Introduction

The global statistics indicate that over the past three decades, there has been a persistent increase in the incidence of breast cancer, which has now surpassed lung cancer to become the most prevalent malignant tumor worldwide. Despite the advancements in early detection and treatment, some breast cancer patients experience progression to the metastatic stage without a known cause1. Therefore, it is crucial to identify new biomarkers and unravel molecular mechanisms that can enhance early detection, inform treatment decisions, and predict prognosis in breast cancer.

Circular RNA (circRNA) is a unique class of non-coding RNA molecules. Unlike traditional linear RNA, circRNA molecules form a closed circular structure through a process known as “back splicing”2 and are not susceptible to RNA exonucleases3. This distinctive structure lends to their enhanced stability and resistance to degradation. As a result, circRNA presents notable advantages as a novel clinical diagnostic marker. Increasing evidence indicates that circRNAs exhibit dysregulated expression patterns in various malignant tumors and exert critical regulatory functions in the initiation and progression of tumorigenesis4,5. circRNAs exert biological functions through various means, such as acting as miRNA sponges6, sequestering RNA-binding proteins and serving as protein scaffolds7, functioning as transcriptional regulators8, regulating alternative splicing9 and even serving as templates for protein translation10. Among them, the former represents the predominant mechanism of action for cytoplasmic circRNAs. For instance, circTMEM59 functions as a competing endogenous RNA (ceRNA) by interacting with miR-668-3p, resulting in the enhancement of inhibitor of DNA binding 4 (ID4) expression in colorectal cancer (CRC)11. By acting as a ceRNA, circAGAP1 promotes the expression of E2F3 by sponge of miR-15-5p, thereby exerting its oncogenic functions12. Analogously, circPVT1 functions as a competing endogenous RNA (ceRNA) by sequestering miR-181a-2-3p, leading to the upregulation of ESR1 and downstream ERα-target genes, thereby facilitating breast cancer cell proliferation13. However, the clinical significance and functional mechanism of the majority of circRNAs implicated in the advancement of breast cancer are still poorly understood.

In the present study, we utilized R language to analyze circular RNA microarray data of breast cancer from the GEO database, and identified 6 circRNAs that were downregulated in breast cancer tissues. Then, we conducted predictions to identify the miRNAs that potentially interacted with the differentially expressed circRNAs, as well as the target genes of these miRNAs, and constructed a regulatory network consisting of circRNAs, miRNAs, and mRNAs. Subsequently, we constructed a protein–protein interaction (PPI) network and selected six hub genes. The expression levels and prognostic value of the six hub genes in the PPI network were subsequently assessed using the GEPIA and Kaplan–Meier plotter databases. Furthermore, based on the relationship between circRNAs, miRNAs, and hub genes, a circRNA-miRNA-hub gene network was constructed. Further experimental validation revealed the significant downregulation of hsa_circ_0059665 in breast cancer, which was also associated with poor prognosis in breast cancer patients. The upregulated expression of hsa_circ_0059665 exerts a functional role in suppressing the proliferation, invasion, and migration capacity of breast cancer cells. Mechanically, hsa_circ_0059665 can act as a sponge for miR-602 to suppress the progression of breast cancer. This study offers new perspectives on the circRNA landscape in breast cancer and suggests that hsa_circ_0059665 could potentially be utilized as a promising target for molecular therapy in the treatment of breast cancer.

Materials and methods

Collection and analysis of breast cancer circRNA microarray and mRNA sequencing data

The circRNA expression profile data from GSE182471, which includes 5 breast cancer tissues and 5 adjacent non-tumor tissues, was downloaded from the GEO database (https://www.ncbi.nlm.nih.gov/geo/). The differential expression analysis of circRNAs was performed using the limma package in R, with a significance threshold set at P-adjust < 0.05 and |log2FC|≥ 1. Moreover, the breast cancer RNA sequencing data from the TCGA database (https://portal.gdc.cancer.gov) consisted of 1109 breast cancer tissues and 113 normal tissues. The Deseq package was employed to identify mRNAs exhibiting differential expression, using a criterion of FDR < 0.05 and |log2FC|≥ 2.5.

