Introduction

CircRNAs were originally thought to be splicing errors or byproducts of RNA transcription because they cannot encode proteins [1,2,3]. Developments in bioinformatics, RNA sequencing and genome sequencing, have led to preliminary discoveries about the production and function of circRNAs in recent years [4, 5]. In contrast to traditional linear RNAs, circRNAs lack a 3' poly(A) tail and a 5' cap structure, are closed into a ring in the form of covalent bonds and are not sensitive to exonucleases. This unusual structure provides circRNAs with an elevated level of stability. Most circRNAs are generated via back-splicing, which occurs mainly by base pairing between the intron reverse repeats of Alu transposition factors, which results in the formation of abnormal secondary structures in pre-mRNAs [6]. Dysregulation expression of circRNAs due to disorder of conventional splicing and back-splicing may cause tumors progression [7].

CircRNAs offer excellent preclinical diagnostic and therapeutic potential for a wide range of tumors owing to their universality, conservation, stability and differential expression in tumor patients and healthy people. The contribution of some circRNAs to tumor progression or immunotherapy have been preliminary demonstrated in preclinical models [8,9,10]. For example, circHIPK3, identified as an oncogene and autophagy regulator, significantly promotes malignant cell properties and inhibits autophagy in non-small cell lung cancer (NSCLC) and may be a prognostic marker and therapeutic target [11]. Circ-ZKSCAN1 is downregulated in bladder cancer (BCa) and hepatocellular carcinoma (HCC), and its absence results in malignant characteristics, such as cancer stemness, and is highly associated strongly correlated with low rates of survival overall and without recurrence, thus, circ-ZKSCAN1 may be a prognostic factor of BCa recurrence and a target for HCC treatment [12,13,14]. Circ-CPA4 is upregulated in NSCLC and can inhibit the functions of tumor immune cells (such as CD8+ T cells) via the let-7/programmed death ligand 1 (PD-L1) axis, suggesting a new approach for exploring NSCLC therapeutic agents [15]. Exosomes contain some circRNAs, and exosomal circRNAs (exo-circRNAs) possess the capability of original biology roles of circRNAs and transferring to target sites [16, 17]. CircSKA3 in tumor-derived exosomes can mediate breast cancer metastasis, and targeting circSKA3 can suppress tumor formation capacity and tumor cells transfer [18]. CircRNAs can be found in liquid biopsy samples (such as plasma, urine and saliva) and have significant potential as diagnostic and prognostic biomarkers [19,20,21,22]. These results reveal that circRNAs constitute a novel and individual class of epigenetic regulators that are engaged in tumor immunity and immunotherapy.

Cancer immunotherapy is currently the fifth most popular type of tumor treatment, behind surgery, chemotherapy, radiation therapy and targeted therapy [23, 24]. However, only some patients can achieve effective treatment outcomes [25]. Thefunction of circRNAs in tumor microenvironment (TME) and immunotherapy has drawn a lot of attentions recently [26,27,28,29]. Our focus is on analyzing the tumor immune microenvironment (TIME) and enhancing the effectiveness of immunotherapy. For example, the cancer-derived exosome circUHRF1 induces malfunction in natural killer (NK) cells and drives anti-PD1 resistance in HCC, hence promoting tumor immunosuppression [30]. Increased circFAT1 expression can bind to STAT3 and promote its activation, then inhibit CD8+ T cells infiltration into TME and reduce the sensitivity to PD1 blockade immunotherapy, which in turn stimulates cancer stemness and immune escape [31]. Focusing on circRNAs or managing alterations in immune cells and TME may can enhance the effectiveness of immunotherapy and survival of tumor patients.

In this review, understanding the roles of circRNAs in immunity and their application in immunotherapy may provide a new approach for guiding targeted clinical therapy.

The glance of circRNA

Researchers discovered circRNAs for the first time in a plant RNA virus in 1976 [32]. In 1979, the circRNA in eukaryotic cytoplasm was observed for the first time via electron microscopy, suggesting that RNA may occur in a circular form in the eukaryotic cells [33]. In 1991, endogenous spliced circRNAs derived from the human DCC gene were initially discovered by researchers. Numerous investigations conducted in the late 1990s and early 2000s revealed the ubiquitous presence of genes producing circRNAs in the eukaryotic cells, which include human cells and those of mammals and flies [34]. However, due to the limitations of scientific knowledge and technological development, circRNAs were commonly thought to be mis-spliced RNAs or byproducts of pre-mRNA processing for several decades. Until 2012, scientists found through high-throughput sequencing that hundreds of human genes can transcribe circRNAs and they are closely associated with a range of disease conditions, thus circRNA-related researches have gradually increased.

