Promising biomarkers for diagnosis and prognosis
CircRNAs are widely and conservatively expressed in hematopoietic cells [
97]. As a result of their abundancy and accessibility, circRNAs are expected to be ideal biomarkers in the diagnosis and prognosis of hematological malignancies. Among all the circRNAs, PVT1 has been considered to participate in the pathogenesis of hematological malignancies [
98,
99]. CircPVT1 showed increased expression in ALL, pediatric B-precursor ALL and AML cases harboring MYC amplifications in the form of dmin, hsr, or ring chromosomes (AML-Amp) [
27,
100,
101]. Silencing circPVT1 was validated to inhibit cell proliferation and induce apoptosis in ALL [
27]. Additionally, circPVT1 was overexpressed in AML-Amp cases leading to the identification between various karyotypes of AML [
101].
Due to the high heterogeneity of hematological malignancies and the cell-type specificity of circRNAs, there are specific expressions of circRNA in different types of hematological diseases (Table
1). At present, a variety of circRNAs have performed promising function for evaluating prognostic model, such as circ_0003602, circ_0005571, circ_0074371, circ_0007609, circ_0012152, hsa_circ_0001857 and circ_0001247 [
102,
103]. Circ-Foxo3, Circ-RPS6KB1, circ-CSMD2, circ-ANXA2, circ-PWP2, circ-RBM5, circ-ZZEF1, circ-GSK3B and circ-FOXP1 could potentially identify AML patients from healthy groups [
104,
105]. Among them, circ-ANXA2 overexpression was related to shorter event-free survival (EFS) and OS of AML patients. Meanwhile, AML patients achieved complete remission (CR) presented lower level of circ-ANAX2 than those did not reach CR, accompanied by longer EFS and OS [
105]. Receiver operating characteristic (ROC) curve analysis revealed that the expression of circ-VIM could distinguish AML patients from healthy groups. Highly expressed circ-VIM acted as an independent prognostic factor for OS and leukemia-free survival (LFS) in AML [
106]. Interestingly, the expression level of the same circRNA differs in subtypes of the same disease, highlighting the specificity of circRNAs as biomarkers. Circ_0075001 was overexpressed in M0 or M1 subtype of AML patients and significantly downregulated in M2, M4 and M5 subgroups, showing the potential to distinguish the differentiation degree of the AML [
107].
Table 1
Circular RNAs implicated in hematological malignancies
AML | circRNA-DLEU2 | Up | proliferation ( +), apoptosis (−), tumor formation ( +) | / | miR-496/PRKACB | |
circ_100290 | Up | proliferation ( +), apoptosis (−) | / | miR-203/Rab10 | |
circPAN3 | Up | autophagy ( +), apoptosis (−), ADM-resistance ( +) | / | miR-153-5p/miR-183-5p/XIAP | |
circ-ANXA2 | Up | proliferation ( +), apoptosis (−) | high disease risk, poor risk stratification, low CR level, short EFS and OS | miR-23a-5p/miR-503-3p | |
circ_0009910 | Up | proliferation ( +), apoptosis (−) | poor risk, short OS | miR-20a-5p | |
circ_0000370 | Up | proliferation ( +), apoptosis (−), cell cycle ( +) | FLT3-ITD + | miR-1299/S100A7A | |
circMYBL2 | Up | proliferation ( +), quizartinib resistance ( +) | FLT3-ITD + | PTBP1, FLT3 kinase translational ( +) | |
circ-VIM | Up | / | distribution of WBC count, FAB subtypes, short OS and LFS | / | |
circ-0004136 | Up | proliferation ( +) | / | miR-142 | |
circ-HIPK2 | Down | differentiation ( +) | ATRA-induced differentiation | miR-124-3p | |
circ_0001947 | Down | proliferation (−), apoptosis ( +) | white blood cell, hemoglobin, diagnosis, prognosis | miR-329-5p/CREBRF | |
circ_0121582 | Down | proliferation (−) | / | miR-224/GSK3β, TET1/GSK3β/Wnt/β-catenin | |
CML | circ_0080145 | Up | proliferation ( +) | / | sponge miR-29b | |
circ_0009910 | Up | proliferation ( +), autophagy ( +), apoptosis (−) | imatinib resistance, short OS | miR-34a-5p/UKL1 | |
circBA9.3 | Up | proliferation ( +), apoptosis (−) | TKI-resistance | c-ABL1 & BCR-ABL1 level ( +) | |
circ_100053 | Up | / | clinical stage, BCR/ABL mutant status, short OS, imatinib resistance | / | |
circ_0080145 | Up | proliferation ( +), glycolysis ( +), apoptosis (−) | IM-resistance | miR-326/PPFIA1 | |
