Introduction

Epithelial ovarian cancer is the most fatal kind of malignancy among females [1]. Early detection of ovarian cancer is hindered by the lack of suitable tumor biomarkers, thus disease is usually diagnosed in advanced stages. Due to late diagnosis, this malignancy has the highest fatality-to-case ratio among gynecological cancers [2]. Malignant progression and prompt development of drug resistance are other problems encountered in clinical management of ovarian cancer [3]. The vast majority of ovarian tumors originate from the epithelial surface of the ovary. Others arise from germ cells or stromal cells. The main subclasses of epithelial cancers are serous, endometrioid, mucinous, clear cell, and undifferentiated cancers. These subclasses have different risk factors, clinical behaviors, and treatment responses [3]. From a molecular point of view, both genetic alterations in epithelial cells and reprogramming of the tumor microenvironment contribute in the evolution of ovarian cancer [3]. Understanding the molecular mechanisms participating in these processes would facilitate design of therapeutic modalities and identification of novel tumor markers [4, 5].

Circular RNAs (circRNAs) are a group of non-coding RNAs with a covalently closed configuration [6]. These transcripts have been initially regarded as a splicing error. However, their roles in the regulation of gene expression have been recognized during recent years. These transcripts are produced through back-splicing or exon skipping of precursor mRNAs [7]. These evolutionarily conserved transcripts have a high abundance in the cytoplasm and are more stable than linear transcripts. They can regulate expression of parental genes, modulate alternative splicing events or mRNA translation and act as molecular sponges for miRNAs or RNA-binding proteins. Moreover, they can occasionally produce peptides or proteins [7]. Recent studies have shown contribution of circRNAs in the pathogenesis of cancers [8]. In the current review, we explain the roles of circRNAs in the evolution and progression of ovarian cancer.

Up-regulated circRNAs in ovarian cancer

CircMUC16 is among up-regulated circRNAs in ovarian cancer tissues whose up-regulation in these tissues has been correlated with higher stage and grade. Down-regulation of circMUC16 in ovarian cancer cells has inhibited autophagy flux, while its forced over-expression has increased autophagy flux of cells. The impact of circMUC16 on autophagy has been shown to enhance invasion and metastasis of ovarian cancer cells. This effect has been exerted through binding to miR-199a-5p and releasing Beclin1 and RUNX1 from its suppressive roles. Moreover, RUNX1 has been found to elevate circMUC16 levels through increasing its transcription. Notably, circMUC16 can also directly bind to ATG13 and enhance its expression [9].

circRNA_MYLK is another up-regulated circRNAs in ovarian cancer tissues. Patients with over-expression of circRNA_MYLK have been found to have a more advanced stage and a lower overall survival time. In vitro studies have shown that circRNA_MYLK silencing attenuates proliferation ability of cells. Functionally, circRNA_MYLK can enhance the malignant progression of ovarian cancer cells through regulation of miR-652 levels [10].

Besides, circRNA-UBAP2 has been shown to be up-regulated this type of cancer. CircRNA-UBAP2 silencing has suppressed proliferation of ovarian cancer cells and induced their apoptosis. Mechanistically, circRNA-UBAP2 can target miR-382-5p and down-regulate its expression to release PRPF8 from its inhibitory effects [11].

CircUBAP2 is another circRNA whose over-expression in ovarian cancer tissues has been correlated with clinical stage and survival of patients. This circRNA is mainly located in the cytoplasm. Up-regulation of circUBAP2 could enhance proliferative and migratory capacities of ovarian cancer cells. This circRNA acts as a sponge for miR-144 to release CHD2 from its inhibitory effects [12]. Figure 1 shows the effects of some oncogenic circRNAs in the progression of ovarian cancer.

Fig. 1
figure 1

A schematic representation of the effects of some oncogenic circRNAs in the progression of ovarian cancer. These circRNAs can sponge tumor suppressor miRNAs such as miR-147a, miR-1205, miR-29a, miR-382, miR-637, miR-145, miR-1182, miR-129 and miR-1243, thus increasing expression of certain oncogenes that affect activity of cancer-related signaling pathways

Full size image

In order to find the impact of circRNAs in autophagy, Zhang et al. have assessed expression profile of circRNAs, miRNAs, and mRNAs in ovarian cancer cells after induction with Torin 1. They have reported up-regulation of 504 circRNAs and down-regulation of 478 ones. CircRAB11FIP1 has been among differentially expressed circRNAs. Expression of this circRNA has been found to be higher in epithelial ovarian cancer samples compared with normal ovarian tissues. Its silencing has suppressed the autophagic flux of SKOV3 cells. CircRAB11FIP1 has been shown to directly bind to miR-129 and regulate expression of miR-129 targets ATG7 and ATG14. CircRAB11FIP1 could also bind with DSC1 to assist its interaction with ATG101 [13]. Table 1 summarizes the results of studies that reported up-regulation of circRNAs in ovarian cancer.

