Background
Triple-negative breast cancer (TNBC) is defined as a subtype of aggressive breast cancer, accounting for 10–20% of all breast cancer cases [
1]. TNBC subjects lack expression of the estrogen receptor (ER) and progesterone receptor (PR) and does not amplify the human epidermal growth factor receptor 2 (HER2) [
2]. TNBC is more commonly diagnosed among young women and is more prone to relapse and visceral metastasis, compared with other breast cancer subtypes [
3‐
5]. Due to the absence of molecular targets, patients diagnosed with TNBC cannot receive endocrine or HER2 targeted therapy [
6], increasing the difficulty of treatment for them [
7]. Chemotherapy is still the main adjuvant treatment option for patients with TNBC [
8]. TNBC remains a disease associated with poor prognosis and limited treatment options because many tumours are resistant to chemotherapy and rapidly relapse or metastasize after adjuvant therapy [
9]. The identification of uniform targets can help achieve more effective and less toxic treatment. Hence, it is imperative and urgent to explore new therapeutic targets for TNBC [
10].
Recently, many biomarkers have been developed for breast cancer. For example, CD82, a potential diagnostic biomarker for breast cancer [
11]. Furthermore, seven lncRNAs (MAGI2-AS3, GGTA1P, NAP1L2, CRABP2, SYNPO2, MKI67, and COL4A6) detected to be associated with TNBC prognosis, can be promising biomarkers [
12]. Advancements in microarray and high throughput sequencing technologies have provided efficient tools to help in developing more reliable biomarkers for diagnosis, survival and prognosis [
13,
14]. However, the predictive power of a single gene biomarker may be insufficient. Emerging studies have found that gene signatures, including several genes, may be better alternatives [
15]. To the best of our knowledge, the studies about multi-gene prognostic signatures in TNBC are very few, and the functions and mechanisms of mRNAs in TNBC remain to be further explored. Thus, it is necessary to identify more sensitive and efficient mRNA signatures for TNBC prognosis.
In this study, we first identified differentially expressed genes (DEGs), using 1109 BC samples and 113 matched non-cancerous samples from The Cancer Genome Atlas (TCGA). We identified ten hub genes associated with the cell cycle by functional enrichment analysis, protein–protein interaction (PPI) network and survival analysis. In addition, we developed a novel six-gene signature that could effectively predict TNBC survival.
Discussion
TNBC is characterized as a complex and aggressive disease with poor survival rates compared with other subtypes. Only 30% to 45% of TNBC patients achieve a complete pathological response and survival rates similar to other breast cancer subtypes [
21]. The poor prognosis of patients diagnosed with TNBC is mainly due to a lack of effective targets for treatment. Therefore, there is an urgent need for more effective therapeutic targets to improve TNBC prognosis.
Misregulation of the cell cycle is a hallmark of cancer [
22], disorders in mechanisms of cell cycle monitoring and proliferation cause tumour cell growth and tumour cell-specific phenomena. However, it remains unclear if misregulation of periodic mRNAs bears significance in TNBC patient pathogenesis. In this study, a total of 755 DEGs involved in TNBC were screened out from TCGA database, including 590 up-regulated and 165 down-regulated genes. We then built related PPI networks of these DEGs and identified a significant module related to cell cycle, including several key DEGs in the regulatory network of TNBC patients. Subsequently, we identified eight periodic core genes (CCNA2, CCNB2, CDC20, BUB1, TTK, CENPF, CENPA, and CENPE) in the PPI network with higher capacity for PPIs. Coincidentally, all of them were up-regulated genes in TNBC (Fig.
5). CCNA2 (CyclinA2) and CCNB2 (CyclinB2) are members of the cyclin family of proteins that play key roles in the progression of G2/M transition, and have been reported to be the risk factors for resistance and recurrence [
23‐
25]. Importantly, CCNA2, CCNB2, CDC20, BUB1, TTK, CENPA, and CENPE have been reported to be potential therapeutic targets for TNBC [
26‐
29], and TTK inhibitors are currently being evaluated as anticancer therapeutics in clinical trials. These trends are highly consistent with our findings. However, there is no relevant report on CENPF in relation to TNBC; CENPF may be related in patient pathogenesis and as a novel potential therapeutic TNBC target.
Clinical pathological features (Additional file
2: Table S4) are the proper prognostic references for TNBC patients. However, recent studies have demonstrated that clinical predictors are insufficient to precisely predict patient disease outcomes. The mRNA prognostic biomarker has the robust capacity of predicting the survival status of cancer patients. For example, Papadakis et al. [
30] confirmed that mRNA BAG-1 acts as a biomarker in early breast cancer prognosis, Zheng et al. [
31] found that CBX2 is a potential prognostic biomarker and therapeutic target for breast cancer.
However, it is insufficient as the single gene marker to independently predict patient survival. Because a single gene is easily affected by various factors, it is difficult to provide a stable and effective prediction effect. Therefore, we used Cox model analysis to construct a gene signature that includes several genes to enhance prognostic prediction efficiency and sensitivity to TNBC. It has been widely confirmed that combined genetic models are superior to previous single gene markers in disease prediction and diagnoses [
32].
In this study, we constructed a six-mRNA (TMEM252, PRB2, SMCO1, IVL, SMR3B and COL9A3) signature for efficient and sensitive prognosis of TNBC patients. A previous study reported that COL9A3 potentially contributes to the pathogenesis of canine mammary tumours [
33]. In another study, using RNA-seq to identify diabetic nephropathy, the expression of TMEM252, increased in diabetic patients relative to wild-type controls [
34], but we have not found any relevant studies of TMEM252 in tumours. PRB2 is a key factor in regulating ER gene expression. In MCF-7 cells, PRB2 can interact with ER-beta to interfere with ER-beta shuttle between nuclear and cytoplasm [
35], whereas ER-α gene inactivation is mediated by PRB2 in ER-negative breast cancer cells [
36]. These findings suggest that PRB2 may be considered a promising target for TNBC therapy. Only one NCBI article was found to study the function of the single-pass membrane protein with coiled-coil domains 1 (SMCO1), which may contribute to hepatocyte proliferation and have the potential to promote liver repair and regeneration [
37]. However, we have not found any research on SMCO1 in breast cancer; we speculate that it may also play an important role in breast cell proliferation. Additionally, we are not aware of any specific study on SMR3B in tumours, but SMR3B amplification has been detected in osteopontin (OPN)-positive hepatocellular carcinoma [
38]. Involucrin (IVL), a component of keratinocyte crosslinked envelope, is found in the cytoplasm and crosslinked with membrane proteins by transglutaminase. This gene is mapped to 1q21, among calpactin I light chain, trichohyalin, profillaggrin, loricrin, and calcyclin. However, to our knowledge, there is no research on IVL in TNBC.
As far as we know, this is the first established 6-mRNA signature for the prediction of OS time in TNBC, and we have demonstrated the independent prognostic value of this 6-mRNA signature in TNBC.
Conclusions
In summary, through bioinformatic analysis, we identified eight hub genes, correlated with cell cycle, that might be tightly correlated with TNBC pathogenesis. Besides, we constructed a 6-mRNA signature which may act as a potential prognostic biomarker in patients with TNBC, and the prognostic model presented a good performance in OS prediction at 3 and 5 years. These findings will provide some guidance for future TNBC prognosis and molecular targeted therapy. However, our research is based on data analysis, and biological experiments are urgently needed to verify the biological roles of these predictive mRNAs in TNBC.
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.