Potential roles of hsa_circ_000839 and hsa_circ_0005986 in breast cancer

Abstract Background Breast cancer (BC) is one of the leading causes of death among women around the world. Circular RNAs (circRNAs) are a newly discovered group of non‐coding RNAs that their roles are being investigated in BC and other cancer types. In this study, we evaluated the association of hsa_circ_0005986 and hsa_circ_000839 in tumor and adjacent normal tissues of BC patients with their clinicopathological characteristics. Materials and methods Total RNA was extracted from tumors and adjacent non‐tumor tissues by the Trizol isolation reagent, and cDNA was synthesized using First Strand cDNA Synthesis Kit (Thermo Scientific). The expression level of hsa_circ_0005986 and hsa_circ_000839 was quantified using RT‐qPCR. Online in silico tools were used for identifying potentially important competing endogenous RNA (ceRNA) networks of these two circRNAs. Results The expression level of hsa_circ_0005986 and hsa_circ_000839 was lower in the tumor as compared to adjacent tissues. The expression level of hsa_circ_0005986 in the patients who had used hair dye in the last 5 years was significantly lower. Moreover, a statistically significant negative correlation between body mass index (BMI) and hsa_circ_000839 expression was observed. In silico analysis of the ceRNA network of these circRNAs revealed mRNAs and miRNAs with crucial roles in BC. Conclusion Downregulation of hsa_circ_000839 and hsa_circ_0005986 in BC tumors suggests a tumor‐suppressive role for these circRNAs in BC, meriting the need for more experimentations to delineate the exact mechanism of their involvement in BC pathogenesis.


| BACKG ROU N D
Breast cancer (BC) is the most frequent malignancy among women worldwide. About 70%-80% of patients diagnosed in the early stage will be cured. 1 Although its mortality has reduced due to the emergence of currently available medications, it is estimated that BC is the leading cause of about 450,000 deaths per year. 2,3 Various factors affect the progression and response to treatment of BC patients, such as the presence or lack of human epidermal growth factor receptor 2 (HER2/neu), progesterone receptor (PR) or estrogen receptor (ER), and lymph node metastasis as well as tumor size. 2,4,5 Also, it is revealed that various environmental factors and dysregulation of several genes, contribute to the BC progression and metastasis in different ways. [6][7][8][9] Recently, it has been shown that non-coding RNAs, including long non-coding RNAs (lncRNAs), microRNAs (miR-NAs), and circular RNAs (circRNAs) are associated with the pathobiology of breast tumors. 10,11 Among these, circRNAs are less studied and emerging evidence supports their crucial role in BC tumorigenesis. These most recently discovered ncRNAs are associated with the occurrence and development of different abnormalities such as cardiovascular diseases, 12,13 central nervous dysfunctions, 14,15 and different types of cancers. 16,17 CircRNAs may expose a sponge effect on miRNAs and thereby regulate their target mRNAs. They are also involved in adjusting the activity of RNA binding proteins (RBP) and transcription regulation. [18][19][20] Due to their stability and abundance in body fluids, circRNAs are promising biomarkers candidates in cancer. [21][22][23][24] Increasing evidence proposed that circRNAs are involved in proliferation, migration, invasion, and apoptosis of BC. 10,17,25 For example, hsa_circ_001783 has been demonstrated to enhance the proliferation and invasion of BC by sponging miR-200c-3p. Moreover, Liang HF et al. have shown that upregulation of hsa_circ_0008717 can significantly increase the progression and proliferation of BC by sponging miR-1271. 17,26 This study set out to investigate the potential contribution of two circRNAs, hsa_circ_0005986 and hsa_circ_000839, in BC pathogenesis. Hsa_circ_0005986, originated from the PRDM2 (PR/ SET Domain 2) gene, 27 is a promising biomarker for hepatocellular carcinoma (HCC), possibly regulating NOTCH1 expression through sponging and inhibiting miR-129-5p. 28 Hsa_circ_000839, also known as hsa_circ_0000497, is a circRNA derived from the SLAIN gene 27 regulating the expression of miR-200b in HCC. It has been shown that decreased levels of RhoA and hsa_circ_000839 by mir-200b inhibited the migration and invasion of HCC cells. This evidence suggests that hsa_circ_000839/miR-200b/RhoA axis can regulate the invasion and metastasis in HCC. 29 Besides, microarray studies have shown that hsa_circ_000839 could also be a potential biomarker in multiple myeloma. 30 Although studies evaluated the potential roles of these circRNAs in a variety of cancers, to the best of our knowledge, the expression pattern and roles of these circRNAs in BC are still unknown. Therefore, this study is aimed at evaluating the expression pattern of hsa_circ_0005986 and hsa_circ_000839 in BC.

