Circular RNAs (circRNAs) are a special class of non-coding RNA. Different from traditional linear RNA, circRNAs have covalently closed circular structure and lack 5′-3′ polarity or 3′ poly-adenylated tails (Sanger et al., 1976; Jeck et al., 2013). In the beginning, circRNAs were rarely reported by scientists as being low in abundance and might represent splicing errors because of the error-prone mechanism of exon juxtaposition (Bailleul, 1996; Jeck et al., 2013). With the development of high-throughput sequencing technology, circRNAs were discovered in large quantities. They are abundant in various eukaryotic organisms and have certain tissue, timing and disease specificity, which implies that the expression of circRNAs is related to the cellular microenvironment (Salzman et al., 2013; Rybak-Wolf et al., 2015; Han et al., 2018). Moreover, circRNAs have been found in blood, saliva and other body fluids (Bahn et al., 2015; Preusser et al., 2018; Li Y. et al., 2019), even many exosomes have higher ratio of circRNAs than cells (Li Y. et al., 2015). Furthermore, circRNAs are resistant to the degradation of exonuclease RNase R due to their special closed circular structure, which makes them greater stability and longer half-life (Jeck et al., 2013; Chen, 2016; Panda et al., 2017; Xiao and Wilusz, 2019). These characteristics and the development of bioinformatics make circRNAs popular biomarkers for disease diagnosis and the research on drug therapeutic targets.

Circular RNAs are mainly classified into three types: exonic circRNAs (EcircRNAs), exon-intron circRNAs (ElciRNAs), and circular intronic RNAs (ciRNAs). EcircRNAs exist in cytoplasm and can regulate gene expression by limiting the roles of miRNAs (Salzman et al., 2012; Hansen et al., 2013). ElciRNAs and ciRNAs mainly exist in nucleus to act as transcription regulators (Zhang et al., 2013; Li Z. et al., 2015). In addition, it has also been reported that viral RNA genome, transfer RNA, ribosomal RNA, and small nuclear RNA can be cyclized into circRNAs (Lasda and Parker, 2014; Schmidt et al., 2019). CircRNAs have numerous biological functions, including: (1) MiRNAs sponge. CircRNAs have complementary miRNAs binding sites and can competitively bind miRNAs to inhibit their functions. This mechanism is the focus of current research and mainly exercised by ecircRNAs (Hansen et al., 2013; Li R. et al., 2020). (2) Combining with RNA-binding proteins (RBPs). CircRNAs can competitively bind with RBPs to regulate the function of RBPs and have an influence on mRNA stability and splicing patterns (Zang et al., 2020). (3) Regulating Transcription. Although ecircRNAs, as a major part of circRNAs, play an important role in the cytoplasm, elciRNAs and ciRNAs are mainly located in the nucleus and can regulate gene transcription by combining with RNA polymerase or other transcription-related components (Hansen et al., 2013; Chen, 2016). (4) Translation. Some studies demonstrate that circRNAs are associated with translation of ribosomes and N6-Methyladenosine can drive extensive translation of circRNAs (Legnini et al., 2017; Pamudurti et al., 2017; Yang et al., 2017). (5) Interacting with proteins. For example, CircRNA Forkhead box O3 (circFOXO3) functions as a dynamic scaffolding molecule that regulates the interaction between cyclin-dependent kinase 2 (CDK2) and cyclin-dependent kinase inhibitor 1 (P21; Du et al., 2016, 2017b). CircACC1 can directly bind to the β and γ subunits of AMPK, promoting its stability and activity (Li Q. et al., 2019). (6) Functions in exosomes. CircRNAs can enter body fluids under the protection of exosomes to transmit biological information and substances to target cells, and regulate cell growth, epithelial mesenchymal transformation, angiogenesis, and other aspects (Wang Y. et al., 2019; Shi et al., 2020). In conclusion, circRNAs have various functions and participate in the regulation of physiological activities through different pathways. In recent years, circRNAs have often been used in basic research and bioinformatics analysis, reflecting the potential of circRNAs as biomarkers (Li Y. et al., 2019; Wang J. et al., 2020).

