Identification of Biomarkers for Breast Cancer Using Databases

Abstract

Breast cancer is one of the major causes of cancer death in women. Many studies have sought to identify specific molecules involved in breast cancer and understand their characteristics. Many biomarkers which are easily measurable, dependable, and inexpensive, with a high sensitivity and specificity have been identified. The rapidly increasing technology development and availability of epigenetic informations play critical roles in cancer. The accumulated data have been collected, stored, and analyzed in various types of databases. It is important to acknowledge useful and available data and retrieve them from databases. Nowadays, many researches utilize the databases, including The Cancer Genome Atlas (TCGA), Gene Expression Omnibus (GEO), Surveillance, Epidemiology and End Results (SEER), and Embase, to find useful informations on biomarkers for breast cancer. This review summarizes the current databases which have been utilized for identification of biomarkers for breast cancer. The information provided by this review would be beneficial to seeking appropriate strategies for diagnosis and treatment of breast cancer.

INTRODUCTION

Breast cancer represents one of the most important health problems worldwide among women.1 A number of studies have shown the primary causes of breast cancer and specific molecules involved in breast cancer. Technologies have been developed to improve early detection of breast cancer. Nowadays, therapeutic compounds, synthetic or natural, that can effectively inhibit or control potential molecular targets are available to increase the survival rate of patients suffering from breast cancer.2–5 However, mortality figures remain high and thus researchers continue to seek new therapies against breast cancer.

Identification of novel biomarkers would be one of the promising approaches for developing new diagnostic and therapeutic strategies. Many biomarkers are used to diagnose and to deeply understand diseases. These are easily measurable, dependable, and inexpensive, with a high sensitivity and specificity. These help not only in screening but also in recurrence detection, as they vary with different stages of disease and have diagnostic and predictive value.6 Diverse techniques have been developed for identifying novel biomarkers in many diseases. One of these techniques is biological information indexing and database provision, which helps to find biomarkers and to better understand biological reactions including invasion, metastasis, and proliferation. Different technologies and techniques have been introduced to find biomarkers and to contribute to diseases characterization. Bioinformation is recently growing in the field of cancer biology. The present review summarizes the current meta-data including The Cancer Genome Atlas (TCGA), Gene Expression Omnibus (GEO), Embase, and Surveillance, Epidemiology and End Results (SEER) (Figure 1).

Figure 1.

A scheme for identification of biomarkers for breast cancer using databases including The Cancer Genome Atlas (TCGA), Genome Expression Omnibus (GEO), Embase, Surveillance, Epidemiology, and End Results (SEER).

THE CANCER GENOME ATLAS

TCGA is a pilot project started in 2006, in collaboration with the National Cancer Institute (NCI) and the National Human Genome Research Institute. The aim of TCGA is to accelerate understanding of diseases by collecting high-dimensional and overall genomic changes in 33 types of cancer such as ovarian, brain, and breast cancer. The TCGA improves prevention, diagnosis, and treatment of cancer via the application of genome analysis and characterization technologies. Tumor DNA and RNA were characterized through a number of approaches in the epigenetic level. The TCGA database has tumor and normal tissues from over 11,000 patients. Researchers who find biomarkers of breast cancer can use TCGA data because it provides accesses without limitation.

Researchers have sought many gene targets from TCGA data, which has been used to find many biomarkers in breast cancer. For example, long non-coding RNAs (lncRNAs) were identified as biomarkers in human breast cancer. A study using 1,000 cases of TCGA data found that LINC00657 plays a crucial role in tumor cell growth and proliferation.7 In addition, FGF 14 antisense RNA2, a novel lncRNA, might act as a tumor suppressor gene and the slow progress of breast cancer.8 The Piwi-interacting RNAs, that play germline maintenance, have been established, and PIWI proteins are potential biomarkers for cancer. Using TCGA dataset, a study found PIWIL3 and PIWIL4 genes using related prognostic relevance.9 MiRNA molecules include many potential biomarkers. MiR-660-5p and miR-574-3p are candidates in breast cancer related to overall survival and recurrence-free survival. BRCA1 expression is regulated by miR-10b, miR-26a, miR-146a, and miR-153 in triple negative breast cancer (TNBC).10,11