Table 1 Up-regulated circRNAs and Ovarian cancer
Full size table

Down-regulated circRNAs in ovarian cancer

A number of studies have reported down-regulation of certain circRNAs in ovarian cancer. For instance, circular RNA-ITCH has been shown to exert tumor suppressor role in this cancer. Down-regulation of circRNA-ITCH in this type of cancer has been associated with up-regulation of lncRNA HULC. Up-regulation of circRNA-ITCH has led to inhibition of cell proliferation, while up-regulation of HULC has resulted in opposite effects. Moreover, up-regulation of circRNA-ITCH has suppressed expression of HULC in these cells. While up-regulation of HULC has not affected expression of circRNA-ITCH, it has decraesed the inhibitory effect of circRNA-ITCH overexpression. Taken together, circRNA-ITCH can suppress proliferation of ovarian cancer cells through down-regulating HULC [77]. Moreover, circRNA-ITCH has been shown to suppress proliferation, invasiveness, and glycolysis of ovarian cancer cells through enhancing expression of CDH1 due to its sponging effect on miR-106a [78].

An RNA sequencing experiment has identified circPLEKHM3 as one of the utmost considerably down-regulated circRNAs in ovarian cancer samples versus normal tissues. Moreover, this circRNA has been found to be down-regulated in peritoneal metastatic ovarian cancers compared with primary cancers. Down-regulation of circPLEKHM3 has also been associated with poor prognosis. Mechanistically, up-regulation of circPLEKHM3 can inhibit cell growth, migration and epithelial-mesenchymal transition, while its silencing has led to opposite consequences. This circRNA acts through sponging miR-9 and regulation expressions of BRCA1, DNAJB6 and KLF4, and activity of AKT1 signaling. Moreover, the tumor-promoting effects of circPLEKHM3 silencing could be blocked by AKT inhibitor MK-2206 [79]. Another study has shown that the tumor suppressor role of curcumin in ovarian cancer is exerted through regulation of circ-PLEKHM3/miR-320a/SMG1 axis [80].

Hsa_circ_0078607 is another tumor suppressor circRNA whose inhibitory roles in ovarian cancer have been verified by different studies. This circRNA has been found to suppress progression of ovarian cancer through regulation of miR-518a-5p/Fas [81] and miR-32-5p/SIK1 [82] pathways. Moreover, down-regulation of this circRNA has predicted poor clinical outcome in high-grade serous ovarian cancer [83]. Figure 2 shows a number of tumor suppressor circRNAs in ovarian cancer.

Fig. 2
figure 2

A schematic representation of the effects of some tumor suppressor circRNAs in the progression of ovarian cancer. The sponging effects of tumor suppressor circRNAs on oncogenic miRNAs such as miR-182, miR-32, miR-145 and miR-740 decrease proliferation and induce apoptosis of ovarian cancer cells. Thus, down-regulation of these circRNAs promotes progression of ovarian cancer

Full size image

CircEXOC6B is another tumor suppressor circRNA that inhibits proliferation and migratory potential of ovarian cancer cells and enhances their sensitivity to paclitaxel via modulation of miR-376c-3p/FOXO3 axis [84]. Moreover, it could progression of this cancer through influencing miR-421/RUS1 axis [85]. Notably, the tumor suppressor circRNA-9119 has been shown to affect miR-21-5p/PTEN/Akt axis [86]. Finally, circ-CDR1as could sequester miR-135b-5p to inhibit progression of ovarian cancer [87]. Moreover, it could up-regulate expression of SCAI to attenuate resistance of ovarian cancer cells to cisplatin through suppression of miR-1270 levels [88].

CircBNC2 is another tumor suppressor circRNA with potential biomarker role. It has been shown to perform better than HE4 and CA125 in differentiating patients with ovarian cancer from those with benign lesions or healthy subjects. Most notably, it could also separate early stage ovarian cancer from benign and healthy conditions. The performance of circBNC2 levels has been similar among pre- and postmenopausal subjects [89].