Prediction and screening of circRNA‑miRNA and miRNA-mRNA binding sites

The prediction of circRNA-miRNA interactions involved the usage of the CircInteractome database (https://circinteractome.irp.nia.nih.gov/). Eventually, miRNAs with a context score percentile of 98 or higher were selected. The miRWalk database (http://mirwalk.umm.uni-heidelberg.de/) was utilized to predict the mRNA targets of the miRNAs.

Construction of the competitive endogenous RNA (ceRNA) regulatory network

The ceRNA regulatory network was constructed by considering the potential interactions among three differentially expressed circRNAs, seven miRNAs predicted to be targeted by these circRNAs, and 161 overlapped mRNAs derived from the list of predicted target genes and downregulated genes in breast cancer. The visualization of the constructed network was achieved using Cytoscape 3.7.1 software.

Construction of the protein–protein interaction (PPI) network and identification of hub genes

A PPI network was constructed using the STRING database (https://string-db.org/), and it was visualized using Cytoscape software. Moreover, the CytoHubba plugin in Cytoscape was employed to screen and identify the hub genes in the constructed PPI network.

Validation of hub gene expression and survival analysis

The GEPIA database (http://gepia.cancer-pku.cn/) was utilized to analyze the mRNA expression levels of the hub genes, and the association between the hub genes and overall survival was evaluated utilizing the Kaplan–Meier plotter database (http://kmplot.com/analysis/).

Cell culture and transfection

Five human breast cancer cell lines, namely MDA-MB-231, MDA-MB-453, SK-BR-3, BT-549, and MCF-7, were selected for investigation, all of which were procured from Procell Life Science & Technology Co., Ltd. (Wuhan, China). All cell lines were cultured in RPMI1640 medium (Gibco, USA) supplemented with 10% Fetal Bovine Serum (FBS) (Gibco, USA) and penicillin/streptomycin. The cells were incubated in a temperature-controlled environment at 37 °C with 5% carbon dioxide (CO2). The overexpression of hsa_circ_0059665 and the empty vector were acquired from Geneseed Biotech, a company based in Guangzhou, China. miR-602 mimics and a miR-negative control (miR-NC) were obtained from RiboBio, a company based in Guangzhou, China. For circRNA transfection, a plasmid of 3 µg was transfected using FuGENE HD transfection reagent (Promega, USA). For miRNA transfection, breast cancer cells were transfected with HiPerFect transfection reagent (Qiagen, Germany) following the instructions provided by the manufacturer.

DNA/RNA extraction, treatment with RNase R, and the Actinomycin D assay

The isolation of total RNA from breast cancer cells was performed using TRIzol reagent (Invitrogen, USA). The extraction of genomic DNA (gDNA) from BC cells was performed using a TIANGEN DNA extraction kit (TIANGEN, China). To perform the actinomycin D assay, the BC cells were pretreated with actinomycin D prior to RNA extraction. In the RNase R assay, the total cell RNA (2 µg) was subjected to digestion with RNase R reagent (Biosearch Technologies, USA) at a concentration of 3 U/µg. The digestion process was carried out at 37 °C for 15 min followed by heating at 85 °C for 3 min. qRT-PCR was performed to investigate the expression of hsa_circ_0059665 and linear ABHD12 after the treatment with RNase R and actinomycin D.

qRT-PCR and nucleic acid electrophoresis

Cells were treated with the TRIzol reagent (Invitrogen, USA) to isolate total RNA, followed by the generation of corresponding cDNA using GoScriptTM (Promega, USA). The expression levels were measured using qRT-PCR with SYBR Green Reagent (Promega, USA). The primer sequences used for PCR amplification can be found in Table S1. The amplified products from RT-PCR were separated by electrophoresis on a 2% agarose gel and visualized using UV irradiation.

Tissue microarray

The HBreD140su03 tissue microarray was the source of 140 breast cancer tissue samples, which were purchased from Shanghai Outdo Biotech. The research protocol received approval from the Ethics Committee of the Fourth Hospital of Hebei Medical University.