Many circRNAs are exhibited stable expression in saliva, blood, and exosomes, and they are more prevalent and more conserved than mRNAs in prokaryotes and eukaryotes [35]. CircRNA expression is 10% higher than mRNA expression in the majority of genes, and the expression abundance of some circRNAs is ten times greater than that of mRNAs [36]. According to other studies, the amount of circRNAs in human platelets is 17 ~ 188 times higher than that in nucleated tissues. CircRNAs are more stable in comparison to linear mRNAs and are not degraded easily. Additionally, circRNAs are easier to detect than miRNAs, which are present in smaller amounts [37]. These results demonstrated the enormous potential of circRNAs as diagnosis, prognostic, and predictive biomarkers. Unlike other noncoding RNAs, circRNAs are produced via the process of back-splicing, which creates a back-splice junction by looping the introns around the upstream and downstream splice acceptor and donor sites. Sequences with inverted intronic repeats (such as Alu elements) can help form loop structures [7]. Research has also revealed that RNA binding proteins (RBPs) may play important roles in the production of certain circRNAs. On the one hand, Muscleblind and Quaking can bind to the flanking introns of circularized exons in the mRNA precursors of their host genes to promote circRNAs production [38,39,40]. On the other hand, DHX9, nuclear RNA helicase, by explicitly binding to the inverted repeat Alu sequences, suppresses the synthesis of circRNAs [6]. Adenosine deaminase 1 cleaves circRNAs by binding double-stranded RNA to the stem‒loop structure [4].

There are three primary forms of circRNAs, exonic circRNAs (ecircRNAs), circular intronic RNAs (ciRNAs), and exonic-intronic circRNAs (EIciRNAs).EcircRNAs, which are generated from single or multiple exons, account for more than 80% of known circRNAs; the majority of ecircRNAs reside in the cytoplasm, they typically act as miRNA response elements and binding protein elements [41,42,43]. EcircRNAs can be formed in three ways: lasso-driven cyclization, intron pairing-driven cyclization and RNA-binding protein-promoted cyclization. Studies have shown that the ATP-dependent RNA helicase DDX39A and the spliceosome RNA helicase DDX39B can promote the release of ecircRNAs from the nucleus to the cytoplasm [44]. CiRNAs consist only of introns and thus are formed in a different manner than EcircRNAs. Under normal circumstances, the intron lassos released during the classical splicing process are degraded. However, ciRNAs are formed when the 5' end splicing site contains a 7-nt GU-rich base sequence and a nearby 3' end contains an 11-nt C-rich base sequence so that the intron lasso is protected, which promotes the formation of stable ciRNAs and functions as a transcription regulator in the nucleus [45, 46]. Exons and introns are present in EIciRNAs. In fact, lasso-driven and intron-driven cyclization can produce not only EcircRNAs but also EIciRNAs. EIciRNAs interact with U1 snRNPs, are mostly located in the nucleus, and stimulate the transcription of associated genes [16]. Some circRNAs that are enriched and highly expressed in exosomes can be packaged and released into the extracellular matrix to execute their functions. Exo-circRNAs not only can mediate tumor growth, invasion and metastasis and regulate antitumor immunity, but also have the potential to become efficient biomarkers for tumor diagnosis, monitoring and therapeutic targeting [47]. However, investigations into circRNAs in immunocytes are still in their early stages, and many new circRNAs have not yet been discovered to confirm their function.

Biogenesis and functions of circRNAs in tumors

CircRNAs can carry out their functions to affect the development and metastasis of malignancies tumors by various mechanisms [48,49,50]. Numerous studies have shown that abnormally expressed circRNAs may be crucial regulators of the development of tumors by sustaining proliferative signals, affecting invasion and metastasis abilities, regulating the formation of cancer stem cells, inducing angiogenesis and radioresistance and influencing the progression of antitumor immunity [51].

CircRNAs act as miRNA sponges

Many circRNAs contain conserved miRNA target sites and can act as miRNA sponges that bind miRNAs [52]. By acting as competing endogenous RNAs (ceRNA), circRNAs competitively bind miRNAs and play essential parts in the progression and spread of malignancies. For example, the first circRNA proved to be a miRNA sponge was the antisense transcript of cerebellar degeneration-related protein-1 (CDR1as or ciRS-7), including 74 specifically conserved miRNA binding sites [41]. ciRS-7 can downregulate miR-7 expression by directly binding to miR-7 and inhibiting its activity, which is postulated to regulate tumor growth and metastasis. ciRS-7 is likely to become a clinical biomarker for tumor diagnosis and therapy [53]. Hsa_circ_0136666 droved PD-L1 phosphorylation via the miR-375/PRKDC axis, which promoted TME formation, suppressed immune function, and lead to tumorigenesis immune escape in gastric cancer (GC) [54]. Circ-CPA4 can decrease PD-L1 expression and CD8+ T cells in TIME by sponging let-7, regulate stemness and drug resistance of NSCLC cells [15] (Fig. 1A). The expression of circ_0001005 can be upregulated by targeting androgen receptor and competitively sponged miR-200a-3p and PD-L1, which can enhance natural killer (NK) cell-mediated tumor-killing efficacy in breast cancer (BC) [55]. Has_circ_0007456 can significantly influence the NK cell-mediated cytotoxicity by miR-6852-3p/ICAM-1 axis, and regulate process of tumorigenesis and immune evasion in HCC [56] (Fig. 1B). Exo-circGSE1 can regulate the expansion of regulatory T cells by miR-324-5p/TGFBR1/Smad3 axis and promote immune escape of HCC [57]. Exo-circPOLQ can active interleukin-10 and STAT3 axis by targeting miR-379-3p to promote polarization of M2 macrophages in a colorectal cancer [58] (Fig. 1D).