Various DNA molecules have been identified as biomarkers for breast cancer. Through TCGA data and patients in the Breast Cancer Care in Chicago, high levels of promoter methylation were found to be strongly associated with hormone-receptor-positive status of breast tumors.12 Matrix metalloproteinase (MMP)-7 expression is related with methylation and it was confirmed from TCGA that hypomethylation of MMP-7 promoter is a prognosis marker.13 Methylation state of DSC2, KCNK4, GSTM1, AXL, DNAJC15, HBII-52, TUSC3, and TES genes may be correlated with worse breast cancer survival in African American.14

RNA-protein complex can be identified as markers. Musashi RNA-binding protein 2 is upstream of ER1 and associated with clinical outcomes.15 The RNA-binging protein Tristetraprolin (TTP, ZFP36) functions as a tumor suppressor, and is related to cAMP response element-binding protein activity regulation. Reduced TTP indicated a poor prognosis in breast cancer.16

Researchers have studied proteins in terms of resistance of anti-cancer drug and ligand. Overexpression of programmed cell death ligand 1 in cell surface generated by phosphatase and tensin homolog loss reduced T-cell proliferation and increased apoptosis, and agents targeting the PI3K pathway might increase the antitumor adaptive immune response.17 Overexpression of pSTAT3 is associated with trastuzumab-resistance in HER2-positive primary breast cancer.18 KLK10 expression was found to contribution to trastuzumab-resistant breast cancer. KLK10 and pSTAT3, associtated with trastuzumab resistance molecules, make them potential marker for diagnosis in breast cancer. The triple-negative breast cancer subtype has a poor survival and high resistance to chemotherapy.19

Taken together, the comprehensive molecular analyses of breast cancer by TCGA Network have significantly broadened our knowledge, which may result in improved therapeutic strategies.20 Additional data and analyses are expected from the breast cancer TCGA studies.21 TCGA is coming to a close in early 2017. New NCI genomics initiatives, which run through the Center for Cancer Genomics, will continue to build upon the success of TCGA by using the same model of collaboration for large-scale genomic analysis.22

GENE EXPRESSION OMNIBUS

GEO is a worldwide data repository that distributes next-generation sequencing, microarray, and other forms of high-throughput functional genomics data. Approximately 90% of the data in GEO are gene expression studies in a broad range of biological themes including ecology, disease, metabolism, toxicology, development, evolution, and immunity. The non-expression data in GEO represent other categories of functional genomic and epigenomic studies, including those that examine chromatin structure, genome copy number variations, genome-protein interactions, and genome methylation. The large volume of data may be effectively explored, queried, and visualized using user-friendly Web-based tools. GEO currently stores approximately a billion individual gene expression measurements, derived from over 100 organisms.23

Several studies have identified novel markers using GEO. Several lncRNAs have shown to be good markers. For example, four lncRNAs genes (U79277, AK024118, BC040204, and Ak000974) have been found by random survival forest algorithm on expression signature, indicating that lncRNAs is involved in breast cancer pathogenesis.24 Several lncRNA signatures can be effective biomarkers for metastatic risk in breast cancer patients, improving our understanding of molecular mechanisms in breast cancer invasion.25 Overexpression of cancer-secreted-miR-105, found in metastasis cancer, is related with metastatic progression in early-stage breast cancer. Down-regulated miR-126/miR-126(*) is correlated with poor metastasis-free survival of breast cancer patients.26,27 The DNA methylation pattern is also a marker of breast cancer and breast cancer molecular subtype survival.28

GEO database has been used to identify EMT and metastasis molecules. Polyomavirus enhancer activator 3 protein (Pea3), a member of the Ets-transcription factor family, is elevated in metastatic progression in various type of solid cancer. Overexpression of Pea3 enhances EMT in human breast epithelial cells via transactivation of Snail, an EMT activator.29 WNT5A and B were found to be overexpressed in MDA-MB-231 triple-negative breast cancer cells, compared to those in MCF-7 ER-positive breast cancer cells. WNT5B destruction by WNT and Jun-N-terminal kinase antagonists enhanced MCF-7 invasion. Using GEO, the same study found alternative WNT receptors ROR1 and 2. In MDA-MB-231 cells, WNT signaling is highly related to breast cancer metastasis to brain and ROR1-2 β-catenin-independent WNT signaling showed WNT5A/B expression.30