Cellular and tissue fluorescence in situ hybridization (FISH) analysis

The hsa_circ_0059665 and miR-602 FISH probes were purchased from GimaPharma Biotech (Guangzhou, China). The detailed steps of this experiment can be found in Zhao’s previous article14. To perform FISH on breast cancer cells, the cells were cultivated on coverslips, fixed with 4% formaldehyde, and then permeabilized using a solution of PBS containing 0.5% Triton X-100. The FISH probes were diluted, denatured, equilibrated, and subsequently added to the breast cancer cells, which were then incubated overnight at 37 °C. In the case of tissue FISH, the tissue microarray slides underwent dewaxing using xylene and ethanol solutions. The subsequent procedures were carried out in accordance with the previously described protocol for FISH in cellular experiments. Following hybridization, the nuclei were counterstained with DAPI. Afterward, imaging was performed using a confocal laser scanning microscope (Zeiss, Germany). The FISH probe sequences were as follows: hsa_circ_0059665 5’-agactgcagggacggtgtgccactgaaaatggatggctct-3’; hsa-miR-602 5’-gggccgcagctgtcgcccgtgtc-3’. Based on the intensity of cytoplasmic expression of hsa_circ_0059665, samples were categorized into the following groups: absence or minimal expression in the majority of cells was classified as the negative group; mild expression in less than 50% of cells or weak expression in most cells was classified as the low expression group; and moderate to strong expression in the majority of cells was classified as the high expression group.

Cell Counting Kit‑8 (CCK‑8) assay

MDA-MB-231 and MDA-MB-453 cells were seeded at a density of 2 × 103 cells per well into 96-well plates, and then CCK-8 (10 μL) was added to each well at 24, 48, 72, and 96 h. Following a 1-h incubation at 37 °C, the absorbance readings were measured at 450 nm using a microplate reader for each well (Tecan Group, Ltd.).

Transwell cell invasion and migration analysis

The Transwell chambers were used to perform migration assays, while the invasion assays were conducted using Matrigel (BD, USA) pre-coated chambers according to the manufacturer’s instructions. Add 200 μl of cell suspension from various groups, with a cell density of 2 × 105 cells per well, to the upper chamber. Incubate the cells for 24 h. Stain the BC cells that have invaded the bottom membrane with crystal violet, capture photographs using an optical microscope at a magnification of 200x, and perform cell counting.

Dual luciferase reporter gene analysis

MDA-MB-453 cells were transfected with Wild-type circRNF10 plasmids using FuGENE-HD transfection reagent in 6-well plates, along with either miR-602 mimic or miR-NC. At 48 h post-transfection, luciferase activity was measured using a dual-luciferase reporter assay system (Promega, USA) according to the manufacturer’s instructions.

RNA Immunoprecipitation (RIP) analysis

MDA-MB-453 cells were co-transfected with miR-602 mimics and Myc-AGO2 vector. The Magna RIP Kit (Millipore, USA) was used to perform the RIP assay. RIP lysis buffer, supplemented with proteinase and RNase inhibitors, was used to lyse MDA-MB-453 cells. Anti-Myc (Abcam, UK) or anti-IgG (Abcam, UK) antibodies were incubated with magnetic beads for 1 h to achieve conjugation. Following that, cell lysates were subjected to overnight immunoprecipitation at 4 °C using magnetic beads. After purification, the expression levels of the targeted circRNAs were quantified using qRT-PCR.

Statistical analysis

The SPSS 26.0 statistical software and GraphPad Prism software were used to analyze the statistical significance of all the experiments. Quantitative data was analyzed using either ANOVA or the Student’s t-test, with a significance level of p < 0.05 determining statistical significance.

Results

Screening for differentially expressed circRNAs in breast cancer

The GSE182471 dataset was obtained from the GEO database, and subsequent differential gene expression analysis was conducted using the R package limma. By applying a threshold of P-adjust < 0.05 and |log2FC|≥ 1.5, a total of 252 differentially expressed circRNAs were identified. Among them, 246 circRNAs were up-regulated, while 6 circRNAs were down-regulated. The volcano plot was used to visualize the differentially expressed circRNAs (Fig. 1A). The heat map was employed to visualize the low expression of six circRNAs in breast cancer (Fig. 1B). The fundamental structural features of the six circRNAs were depicted in Fig. 1C.