Fig. 1
figure 1

Roles of circRNAs in the tumor microenvironment. A CircRNAs can as miRNAs sponge to regulate the expression of PD1/PD-L1 and inhabit activation of T cells. B CircRNAs can bind to miRNAs to influence the NK cell-mediated killing ability and cytotoxicity. C Exogenous circRNAs can regulate the function of antitumor immunity in RIG-dependent manner. D Exo-circRNAs can regulate antitumor immunity. ciRNAs: circular intronic RNAs; ecircRNAs: exonic circRNAs; EIciRNAs: exonic-intronic circRNAs; exo-circRNAs: exosomes circRNAs

Full size image

CircRNAs interact with proteins as sponges, scaffolds or protein recruitment

Researches have demonstrated that circRNAs are capable of binding to RBPs and function as protein decoys or antagonists to regulate protein activity, which are crucial for the development of tumors [59]. RBPs are important regulators in tissue-specific circRNAs expression [60]. Many circRNAs have been reported to bind human antigen R (HuR) as sponges [61]. CircAGO2, which is derived from AGO2, interacts with the HuR protein to enable HuR-mediated inhibition of AGO2-miRNA complexes function and promote the GC progression [62]. CircDLC1 can bind to HuR, which decreases the interaction between HuR and MMP1 and consequently restricts MMP1 expression, thus leading to HCC [63]. Moreover, circRNAs are linked to proteins and act as scaffolds to form RNA–protein complexes, promoting enzyme–substrate colocalization to alter reaction kinetics [64]. For example, circ-Foxo3 facilitates p53 ubiquitination and degradation in an MDM2-dependent way, which can increase the amounts of the protein Foxo3 due to its low binding affinity and cause cancer cells to undergo apoptosis [65]. CircATG7 is located in pancreatic cancer cells of cytoplasm and nucleus. Cytoplasmic circATG7 plays a role by increase miR-766-5p/ATG7 axis. Nuclear circATG7 acts as a scaffold to enhance the interaction between HuR and ATG7, and regulate the growth and spreading of pancreatic cancer cells [66]. circREEP3 can accelerate tumorigenesis and metastasis of colorectal cancer by recruiting the CHD7 to FKBP10 promoter, and inhabit function of antitumor immunity in RIG-1-dependent manner [67]. CircKEAP1 upregulates of immune response genes and activates antitumor immunity of oral squamous cells by interacting with RIG-I [68] (Fig. 1C). circEIF3I, serving as a molecular scaffold, binds to the MH2 domain of SMAD3 and recruit SMAD3 to early endosomes, subsequently promotes pancreatic cancer progression via TGF-β signaling pathway [69]. Finally, circRNAs have ability to recruit particular proteins to certain cellular sites. For example, circular RNA FECR1,consisting of FLI1 exons 4-2-3 in FLI1 promoter chromatin complex, recruits TET1 demethylase to the FLI1 promoter region and downregulates DNMT1 in trans, which may coordinate the methylation and demethylation of DNA for regulating the metastasis of BC cells [70]. CircRHOT1 can bring TIP60 to the NR2F6 promoter and promote the transcription of the NR2F6 mRNA in HCC [71].

CircRNAs act as translation templates for peptides

Pamudurti first reported that circRNAs can be translated in 2017 [64]. This type of circRNA has an open reading frame, and its translation template mechanisms can be categorized as internal ribosome entry site (IRES)-dependent or IRES-independent ways [48, 65]. The translation of circ-SHPRH into the SHPRH-146aa protein is dependent on IRES [72]. Circ-ZNF609 has a translational function, and can influence myoblast proliferation through its open reading frames and the IRESs [73]. The 5'-cap and poly(A) tails, which are necessary for cap-dependent translation, are absent from circRNAs. However, circRNAs can form through IRESs or following the incorporation of m6A RNA modifications in the 5’-untranslated region [74]. CircRNAs can translated into peptides by an m6A-driven progress, which may be involved in physiological reactions in tumors. Circ-FBXW7 is translated into circ-FBXW7-185AA by m6A modification and inhibits β-catenin ubiquitination, effectively inhibiting the proliferation of lung adenocarcinoma (LUAD) stem cells and reversing resistance to tyrosine kinase inhibitors by regulating Wnt pathway function [75]. To date, only a small number of endogenous circRNAs that function of protein templates have been discovered. The majority of peptides produced from circRNAs are still unknown to have any functional significance, despite the prediction that certain circRNAs will have a putative open reading frame with an upstream IRES.

