A Concise Review on the Role of CircPVT1 in Tumorigenesis, Drug Sensitivity, and Cancer Prognosis

CircPVT1 (hsa_circ_0001821) is a cancer-related circular RNA (circRNA) that originated from a genomic locus on chromosome 8q24. This locus has been previously found to encode the oncogenic long non-coding RNA PVT1. Expression of this circRNA has been found to be upregulated in diverse neoplastic conditions. CircPVT1 acts as a sponge for miR-125a, miR-125b, miR-124-3p, miR-30a-5p, miR-205-5p, miR‐423‐5p, miR‐526b, miR-137, miR-145-5p, miR-497, miR-30d/e, miR-455-5p, miR-29a-3p, miR-204-5p, miR-149, miR-106a-5p, miR-377, miR-3666, miR-203, and miR-199a-5p. Moreover, it can regulate the activities of PI3K/AKT, Wnt5a/Ror2, E2F2, and HIF-1α. Upregulation of circPVT1 has been correlated with decreased survival of patients with different cancer types. In the current review, we explain the oncogenic impact of circPVT1 in different tissues based on evidence from in vitro, in vivo, and clinical investigations.


INTRODUCTION
Circular RNAs (circRNAs) are single-stranded covalently enclosed uninterrupted loops with no free end or polyadenylated tail (1). These transcripts are prevalent in human transcriptome since approximately 20% of active genes have the potential to produce circRNAs (1,2). In fact, circRNAs are a group of long non-coding RNAs (lncRNAs). Compared with linear ncRNA, circRNAs have more stability, since their circular structure protects them from degradation by the majority of RNA decay mechanisms (3,4).
Being mainly produced by back-splicing, circRNAs consist of exonic and/or intronic regions. Back-splicing is a non-canonical alternative RNA splicing process that is facilitated by the spliceosomes and contribution of a number of cis-or trans-acting factors (5). Biogenesis of circRNA is under control of numerous cis-and trans-acting factors (6). The splice sites, enhancers, and silencers, particularly element adjacent to the junction sites, including the inverted Alu repeat segments are examples of the cis-regulatory factors (7). Spliceosome elements, RNA helicases, and RNA-binding proteins are among trans-regulatory factors in regulation of circRNA biogenesis (8).
CircRNAs have functional roles in the regulation of gene expression through competitively binding and sponging miRNAs. This action of circRNAs leads to the stabilization of miRNA targets (5). This mode of action of circRNAs is well assessed. In fact, a number of circRNAs have numerous binding sites for single or numerous miRNAs (5). In addition, a number of circRNAs can sponge proteins and block their activity (9). Some circRNAs can bind to numerous proteins and keep them together. These circRNAs serve as a scaffold to enable interactions of these proteins (5). Thus, in addition to sponging miRNAs, they regulate gene expression via interacting with several proteins. There is also evidence that certain circRNAs can produce proteins (10).
Three major categories of circRNA have been identified: exonic circRNAs, circular intronic RNAs, and exon-intron circRNAs (11). Exonic circRNAs mostly serve as miRNA sponges. Thus, they increase expression of miRNA targets through adsorption of miRNAs. However, intron-containing circRNAs (including both circular intronic RNAs and exonintron circRNAs) are mainly located in the nucleus where they modulate transcription of certain genes (12,13).
CircRNAs partake in the regulation of all principal hallmarks of malignancy and are considered as promising markers for diagnosis and prediction of course of cancer (5). CircPVT1 (hsa_circ_0001821) is an example of a cancer-related circRNA that originated from a genomic locus on chromosome 8q24 (14). This locus has been previously found to encode the oncogenic lncRNA PVT1. The CircInteractome Database (https:// circinteractome.nia.nih.gov/index.html) has listed 26 isoforms for circPVT1 (15). The spliced length of these isoforms ranges from 113 to more than 11,000 nucleotides, with the most frequent isoform being 410 nucleotide long. CircPVT1 is produced by back-splicing and encompasses the entire length of exon 2 of PVT1 (16). On the other hand, some of the identified 26 isoforms of lncRNA PVT1 do not have exon 2 (17). These alternatively spliced variants have 5′ cap and polyadenylated tail at 3′ end (18).
Expression of circPVT1 has been assessed by different methods. Current methods usually use simple statistical methods or differential expression analysis strategies developed for linear RNAs. As the majority of circRNAs have very low levels of expression, RNase R treatment is typically used for enrichment of circRNAs. When RiboMinus/RNase R-treated RNA-seq libraries are used, alterations in enrichment coefficient in the RNase R treatment phase might lead to bias in estimation of circRNA expression (19).
A high-throughput RNA sequencing experiment for comparison of circRNA signature in proliferating versus senescent human fibroblasts has identified circPVT1 as a downregulated transcript in senescent fibroblasts. Further experiments have indicated that downregulation of circPVT1 expression in proliferating fibroblasts induces their senescence, as being evident by upregulation of senescence-related bgalactosidase level, upregulation of CDKN1A/P21 and TP53, and decrease in proliferation rate. These effects are most probably mediated through modulation of let-7 levels and consequent alteration in levels of let-7-regulated transcripts, including IGF2BP1, KRAS, and HMGA2 (16).
Among different cancer types, circPVT1 has been primarily identified as an upregulated circRNA in gastric cancer specimens compared with corresponding normal tissues (14). Subsequently, overexpression of this circRNA has been verified in other types of malignancies. In this review, we explain the oncogenic roles of circPVT1 in different tissues based on evidence from in vitro, in vivo, and clinical investigation.

