Identification of a six‐microRNA signature as a potential diagnostic biomarker in breast cancer tissues

Abstract Background Breast cancer (BC) is by far the most common malignancy among women. Epigenetic modulators, microRNAs in particular, may set stages for BC development and its progression. Herein, we aimed to assess the diagnostic potentiality of a signature of six miRNAs (i.e., hsa‐miR‐25‐3p, ‐29a‐5p, ‐105‐3p, ‐181b1‐5p, ‐335‐5p, and ‐339‐5p) in BC and adjacent non‐tumor tissues. Methods A pair of 50 tumor and adjacent non‐tumor samples were taken from BC patients. The expression of each candidate miRNA was measured using quantitative reverse transcription PCR. To investigate the possible roles of each miRNA and their impressions on BC prognosis, in silico tools were used. Receiver operating characteristic (ROC) curves were performed to determine the diagnostic accuracy of each miRNA and the possible association of their expression with clinicopathological characteristics was analyzed. Results Our findings showed the upregulation of hsa‐miR‐25‐3p, ‐29a‐5p, ‐105‐3p, and ‐181b1‐5p, and the downregulation of hsa‐miR‐335‐5p and ‐339‐5p in BC tumor compared to corresponding adjacent tissues. Except for hsa‐miR‐339‐5p, the up‐/down‐regulation of the candidate miRNAs was associated with TNM stages. Except for hsa‐miR‐105‐3p, each candidate miRNA was correlated with HER‐2 status. ROC curve analysis showed that the signature of six‐miRNA is a potential biomarker distinguishing between tumor and non‐tumor breast tissue samples. Conclusion We showed that the dysregulation of a novel signature of six‐miRNA can be used as a potential biomarker for BC diagnosis.


| INTRODUC TI ON
Breast cancer (BC) is responsible for about 25.2% of all women cancers worldwide, suggesting an increasingly rising trend. 1 BC develops as localized disease, but it can metastasize to distant organs (e.g., bone, lung, liver, and brain) and pose the patients' life danger. 2 Accordingly, patients whose disease is diagnosed late often have a low rate of prognosis. Needless to say that if BC is diagnosed as early as possible, it improves the patients' life. 3,4 Among different imaging techniques that are used for BC diagnosis, mammography is still a gold standard technique. 5 This method, on the other hand, is problematic due to false-negative and/ or -positive diagnoses, required biopsy, low sensitivity, and imposing psychological stresses. [6][7][8] Hence, introducing noninvasive methods is necessary that can discriminate tumor and healthy markers as early as possible with good enough sensitivity. Different attempts have been pushed back the frontiers of knowledge so far to develop diagnostic and therapeutic resources for BC patients; for instance, it has been recently identified that non-coding RNAs (e.g., miRNAs) can be used as a biomarker to diagnose the early stages of cancer and also follow-up the patients to determine treatment efficacy. [9][10][11][12] As a class of small non-coding RNAs that are composed of <22 nucleotides, microRNAs (miRNAs and miRs) regulate gene expression at the post-transcriptional level. [12][13][14] These molecules play in a variety of biological processes such as cell proliferation, differentiation, and apoptosis, 12,14 and their dysregulated expression-i.e., up-regulation or down-regulation, 9 has been identified in different tumors. According to their functions, miR-NAs are classified into oncogenes or tumor suppressors. 15,16 It has been also suggested that miRNA expression profiling is of importance because it paves the way toward using these molecules to determine the diagnosis, staging, prognosis, and response to treatment in BC. 10 In this study, we aimed to examine whether a signature of six-miRNA (i.e., hsa-miR-25-3p, -29a-5p, 105-3p, -181b1-5p, -335-5p, and -339-5p) can be used as a biomarker to distinguish BC from adjacent non-tumor tissues (ANT). We selected the candidate miRNAs based on literature reviews and data mining for those that were actively involved in BC pathogenesis. Moreover, we looked into the possible relationships between the expression of these miRNAs and patients' clinicopathological features and HER-2 expression status.
To assess which miRNA discriminates BC from ANT tissues, ROC curve analysis was used. Besides, to show targets and evaluate the prognostic value of each candidate miRNA, in silico analysis was performed.