Roles and functions of circRNAs in tumor immunity and immunotherapy

The immune response is an important line of defense that monitor tumor occurrence and eliminate tumor cells, and tumor progression is accelerated after immune evasion of tumor cells [76, 77]. Immune evasion mechanisms continue to develop and become more diverse and complex in the process of tumor progression, leading to poor treatment efficacy in the clinic [78, 79]. CircRNAs mediate many signaling pathways, recruit and rewire essential TME components, thus influence carcinogenesis, the immune system, and the effectiveness of antitumor immunotherapy [76] (Table 1).

Table 1 The roles and functions of circRNAs in tumor immune
Full size table

CircRNAs regulate cancer stemness and affect the antitumor immune response

Cancer stem cells (CSCs) are a special class of cells that can self-renew and maintain stemness; can induce tumorigenesis, tumor recurrence and progression in the TME; and are usually resistant to chemotherapy, radiotherapy, targeted therapy, and even immunotherapy [80]. Heterogeneous CSCs can interact with TME components and immune cells, thus playing an antitumor role and leading to treatment failure [81]. In the immunosuppressive TME, CSCs evade immune system surveillance by affecting immune cells and producing immunoregulatory factors, cytokines, and even exosomes [82]. CSCs can regulate immune checkpoints and immune evasion by producing extracellular vesicles (EVs) or exosomes. CSC-derived exosomes not only express PD-L1 but also increase the expression of PD1 on CD8+ T cells, which promotes CD8+ T cells exhaustion [83]. Additionally, CSCs can directly promote tumor resistance to immunotherapy via adoptive cytotoxic T cells transfer [84] (Fig. 2A). For example, circFAT1 plays an important role in sustaining CSC self-renewal and tumorigenicity in squamous cell carcinoma. Disruption of circFAT1 expression can impair the stemness of cancer cells and reverse tumor cell-intrinsic immunity in the tumor immunosuppressive microenvironment by binding STAT3 in the cytoplasm and promoting STAT3 activation, which significantly increases resistance to PD1 blockade immunotherapy by reducing CD8+ T cells infiltration into the TME [31]. These findings suggest a significant relationship between circRNAs modulation of stemness in cancer and the establishment of an immunosuppressive microenvironment, suggesting that circFAT1 is a significant therapeutic target for overcoming immunotherapy resistance to PD1 blockade in head and neck squamous cell carcinoma. The TIP60 complex is recruited to the MAFF promoter by cia-MAF, which binds to the promoter, and induces liver tumor-initiating cell proliferation and self-renewal [85]. The miR-3908/TRIM59/p53 axis is the target of circLMP2A via an EBV-encoded circRNA, whose expression is increased and is essential for generating and sustaining stemness phenotypes in GC [86]. CircIPO11 binds TOP1 to the GLI1 promoter to stimulate its transcription, leading to the activation of Hedgehog signaling to maintain self-renewal of liver CSCs [87]. CircRGPD6 maintains the stem cell-like features of BC, resulting in the establishment of a TV-circRGPD6 nanoparticle that eliminates breast cancer metastases by specifically expressing circRGPD6 in metastatic BC [88]. In addition, these studies provide theoretical support for further investigations about the roles of circRNAs in CSC stem maintenance. However, the molecular mechanisms underlying the interactions among immune cells, CSCs, and immunotherapy response are poorly understood. Studying the interaction between tumor cells stemness and immune evasion according to circRNAs expression in CSCs will help improve a more knowledge of the underlying immunological mechanisms and design novel circRNAs CSC-targeted immunotherapy strategies to improve cancer prognosis.

Fig. 2
figure 2

Roles of circRNAs in tumor immunity. A CircRNAs regulate cancer stemness and affect the antitumor immune response. B CircRNAs regulate immune checkpoints. C CircRNAs influence immune cells infiltration and recruitment by cytokines and chemokines to shape the TIME. D Exo-circRNAs induce NK cells/T cells exhaustion. circRNAs: Circular RNAs; CSCs: Cancer stem cells; TME: tumor microenvironment

Full size image

CircRNAs regulate immune checkpoints

Immune checkpoints are a group of molecules located on immune cells that limit the level of immunological activation, act as gatekeepers during the immune response and prevent the body's defense system from becoming active. Increasing amounts of researches in molecular oncology have revealed the intricate regulatory mechanisms affecting immune checkpoint expression in recent years. A strong correlation exists between the expression of immunological checkpoints and circRNAs, which are abundant noncoding RNAs in the human transcriptome that are involved in all cancer hallmarks. CircRNAs can be used not only as targets for regulating tumor progression but also as potential mediators of antitumor immunity and cancer immunotherapy targets. Recent studies have demonstrated that cancer-related circRNAs play crucial roles in the formation of the immunosuppressive tumor microenvironment and immunotherapy resistance and may regulate the expression of immune checkpoints, such as PD1/PD-L1, cytotoxic T-lymphocyte-associated protein 4 (CTLA-4), lymphocyte-activation gene 3 (LAG-3), T cells immunoglobulin and mucin domain-3 (TIM-3), and CD47, through the ceRNA mechanism [89,90,91] (Fig. 2B).