Fig. 1
figure 1

Expression profiles of circRNAs in breast cancer. (A) Volcano plot of the differentially expressed circRNAs, green dots represent downregulated circRNAs, and red dots represent upregulated circRNAs. (B) Heatmap of the low expression of six breast cancer-specific circRNAs. (C) The basic structural characteristics of the six circRNA predicted by the circPrimer software.

Full size image

The identification of miRNAs that exhibit binding interactions with the circRNAs was conducted

Growing evidence suggests that several circRNAs, which originate from exons, play vital roles by functioning as “sponges” for specific miRNAs15,16. To assess whether these six circRNAs exhibit similar functionalities in breast cancer, the CircInteractome database was used to screen for miRNAs that have the potential to interact with these circRNAs. A total of seven target miRNAs (miR-585, miR-1281, miR-602, miR-671-5p, miR-495, miR-502-5p, and miR-607) were predicted for the three circRNAs with context scores in the percentile of ≥ 98 (Fig. 2A). The predicted binding sites between circRNAs and miRNAs are depicted in Fig. 2B. Reports indicate that miR-58517, miR-128118, miR-502-5p19, and miR-60720 exert tumor-suppressive roles in cancers, whereas miR-60221, miR-671-5p22, and miR-49523 play oncogenic roles in tumors. The context scores of the remaining three circRNAs are all < 98 (Fig. S1).

Fig. 2
figure 2

Identification of circRNA-bound targeted miRNA. (A) The CircInteractome database was utilized to screen for miRNAs that can potentially interact with the three circRNAs. (B) The putative binding sites between the circRNAs and miRNAs.

Full size image

Establishment of the regulatory network encompassing circRNAs, miRNAs, and mRNAs in breast cancer

A total of 8,243 target genes of the seven abovementioned miRNAs were acquired from the miRWalk database. Moreover, 553 differential expression genes showing downregulation in breast cancer were obtained from TCGA using the criteria of FDR < 0.05 and |log2FC|≥ 2.5 (Fig. 3A). Then, we employed the Venn diagram to identify the overlapping genes between the target gene and the downregulated differentially expressed gene, resulting in the identification of 161 overlapping genes (Fig. 3B). Based on this, we constructed a network consisting of circRNA-miRNA-mRNA interactions by integrating the interactions between circRNA and miRNA, as well as miRNA and mRNA (Fig. 3C). This network shows the potential competitive endogenous RNA (ceRNA) relationships involving three differentially expressed circRNAs (hsa_circ_0059665, hsa_circ_0077930 and hsa_circ_0015962), seven miRNAs (hsa-miR-585-3p, hsa-miR-1281, hsa-miR-602, hsa-miR-671-5p, hsa-miR-495-3p, hsa-miR-502-5p and hsa-miR-607), and 161 differentially expressed genes.

Fig. 3
figure 3

The ceRNA regulatory network of circRNA-miRNA-mRNA in breast cancer. (A) Volcano plot of differentially expressed genes between breast cancer and normal breast tissues in TCGA. (B) Venn diagram indicating 161 overlapping differentially expressed genes determined from TCGA in breast cancer and the predicted target genes. (C) The ceRNA regulatory network of circRNA-miRNA-mRNA in breast cancer. Hexagon represent circRNAs, triangles indicate miRNAs, and ovals indicate miRNAs.

Full size image

Construction of protein–protein interaction (PPI) network and selection of hub genes

To investigate the interactions among the 161 overlapping genes, we utilized the String online database to construct the PPI network. The resulting network was then visualized using Cytoscape software. As shown in Fig. 4A, after eliminating the unconnected nodes, a total of 97 nodes and 165 edges were filtered in this network. Following that, a core module consisting of six hub genes, Kit proto-oncogene, receptor tyrosine kinase (KIT), fibroblast growth factor 2 (FGF2), neurotrophic receptor tyrosine kinase 2 (NTRK2), caveolin-1 (CAV1), leptin (LEP) and lipase E (LIPE), were screened from the PPI network using the CytoHubba plugin in Cytoscape (Fig. 4B).

