Expanding roles of circRNA in cancer
circRNAs are the perfect loop of non-canonical RNAs that are formed by alternative splicing. Although they have been ignored over a long period as splicing abnormalities, they have now come to the limelight as dynamic molecules with a wide range of regulatory functions in normal as well as disease physiology, that are distinct from their canonical counterparts. According to recent NGS data, there are about 3 million circRNAs and over 1 million cancer-specific circRNAs in humans. While studies show that over 95% of these are produced as a result of splicing errors, it has also been reported that orthologous circRNAs are evolutionarily conserved in the neural tissue.
The hollow fibre (HF) model in mice provides a platform to rapidly screen anti-cancer compounds simultaneously for PK and efficacy against multiple cell lines to identify promising compounds for proof of concept studies in xenograft models. Efficacy in Hollow Fibre models has been shown to be predictive of efficacy in Xenograft models.
Types of circRNAs
CircRNAs have been classified into 4 types: Exonic circRNAs (ecircRNA), circular intronic RNAs (ciRNAs), exon–intron RNAs (EIciRNAs) and tRNA intronic circRNAs (tricRNAs) and the biogenesis of these RNAs differs accordingly. Among these ecircRNAs account for >80% of circRNAs and are by far the most studied. One of the major mechanisms of circRNA formation is backsplicing, in which a 3′ splice donor site (commonly demarcated by the intronic dinucleotide GU) of an exon is covalently bound to a 5′ splice acceptor (common AG dinucleotide) of the same or upstream exon. Each circRNA therefore possesses a unique sequence of bases created at the ligation point, termed the backsplice junction (bsj).
Formation of circRNAs
Major types of biogenesis mechanisms include a) RNA-binding protein (RBP)-induced circularization: RBPs such as Quaking (QKI), Muscleblind (MBL) and Fused-in sarcoma enhance circularization by bridging related intronic sequences and this in turn induces a closer link between 3’ and 5’ ends of the circularized exons and facilitates splicing. b) intron-pairing driven circularization: In intron pairing, complementary inverted sequence and competition of reverse complementary sequences results in one gene producing different circRNA isoforms, c) lariat-driven circularization: Internal splicing leads to lariat formation, which consequently releases ecircRNAs or EIciRNAs and d) tricRNA splicing enzyme mediated circularization: tRNA splicing enzymes lead to the splicing of pretRNA into tricRNA.
Functions of circRNAs
Formulation of circRNA often limits or interferes with mRNA formation and therefore, there is competition between circRNA and mRNA leading to changes in gene expression and protein production. Studies have shown several conserved miRNA binding sites in circRNAs alluding to the possibility of direct interaction of these RNAs leading to changes in expression of several proteins, especially in pathological conditions.
miRNA sponge: circRNA functions as protein antagonist to inhibit the function of proteins (Fig. 1). circRNAs contain miRNA binding sites and function as miRNA binding sponges to regulate their expression. When splice factor muscleblind (MBL) is highly expressed, it promotes the synthesis of circMBL and prevents the formation of linear MBL Similarly, expression of circ-ITCH is lower in bladder cancer, where it acts as a mRNA sponge and binds to miR-17 and miR-22, to prevent the development of bladder cancer. Another example is the circ-TFRC which is a promoter of bladder cancer where it acts a miRNA sponge by binding to miRNA-107.
Transcriptional regulation and mRNA stability: circRNAs also regulate transcription by binding to RNA-binding proteins (RBPs) or mRNAs or forming a part of the transcriptional scaffold thereby governing the stability of the complex. Some proteins that are relevant in cancer progression such as Yap and eIF4G are known to have corresponding circRNAs that regulate their expression. circRNAs have been known to promote the stability of certain lncRNA and mRNAs. For e.g. CiRS-7/CDR has been shown to stabilize its cognate mRNA by forming an RNA duplex. Similarly, circRNAs also bind to proteins to regulate the stability of other proteins as in the case of circXPO1, which binds to IGFBP2, sequesters it to enhance the stability of CTNNB1 mRNA. Further circRNAs bind to RNA binding proteins to modulate protein translation and stability (Fig. 1).
In addition, circRNAs also play a role in cytoplasm to nuclear translocation of some proteins; CircAmotl1 in samples from tumor patients assists the oncogenic c-MYC (MYC) protein to translocate to the nucleus, and the elevated amount of c-myc enhances its stability and target binding.
CircRNA proteins: Although circRNAs are in general non-coding, some of them are translated in to small peptides or proteins in a cap-independent manner. They are also associated with some cellular function. Again, as with circRNA, they typically antagonize their protein counterpart produced by mRNA thereby modulating cell signalling, usually under pathological conditions.