Circular RNAs are closely related to the occurrence and development of various human diseases (Aufiero et al., 2019; D’Ambra et al., 2019; Liu et al., 2019; Haque et al., 2020; Li R. et al., 2020). In particular, circRNAs are present in cancer diagnosis, development, drug resistance and circFOXO3 is one of the important ones (Zhang and Xin, 2018; Greene et al., 2019; Chen et al., 2020). The FoxO subfamily of forkhead transcription factors (Fox) widely exists in eukaryotic cells and is involved in the regulation of cell cycle, energy metabolism and tumorigenesis through specific activation of transcription process (Link, 2019). The mammalian system consists of four members, FOXO1, FOXO3, FOXO4, and FOXO6, which are regulated by the PI3K-PKB signaling pathway (Liu et al., 2018). Among them, FOXO3 is widely expressed and highly correlated with a series of malignant tumors such as breast cancer, prostate cancer (PCa) and acute myeloid leukemia (AML; Du et al., 2017a; Zhou et al., 2019; Kong et al., 2020). CircFOXO3 is formed from the exon 2 of FOXO3 and it can not only increase the protein level of FOXO3, but also participate in the post-transcriptional regulation of transcription products through the competitive endogenous RNAs (ceRNAs) network, thus having a dual effect on the development of cancers (Tay et al., 2014; Du et al., 2017a; Zhou et al., 2019). To sum up, the functions of circFOXO3 are complex and important for cancer research. In the following chapters, we will summarize the characteristics of circFOXO3 and its role in cancer through existing studies, so as to provide some basic knowledge for subsequent studies and some inspiration for its research direction in cancers.

CircRNA Forkhead box O3 is a closed circular RNA that contains 1435 nucleotides, formed from the exon 2 of FOXO3 gene (Figure 1A) and the biological functions of it overlap with that of FOXO3 partly. Just like other circRNAs, circFOXO3 has extensive and complex biological functions, which are currently known to be related to cell differentiation, apoptosis and cell cycle progression. For example, Li et al. found that the expression of myogenin (MyoG) and myosin heavy chain (MyHC) was significantly increased by interfering with the expression of circFOXO3. MyoG and MyHC are important marker genes for muscle differentiation, and the effect of circFOXO3 on them can inhibit the differentiation of myoblast cells (Li X. et al., 2019). Furthermore, the overexpression of circFOXO3 is also associated with glutamate-induced oxidative damage in HT22 cell line (Hippocampal neurons from mice). Silencing circFOXO3 can protect HT22 cells by reducing the loss of glutamate-induced mitochondrial membrane potential (Lin et al., 2020). In terms of the effect on cell cycle, circFOXO3 forms ternary complex by combining with CDK2 and P21 (Figure 1B). CDK2 can promote the entry of cell cycle by interacting with cyclin A and cyclin E, while p21 has an opposite effect on cell cycle (Bivik Stadler et al., 2019). CircFOXO3-p21-CDK2 ternary complex blocks the formation of cyclin E/CDK2 complex and eliminates the inhibition of cyclin A/CDK2 complex by p21. As a result, cell cycle is blocked in G1 phase and the process is delayed (Du et al., 2016). By the way, it has also been observed that circFOXO3 in peripheral blood is specifically expressed in the elderly compared to young people, which is related to cellular senescence and has certain predictive significance for human senescence phenotype (Haque et al., 2020).

Sun et al. (2018) fount that ipatasertib, a novel ATP-competitive pan-AKT inhibitor, was used in the treatment of colon cancer to inhibit AKT activity, and then FOXO3 was activated, which up-regulated p53 up-regulated modulator of apoptosis (PUMA), leading to PUMA/Bax-dependent endogenous apoptosis, thereby exerting the anticancer effect of ipatasertib. Similarly, Du et al. (2017a) found that the overexpression of circFOXO3 inhibited the interaction between FOXO3 and mouse double minute 2 (MDM2), prevented MDM2 from inducing ubiquitination and degradation of FOXO3, thus increasing the activity of FOXO3 and promoting the expression of PUMA to induce apoptosis of cancer cells (Lin et al., 2020; Figure 1C). In addition, the main function of circRNAs in cancers is to influence the post-transcriptional regulation of other genes by acting as miRNA sponge (Figure 1D). Using database such as RegRNA 2.0 or Circinteractome, it is very convenient to predict miRNA binding sites through circFOXO3 sequence and some of them have been experimentally confirmed (Dudekula et al., 2016; Liu et al., 2016). For example, circFOXO3 promotes solute carrier family 25 member 15 (SLC25A15) transcription by acting as a miR-29a-3p sponge, affecting the apoptosis and cell cycle of PCa and showing carcinogenic activity (Kong et al., 2020). In glioblastoma (GBM), circFOXO3 also plays a pro-tumor role to adsorb both miR-138-5p and miR-432-5p. It is worth mentioning that the inhibition of circFOXO3 downregulation on GBM can be reversed by miR-138-5p and miR-432-5p inhibitors (Zhang et al., 2019).