IN VITRO STUDIES
CircPVT1 has been found to increase proliferation of gastric cancer cell through serving as a molecular sponge for miR-125 family members (14). Moreover, expression of circPVT1 has been higher in paclitaxel-resistant gastric cancer cells. CircPVT1 silencing has improved sensitivity of gastric cancer cells through modulating miR-124-3p levels. Since ZEB1 is a direct target of miR-124-3p, circPVT1 enhances expression of ZEB1 through sequestering this miRNA (20). Exosomal levels of circPVT1 have also been higher in cisplatin-resistant gastric cancer cells parallel with downregulation of miR-30a-5p. CircPVT1 silencing has suppressed cisplatin resistance of gastric cancer cells through inducing apoptosis and reducing invasion or autophagy. Functionally, circPVT1 modulates expression of YAP1 through influencing expression of miR-30a-5p (21).
In breast cancer cells, upregulation of circPVT1 has been accompanied with underexpression of miR-29a-3p. Suppression of circPVT1 or upregulation of miR-29a-3p could block proliferation, invasiveness, and migratory potential of breast cancer cells while promoting their apoptosis. Mechanistically, circPVT1 binds with miR-29a-3p to release AGR2 from its inhibitory effect. AGR2 has been found to increase expression of HIF-1a and then accelerated malignant features of breast cancer cells (22). CircPVT1 has also been shown to promote invasiveness and epithelial-mesenchymal transition (EMT) of neoplastic breast cells through sequestering miR-204-5p (23).
Expression of circPVT1 has also been upregulated in human epithelial ovarian cancer cells. In both SKOV3 and CAOV3 cells, suppression of circPVT1 has decreased cell proliferation and enhanced cell apoptosis. CircPVT1 has been found to negatively regulate miR-149 (24).
CircPVT1 has been demonstrated to be upregulated in osteosarcoma cells parallel with upregulation of c-FLIP and downregulation of miR-205-5p. CircPVT1 silencing has suppressed proliferation, migration, and invasiveness of osteosarcoma cells through inhibiting EMT. This circRNA sponges miR-205-5p and increases expression of c-FLIP (25). Another study in osteosarcoma has revealed downregulation of miR-423-5p while upregulation of Wnt5a/Ror2 and circPVT1. MiR-423-5p has a role in inhibition of glycolysis and suppression of cell proliferation, migration, and invasiveness through influencing expressions of Wnt5a and Ror2. CircPVT1-mediated silencing of miR-423-5p leads to activation of Wnt5a/Ror2 signaling (26). CircPVT1 also enhances metastasis of osteosarcoma through modulation of miR-526b/FOXC2 axis (27).
In addition, circPVT1 affects response of osteosarcoma cells to chemotherapeutic medications since its silencing has decreased chemoresistance of osteosarcoma cells to doxorubicin and cisplatin through reducing levels of ABCB1 (28). CircPVT1 also participates in doxorubicin resistance of these cells through miR-137-TRIAP1 axis (29).
CircPVT1 contributes in the malignant behaviors of lung cancer via different routes. It induces chemoresistance via modulation of the miR-145-5p/ABCC1 signals (30). In addition, it enhances proliferation and invasion of lung cancer cells via sequestering miR-125b and enhancing E2F2 signals (31). CircPVT1 also serves as a sponge for miR-497 to increase levels of Bcl-2 lung cancer cells (32).
In oral squamous cell carcinoma, circPVT1 has been found to sponge miR-125b and miR-106a-5p to release STAT3 and HK2 from their inhibitory effects (33,34). In acute lymphoblastic leukemia, the oncogenic role of circPVT1 is mediated through upregulation of Bcl-2 and c-Myc (35).
Finally, miR-199a-5p and miR-145-5p have been identified as targets of circPVT1 in glioblastoma (39) and clear cell renal cell carcinoma (40), respectively. Figure 1 shows the oncogenic roles of circPVT1 in different cancer types.