CircRNAs not only can regulate the expression of PD-L1 directly but can also affect the expression of downstream molecules to regulate the expression of PD-L1 via a sponge mechanism. CircFGFR1 can directly interact with miR-381-3p to increase the expression of target gene CXCR4, which promotes NSCLC progression and resistance to anti-PD1-based therapy [92]. Hsa-circRNA-002178 can cause T cells exhaustion in cancer cells by sponging miR-34, which increases PD-L1 expression [93]. circCHST15 can directly suppress the expression of PD-L1 as a sponge for miR-155-5p and miR-194-5p to increase tumor development and immune evasion in lung cancer [94]. Overexpression of circTMTC3 and circFAM117B can increase PD-L1 expression by sponging miR-142–5, thereby reducing the cytotoxicity of CD8+ T cells and promoting melanoma immune escape [95]. circRNA_001678 actives the PD1/PD-L1 pathway and triggers CD8+ T cells apoptosis via the miR-326/ZEB1 axis, thus promoting NSCLC immune escape progression [96].

CircRNAs also regulate PD1/PD-L1 through interacting with proteins or other unclarified mechanisms. Has_circ_0067842 interacts with HuR, altering its nuclear translocation and thus increasing the stability of CMTM6, influencing the ubiquitination of PD-L1, preventing its degradation and promoting immune escape in BC [97]. Hypoxia-related circPRDM4 is highly expressed in HCC, inhibits CD8+ T cell infiltration and promotes PD-L1 expression in the TME by regulating the interaction between HIF-1α and the promoter of CD274 [98]. CircBART2.2, encoded by Epstein‒Barr virus (EBV), can bind to RIG-I and activate the transcription factors IRF3 and NF-κB, which are essential for the regulation of PD-L1, thus inhibiting T cells function and promoting tumor immune escape in nasopharyngeal carcinoma [99]. N6-methyladenosine-modified circIGF2BP3 stabilizes the OTUB1 mRNA in a PKP3-dependent way to reduce PD-L1 ubiquitination, which in turn reduces CD8+ T cells responses to promote NSCLC tumor immune evasion [100]. M6A-activated circNFIX increases ovarian cancer development and immune evasion by modulating the miR-647/IL-6R/PD-L1 pathway [101].

In addition to PD1/PD-L1, circRNAs regulate numerous other immune checkpoints. Other immunomodulatory receptors, such as CTLA-4 and TIM-3, bind to ligands on tumor cells and prevent antitumoral immunity in the surrounding tissue [102]. CircQSOX1 facilitates Treg cell-based immune escape in colorectal cancer by activating the PGAM1/miR-326/miR-330-5p pathway to induce glycolysis and inactivate the anti-CTLA-4 therapeutic response [103]. CircUHRF1 inhibits NK cell function and upregulates the expression of TIM-3 via the degradation of miR-449c-5p, and circUHRF1 may drive resistance to anti-PD1 immunotherapy in HCC patients [30]. circRERE (has_circ_0009581) promotes resistance to the chemotherapy drug bortezomib in multiple myeloma by sponging miR-152-3p to upregulate the expression of CD47 [104]. CircHMGCS1-016 increases intrahepatic cholangiocarcinoma cells invasion and reduces CD8+ T cells infiltration via miR-1236-3p to regulate CD73 and GAL-8 expression, thereby reshaping the TIME [105].

Immune checkpoint inhibitors (ICIs), such as PD1/PD-L1 inhibitors, are crucial strategies for cancer immunotherapy. By targeting immune checkpoints, ICIs can preferentially activate immune cells; however, their practical applicability in cancer immunotherapy is limited by primary and acquired resistance [106]. Combined utilization of piRNAs and circRNAs as innovative sensitizers to augment PD1/PD-L1 inhibitor therapy serves different purposes in regulating PD-L1 expression and modifying the immune response [107]. However, further studies are needed to verify whether circRNAs are involved in the regulation of immune checkpoints at the transcriptional, posttranscriptional, or protein levels or at the protein modification and degradation levels. More researches are need to ascertain whether circRNAs can act as immune checkpoints directly to control immune responses against cancer cells or whether they facilitate ICI resistance by being engaged in cancer hallmarks such as cell death and proliferation; such studies could reveal whether these circRNAs can serve as new potential cancer immunotherapy targets.

