CircRNAs can specifically interact with different partner molecules, including DNA, RNA and proteins, and function as key regulators for mediating a diverse range of phenotypic changes. In this section, we focus on the expanding landscape of circRNAs in cell cycle regulation, including regulation of cyclins, CDKs, CKIs and related pathways (Table 1, Figure 3).

3.1 CircRNAs regulate cyclins and CDKs

Cyclins and CDKs play versatile roles as regulators in G1/S phase and G2/M phase transitions. The current classification of cyclins and CDKs is based on functional relevance and evolutionary conservation.42, 43 In the ‘classical’ model of the mammalian cell cycle, specific CDK–cyclin complexes drive the numerous events that occur during the interphase in a sequential and orderly manner.27 At the G1/S transition point, cyclin E–CDK2 promotes G1/S progression by phosphorylating pRB as well as several DNA replication-related proteins.44 CDK1 and cyclin B mediate the transition from G2 into M phase. Although cyclin B (with subtypes B1 and B2) is expressed in G1, S and G2 phases, its peak expression and function are found in the M phase.45, 46 A plethora of cell cycle constituents manipulate the machinery either directly or are targeted by ncRNAs, including circRNAs. Interestingly, they can perform a few of these tasks alone, without the need for CDK/cyclin complex formation or kinase activity.25

Current studies are focused on the effect of circRNAs on tumour progression via regulation of diverse cyclins/CDKs and their downstream genes. Mounting evidence indicates that circRNAs function as molecular sponges that absorb miRNAs to regulate cell or tumour progression. Numerous lines of evidence have proven that miRNAs may serve as pivotal modulators of the cell cycle, including G1 to S transition. CircRNAs reportedly promote the cell cycle transition by interacting with cyclin D1 and corresponding regulators. For example, Yang et al. identified abnormal overexpression of circRNA circCHFR in the ox-LDL-mediated vascular smooth muscle cells (VSMCs); circ-CHFR knockdown inhibited VSMC proliferation and migration via circCHFR/miR-370/FOXO1/CyclinD1 axis, implicating circRNAs in VSMCs and atherosclerosis.47 Duan et al. found that circMYLK was markedly up-regulated in laryngeal squamous cell carcinoma (LSCC). Gain-of-function experiments indicated that circMYLK facilitated LSCC cell proliferation and G1/S cell cycle transition. Thus, circMYLK elevated cyclin D1 expression by sponging miR-195.48 Furthermore, CCND1-derived circ-CCND1 as a crucial player that was significantly elevated in LSCC and was associated with aggressive clinical features and adverse prognosis. Loss-of-function studies uncovered that circ-CCND1 suppressed LSCC cell proliferation and tumour growth. Mechanically, circ-CCND1 interacts with HuR protein to strengthen CCND1 mRNA stability as well as promotes CCND1 translation by sponging miR-646. This study identified that circ-CCND1 post-transcriptionally promotes LSCC tumourigenesis.49 Another study identified circPUM1 was significantly up-regulated in lung adenocarcinoma and increased cyclin D1 and Bcl-2 expression by sponging miR-138-5p, thereby facilitating cell proliferation, migration and invasion of lung adenocarcinoma.50 Li et al. identified high expression of hsa_circ_000984 in non–small cell lung cancer (NSCLC) tissues and cell lines, which was linked to advanced TNM stage and lymph node metastasis. Functional assays demonstrated that hsa_circ_000984 silencing inhibited NSCLC cell proliferation, migration, invasion and EMT. In addition, hsa_circ_000984 exerts oncogenic effects by regulating the activation of Wnt/β-catenin pathway, as determined by detecting its influence on the expression levels of β-catenin, c-myc and cyclin D1.51 Thus, the circ-CMPK1/miR-302e/cyclin D1 signalling pathway plays an important regulatory function in NSCLC, and targeting this axis may be an effective treatment strategy.52 Another circRNA, circ5615, was significantly up-regulated in colorectal cancer (CRC) and exhibited an oncogenic role via impacting the cell cycle. Silencing circ5615 suppressed CRC proliferation via miR-149-5p sponging, leading to increase in expression of tankyrase, a regulator of β-catenin stabilization, thereby inducing β-catenin and cyclin D1 expression.53 hsa_circ_0035483 promotes gemcitabine-induced autophagy and augments gemcitabine resistance in renal cancer cells. Silencing hsa_circ_0035483 increases gemcitabine sensitivity and inhibits tumour growth via regulation of miR-335/CCNB1 axis.54 hsa_circ_0136666 promotes breast cancer cell proliferation and induces G2/M phase transition by sponging miR-1299 and up-regulating CDK6 expression.55 CircRNAs regulate cell cycle regulators such as cyclins, CDKs, CKIs, crucial cell cycle effectors. These reports suggest that several circRNAs promote cell cycle transition, thereby playing oncogenic roles.