| Patients
Formalin-fixed paraffin-embedded (FFPE) samples were obtained from BC patients who were admitted by Imam Khomeini Cancer Institute, Tehran, Iran. The study protocol was approved by the Ethics Committee of Islamic Azad University, Tehran, Iran, and before doing the experiment, each participant voluntarily gave their written informed consent. They also signed an informed consent form to collect their tissue samples. Breast tissues were collected using standard operating procedures that had been undertaken at National Cancer Center Hospital. In total, a pair of 50 tumor and ANT samples were collected from the patients and were assessed histopathologically based on the World Health Organization (WHO) criteria for the histologic grade, 17 the TNM system for stage classification, and human epidermal growth factor receptor 2 (HER-2) status. Some important clinicopathological features for these tissue samples such as tumor stage, estrogen receptor, and HER-2 status are summarized in Table 1.

| RNA isolation and quality evaluation
Formalin-fixed paraffin-embedded blocks were cut, mounted on slides, and tumor tissue was scraped into 1.5-ml tubes by needle macrodissection for subsequent RNA extraction. Briefly, 1 ml of xylene was added into the 4 pieces of 20µm-thick FFPE sections to remove traces of paraffin. The tissues were digested with protease K at 50°C overnight. Total RNA was extracted from FFPE tissues using TRIzol reagents (Invitrogen, CA, USA). Before extraction, the samples were washed several times using xylene to solubilize and remove the paraffin. The concentration of RNA samples was determined using the NanoDrop 2000c (Thermo Fisher Scientific), while the integrity was confirmed using 2% gel electrophoresis. To eliminate any remaining DNA contaminations, the samples were treated with RNase-free DNase (Ambion, Austin, TX, USA).

| Choosing candidate miRNA for experimental validation
We selected the candidate miRNAs based on literature reviews and data mining for those that were actively involved in BC pathogenesis. The candidate miRNAs were dysregulated and detectable in human BC. Besides, each of the candidate miRNA must have been annotated straightforwardly in miRBase 22.1. Finally, we selected candidate miRNAs that Conclusion: We showed that the dysregulation of a novel signature of six-miRNA can be used as a potential biomarker for BC diagnosis.

| Functional enrichment analysis
To assess the target potential biological processes and pathways of each candidate miRNA, bioinformatics analyses were undertaken using the Gene Ontology (GO) biological process and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway options of the DIANA-miRPath v3.0. 18 In the context of interactive interaction networks, miRTar-getLink Human was used which provided comprehensive information on human miRNA-mRNA interactions. 19  We only just focused on the 3-untranslated regions as the primary base-pairing regions. We considered the miRNA-gene pairs that were in common in at least 5 databases (p-values <0.05).

| Identification of candidate prognostic markers using Kaplan-Meier method
Prognosis is a critical parameter determining medication efficacy and the relationship between gene expression and disease progression. 21 Since the BC patients' following up was impossible, we used the Kaplan-Meier plotter database 22 and Pan-Cancer Tool 23 to assess the possible relationship between the expression of each miRNA and the prognosis. This  (52) Positive 24 (48) database involves the genome-wide gene expression profiles of more than 5,000 BC samples. We divided the samples into two classes based on the upper quartile expression value of each miRNA, and the log-rank test was used to determine differences in 'overall survival' between the high and low expression groups. Hazard ratio, 95% confidence intervals (CI), and log-rank p-values were calculated.

| Statistical analysis
We used SPSS v. 26 Table S1, the amplification efficiency of each miRNA and internal control was roughly equal with a high linear correlation, showing the validity of the assay. Moreover, dissociation curve analysis was used to endorse the uniqueness and specificity of the amplified products. Single and sharp peaks of the melting curves showed no primer-dimer or non-specific products ( Figure S1). can consummately discriminate BC tumor and ANT tissues ( Figure 1M).

| Correlation between miRNA expression levels and the level of malignancy and HER-2 status
We also assessed the possible association between the signature expression and clinicopathological characteristics of the BC patients.

| Estimation of miRNA biomarker potentiality in BC diagnosis
The AUC-ROC curve was used to assess the sensitivity and specificity of each candidate miRNA analyzed for each candidate miRNA

| Functional enrichment analysis
To show an integrated network for the candidate miRNAs and their potential targets, miRTargetLink Human was used ( Figure S2). We

| Survival analysis of the signature
Using the Kaplan-Meier plotter database, we determined that the overall survival BC is low in cases with the aberrant expression of each candidate miRNA. The follow-up time was determined at a maximum of 300 months ( Figure S4).

| DISCUSS ION
Breast cancer is the most common cancer in females worldwide 2 and despite major advances in early diagnosis, the discovery of more reliable diagnostic and prognostic biomarkers and innovative therapeutic strategies remain the primary goal. It has been sug-