| Patients and tissue sampling
In this study, samples including 50 tumor tissues and 50 adjacent normal tissues were taken from BC patients from Shahid Faghihi hospital, Shiraz, Iran. Patients had not been treated with chemotherapy and radiotherapy before surgery, and samples were collected after taking informed written consent from patients. Tumor and adjacent normal tissues were stored at −80°C after surgery for later use. Clinicopathological and demographic characteristics of patients are provided in Table 1 and Table 2. We categorized patients into two groups based on whether they have used hair dye in the last 5 years ago or not. Also, participants were divided into two groups according to the age at the first full-term pregnancy of <25 or ≥25 years. The study was approved by the local ethical committee at the Fasa University of Medical Sciences (FUMS).

| Real-time PCR
The real-time PCR assay was performed using 10 µl RealQ Plus 2x Master Mix Green with High ROX (Ampliqon, Cat. No: A325402-25), along with 2 µl of cDNA, 1 µl of each primer, and 6 µl DNase-free dH2O for per 20 µl reaction volume. circRNAs sequences were obtained from CircInteractome. 31 Specific primer sequences are shown in Table 3. The relative expression was calculated by the 2 −ΔΔC T (fold change) method.

| Statistical analysis
Data analyses have been carried out by IBM SPSS 21, and p-value <0.05 was considered the statistically significant, and graphical presentation was done using GraphPad Prism 6.0.1 and MedCalc 18.2.1. Wilcoxon test was applied to compare circRNAs expression among tumors and adjacent normal tissues. The association of circR-NAs expression with demographic and clinicopathological features of BC patients was assessed by Mann-Whitney and Kruskal-Wallis tests. Samples were categorized into groups of high and low expressions based on the median fold changes of each gene. Differences between these groups in terms of demographic variables were analyzed using either a chi-square test or an independent t test.

| Survival analysis
To evaluate the importance of mRNAs in the ceRNA network of these circRNAs, we set out to evaluate the association of mRNAs with the survival of BC patients in publicly available datasets. The measure of overall survival would compare the number of patients who had died and not died related to these genes in BRCA. To this end, we used the GEPIA database which performs log-rank survival analysis using TCGA data. GEPIA database uses a logrank test for statistical analysis based on gene expression. 39 The threshold to investigate the importance of survival was log-rank p < 0.05.

| The expression level of hsa_circ_0005986 and hsa_circ_000839 in tumors and paired tumoradjacent tissues
The p-value = 0.022, Figure 1B).

| Relation of hsa_circ_0005986 expression with clinicopathological and demographic characteristics in breast cancer patients
Data analysis revealed that the expression level of hsa_circ_0005986 was significantly lower in patients who have used hair dye in the last 5 years than patients who have not used hair dye in this time period (p-value = 0.012, Figure 2). Furthermore, according to the median for hsa_circ_0005986 expression level, BC patients were divided into high and low expression groups ( Table 2). The comparison of clinicopathological and demographic data between low and high expression groups did not show any significant difference.

| Relation of hsa_circ_000839 expression with clinicopathological and demographic characteristics in breast cancer patients
Comparison of hsa_circ_000839 expression level between patients who had used hair dye in the last 5 years with those who had not used hair dye in this period of time did not show a significant difference (p-value = 0.344, Figure 2), but the correlation analysis revealed that hsa_circ_000839 expression level has a significant negative correlation with body mass index (BMI) (r = −0.357, p-value = 0.01, Figure 3). Similar to hsa_circ_0005986, BC patients were divided into high and low expression groups (Table 2). There was not any significant difference between clinicopathological and demographic data in low and high hsa_circ_000839 expression groups except in the BMI case which could affect the expression of hsa_circ_000839.