Fig. 4
figure 4

Constructing a network of protein–protein interactions (PPI) and selecting hub genes. (A) The protein–protein interaction (PPI) network was created utilizing the STRING database and visualized using Cytoscape software. (B) The CytoHubba plugin in Cytoscape was utilized to select the top six hub genes from the PPI network.

Full size image

Survival analysis of hub genes and construction of circRNA–miRNA-hub gene network

The expression levels and prognostic value of the six hub genes in the PPI network were subsequently assessed using the GEPIA and Kaplan–Meier plotter databases. The expression levels of these six hub genes in breast cancer tissues were found to be significantly lower compared to adjacent tissues. Furthermore, the survival analysis results indicated that, except for LIPE, the low expression of the remaining five genes was associated with poorer prognosis in patients (Fig. 5A–F). Then, based on the relationship between circRNAs, miRNAs, and hub genes, a circRNA-miRNA-hub gene network was constructed. As shown in the Fig. 5G, this regulatory network comprises two circRNAs (hsa_circ_0059665 and hsa_circ_0077930), four miRNAs (hsa-miR-1281, hsa-miR-602, hsa-miR-671-5p, and hsa-miR-502-5p), and six hub genes.

Fig. 5
figure 5

The expression levels and prognostic value of the six hub genes in the PPI network were assessed using the GEPIA and Kaplan–Meier plotter databases. (A) Receptor tyrosine kinase(KIT), (B) fibroblast growth factor 2 (FGF2), (C) neurotrophic receptor tyrosine kinase 2 (NTRK2), (D) caveolin-1 (CAV1), (E) leptin (LEP), and (F) lipase E (LIPE). (G) CircRNA–miRNA-hub genes network was constructed. Hexagon represent circRNAs, triangles indicate miRNAs, and ovals represent miRNAs.

Full size image

Characterization of hsa_circ_0059665 in breast cancer

Based on selection criteria such as the length, primer, and probe design for circular RNA, we chose hsa_circ_0059665 for further investigation. Hsa_circ_0059665 derived from exons 4–8 of the host gene ABHD12 with a length of 445nt, and head-to head splicing of hsa_circ_0059665 was established through Sanger sequencing (Fig. 6A). Hsa_circ_0059665 could only be amplified in complementary DNA (cDNA), but ABHD12 mRNA was amplified in both genomic DNA (gDNA) and cDNA (Fig. 6B). Additionally, the stability and subcellular distribution of hsa_circ_0059665 were further evaluated using actinomycin D, RNase R, and fluorescence in situ hybridization (FISH) assays. After treatment with actinomycin D, qRT-PCR analysis revealed that hsa_circ_0059665 exhibited greater stability compared to ABHD12 mRNA (Fig. 6C). The RNase R digestion assay demonstrated that hsa_circ_0059665 displayed resistance to RNase R degradation, whereas the linear ABHD12 was degraded upon RNase R treatment (Fig. 6D). The results of the FISH experiment showed that hsa_circ_0059665 mainly localized to the cytoplasm in MDA-MB-231 and MDA-MB-453 cells. Collectively, the above results indicate that hsa_circ_0059665 is stably present in the cytoplasm of breast cancer cells. Moreover, following the fractionation of nuclear and cytoplasmic RNA in MDA-MB-231 and MDA-MB-453 cells, the qRT-PCR analysis revealed a higher relative expression level of hsa_circ_0059665 in the cytoplasm than in the nucleus (Fig. S2).