ALL | circPVT1 | Up | proliferation ( +), apoptosis (−) | / | c-Myc & Bcl-2 expression ( +) | |
circAF4 | Up | apoptosis (−), leukemogenesis ( +) | risk stratification | miR-128-3p/MLL-AF4 | |
CLL | circ-CBFB | Up | progression ( +), apoptosis (−) | diagnosis, low survival time, independent predictor of prognosis | miR-607/FZD3/Wnt/β-catenin pathway | |
circ-RPL15 | Up | proliferation ( +) | IGHV mutation status | miR-146b-3p/RAS/RAF1/MEK/ERK pathway | |
circ_0132266 | Down | proliferation (−), apoptosis ( +) | / | miR-337-3p/PML | |
DLBCL | circ-APC | Down | proliferation (−), cell cycle (−) | Ann Arbor stage, CHOP-like and rituximab resistance, short OS, independent prognostic factor | miR-888/APC, TET1/APC, inactivate Wnt/β-catenin pathway | |
MCL | circCDYL | Up | proliferation ( +) | diagnosis | | |
T-LBL | circ-LAMP1 | Up | proliferation ( +), apoptosis (−) | / | miR-615-5p/DDR2 | |
MM | circ_0007841 | Up | proliferation ( +) | clinical type, cytogenetic mutation, bone destruction, R-ISS staging, DOX resistance | ABCG2 level ( +) | |
circCDYL | Up | DNA synthesis ( +), apoptosis (-) | ISS and DS stage, diagnosis, short OS | miR-1180/YAP | |
circ_0000190 | Down | proliferation (−), apoptosis ( +), tumor growth (−) | ISS and DS stage, high risk, short PFS, OS | miR-767-5p/MAPK pathway | |
circ-SMARCA5 | Down | proliferation (−), apoptosis ( +) | β2-MG level, ISS stage, short PFS and OS | miR-767-5p | |
The high expression of circHIPK3 in serum of CML patients was related to Sokal relative risk, an independent factor of CML prognosis, and shorter OS, indicating poor clinical outcome [
70]. The level of circ-RPL15 was negatively correlated to the mutation state of immunoglobulin heavy chain (IGHV) gene, predicting poorer OS [
108]. In DLBCL, downregulating plasma circ-APC presented diagnostic potential and was related to advanced Ann Arbor stage, low International Prognostic Index, rituximab resistance, and shorter OS [
33]. The high expression of circRNA_101237 was associated with shorter OS and PFS in MM patients [
109]. What is more, the expression of circRNAs in several diseases exhibited temporal specificity, which indicated that circRNAs were likely to predict clinical outcome [
110]. A total of 508 circRNAs expressed dynamically throughout the treatment of all-trans retinoic acid (ATRA) in NB4 cells, and independently from the parent genes [
111]. The low expression state of circ_0004277 in AML patients was diminished after chemotherapy, while the level of circ_0004277 decreased again when patients relapsed after CR, demonstrating the relationship between the increasing expression and good curative effect [
112]. As consequence, the expression of circRNAs is dynamic during disease progression, which provides new aspects for therapeutic efficacy and prognosis evaluation.
The existing modalities of disease diagnosis and efficacy evaluation are invasive. Liquid biopsy, being non-invasive and repeatable, is becoming a new diagnostic tool. Accumulating evidence discovers the enrichment of circRNAs in exosomes. Exosomes protect inner circRNAs from influences of extracellular substances, making it more possible for detecting the existence of exosomal circRNAs [
113]. Exosomal circRNAs act a significant part mainly in proliferation and tumor metastasis [
114]. Mc-COX2, a mitochondrial genome-derived circRNA, was significantly enriched in exosomes of plasma from CLL patients, and was positively correlated with worse OS [
115]. Associated with deletion 17p, t (4; 14), Durie-Salmon staging and international staging system, the level of exosomal circMYC was higher in bortezomib-resistant patients than non-resistant groups [
116]. Additionally, the exosomal circ_0007841 was validated to enhance proliferation and metastasis and suppress apoptosis via activating PI3K/AKT pathway in MM cell lines [
39]. Although a large number of circRNAs with biomarker value have been discovered by high-throughput sequencing, the targets and mechanisms are still unclear. The constantly emerging circRNA databases provide great convenience for target prediction and expression visualization. Here, we summarize 10 representative circRNA databases (Table
2).