F I G U R E 3
The association of the candidate miRNA expression level and HER-2 status. a higher expression of hsa-miR-25-3p, hsa-miR-29a-5p, and hsa-miR-339-5p was detected in patients who were HER-2 positive. Regarding hsa-miR-105-3p, a statistically insignificant higher expression of this miRNA was detected in HER-2 positive patients. Interestingly, hsa-miR-181b1-5p and hsa-miR-335-5p, on the other hand, were upregulated in HER-2 negative samples. In this figure: *p-value <0.05, **p-value <0.01 Hsa-miR-25-3p promotes tumor cell proliferation by targeting tumor suppressor BTG2; this miRNA is a new diagnostic and therapeutic target in triple-negative BC 28 ; however, its potential to be a reliable biomarker in BC is still unclear. Likewise, exosomal miR-25-3p is involved in pre-metastatic niche formation of colorectal cancer and is used as a biomarker. 29 Herein, we showed that the expression of this miRNA was increased in tumor tissues, especially in the advanced TNM stages and in HER-2 positive patients. MiR-25 cuts both ways because it plays as an oncogenic miRNA in esophageal cancer, 30 cholangiocarcinoma, 31 gastric cancer, 11 and lung cancer, 32 while it is a tumor suppressor in colon cancer 33 and anaplastic thyroid carcinoma. 34 Herein, we showed that hsa-miR25-3p plays an oncogenic role in BC progression, and ROC curve analysis also showed that this miRNA can be used as a biomarker for early diagnosis of BC; in silico analysis demonstrated that this miRNA is dysregulated in BC patients with poor prognosis; however, further studies are needed to unveil the exact molecular mechanisms whereby this miRNA functions in BC development.
Hsa-miR-29a-5p has been detected to play fundamentally in BC development and invasion. 35 Upregulation of this miRNA was also reported in BC serum and tissue samples. 36 MiR-29a inhibits tristetraprolin expression and thus controls BC cell epithelial-mesenchymal transition and metastasis. 37 Given such findings, we hypothesized whether hsa-miR-29a-5p is a BC biomarker; our findings showed stage, and poor overall survival. 44 We also found that the expression of this miRNA is associated with the advanced TNM stages. Li et al. showed that the combined circulating miR-105/93-3p levels are a diagnostic biomarker for early and advanced stages of triple-negative BC. 27 In sum, we showed that hsa-miR-105-3p is a diagnostic Herein, we showed that the hsa-miR-181b1-5p was upregulated in tumor samples in comparison with ANT tissues. This upregulation was also correlated with advanced TNM stages and

HER-2 negative status in BC patients. Different studies in vitro
and in vivo demonstrated the upregulation of miR-181b1-5p in highly metastatic BC cell lines. [45][46][47][48] Overexpression of miR-181b induces breast tumorigenesis and aggressiveness. 45 55 We also demonstrated that the expression of hsa-miR-335-5p increased in the initial stages of tumorigenesis (stages I/II), and was reversely correlated with positive HER-2 status. The AUC-ROC curve analysis suggested that this miRNA is a reliable biomarker with good sensitivity and specificity.
Moreover, aberrant expression of this miRNA was seen in patients with a poor prognosis.
Using in silico available tools, we also showed that hsa-miR-  62 and melanoma. 63 This miRNA has been identified to substantially decrease BC cell migration and invasion capacity. 59 Hsa-miR-339-5p F I G U R E 5 Functional enrichment analysis using in silico databases. (A) The detailed information on miRNA-mRNA interactions in Homo sapiens in the form of interactive interaction networks was obtained using miRTargetLink Human. By selecting the 'Strong Experimental Evidence' option, only three miRNAs had been identified to possibly target specific pathways. These data were confirmed using miRWalk 2.0 ( Figure S2). A high-resolution figure of interaction networks is accessible in Figure S3. In this study, we demonstrated that a signature of six miRNAs (i.e., hsa-miR-25-3p, -29a-5p, 105-3p, -181b1-5p, -335-5p, and -339-5p) effectively distinguishes the tumor and ANT tissues with acceptable sensitivity and specificity; however, further steps should be taken forward to show the underlying molecular mechanisms whereby such miRNAs function in BC development.

ACK N OWLED G M ENTS
We gratefully thank our colleagues for their excellent technical assistance and advice.

CO N FLI C T O F I NTE R E S T
The authors declare no conflict of interest.