Fig. 6
figure 6

Characterization of hsa_circ_0059665. (A) Schematic illustration of hsa_circ_0059665 originating from exons 4–8 of the host gene ABHD12 and Sanger sequencing of hsa_circ_0059665 indicate a back-splice junction. (B) RT-PCR was conducted to validate the presence of hsa_circ_0059665 in breast cancer cells. Divergent primers were used to amplify hsa_circ_0059665 specifically in cDNA, not gDNA. (C) Relative RNA levels of hsa_circ_0059665 and ABHD12 mRNA were measured in breast cancer cells upon treatment with actinomycin D. *** p < 0.001. (D) qRT-PCR was performed to analyze the mRNA expression of hsa_circ_0059665 and ABHD12 in breast cancer cells treated with RNase R. *** p < 0.001. (E) The subcellular localization of hsa_circ_0059665 was assessed using fluorescence in situ hybridization (FISH).

Full size image

The progression of breast cancer is correlated with the expression of hsa_circ_0059665

To investigate the relationship between hsa_circ_0059665 expression and breast cancer prognosis, we conducted a fluorescence in situ hybridization (FISH) analysis on tissue microarrays containing 140 breast cancer tissues to determine the expression of hsa_circ_0059665 (Fig. 7A). The survival analysis results revealed that a lower expression level of hsa_circ_0059665 is strongly associated with a poor prognosis (Fig. 7B). These findings provide evidence for the potential anti-oncogenic role of hsa_circ_0059665 in the progression of breast cancer.

Fig. 7
figure 7

Hsa_circ_0059665 is associated with the prognosis of breast cancer and inhibits the proliferation, migration, and invasion of breast cancer cells. (A) FISH analysis was used to observe the level of hsa_circ_0059665 in breast cancer tissue microarrays. (B) Survival analysis indicates that decreased expression of hsa_circ_0059665 is associated with poor prognosis in breast cancer patients. (C) Expression levels of hsa_circ_0059665 in breast cancer cells. * p < 0.05, ** p < 0.01. (D) Level of hsa_circ_0059665 in breast cancer cells after transfection with hsa_circ_0059665 overexpression plasmid was observed using qRT-PCR. *** p < 0.001. (E, F) CCK8 assay showed that overexpression of hsa_circ_0059665 significantly inhibited proliferation of MDA-MB-231 and MDA-MB-453 cells, respectively. ***p < 0.01. (G, H) The Transwell assay revealed that upregulation of hsa_circ_0059665 inhibited the migratory and invasive abilities of MDA-MB-231 and MDA-MB-453 cells, respectively. * p < 0.05, ** p < 0.01.

Full size image

Upregulated expression of hsa_circ_0059665 inhibits the proliferation, invasion, and migration abilities of breast cancer cells

To investigate the biological effects of hsa_circ_0059665, we first examined the expression levels of hsa_circ_0059665 in five breast cancer cell lines (Fig. 7C). We selected two cell lines with endogenous low expression of hsa_circ_0059665 and transfected them with an overexpression vector of hsa_circ_0059665. The qRT-PCR results demonstrated a significant increase in the expression of hsa_circ_0059665 (Fig. 7D). CCK-8 assay revealed that overexpression of hsa_circ_0059665 inhibited the proliferation ability of MDA-MB-231 and MDA-MB-453 cells (Fig. 7E,F). Transwell cell invasion and migration assay results indicated that overexpression of hsa_circ_0059665 inhibited the invasive and migratory abilities of MDA-MB-231 and MDA-MB-453 cells (Fig. 7G,H). These findings suggest that hsa_circ_0059665 exerts inhibitory effects on the proliferation, invasion, and migration capabilities of breast cancer cells.