ANIMAL STUDIES
Animal studies have shown the role of circPVT1 suppression on enhancement of cisplatin sensitivity of gastric cancer through miR-30a-5p/YAP1 axis (21). Moreover, circPVT1 silencing could increase drug sensitivity in osteosarcoma models (29). Other studies have consistently pointed to the fact that circPVT1 silencing decreases the ability of malignant cells in induction of palpable tumors in animal models. Almost all of these studies have used BALB/c mice as the recipient of cancer cells ( Table 2).

HUMAN STUDIES
CircPVT1 levels have been found to be upregulated in gastric cancer tissues as a result of amplification of its genomic locus. Expression of circPVT1 could be regarded as an independent prognostic marker for prediction of overall and disease-free survival of patients with this type of cancer (14). Serum exosomal levels of circPVT1 have been higher in cisplatinresistant gastric cancer patients, indicating a role for this circRNA in predicting response to cisplatin (21). CircPVT1 has also been found to be upregulated in both osteosarcoma tissues and serum samples of these patients in correlation with poor prognosis of osteosarcoma patients. Moreover, circPVT1 performance as a diagnostic marker for osteosarcoma has been superior to alkaline phosphatase (28). Another study in osteosarcoma patients has shown upregulation of circPVT1 in osteosarcoma tissues compared with normal tissues. Moreover, expression of circPVT1 has been considerably elevated in the chemoresistant patients compared with the chemosensitive ones (29). While the association between overexpression of circPVT1 and lymph node metastasis has been verified in gastric cancer (14) and colorectal cancer (48), Kong et al. reported no correlation between expression levels of circPVT1 and lymph node metastasis in gastric cancer (49). Moreover, they reported downregulation of circPVT1 in gastric cancer (49). Other studies in diverse types of cancers have verified overexpression of circPVT1 in neoplastic tissues versus non-neoplastic tissues adjacent to the tumors ( Table 3). Upregulation of circPVT1 has been correlated with tumor size in non-small cell lung cancer (32) and hepatocellular carcinoma (38). However, this correlation has not been verified in osteosarcoma (27).
In lung cancer, expression levels of circPVT1 could differentiate tumor samples from neighboring non-cancerous   tissues with diagnostic power of 0.803. More importantly, serum levels of circPVT1 could diagnose patients from healthy subjects with diagnostic value of 0.794 (31). The diagnostic value of circPVT1 has also been assessed in oral squamous cell carcinoma tissue specimens through depicting receiver operating characteristic (ROC) curves. The area under this curve has been measured as 0.787 with sensitivity and specificity values of 68.6% and 86.0%, respectively (33) ( Table 4).
Upregulation of circPVT1 has been associated with decreased survival of patients with diverse cancer types, demonstrating the role of this circRNA as a prognostic marker. Two recent meta-analyses have indicated the importance of circPVT1 levels in prediction of malignant behavior of different types of neoplasms (53,54). Further proofs for participation of circPVT1 in the carcinogenesis have come from the observed association between its levels in tumors and clinical data including TNM stage, tumor size, and lymph node positivity. CircPVT1 has also been suggested as a diagnostic marker in lung cancer as well as oral squamous cell carcinoma. The discovery of presence of circPVT1 in the cancerderived exosomes not only highlights the biomarker role of this circRNA but also unravels a less-studied route of promotion of malignant behavior in tumor tissues by this circRNA.
Although expression of circPVT1 has been assessed by different methods, based on the poor reproducibility of assessment of circRNA expression levels (55), precise identification and quantification of circPVT1 expression are crucial.
Cumulatively, circPVT1 is implicated in response of cancer patients to chemotherapeutic agents such as cisplatin, doxorubicin, and paclitaxel. Thus, circPVT1 silencing is a putative modality for improvement of chemotherapy response in patients.