GEO dataset is used to identify biomarkers in specific drugs-resistance breast cancer. One of these is Lapatinib, a dual tyrosine kinase inhibitor that interferes with epidermal growth factor receptor and HER2/neu pathways by treating HER2-positive breast cancer. A study presented a novel strategy for researching the mechanism of laptinib-resistance breast cancers.31 The oncogenic isoform of HER2, HER2Δ16, is expressed with HER2 in nearly 50% of HER2 positive breast tumors, and HER2Δ16 drives metastasis and resistance to drugs including tamoxifen and trastuzumab.32 Other molecules such as thyroid hormone receptor-interacting protein 13, DLDN11, CD44, and midkine protein in breast cancer may be potential prognostic biomarkers and therapeutic targets.33–36

GEO database contributed to development of technologies and provided information on specific disease characteristics. The aim is enhancing, indexing, linking, searching and displaying capacity in order to permit more data mining.37 The search engine enlarges the data and it can lead to understanding of unidentified relationships between variety of data types, facilitating novel hypothesis generation, or assisting in the analysis of available information.37–39

SURVEILLANCE, EPIDEMIOLOGY AND END RESULTS

SEER is an authoritative source of information on cancer incidence and survival in the United States. SEER currently collects and publishes cancer incidence and survival data from population-based cancer registries covering approximately 28% of the US population. These SEER data, combined with Medicare data, are used to evaluate the prospect of surviving or dying from cancer or from other reason based on a given set of patient and tumor characteristics.40 SEER data is annually updated and provided as a public service in print and electronic formats. These are used by thousands of community groups, researchers, clinicians, policymakers, public health officials, legislators, and the public.

Several studies focused on receptors to identify biomarkers using SEER data. To evaluate the association between tumor’s estrogen receptor (ER), progesterone receptor (PR), and HER2-neu (HER2) expression and risk of contralateral breast cancer (CBC), SEER datasets were used, including 482 cases diagnosed with a first primary breast cancer (FBC) and a CBC and control women. The TNBC patients increased risk of CBC development compared to patients with HER-2 overexpressing first cancers. HER2-positive breast cancer patients showed a high prevalence of brain invasion and a poor prognosis.41,42 Loss of androgen receptor (AR) expression caused hypermethylation of the AR promoter and AR-negative patients had worse survival.43 However, metaplastic breast cancer was showed no difference survival between hormone-positive and hormone-negative tumors. This data showed that hormone receptors do not improve diagnosis in metaplastic breast cancer.44

Some of studies are analyzed characteristics of diagnosis, survival, and recurrence through age, drug, race, nation, and breast cancer stages. They addressed that these biomarkers, found through differences characteristic, are useful for better treatment, and surgical or radio or hormone therapy.45–48

Proteins biomarkers have been identified as effective therapeutic or prognosis targets. KU70/KU80, DNA damage repair protein, shows higher expression in the BRCA-1 proficient cell line, compared to BRCA-1 deficient cell line.49 Ki-67, a marker for proliferating cells and overexpression in many breast cancers, is a time-varying biomarker of breast cancer in women with atypical hyperplasia.50 High expression of eukaryotic initiation factor 4E is associated with an increased risk for systemic metastasis in node-positive breast cancer patients.51 SEER dataset is used to classify patient survival, status and cancer stage between tumor tissue, positive or negative. Further SEER data may be used to observe clinical outcomes in many cancers with different molecules.

EMBASE

Embase is a biomedical and pharmacological database from biomedical journals. It is special strong in its coverage of drug and pharmaceutical research. There are various studies identifying markers using Embase. Most studies focused on HER-2-positive breast cancer. HER-2-positive metastatic breast cancer is less susceptible to endocrine treatment. The interaction between the response to endocrine treatment and the overexpression of HER-2 was demonstrated in metastatic breast cancer.52 Blocked HER2 was not associated with risk of treatment-related mortality.53

Other studies using Embase show relations between drugs and breast cancer. There are many factors in treating metastatic breast cancer. COX-inhibiting non-steroidal anti-inflammatory drug has protective effects on breast cancer risk and disease progression. COX-2 is expressed in normal breast epithelium, ductal carcinoma in situ of breast (DCIS), and in invasive breast cancer. COX-2 expressions are comparable to DCIS and invasive breast cancer.54 Analysis of Embase data showed that mixed chemotherapy can give a better prognosis than mono-therapy.55–58 Meta-analysis was conducted to find relationships between positive survivin expression and poor overall survival in breast cancer and lymph node metastasis patients. Overexpression of survivin indicated a significantly higher risk of recurrence and reduced overall survival.59,60