Hsa_circ_0059665 functions as a miR-602 sponge in breast cancer cells

The bioinformatics analysis indicates that miR-602 could potentially serve as a downstream target gene regulated by hsa_circ_0059665. Initially, we employed the CircInteractome database to predict the binding sites between hsa_circ_0059665 and miR-602 (Fig. 8A). The region in the secondary structure of hsa_circ_0059665 that is relevant for the binding of miR-602 is depicted in Fig. 8B. Expression of miR-602 in breast cancer cells, as depicted in Fig. S3. To evaluate the influence of miR-602 on the functioning of hsa_circ_0059665, a luciferase promoter assay was conducted. The findings from this experiment revealed a notable reduction in the luciferase activity of luc-hsa_circ_0059665 following the administration of miR-602, suggesting strong suppression by miR-602 (Fig. 8C). The RIP analysis revealed a significant enrichment of hsa_circ_0059665 with anti-Myc antibodies in MDA-MB-453 cells co-transfected with the miR-602 mimics and Myc-AGO2 vector (Fig. 8D). The survival analysis demonstrated a strong correlation between an increased level of miR-602 and a negative prognosis in breast cancer patients (Fig. 8E). We further investigated the cellular localization of hsa_circ_0059665 and miR-602 using FISH analysis. The results revealed that hsa_circ_0059665 and miR-602 co-localized in the cytoplasm of the cells (Fig. 8F).

Fig. 8
figure 8

Hsa_circ_0059665 serves as a miR-602 sponge in breast cancer cells. (A) Predicted potential binding sites of hsa_circ_0059665 and miR-602. (B) The predicted MFE secondary structure of hsa_circ_0059665, as determined by the RNAfold web server, and the specific region in which this secondary structure is relevant for miR-602 binding. (C) Luciferase promoter assay was conducted to measure the luciferase activity of luc-hsa_circ_0059665 in MDA-MB-453 cells that were transfected with miR-602 mimics. * p < 0.05. (D) RIP assay was conducted to study the interaction between hsa_circ_0059665 and miR-602. * p < 0.05. (E) The Kaplan–Meier Plotter database revealed a correlation between miR-602 expression and overall survival in breast cancer patients. (F) FISH analysis was employed to investigate the cellular localization of hsa_circ_0059665 and miR-602.

Full size image

Hsa_circ_0059665 inhibits proliferation, migration, and invasion of breast cancer cells via miR-602

To investigate whether hsa_circ_0059665 exerts its tumor-suppressive effect by acting as a sponge for miR-602, we employed miR-602 mimics to assess whether the inhibitory effects of overexpressed hsa_circ_0059665 on the proliferation, migration, and invasion of breast cancer cell lines can be reversed upon miR-602 upregulation. The CCK8 results demonstrated that overexpression of miR-602 partially reversed the inhibitory effect of hsa_circ_0059665 overexpression on breast cancer cell proliferation (Fig. 9A). The findings from the Transwell cell invasion and migration assays showed that miR-602 mimics partially reversed the suppressive impact of hsa_circ_0059665 overexpression on breast cancer cell invasion and migration (Fig. 9B,C).

Fig. 9
figure 9

Hsa_circ_0059665 regulates miR-602 to inhibit breast cancer cell proliferation, migration, and invasion. (A–C) The proliferation, migration, and invasion of MDA-MB-453 cells, which were concurrently transfected with empty vectors or hsa_circ_0059665 overexpression vectors, and miR-602 mimics or miR-NC, were observed using CCK-8 assays and transwell assays.

Full size image

Discussion

In recent decades, high-throughput sequencing and bioinformatics data analysis have led to the discovery of numerous circRNAs in various diseases24. Abnormal expression of circRNAs has been observed in different types of carcinomas, and they play significant roles in the biology of these cancers, as indicated by numerous studies. Due to their high tissue-specificity, conservation, and stability, circRNAs are considered promising candidates for early diagnosis, prognosis, and as therapeutic targets in carcinoma25. Furthermore, an increasing number of circRNAs, such as circCAPG26, circEZH227, circTBC1D1428, circRNA-CREIT29, and circUHRF130 have been identified as playing significant roles in tumorigenesis, metastasis, autophagy, and chemoresistance of breast cancer. Nevertheless, there still remains a multitude of unexplored roles that circRNAs play in breast cancer.