CircRNAs shape the TIME

A growing body of research has demonstrated that circRNAs regulate numerous of biological processes within the TME. In addition to immune cells and immune checkpoints, the release of cytokines or chemokines also participates in the regulation of tumor immunity. Cancer-derived circRNAs are transferred into surrounding immune-related cells through transporters, thereby inhibiting the secretion of cytokines or chemokines and influencing immune cell infiltration and recruitment in tumors [108] (Fig. 2C). For example, circSMARCC1 was positively associated with the colonization of CD68+/CD163+/CD206+ tumor-associated macrophages and CD163 expression in macrophages via the CCL20-CCR6 axis, which leads to tumor-associated macrophages infiltration and M2 polarization during prostate cancer progression [109]. CircMAPK1 can act as a scaffold that promotes the occupation of the 3’UTR of CCL5 by IGF2BP1, which significantly promotes CD8+ T cells intratumoral infiltration in LUAD [110]. circNDUFB2 is involved not only in the IGF2BP3 ubiquitination and degradation, but also in the immune response to NSCLC by recruiting CD8+ T cells and DCs to the TME through the secretion and nuclear translocation of chemokines and IFNs [111]. Circ_0020710 can serve as a miR-370-3p sponge to upregulate CXCL12 expression. It not only promotes the malignant activities of melanoma through autocrine and paracrine pathways, but also promotes the recruitment of immune-mediated suppressive cells (foxp3+ regulatory T cells), which may suppress the immune system and reduce the invasion of cytotoxic T lymphocytes [112]. By regulating the expression of proinflammatory and interferon-related genes in sarcoma cells, circCsnk1g3 and circAnkib1 can drive the recruitment of immune cells into the tumor mass, hence promoting carcinogenesis by shaping the TME [113]. Reduced CD8+ T cells infiltration and low CXCL10 levels are linked to circMET upregulation, which can induce a TIME and anti-PD1 therapy resistance in HCC [114, 115]. Chemokines bind to their receptors in the TME and can function as a link between host and tumor cells. These interactions aid in immune cell recruitment and accelerate angiogenesis and tumor progression. Blocking the binding of chemokines to receptors and downstream signaling pathways can improve immunotherapy sensitivity and efficacy, thus slowing tumor progression. Moreover, effector T cells need to respond quickly to produce sufficient cytokines, chemokines, and cytotoxic molecules to respond to inflammation and tumorigenesis [116]. Targeting circRNAs interferes with the TME by regulating cytokines, providing insights and ideas for clinical therapy to improve tumor inhibiting immune processes and drug resistance mechanisms.

Tumor cells secrete exo-circRNAs to regulate tumor immunity

Exosomes are extracellular vesicles with diameters ranging from 30 to 150 nm and have a lipid bilayer membrane structure. Both normal and pathological conditions cause nearly all cell types to produce exosomes, which are extensively distributed in the microenvironment [117,118,119]. The circRNAs expression levels in exosomes far exceed that of the corresponding linear RNAs. Exo-circRNA-mediated tumor cells can maintain antitumor immune processes by inducing immune cell exhaustion [120]. CircRNAs are associated with exosomes and EVs in the TIME. Certain exosomes and EVs can transfer circRNAs from donor cells to immunocytes, where circRNAs can function as tumor antigens to modulate immune responses by interfering with immune checkpoints expression and immune cells activation [121]. Furthermore, exo-circRNAs promote the entry of miRNAs into immunocytes to silence associated target genes when they are transferred from tumor cells to immunocytes (Fig. 2D). For example, cancer cell-derived exo-circUHRF1 inhibits NK cells function and tumor infiltration, inhibits the secretion of NK cell-derived IFN-γ and TNF-α and upregulates the expression of TIM-3 by reducing the expression of miR-449c-5p, which leads to NK cell exhaustion. CircUHRF1 may result in resistance to anti-PD1 immunotherapy, providing HCC patients with a potential therapeutic strategy [30]. Exo-circCCAR1 produced by HCC cells induces immunosuppression by promoting the exhaustion of CD8+ T cells, which results in resistance to anti-PD1 immunotherapy. Targeting exo-circCCAR1 or CCAR1 may be a novel strategy to maximize immunotherapy efficacy in HCC patients [12]. NSCLC cells are the primary producers of exo-circUSP7, which promotes CD8+ T cells exhaustion dysfunction and induces resistance to anti-PD1 immunotherapy, hence contributing to immunosuppression and providing a viable therapeutic approach for NSCLC patients [122]. Exo-circZNF451 can induce the anti-inflammatory function of macrophages and exhaustion of cytotoxic CD8+ T cells and induce macrophage polarization by reshaping the TIME via the TRIM56-mediated pathway of the FXR1/ELF4/IRF4 axis, and one potential novel biomarker for predicting LUAD sensitivity to PD1 blockade could be exo-circZNF451 [123]. Studies have shown that circTRPS1 derived from the exosomes of BC cells can regulate glutamine metabolism through the circTRPS1/miR-141-3p/glutaminase-1 axis, induce redox imbalance and CD8+ T cells exhaustion in the TME, and promote the proliferation, migration and invasion of BC cells [124]. Inhibitory receptors such as PD1, T cell immunoglobulin and mucin domain 3, lymphocyte-activation gene 3, and T cells immunoreceptor with Ig and ITIM domains are expressed at high levels in both exhausted NK cells and exhausted CD8+ T cells. Monotherapies that inhibit these receptors may be beneficial for cancer patients. Exo-circRNA-targeted monotherapies can improve the delivery of drugs and increase inhibitory receptor resistance to immunotherapy. One potentially effective tactic to reduce tumor spread and increase immunotherapy sensitivity is to target exo-circRNAs.