Up-regulated enhancer of zeste homologue 2, epigenetic silencing of tumor suppressor genes in cancer, might be associated with pathological types, histological grade, ER negativity, PR negativity, HER-2 positivity, and high expression of p53 in breast cancer.61 Methylation of adenomatous polyposis coli, is known as an antagonist of the Wnt signaling pathway via the inactivation of β-catenin, was higher in breast cancer than controls and significantly reduced in early-stage compared to late-stage breast cancer patients.62

Several microRNAs are clinically proven targets. MiRNA-21, overexpressed in breast cancer, has been found to be a novel biomarker, while some studies demonstrate that miR-21 has a poor prognosis.63–67 Overexpression of miR-155 was found in HER-2- positive or lymph node metastasis-positive, or p53 mutant-type of breast cancer.68

The expression levels of B-cell-specific moloney leukemia virus insertion site (Bmi)-1 differ by race. Increased Bmi-1 expression was a negative predictor for overall survival in Asian patients, whereas overexpressed Bmi-1 was correlated with better overall survival in Caucasian patients.69

Embase data provide wide-ranging information about drug-cancer relationships, molecular expression between races, stages of breast cancer and characteristics of breast cancer. Further studies would be required to investigate more accurate and crucial evidence for prognostic and therapeutic targets using Embase data.

OTHERS

Ovid provides more than 100 core and niche databases to support the breadth of research needed in a wide range of disciplines including clinical medicine and pharmacology. The powerful combination of Ovid’s rich database implementation with Ovid’s advanced search features, sophisticated linking technology, customizable display options, and natural language processing, offers a unique, integrated database solution that is effective for a wide range of users.70 The examples are as follows. HER2 overexpression shows a worse diagnosis and adjuvant therapy in patients with lymph node negative breast carcinoma.71 Positive MMP-9 expression represents a higher risk of recurrence and a worse survival in breast cancer patients.72 Sphingosine kinase 1 (SK1) is important in the pathological cancer genesis, progression, and metastasis processes. SK1 mRNA and protein expression levels were increased in cancer tissues and related to shorter and overall survival.73 Neutrophil gelatinase-associated lipocalin plays a crucial function in cell proliferation, survival, and morphogenesis in breast cancer, and Decoy receptor 3 is involved in development and prognosis of women reproductive cancers, tumorigenesis and progression.74,75

Gene Expression-Based Outcome (GOBO) is a multifunctional online tool that can use a 181-sample breast cancer dataset and 51-sample breast cancer cell line set. GOBO data can be applied to gene expression level, identification of co-expressed genes of potential metagenes, and relation with outcome for gene expression level.76 Transcriptional regulator megakaryoblastic leukemia-1-induced tenascin-C and SAP domain are key players for breast cancer progression and migration.77 The association between let-7c, miR-125a-5p, miR-125b-5p, and miR-21-5p and breast cancer was revealed.78 E26 transformation-specific (ETS) transcription factor regulate the expression of genes involved in biological processes. Friend leukemia integration 1 transcription factor (FLI1) is an ETS protein aberrantly expressed in retrovirus-induced hematological tumors. Down-regulated FLI1 in breast cancer may enhance tumor progression and is related with poor survival and more aggressive phenotypes of breast cancers.79

Small molecule-miRNA Network-Based Inference (SMiR-NBI) is used for predicting new cancer pharmacogenomics mechanisms for small molecules as well as miRNAs. This model was built using a network-based inference algorithm, based on a comprehensive reported dataset. Eleven onco-miRNAs (e.g., miR-20a-5p, miR-27a-3p, miR-29a-3p, and miR-146a-5p) from the top 20 predicted miRNAs provided in the SIiR-NBI model, were down-regulated after metformin in MDA-MB-231 cells. The SMiR-NBi model can be a powerful tool to find potential biomarkers characterized by miRNAs in the emerging field of cancer medicine.80