circRNA in carcinogenesis
Multiple roles of circRNAs have been identified to be pathologically relevant, especially in cancer. There is increasing evidence to suggest that circRNAs play important roles in carcinogenesis in more ways than one. All the above mentioned functions of circRNAs have been associated with carcinogenesis both as tumor suppressors and oncogenes. One of the most studied circRNA is circHIPK3, which is highly abundant in tumor tissues compared to normal tissues. It harbours multiple binding sites for growth suppressive miRNAs thereby sequestering them and aiding in the process of carcinogenesis. On the other hand, circITCH and circFOXO3 bind to several oncogenic miRNAs thereby acting as tumor suppressors. Another recent discovery has highlighted that circRNA plays a role in chromosome translocation as well by forming a DNA-RNA hybrid. circRNA loops are found abundantly in MLL, promoting chromosomal translocation, driving leukemogenesis. Overall, critical roles of circRNAs have been demonstrated in most if not all the hallmarks of cancer, including proliferation, apoptosis, angiogenesis, immune response, metastasis and so on.
circRNA in therapy resistance
Another important role of circRNA appears to be conferring resistance to cancer therapy. Specific circRNAs have been shown to be associated with resistance to either targeted therapy, chemo or immunotherapies. In this regard, both increased and decreased levels of circRNAs have been associated with therapy response. To name a few, resistance to Sorafenib (circSORE), Imatinib (cBA9.3) radiotherapy (circAKT3), Enzalutamide (circRNA17), anti-PD1 (circUHRF1, circMET) therapies, doxorubicin and cisplatin (circPVT1) has been demonstrated based on preclinical data and clinical samples.
circRNA as biomarkers
As we saw earlier, circRNAs have a differential expression pattern in cancer tissue as compared to normal cells. Besides, circRNAs are distributed widely in cell-free saliva, plasma, urine and other tissues in a cell specific manner. They show selective abundance, they are highly conserved and stable. Interestingly, several circRNAs show abundance in blood as compared to their linear counterparts which could make them an ideal choice to be considered as biomarkers. Accordingly, there are multiple studies, including clinical trials that are evaluating the potential of circRNAs as predictive or diagnostic biomarkers. Several circRNAs have already been identified to have significant prognostic and predictive value as biomarkers in multiple cancers. Lately, it’s becoming increasingly convincing to use a set or signature of circRNAs rather than using a single circRNA for enhanced accuracy of prediction. For example, one study demonstrated the efficient use of 5 different circRNAs from urine to detect high-grade prostate cancer and in another 5 circRNA panel was used to differentiate between early and late stage PDAC. Similarly, a signature of circRNA, ICBcircSig has been established in melanoma patients to robustly determine the treatment response to immune checkpoint inhibitor therapies.
circRNA as therapeutics
Based on the differential expression profile of circRNAs, KD and over-expression of circRNAs, has demonstrated the clear roles of circRNAs as oncogenes as well as tumor suppressors. Therefore, targeting these circRNAs through antisense or over-expression using lenti/adenoviral vector systems or nanoparticle mediated delivery into cells has been tried for therapeutic approaches. But, such approaches are still in experimental stages and further research is needed for clinical translation of these therapeutics.
circRNAs in immunetherapy: One interesting study published earlier this year showed that non-canonical translation of circRNAs drives antitumor immunity. This study showed that circFAM53B in dendritic cells encodes a tumor specific antigen that specifically induces strong immune responses in breast cancer. This circRNA also appears to be a positive prognostic marker as observed by breast cancer sample analysis.
Also, peptide vaccine derived from circFAM53B showed a strong effect on tumor growth inhibition of syngeneic tumors in mice models. Several similar studies have shown the potential of circRNAs as both immunogens as well as vaccine adjuvants in multiple cancers. In the clinical setting, although these antigens may be present, the immune-suppressive milieu makes it difficult of the immune cells to access.
It is also interesting to note that the linear counterparts of circRNAs do not elicit the same response as the circRNAs making them unique and important targets for immunotherapy. In this regard, circRNA based vaccines could enhance the availability of the antigens thereby boosting immune response.
Challenges
Considering the huge potential of circRNAs in disease pathogenesis, there’s a lot that needs to be understood about circRNAs in terms of their biogenesis, degradation, transportation, molecular mechanisms and the wide array of functions that they are involved in. For example, further studies are needed to understand a) the competition between miRNA and circRNA in terms of binding partners, b) sharing of the same 5’ UTR miRNA and circRNA, c) enriching of circRNAs in sub-cellular compartments and d) other functions that are potentially shared among miRNA and circRNAs. Also, low abundance and sequence overlap with miRNA make detection of circRNA a challenge, which needs to be overcome with accurate detection techniques. Assessing the copy numbers of circRNA vs. miRNA, splicing patterns, expression patterns and integration of circRNA transcriptome hold the key to better understanding of their function and therapeutic applications.
Conclusion
Ironically, once considered as errors of splicing, circRNAs have now been accepted as new arsenal in the world of non-coding RNAs, opening up plethora of exciting opportunities in research, particularly cancer research. Further understanding and biological/clinical validation of initial research will pave way for standardized use of circRNA in cancer diagnosis, prognosis and therapeutics.
- 1. He et al., Signal Transduction and Targeted Therapy (2021) 6:185
- 2. Kim, Int. J. Mol. Sci. 2024, 25, 10121
- 3. Su et al., Molecular Cancer (2019) 18:90
- 4. PiSignano et al., Oncogene (2023) 42:2783–2800
- 5. Tang et al., Computational and Structural Biotechnology Journal 19 (2021) 910–928
- 6. Huang et al., Nature | Vol 625 | 18 January 2024 | 593