Applications and prospects of circRNA-based vaccines in cancer immunotherapy

Immunotherapy, including immune checkpoint inhibitors (ICIs), chimeric antigen receptor (CAR) T cells therapy, tumor vaccines, and other innovative immunotherapy drugs, has been a prominent topic of tumor therapy research in the past few years both domestically and abroad [125, 126]. However, with the accumulation of more clinical data, these immunotherapies have been found to have many shortcomings, such as a low overall treatment response rate [127]. In clinical studies of cancer immunotherapies, tumor vaccines have demonstrated considerable promise. By customized design, these RNA vaccines can train the immune system to target specific neoantigens in individual cancer patients, thereby effectively attacking residual tumor cells [128]. Researchers are explored the use of mRNA vaccines and post-operative treatment, these vaccines are expected to train the immune system to identify and eliminated residual cancer cells (such as melanoma, colorectal cancer, and pancreatic cancer), thereby preventing disease from recurrence [129] The rapid development of mRNA technology and its applications in the fields of medicine, pharmacy, material science, may benefit from its advantages of not having to pass to the cell nucleus and not to interfere with the human genome [130]. However, mRNA vaccines still face some challenges. First, the stability of mRNA is an important issue, because mRNA is easily degraded by enzymes in the body, which affects its effectiveness. The large molecular mass and single chain structure of mRNA lead to its instability and therefore requires chemical modification to ensure its function. Secondly, mRNA vaccines may trigger an immune response, leading to side effects. In addition, mass production of mRNA vaccines is also a technical challenge, and production costs and supply chain issues need to be addressed.

The most promising cancer vaccine subtypes are circRNA-based vaccines. circRNA vaccines offer a novel solution to the limitations of mRNAs vaccines because that they have lower immunogenicity/cytotoxicity abilities and can boost the immune response to tumor antigens in terms of strength, breadth, quality, and duration [131, 132]. CircRNA vaccines have been developed to elicit effective both B cells and T cells immune responses against SARS-CoV-2 in mice and rhesus macaques [133]. Conventional CAR T cells therapy requires high-cost, long-term preparation, while having cleansing bone marrow, cell storm, targeted, non-tumor cell destruction, are not widely used. The pre-printed article published by Qian Yuan et. shown that circRNA-based in vivo CAR might enable higher and more durable CAR proteins on the cell membrane of functional immune cells than mRNAs. Combination of circRNA-based in vivo CAR and vaccines exhibited synergistically to enhance antitumor activity, holds the potential to become an upgraded off-the-shelf immunotherapy. According to a recent study, innate immunity and adjuvant action may be inhibited by m6A mutation of circFOREIGN, which is identified as "self" via YTHDF2 [134]. Clinically sophisticated carriers known as lipid nanoparticles (LNPs) can be utilized to package and transport RNA to specific organs [135]. An immune desert tumor model in B16 cells demonstrated that the circRNA-LNP vaccine can significantly suppress the immune-rejected tumor process, cause immune-desert tumors to regress, and prevent the spread of cancer cells. These findings demonstrate the superiority of cancer RNA vaccines over currently available cancer immunotherapies [136]. Chen et.al. developed a circRNA vaccine encoding HCC specific tumor neoantigen Ptpn2_I383T peptide (namely PTPN2), and proved the stability of the circRNA vaccine, and can continuously express the antigen protein, effectively stimulate a strong anti-tumor immune response, and significantly inhibit tumor growth [137]. The newly discovered circRNA vaccine represents a new safe, stable and simple immunotherapy option. CircRNA vaccines can increase the immunogenicity of cancer vaccines, which allows immunostimulatory proteins to more effectively activate the immune response during cancer immunotherapy. The potential of circRNAs as a source of new antigens in immunological agents was shown by an analysis of the candidate peptides from circRAPGEF and circMYH9, which are elevated in colorectal cancer and can trigger antigen-specific T cells reactions and proliferation [138]. CircRNAs also preserve antigen molecules (such as cytokines and small antibody fragments) for extended periods, which may stimulate the infiltration of immune cells to the tumor location. Specific circRNA vaccines can be developed for individualized and precise cancer treatment. These findings provide crucial insights into the use of circRNAs as key therapeutic approaches and synergistic inhibitors of tumor development in conjunction with established tumor immunotherapies. Moreover, circRNAs have longer sequences than miRNAs and siRNAs, and off-target effects are greatly reduced when circRNAs are used for targeted tumor immunotherapy. The targeting of circRNAs could increase the antitumor response. Treatments that directly control the levels of certain circRNAs in tumors via gene-editing therapy may have fewer side effects and may be more effective than chemoradiotherapy or chemical drugs.