The Georgetown database of cancer (G-DOC) is a precision medicine platform containing molecular and clinical data from thousands of patients and cell lines, along with tools for analysis and data visualization. The platform enables the integrative analysis of multiple data types to understand disease mechanisms. It integrates molecular and clinical data with patient data and outcomes with patient as the central focus. Increased expression of Ly6 family members was observed in several cancer types such as breast, ovarian, and lung cancer, and it is related with poor outcomes for patient survival using G-DOC and GEO.81

CONCLUSION

This review provides information about biomarkers in breast cancer using databases (Table 1). Most databases offer new comprehensive information on the molecular biology of cancer. The application of updated technology and bioinformatics tools contributes to identification and differentiation of molecules and genome structures, expression levels, and responses to drugs in cancers. Most of the databases provided large volumes of information to researchers without any limitations. The researchers have used the databases to develop candidate breast cancer biomarkers, drug and therapeutic targets, to analyze overall survival and recurrence-free survival and to understand cancer genetic and epigenetic profiles. The targets identified through analysis of the databases undergo clinical trials, and are subsequently used of treatment, prognosis and breast cancer prevention, leading to improvements in personal therapy. In this review, we summarized several databases that are utilized for curing breast cancer patients in various clinical settings, which would be beneficial to the diagnosis and treatment of breast cancer.

Table 1.

Classification of biomarkers and their functions by utilizing databases

Database	Biomarkers	Functions	References No.
The Cancer Genome Atlas (TCGA)	LIC00657, FGF14-AS2	Growth, proliferation	7, 8
	PIWIL3, PIWIL4	Related with overall survival, recurrence-free survival	9
	miR-574-3p, miR-660-5p	Related with overall survival, recurrence-free survival	10
	miR-10b, miR-26a, miR-146a, miR-153	Regulation BRCA1 expression in triple negative breast cancer	11
	DNA methylation	High expression in breast cancer	12–14
	RNA-protein complex (MSI2, TTP)	Related with clinical outcome	15, 16
	PD-L1	Increased the antitumor adaptive immune response	17
	pSTS3, KLK10	Association with trastuzumab-resistance in breast cancer	18, 19
Genome Expression Omnibus (GEO)	U79277, AK024118, BC040204, Ak000974	Metastasis, breast cancer pathogenesis	24, 25
	miR-105, miR126/miR-126(*)	Metastasis	26, 27
	Pea3, WNT5A/B	EMT	29, 30
	HER2Δ16	Resistance-tamoxifen and trastzumab	32
Surveillance, Epidemiology and End Results (SEER)	ER, PR, HER2-neu	Development, invasion, metastasis, worse survival	42
	DNA damage repair protein (KU70/80)	High expression of BRCA-1 proficient cell line	49
	Ki-67	Proliferation	50
	eIF4E	Metastasis	51
Embase	HER2	Metastasis	53
	COX-2	Expression in invasive breast cancer	54
	Survivin	Overexpression in breast cancer	59, 60
	EZH2	High expression of p53	61
	APC	Low expression in early-stage breast cancer patients	62
	miR-21, miR-155	Over expression in breast cancer	63–68
	Bmi-1	High expression in Caucasian patients	69
Ovid	MMP-9	Invasion, metastasis	72
	SK1	Cancer genesis, progression, metastasis	73
	NGAL	Proliferation, survival, morphogenesis	74
	DcR3	Tumorigenesis, progression	75
Gene Expression-Based Outcome (GOBO)	MKl1	Progression, migration	77
	FLI1	Aggressive phenotype in breast cancer	79
Small Molecule-miRNA Network-Based Inference (SMiR-NBI)	11 onco-miRNAs	Up-regulated in MDA-MB-231	80
Georgetown Database of Cancer (G-DOC)	Ly6 family members	Poor outcome	81

FGF14-AS2, FGF 14 antisense RNA2; MSI2, Musashi RNA-binding protein 2; TTP, Tristetraprolin; PD-L1, programmed cell death ligand 1; Pea3, polyomavirus enhancer activator 3 protein; ER, estrogen receptor; PR, progesterone receptor; HER2, HER2-neu; EZH2, enhancer of zeste homologue 2; APC, adenomatous polyposis coli; Bmi, B-cell-specific moloney leukemia virus insertion site; MMP, matrix metalloproteinase; SK1, sphingosine kinase 1; NGAL, neutrophil gelatinase-associated lipocalin; DcR3, Decoy receptor 3; Mkl1, megakaryoblastic leukemia-1.