On the other hand, several circRNAs inhibit the cell cycle transition via diverse molecular mechanisms. For example, circNR3C1 significantly suppresses cell proliferation and cell cycle progression by sponging miR-27a-3p, thereby inhibiting cyclin D1 expression; it was found to be significantly down-regulated in bladder cancer.56 Another study found reduced levels of hsa_circ_0000096 in gastric cancer tissues and cell lines, and its correlation with invasion and TNM stage. Functional assays showed that hsa_circ_0000096 knockdown inhibited cell proliferation by arresting the cell in the G0/G1 phase. Moreover, hsa_circ_0000096 alters protein levels of CDK6, cyclin D1, MMP-2 and MMP-9.57 Taken together, these reports indicate that circRNAs regulate the G1/S transition by directly or indirectly affecting cyclins.

There are, of course, other ways to influence tumour progression, such as by modulating activation of transcription factors, affecting DNA damage or overcoming chemoresistance, or through yet unknown mechanisms.58-62 Triple-negative breast cancer is a lethal disease.63, 64 Yang et al. found that high expression of circAGFG1 in triple-negative breast cancer (TNBC) was linked to clinical stage, pathological grade and poor prognosis. Functional assays confirmed that circAGFG1 could promote TNBC cell proliferation and tumourigenesis in vivo. CircAGFG1 acts as an miRNA sponge; its binding to miR-195-5p increases cyclin E1 expression, indicating that circAGFG1 could serve as a diagnostic biomarker and therapeutic target in TNBC.9 Xie et al.65 proved that the elevated expression of has_circ_0078710 promoted cell proliferation, migration and tumour growth in hepatocellular carcinoma (HCC), which acts as a ceRNA for miR-31 to increase HDAC and CDK2 expression and regulate expression of other cell cycle components, including CDK4, cyclin D1, cyclin A and p21. Zhang et al.66 reported that the adipocyte-derived exocirc-deubiquitination (exo-circ-DB) could sponge miR-34a to activate the USP7/CyclinA2 signalling pathway, thus promoting HCC growth and metastasis and reducing DNA damage.

Accumulating evidence indicates that tumour cells require specific interphase CDKs for proliferation.27 Some of these CDKs are directly up- or down-regulated by circRNAs, while others are indirectly involved. For example, Wang et al. identified circMMP9 up-regulation and its oncogenic role in glioblastoma multiforme (GBM). They found that eukaryotic initiation factor 4A3 (eIF4A3) physically interacts with matrix metallopeptidase 9 (MMP9) mRNA transcript to induce its cyclization, promoting circMMP9 production. CircMMP9 enhanced GBM cell proliferation and tumourigenesis via CDK4 and aurora kinase A (AURKA) by sponging miR-124.67 Similarly, circ_001621 facilitated cell proliferation and migration in advanced osteosarcoma via the miR-578/CDK4/MMP9 axis, highlighting its role as an oncogene by activating VEGF-dependent progression.68 In oral squamous cell carcinoma (OSCC), circRNAs_100290 was significantly up-regulated and repressed CDK6 expression and cell proliferation by sponging miR-29b.69 Circ-ZEB1.33 was found to promote HCC proliferation by regulating CDK6 expression via competitively binding miR-200a-3p. Circ-ZEB1.33 facilitated HCC proliferation by elevating the percentage of S phase mediated by CDK6/Rb. The identification of this key circ-ZEB1.33/miR-200a-3p/CDK6/ axis provides new insight into the potential of circ-ZEB1.33 as an indicator of prognosis in HCC patients.70 Zhong et al.71 screened differential circRNA expression profiles and identified that the circTCF25-miR-103a-3p/miR-107-CDK6 axis facilitates proliferation and migration in bladder carcinoma, thus highlighting the crucial role of circTCF25 and its potential as prospective molecular markers in bladder cancer. cZNF532 regulates pericyte biology by sponging miR-29a-3p and inducing CDK2 up-regulation in pericytes under diabetic stress. Silencing cZNF532 or overexpressing miR-29a-3p aggravated streptozotocin-induced retinal pericyte degeneration and vascular dysfunction.72 Thus, circRNAs regulate cell cycle-related genes in several diseases, suggesting their potential for therapeutic intervention in disease progression.