Conclusions and outlook

Increasing researches have demonstrated that circRNAs are essential for several in processes in tumors. CircRNAs have been confirmed to be critical for tumor-related processes, including growth, angioplasty and metastasis, and can further regulate the TME and promote antitumor immunity. Recently, circRNAs were discovered to be involved in antitumor immune responses in several ways. For example, circRNAs can activate immune cells, especially CSCs, or shape the TIME by regulating the expression of immune checkpoints and secretion of cytokines and chemokines to regulate antitumor immunity. Tumor cells that secrete exo-circRNAs can also induce NK cells/T cells exhaustion in the TME to cause an immune evasion. The broad importance of circRNAs in cancer immunotherapy is gradually becoming known. The combination of circRNAs and ICIs for sensitizing tumors to immunotherapy or the use of circRNA vaccines alone represent key steps forward in our understanding of tumor immunity and immunotherapy resistance. CircRNA vaccines form bench to bedside will bring benefit to more tumor patients.

From a clinical perspective, circRNAs have shown promise as valid biomarkers for diagnosis and treatment in various tumors, including early detection, diagnosis, prognosis, and therapeutic response. Due to their stability in circulation, circRNAs may serve as early indicators of tumor, allowing for timely diagnosis before the onset of symptoms. Certain circRNAs may be specific to particular tumor, helping to confirm diagnoses when traditional methods yield ambiguous results. Expression levels of specific circRNAs may correlate with tumor aggressiveness, offering prognostic insights and helping to stratify patients according to risk. Some circRNAs may predict how tumors respond to specific therapies, allowing for more personalized treatment approaches. Engineered circRNAs could be designed to restore or enhance tumor-suppressive pathways, potentially counteracting oncogenic signals. However, there are some questions and challenges related to circRNAs as therapeutic target.

First, tumor immune escape is a difficult problem in tumor therapy. Several treatments, including ICIs, have failed to address this problem. By making full use of the functions of circRNAs in tumor immunity, the combination of circRNAs with existing clinical treatment methods (such as PD1/PD-L1 ICIs to enhance cancer immunotherapy sensitivity), more effectively activates the host immune system and prevents the immune escape of tumors, highlighting a direction worthy of further research. Second, the stability, conservation and tissue specificity of circRNAs can be fully utilized to produce targeted drugs against certain components of tumor cells or the TME to greatly limit adverse effects on normal cells in the body, improve the recovery rate of patients, and reduce adverse effects for treatment while improving patients' quality of life. For example, LNP-coated circRNA-targeting vectors regulate circRNA expression in tumors. CircRNAs are persisitent markers producedin response to immunotherapy and can also be used as tumor vaccine targets to improve the antitumor immune response.

Although the circRNA sector shows significant prospects for clinical transformation, several scientific problems and challenges in immune processes must be solved before this translational promise can be fulfilled. 1) Only 10%-20% of circRNAs are conserved between humans and mice. This property limits the translation of findings on poorly conserved circRNAs in animal models to patients. 2) Current technical restrictions on circRNA synthesis include the low cyclization efficiency and costly enzymes and reagents. The special technical problems related to the cyclic structure include the low reproducibility of biomarkers and the increased complexity of cancer immunotherapy development. 3) Many circRNA biomarkers have been poorly developed. At present, the determination of tumor markers mostly depends on comparisons of tumor-adjacent tissues or normal tissues, while changes in the TME, such as the secretion of immune cells, inflammatory factors and cytokines, are rarely comprehensively considered. Whether the presence of ferroptosis-related circRNAs or immune-related circRNAs, similar to long non-coding RNAs, can predict tumor prognosis, immune cell infiltration, and sensitivity to immunotherapy remains unclear.

Whether circRNAs are engaged in the regulation of tumor immune checkpoints and the TIME at multiple levels (transcription, posttranscription, translation, and protein modification), as well as their effects on the evolution of immune cells themselves, remains to be further explored. Strategies for utilizing these characteristics and roles of circRNAs to boost the immune response, prevent the immune escape of tumors, improve the efficacy of immunotherapy, and ultimately realize the translation of clinical still need to be further explored.