrated that alcohol promotes TNBC cell proliferation, migration and invasion (30). However, to the best of our knowledge, alcoholism in TNBC has not been reported by other datasets enrichment analyses. Cell-cell adhesion alterations and attachment to the ECM are common events in diverse epithelial malignancies, which are associated with cellular invasion and metastasis (22). Excessiveness of ECM deposition may enhance tumor cell invasion in the breast cancer (31). Dysregulated microRNAs (miRNAs) associated with progesterone-mediated oocyte maturation may have an impact on follicular growth arrest and metabolic disorders (32). Furthermore, oocyte meiosis and progesterone-mediated oocyte maturation pathways are enriched in survival associated miRNAs of ovarian carcinomas (33). Taken together, these results suggest that these DEGs may be associated with the pathogenesis and development of TNBC.

To investigate the interactions between these DEGs, the PPI network complex was constructed. Following centrality analysis, the results demonstrated that the degree and betweenness parameters displayed power-law distributions. It is well-known that power-law distributions frequently appear in several disease or metabolic biological networks (34), suggesting that the PPI network in the present study has similar scale-free characteristics with other biological networks. Modules analysis identified four sub modules, including the first-ranked module, which contained 96 nodes. Increasing evidence suggest that modules analysis has been extensively applied for identifying hub genes in diverse cancers, such as colorectal cancer (35), oral cancer (36) and renal carcinoma (37). Thus, these DEGs and interactions in the first-ranked module may be the core of the network. A total of 66 candidate DEGs with high degree and betweenness values among the 96 nodes in the first-ranked module were selected. Collectively, the results of the present study suggest that the 66 candidate genes may be pivotal in regulating the occurrence and progression of TNBC.

Survival analysis demonstrated that high expression levels of the five hub genes (CCNB1, GINS2, NCAPG, MCM4 and RRM2) among the 66 candidate genes were significantly associated with shorter RFS times in patients with TNBC (P<0.05). This suggests that the five hub genes may be indispensable to tumorigenesis and progression in TNBC. Reverse transcription quantitative PCR analysis was performed to validate the expression levels of the five hub genes in TNBC clinical samples and their matched adjacent normal controls. The results demonstrated that CCNB1, NCAPG, MCM4 and RRM2 expression levels were significantly upregulated in TNBC samples compared with the controls, and no significant difference in GINS2 expression was observed between the two groups. GSEA was performed to investigate the potential biological functions of the four oncogenes, which revealed significant enrichment of cell proliferation markers for high expression levels of CCNB1, NCAPG, MCM4 and RRM2.

CCNB1 is a checkpoint protein in the G₂-M transition phase during cell cycle (38). CCNB1 protein is upregulated in TNBCs compared with other subtypes, which is closely associated with adverse clinical prognosis in patients with breast cancer (7,39). In clinical practice, CCNB1 has been applied as a cell proliferation biomarker to evaluate breast cancer recurrence risk in a genetic test called the 21-gene expression assay (40). Recent studies have reported that some drugs, such as Dipalmitoylphosphatidic acid (41) and F1012-2 (a material isolated from Eupatorium lindleyanum DC) (42), inhibited TNBC tumor growth by suppressing CCNB1 expression. NCAPG is a constituent of the condensin complex, which serves as a major molecular effector of chromosome condensation and segregation during mitosis (43). Upregulated NCAPG expression is significantly associated with adverse prognosis in various malignant tumors, particularly in hepatocellular cancer (44). In TNBC, upregulated NCAPG expression is associated with Ki67 index, a biomarker of mitosis and proliferation of tumor cells (45). MCM4 is part of the MCM2-7 heterohexameric complex, which is important for DNA replication initiation, elongation and replication licensing (46). It has been reported that overexpression of MCM4 is associated with tumor progression, high histological grade and poor survival outcomes in patients with breast cancer (47). Both elevated mRNA and protein expression levels of MCM4 have been observed in TNBC tissues (48). Overexpressed mutant p53 shows a protein-protein interaction with MCM4, and after inhibiting this interaction with the poly ADP-ribose polymerase, TNBC cells with mutant p53 undergo apoptosis (49). RRM2 is an important component of ribonucleotide reductase, which catalyzes the rate-limiting step for DNA synthesis and repair (50). RRM2 expression is elevated in the TNBC subtype, with respect to non-TNBC subtypes (51). Notably, RRM2 expression is upregulated in tamoxifen-resistant breast cancer cells, the effects of which are reversed following inhibition of RRM2 (52), which suggests that RRM2 promotes the conversion of ER-positive to ER-negative subtype. Several studies have demonstrated that upregulated RRM2 expression is associated with oncogenic cellular activities, such as anti-apoptotic, cell proliferation and invasiveness, as well as angiogenesis in breast cancer (53,54). GINS2, a subunit of the DNA replication complex GINS, is crucial to initiation of DNA replication (55). Zheng et al (56) reported that upregulated GINS2 expression is associated with histological grade, metastasis and endocrine therapy resistance in patients with breast cancer. Peng et al (57) confirmed that GINS2 mediates cell cycle progression and proliferation, and that GINS2 knockdown inhibits the migratory and invasive abilities of TNBC cells. The results of the present study demonstrated that GINS2 expression was not significantly elevated in TNBC tissues compared with adjacent normal tissues. This may be attributed to limited samples and imprecise primer extension reaction temperature, which require confirmation with large sample size and perfect reaction conditions.

Taken together, the results of the present study suggest that CCNB1, NCAPG, MCM4 and RRM2 may be potential prognostic factors and therapeutic targets for TNBC. However, further studies, including in vivo and in vitro experiments are required to determine the molecular mechanisms of these genes.

In conclusion, based on bioinformatics analysis of three independent datasets, the present study filtered 377 DEGs of TNBC primarily, which were significantly enriched in the cell cycle process, p53 pathway and DNA replication. Furthermore, the TNBC related PPI network was constructed, consisting of 335 nodes and 6,026 edges. A total of 66 candidate genes with high centrality values in a significant module were identified. Reverse transcription-quantitative PCR analysis revealed that CCNB1, NCAPG, MCM4 and RRM2 were upregulated in TNBC tissue samples, and high expression levels of these oncogenes were associated with unfavorable survival outcomes. In addition, the four oncogenes were significantly associated with tumor cell proliferation. Collectively, the results of the present study provide theoretical guidance for TNBC prognosis evaluation and prospective molecular targeted therapy.

Funding Statement

The present study was supported by the National Natural Science Foundation of China (grant no. 81700639).

Funding

The present study was supported by the National Natural Science Foundation of China (grant no. 81700639).

Availability of data and materials

The gene expression profiles of TNBC included in the present study are accessible through GEO accession numbers GSE38959 and GSE65212. Gene expression data and clinical information of TCGA breast cancer cohort can be accessed at http://gdac.broadinstitute.org/runs/analyses_2016_01_28/data/BRCA.

Authors' contributions

XX and XC contributed toward study conception and design. XX drafted the initial manuscript and acquired the data. ZZ improved the study design and revised the manuscript for important intellectual content. RL and YS performed statistical analysis. RP and JW helped analyze and interpret the data. XX and XC confirmed the authenticity of all the raw data. All authors have read and approved the final manuscript.

Ethics approval and consent to participate

The present study was approved and supervised by the Ethics Committee of the First Affiliated Hospital of Chongqing Medical University (Chongqing, China; approval no. 2020-124). Written informed consent was provided by all patients prior to the study start.

Patient consent for publication

Not applicable.

Competing interests

The authors declare that they have no competing interests.