3.2 CircRNAs regulate CKIs

Considering that most cyclins facilitate CDK activity, CKIs repress CDK activity.25 Based on homology and CDK specificity, CKIs are subdivided into two families: Cip/Kip (CDK-interacting/kinase-inhibiting protein) family and INK4 CKI family. The Cip/Kip family comprises p57 (Cdkn1c), p27 (Cdkn1b) and p21 (Cdkn1a) that mainly inhibit CDK2 activity, thereby inducing cell cycle arrest. INK4 CKI family includes p15 (CDKN2B), p16 (CDKN2A), p18 and p19.73 Recent studies showed that circRNAs orchestrate cell cycle progression by mediating cell cycle inhibitors, such as p21 and p27, in cancers. p21 serves as a crucial regulator of p53-induced cell cycle arrest during distinct checkpoints. p27/Kip1 modulates the cell cycle by repressing the checkpoint kinase CDK2/cyclin E and obstructing cell cycle progression via the G1/S transition.

Emerging studies suggest that circRNAs regulate cell proliferation and progression by affecting the circRNAs/miRNA/p21 axis. Bi et al. identified down-regulation of circ-ZKSCAN1 in bladder cancer tissues and cell lines; circ-ZKSCAN1 suppresses cell proliferation by sponging miR-1178-3p to increase p21 expression.74 Similarly, circ-ITCH that acts as a tumour suppressor to inhibit cell proliferation and tumourigenesis was also down-regulated in bladder cancer. Circ-ITCH promotes the expression of miRNA target gene p21 and PTEN by competitively binding miR-17 and miR-224.75 Recent studies indicate that p27 is also regulated by circRNAs. CircBCRC-3 inhibits bladder cancer proliferation through miR-182-5p sponging to induce p27 expression. Furthermore, methyl jasmonate significantly elevated the expression of BCRC-3, leading to a marked up-regulation of p27.76 CircRNAs impact cell cycle control in other cancer types as well. For example, reduced circMTO1 expression could serve as a prognosis predictor and poor survival in HCC patients. Silencing circMTO1 could reduce the levels of p21, an oncogenic miR-9 target, thereby facilitating HCC cell proliferation and invasion.77 Another loss of function study showed that circLARP4 inhibits HCC cell proliferation and regulates cell cycle arrest. Mechanistic studies revealed that circLARP4 binds to miR-761 to activate downstream p53/p21 targets and promote RUNX3 expression, thereby regulating disease progression.78 Overexpression of circPCNX specifically interfered with the binding between AUF1 and p21 (CDKN1A) mRNA, promoting p21 mRNA stability and elevating its production. Silencing circPCNX increased AUF1–p21 mRNA binding, thus reducing p21 production and promoting cell division.79

Furthermore, circRNAs regulate CKIs through other mechanisms. CircRNAs target miRNAs that in turn target tumour suppressor genes. This axis is used to regulate cell cycle progression and considered an indicator of healing. For example, circ_0021977 was found to suppress CRC proliferation, migration and invasion by modulating the miR-10b-5p/p21 and p53 axis. Low circ_0021977 expression in CRC patients was correlated with higher TNM stage and poorer prognosis.80 Additionally, circYAP1 acts as a tumour suppressor to restrain cell growth and invasion via targeting the miR-367-5p/p27 Kip1 axis, which may offer a promising prognostic indicator of survival in patients with gastric cancer.81

3.3 CircRNAs mediate other pathways

In addition to the p53 pathway, many other signalling pathways are regulated by a range of circRNAs in the cell cycle. Here, we discuss two circRNA-mediated signalling pathways involved in cell cycle regulation: RB-E2F pathway and PI3K/AKT pathway.