In this study, we conducted an analysis of circRNA expression profiles obtained from the GEO database to identify key circRNAs involved in breast cancer progression. The study indicated that markers with low expression may be more easily detected in the early stages of tumor development compared to those with high expression. Investigating these low-expression markers can aid in early diagnosis and treatment. Additionally, low-expression markers could potentially serve as new drug targets, and researching these markers may lead to the discovery of novel therapeutic strategies and drugs. Previous studies have shown that low-expression circular RNAs play significant roles in the occurrence and development of tumors6,31. Therefore, this study selected low-expression circular RNAs in breast cancer for further investigation. Ultimately, we discovered six circRNAs that showed down-regulation in breast cancer tissues. Given the abundance of miRNA binding sites within a wide range of circRNAs, circRNAs have the potential to function as molecular sponges by selectively binding to specific miRNAs. This interaction enables them to modulate gene expression through the ceRNA mechanism32,33. To gain a better understanding of the interactions between circRNAs, miRNAs, and mRNAs involved in breast cancer, we constructed a circRNA-miRNA-mRNA regulatory network. In order to gain a comprehensive understanding of the functional mechanism underlying the ceRNA network, we constructed a protein–protein interaction (PPI) network and identified six key genes, namely KIT, FGF2, NTRK2, CAV1, LEP, and LIPE, which play pivotal roles in orchestrating this intricate process. The survival analysis results indicated that, except for LIPE, the low expression of the remaining five genes was associated with poorer prognosis in patients. Then, based on the relationship between circRNAs, miRNAs, and hub genes, a circRNA-miRNA-hub gene network was constructed.

According to selection criteria such as the length, primer, and probe design for circular RNA, we select hsa_circ_0059665 for further investigation. Previous studies have shown that hsa_circ_0059665 is significantly downregulated in patients with acute myocardial infarction34. However, there have been no reports on the expression of hsa_circ_0059665 in breast cancer to date. According to the results of our research, breast cancer is associated with a notable decrease in hsa_circ_0059665 expression, which in turn correlates with an adverse prognosis in patients. Overexpression of hsa_circ_0059665 resulted in significant inhibition of breast cancer cell proliferation, invasion, and migration capabilities. Based on this biological phenomenon, we subsequently explored the potential role and mechanism of hsa_circ_0059665 in breast cancer cells. The bioinformatics analysis presented above suggests that miR-602 may serve as a downstream target gene of hsa_circ_0059665. We have conducted several experimental studies, such as FISH, luciferase reporter gene assay, and RIP, to further validate the accuracy of our prediction. Fortunately, our results provide substantial evidence in favor of the hypothesis that hsa_circ_0059665 functions as a sponge for miR-602.

Previous studies have indicated that miR-602 plays a significant role in various diseases35,36,37. Recent studies have also indicated that miR-602 exerts an oncogenic role in various tumors. For instance, miR-602 has been found to potentially inhibit the JNK signaling pathway by suppressing the expression of RASSF1A. This inhibition of RASSF1A by miR-602 may contribute to the promotion of liver cancer recurrence after surgery38. Moreover, the upregulation of miR-602 has been demonstrated to inhibit the apoptosis of gastric cancer cells39. In addition, the upregulation of miR-602, induced by promoter hypomethylation, is functionally associated with the promotion of ESCC growth and metastasis through the modulation of FOXK221. Our rescue experiment data indicated that overexpression of miR-602 partially reversed the inhibitory effect of hsa_circ_0059665 overexpression on breast cancer cell proliferation, invasion and migration. These findings provide evidence that hsa_circ_0059665 exerts its biological functions by acting as a sponge for miR-602, thereby elucidating the intricate molecular mechanisms underlying breast cancer initiation and progression.

This study has some limitations. Although it suggests that the tumor suppressor hsa_circ_0059665 may regulate breast cancer progression by functioning as a miR-602 sponge, and bioinformatics analysis indicates that hsa_circ_0059665 could influence the expression of downstream target genes LEP and KIT by sequestering miR-602, further experimental studies are required to validate these effects in breast cancer.

Conclusion

In conclusion, our study successfully constructed and comprehensively analyzed a circRNA-miRNA-hub gene regulatory network, demonstrating the inhibitory role of hsa_circ_0059665 as a miR-602 sponge in restraining breast cancer cell proliferation, invasion, and migration. These findings suggested that hsa_circ_0059665 could potentially serve as a valuable biomarker for prognostic assessment and as a target for therapeutic interventions in breast cancer.