CircRNA Forkhead box O3 plays an important role in the occurrence, progression, and prognosis of many cancers. Understanding the expression pattern of circRNAs in cancers is of great significance for predicting the prognosis of patients. However, the expression of circFOXO3 varies greatly in different cancers. For example, circFOXO3 is up-regulated in HCC, GC, and GBM, but down-regulated in ESCC, BC, AML, NSCLC, and breast cancer. The exact reasons for the discrepancy are unclear. This difference may be related to the timing and tissue specificity of circRNAs (Salzman et al., 2013). As Li X. et al. (2019) found in mice, circFOXO3 was most expressed in hearts and least expressed in the kidneys among all organs. Moreover, it is known that circFOXO3 is formed by reverse splicing of the exon 2 of FOXO3, and the regulatory mechanism in this process hasn’t been mentioned yet. There may be some regulatory process that affects the expression of circFOXO3 in cancers. In addition, circFOXO3 is inhibited by miRNAs. MiRNAs can promote the splicing or cleavage of circFOXO3, which also has certain influence on the expression of circFOXO3 in cells (Hansen et al., 2011; Pan et al., 2019). Furthermore, circFOXO3 has its own characteristics in various cancers and is related to many clinical features of patients. The expression of circFOXO3 in PCa is correlated with Gleason score and chemotherapy resistance (Kong et al., 2020; Shen et al., 2020). In GBM, the expression level of circFOXO3 in high-grade tumor tissues is significantly higher than that in low-grade tumor tissues, which suggests a poor prognosis (Zhang et al., 2019). What’s more, circFOXO3 is sensitive to the state of tumor cells, and may exhibit an opposite expression state when stimulated by apoptosis (Du et al., 2017a; Wang C. et al., 2019). In addition, although peripheral blood tests show specific expression of circFOXO3 in the elderly, the expression of circFOXO3 in tumors does not appear to be affected by age (Haque et al., 2020; Kong et al., 2020). These evidences suggest that circFOXO3 has sufficient sensitivity and specificity for the clinical status of patients and can be a promising biomarker.

For a long time, scientists have never stopped searching for a cure for cancers and the discovery of circRNAs, in particular, has provided scientists with a new way to tackle cancers. Among them, circFOXO3 is extensively studied for its diverse functions in cancers. Overall, the expression of circFOXO3 varies among different types of cancer, and even the expression level detected in the same type of cancer remains controversial. However, it is clear that circFOXO3 is involved in the development of cancers by regulating the expression of multiple downstream genes. In addition to affecting the apoptosis, proliferation, migration and invasion of cancer cells, circFOXO3 is also associated with clinical characteristics, chemotherapy resistance, etc. Since the roles of circFOXO3 in different cancers are not consistent, which poses a challenge for circFOXO3 as a potent biomarker, a timely summary of the expression and roles of circFOXO3 in these cancers is necessary.

At present, our understanding of circFOXO3 is insufficient due to the lack of repeated experiments and additional samples. Insufficient experimental data and researchers’ overemphasis on some downstream genes may obscure the truth. Therefore, more experiments are needed to determine the expression characteristics of circFOXO3 in various cancers and at different stages of the same cancer. At the same time, we need to explore a deeper and broader functional mechanism of circFOXO3, which is conducive to discover the dominant mechanism by which circFOXO3 acts on cancer. Furthermore, referring to the roles of circFOXO3 in non-neoplastic diseases, such as its effect on heart disease, may be useful to our exploration of cancer. In conclusion, according to the types of cancer, circFOXO3 is a potential and specific biomarker for predicting the occurrence of cancers and guiding clinical practice. With the development of the technology and the efforts of the researchers, it is expected to discover its potential value and contribute to the treatment of cancer.

DR drafted and revised the manuscript. WH designed the review. CY, JS, and EL participated in the design of the review and helped to draft and revise the manuscript. All authors read and approved the final manuscript.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.