3.3.1 RB-E2F Pathway

The transition from G1 to S phase is regulated by the interplay between CDKs and retinoblastoma protein (Rb) phosphorylation, thereby releasing E2F transcription factors to facilitate the expression of S phase genes. It is widely recognized that pRB targets members of the E2F family.82 As subgroups of E2F family, E2F1, E2F2 and E2F3 are the ‘activating’ E2Fs that function as transcriptional activators to induce the G1/S transition.83, 84 The pRB-E2F pathway plays significant roles in manipulating the mammalian cell cycle progression. Moreover, emerging studies report the participation of dysregulated circRNAs in cell cycle progression via maintenance of proliferative signalling in a plethora of diseases. For example, circCAMSAP1 (hsa_circ_0001900) was dramatically increased in CRC tissues and facilitated CRC malignant behaviour. Mechanistical analysis indicated that circCAMSAP1 cyclization was regulated by epithelial-splicing regulatory protein 1 (splicing factor). Moreover, circCAMSAP1 functioned as an miR-328-5p sponge and up-regulated E2F1 expression in CRC.85 Cen et al. reported aberrant circSDHC expression profile in renal cell carcinoma (RCC) patients; it promoted RCC proliferation and invasion by competitively binding miR-127-3p to up-regulate the downstream gene, CDKN3 and the E2F1 pathway.86 Similarly, circ-Foxo3 was highly expressed in ageing cardiac tissues and facilitated cellular senescence by interacting with the anti-senescence proteins E2F1 and ID1 to influence impacted their subcellular translocation. Moreover, overexpression of E2F1 in miR-205-expressing cells could partly reverse the senescent phenotype.87 Another study reported that circRNA CDR1 promotes E2F3 expression by binding miR-7-5p, thereby facilitating nasopharyngeal carcinoma growth and glucose metabolism.88 Highly expressed hsa_circ_0008039 exerted oncogenic effects in breast cancer, while its depletion markedly inhibited cell cycle progression and proliferation via sponging miR-432-5p and elevating E2F3 expression.89

3.3.2 PI3K/AKT Pathway

The phosphoinositide 3-kinase (PI3K) pathway plays an integral role in many cellular processes and is frequently altered in cancer, contributing to tumour growth and survival.90, 91 Small molecule inhibitors have been developed that target the three major nodes of this pathway: PI3K, AKT and mammalian target of rapamycin (mTOR).92-95 CircRNAs have also been shown to target this axis. For example, a study identified that silencing of circ-ZNF609 specifically blocks the G1/S transition via inhibiting phosphorylated Rb:Rb ratio. By contrast, circ-ZNF609 expression promotes cell proliferation in rhabdomyosarcoma, highlighting circ-ZNF609 as a new regulator of cell proliferation-related pathways that abrogates p-Akt proteasome-dependent degradation.96 Similarly, circ-IGF1R was significantly up-regulated, acting as an oncogene in HCC. siRNA-mediated knockdown of circ-IGF1R inhibited cell proliferation and triggered cell apoptosis. Further analysis indicated that circ-IGF1R activates PI3K/AKT signalling pathway, contributing to cell cycle.97 Another study confirmed that exosomal circNRIP1 could be transported between gastric cancer cells and promoted the proliferation, migration and invasion by sponging miR-149-5p to activate the AKT1/mTOR signalling pathway.98 Despite being the most well-studied signalling pathway implicated in cancer molecular mechanisms, the physiological functions of these kinases in cells and organisms are far more complex than previously assumed.99 Therefore, a comprehensive investigation of underlying mechanism may also provide novel insights into the crosstalk in PI3K/AKT pathway and cell cycle regulation.