Table 2
CircRNA databases
circBase | | A public dataset of thousands of circRNAs in eukaryotic cells | |
circInteractome | | A computational tool enabling the prediction and mapping of binding sites for RBPs and miRNAs on reported circRNAs | |
CIRCpedia V2 | | An updated comprehensive database containing circRNA annotations from over 180 RNA-seq datasets across six different species | |
circRNADb | | A comprehensive database of circular RNA molecules in humans | |
CSCD | | An integrated interactional database of cancer-specific circRNAs | |
exoRBase | | A repository of circRNA, lncRNA and mRNA derived from RNA-seq data analyses of human blood exosomes | |
MiOncoCirc | | A compendium of circular RNAs compiled from cancer clinical samples | |
CircAtlas 2.0 | | A database of over one million of circRNAs across 6 species (human, macaca, mouse, rat, pig, chicken) and tissues | |
CircBank | | A comprehensive database of human circRNA including more than 140,000 human annotated circRNA from different source | |
NoncoRNA | | A database for experimentally supported ncRNA and drug target associations in cancer | |
The circRNA-miRNA-mRNA axis has become a vital mechanism in hematological tumorigenesis. As circRNAs contain multiple miRNA binding sites, targeted inhibition of circRNAs exerts more therapeutic advantages and potential than targeted inhibition of single miRNA/gene. RNA interference (RNAi) is one of the most common methods to determine the function of circRNA through loss-of-function approach. Transcripts of circRNAs could be packaged into viral vectors or oligonucleotide and then delivered to target cells to mediate their therapeutic effects [
96]. Inhibiting the expression of specific circRNA could enhance the protective function of the relevant miRNAs in inhibiting oncogenes, such as XIAP, β-catenin, GSK3β and YAP [
24,
31,
35,
43]. Recently, the CRISPR/Cas9-mediated genetic engineering technology provides a robust tool for circRNAs investigation. The CRISPR/Cas-assisted homologous recombination method can replace circRNA gene with a marker gene, thereby consuming circRNAs without affecting the existing gene [
117]. Future investigations fueled by the well-defined guide RNA (gRNA) libraries designed for circRNA will promote the targeted therapy based on circRNA screening.
At present, a practical artificial circRNA sponge could be synthesized using simple enzymatic ligation approach. The artificial circRNA molecule is applied as an exogenous miRNA inhibitor to effectively bind and block mature miRNA, providing a promising strategy for cancer therapy [
18]. Jost et al
. engineered the artificial circRNA sponges into customized miRNA to isolate miR-122 from hepatitis C virus (HCV). In addition, circRNAs can also be used as protein sponges, and the binding sites obtained from SELEX or CLIP data can be used for many RBPs [
118]. The anti-HCV circular miRNA-122 RNA sponge can be used in combination with the sequence of host factors necessary to isolate the propagation of HCV, such as hnRNP L and HuR [
118]. Therefore, the artificial circRNA sponge is a promising tool in circRNA-based anti-tumor therapy, which has potential value in clinical application.
In addition, emerging evidence indicated the potential therapeutic value of tumor-related functional peptides encoded by circRNAs, especially cancer-inhibiting peptides/proteins, such as β-catenin-370aa encoded by circβ-catenin, circPPP1R12A-73aa by circPPP1R12A, and AKT3-174aa by circ-AKT3 [
119]. These functional peptides can play important roles in tumorigenesis, which made them potential novel targets for drug development [
119]. Due to the potential development value and clinical utility of functional peptides encoded by circRNAs, the functional peptides may be used in the research and treatment of hematological malignancies in the future.
Both the artificial circRNA and functional peptides need to be transported to the cell through an appropriate delivery system. Nanoparticles could be used to treat tumor in a variety of ways, such as intravenous injection and tail vein injection, and have become promising tools for cancer treatment. Recently, Wang et al
. established a new plasmid delivery system, Micropoly-transfecter, which can deliver circ-1073 plasmid through intratumoral injection, thereby inhibiting tumor progression [
120]. Moreover, accumulating evidence has indicated the potential value of exosomal circRNAs in clinical application [
121]. Exosomes could carry circ-0051443 from normal cells to HCC cells, and inhibit the malignant biological behaviors through inducing apoptosis and cell cycle arrest [
122].
CircRNAs play vital roles in the tumor microenvironment (TME) by regulating the immune surveillance and remodeling the extracellular matrix [
9,
123]. CircRNA-002178 was indicated to promote the expression of PD-L1 in tumor cells through the ceRNA mechanism. Meanwhile, circRNA-002178 in tumor cells was delivered from exosomes to CD8 + T cells to achieve immune evasion of tumor cells by promoting PD-1 expression [
124]. The regulation of PDL-1/PD-1 pathway by circRNA-002178 may also provide a new direction for the development of tumor-targeted drugs. Currently, the circRNA-based targeted therapy in hematological malignancies is still in its infancy. Therefore, regulation of PD-1/PD-L1 by targeting relevant circRNA may be a promising direction of future immune therapy.