| Potential circRNAs-mediated sponge regulatory network
The Regarding the hsa_circ_000839 network, we found the lower expression of hsa-mir-654-3p, hsa-mir-409-3p, and hsa-mir-944 and higher expression of hsa-mir-331-3p and hsa-mir-590-5p (Table 4) in TCGA-BRCA. Among the mRNAs in this network, we observed altered expression of several genes given in Table 5. Additionally, some of these mRNAs showed positive or negative significant correlations, from very weak to fairly strong, to the above-mentioned miRNAs terms for these genes are also given in Figure 6. These GO terms data show their involvement in cancer pathogenesis and pathways. The significant term. All KEGG, BP, and MF GO terms information for these genes are also given in Figure 6.
Concerning the hsa_circ_0005986, our data revealed the significant upregulation of hsa-mir-421 and hsa-mir-188-3p and downregulation of hsa-mir-326 in TCGA-BRCA (Table 6). Moreover, several mRNAs of this network showed altered expression in TCGA-BRCA (Table 7). In this network, then we investigated the correlation between mRNAs and miRNAs and found several significant correlations between them.
However, most of these correlations were very weak (Appendix S2).
Therefore, our data of expression profile and correlation analysis of mRNAs and miRNAs in hsa_circ_000839 and hsa_circ_0005986 indicated the more important role of hsa_circ_000839 in BC with the most important mRNA and miRNAs as follow: RECK, ZEB1, and BTG2 as mRNAs; and hsa-mir-409-3p and hsa-mir-590-5p as miRNAs.

| Survival analysis
We performed survival analysis for 73 mRNAs in the hsa_ circ_000839/miRNA/mRNA network. Two mRNAs (IFNG and IGF1R) were significantly associated with survival prognosis in patients with BC in the TCGA-BRCA dataset (Figure 7). In the hsa_circ_0005986/ miRNA/mRNA network, no mRNA was significantly associated with survival prognosis in patients with BC. clinicopathological and demographic characteristics. To the best of our knowledge, this is the first study that evaluates the association of these two circRNAs in the BC context. In addition, we constructed the ceRNA regulatory network of these circRNAs based on online databases to better figure out their importance in BC circumstances.

| DISCUSS ION
Our results show that the expression level of hsa_circ_000839 is significantly lower in tumor tissues than in normal adjacent ones  patients. Our correlation analysis shows a significant negative correlation between the expression level of hsa_circ_000839 in tumor tissues and BMI, but we could not find any significant association between this circRNA and other clinicopathological characteristics.
In contrast with hsa_circ_000839, the hsa_circ_0005986 did not  Also, the expression of mRNAs and miRNAs in two networks was evaluated by using TCGA data and Gene Ontology (GO) analysis, and the results of these analyses were shown the significance of mRNAs and miRNAs related to these pathways that this issue provides another evidence for the importance of related circRNAs as potential biomarkers in BC.
Also, in addition to several coding and non-coding genes which were shown that have an important role in the pathophysiology of BC, evaluation of genes that were involved in the hsa_circ_000839/ miRNA/mRNA network revealed that IFNG and IGF1R upregulation and downregulation affect the survival of BC patients, respectively, and this can be another evidence that emphasizes the importance of this circRNA in BC ( Figure 7A,B).

| CON CLUS ION
In conclusion, a significant decrease of hsa_circ_000839 and hsa_ circ_0005986 in tumor tissues compared to adjacent normal tissues support the hypothesis that these circRNAs may function as tumor suppressors in BC. Although our bioinformatics analysis reveals the importance of these two circRNAs and their ceRNA regulatory networks. There is little and controversial experimental evidence about the role of both hsa_circ_000839 and hsa_circ_0005986 in tumors, and more functional studies are needed to confirm the exact roles of these circRNAs.

ACK N OWLED G M ENTS
This study was financially supported by 97181(grant number) ethic code IR.FUMS.REC.1397.143 The authors are thankful to patients who took part in this study.

CO N FLI C T O F I NTE R E S T
None.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data used to support the findings of this study are available from the corresponding author upon request.