Table 2 shows the list of down-regulated circRNAs in ovarian cancer.

Table 2 Down-regulated circRNAs in ovarian cancer
Full size table

Discussion

Ovarian cancer is a malignancy with highly variable clinical behavior ranging from good prognosis and high chance of cure to fast progression and poor clinical outcome [3]. This variable clinical manifestation most probably reflects dissimilarity in the biological characteristics of tumors [3]. Recent studies have used bioinformatics tools for identification of dysregulated genes in this kind of cancer to find the most important pathways, targets for treatment and candidate drugs [104].

CircRNAs with prominent roles in determination of cancer cells malignant behavior [105] and response to therapeutic options can explain at least some parts of this variability. These transcripts have critical roles in the regulation of expression of known tumor suppressor genes or oncogenes, since they can sequester miRNAs that suppress expression of these genes [106, 107].

CircRNAs have been shown to participate in the pathogenesis ovarian cancer through sponging miRNAs. CircMUC16/miR-199a-5p, circRNA_MYLK/miR-652, circRNA-UBAP2/miR-382-5p, circRNA-UBAP2/miR-144, circWHSC1/miR-145, circ_0013958/miR-637, circFGFR3/miR-29a-3p, hsa_circRNA_102958/miR-1205, circ_0072995/miR-147a, circ_0072995/miR-122-5p and circEPSTI1/miR-942 are examples circRNAs/miRNA axes in which an oncogenic circRNA acts as a sponge for a tumor suppressor miRNA. On the other hand, circPLEKHM3/miR-9, circPLEKHM3/miR-320a, circ_100395/miR-1228, circ_0078607/miR-518a-5p, circ_0078607/ miR-32-5p, circATRNL1/miR-378 and circEXOC6B/miR-376c-3p are examples of tumor suppressor circRNAs/oncogenic miRNA axes.

Since expression of circRNAs is influenced in the process of carcinogenesis and they are stable in the circulation of patients, circRNAs can act as diagnostic and prognostic markers in ovarian cancer. The former application is highlighted by the stability of these transcripts in the circulation of affected individuals which potentiates them as candidates for non-invasive methods of cancer detection. It is expected that therapeutic modalities affect expression of circRNAs, thus evaluation of expression of these transcripts in the peripheral blood might reveal response to therapy or tumor recurrence. Thus, they might replace the conventional nonspecific ovarian cancer biomarkers. Application of circRNAs as prognostic markers is supported by the studies that reported correlations between their levels and clinical as well as pathological parameters related to cancer prognosis. Future studies are needed to elaborate the association between expression levels of circRNAs and standard staging and grading systems of ovarian cancer. High throughput sequencing techniques would pave the way for identification of stage-/grade-specific panels of dysregulated circRNAs in ovarian cancer.

Moreover, circRNAs can affect response of ovarian cancer cells to paclitaxel. CircCELSR1, circ_CELSR1, circATL2, circNRIP1, circTNPO3 and hsa_circ_0000714 are examples of circRNAs that have important roles in either determination or modulation of chemoresistant phenotype. Since expression levels of these circRNAs affect responses of ovarian cancer cells to chemotherapy, they are putative markers that could be useful for monitoring molecular responses. Epithelial-mesenchymal transition of ovarian cancer cells has also been shown to be affected by a number of oncogenic circRNAs such as circ_100395, circFGFR3, circ_0000554, circCELSR1, circ-PTK2, circLNPEP, circ-CSPP1 and circ_0000745 as well as tumor suppressor ones such as circ_100395 and circPLEKHM3. The impact of non-coding RNAs on activity of cancer-related signaling is a crucial element in the carcinogenesis [108].

Conclusion

Taken together, circRNAs can represent suitable candidate tumor markers in ovarian cancer and therapeutic targets to enhance response of cancer cells to conventional therapies. Moreover, results of in vitro and animal studies have proposed that targeting circRNAs can decrease malignant phenotype of ovarian cancer cells. A prominent limitation of studies conducted in this field is lack of verification of the obtained results in the clinical settings. Future studies are needed to verify these results in the clinical settings. Moreover, the importance of circRNAs in the determination of chemoresistance and possible targeted therapies for combating this phenotype should be assessed in future studies.