Possible mechanisms of circRNAs underlying SARS-COV-2 infection

CircRNAs sponge miRNAs to influence viral replication. CircRNAs, which show a high evolutionary conservation degree, exhibit tissue- and cell-specific expression patterns.^117-119 CircRNAs containing miRNA-responsive elements are the vital components of ceRNA network, which can modulate the expression of downstream target genes through sponging microRNAs (miRNAs), swiftly binding to specific miRNAs, and alleviating their inhibition against messenger RNAs (mRNAs) translation. Clearly, circRNAs containing miRNA-responsive elements can effectively sponge miRNAs in the ceRNA network in order to impede the miRNA-dependent target gene modulation.^{33, 120} CircRNAs are verified to have higher sponging activities than long noncoding RNA (lncRNA) transcripts and linear miRNAs.^{70, 121} Arora et al.¹²² recently detected the ceRNA network within SARS-CoV-1-infected cells, which comprised a lncRNA (Gm26917), a miRNA (MMU-miR-124-3p), a mRNA (Ddx58), a transcription factor (TF) (Stat2), together with two circRNAs (C330019G07Rik, Ppp1r10). The RIG-I/Ddx58 receptor in this ceRNA network interacts with SARS-CoV-1 nsp13, which initiates the viral life cycle (Figure 3A). Ddx58 also participates in mRNA splicing and miRNA biosynthesis. The increased expression can result in miRNA splicing reprogramming, which decreased miRNA expression. miR-124-3p upregulation induces Ddx58 degradation, leading to decreased viral replication. Moreover, miR-124-3p can regulate Toll-like receptor (TLR)-dependent innate immunity via Stat3, therefore downregulating TNF-α and IL-6.¹²³ The above findings indicate a potential similar role of miR-124-3p in modulating Stat2, impacting viral life cycle. In the ceRNA network, two circRNAs (C330019G07Rik and Ppp1r10) execute crucial modulating functions as the miR-124-3p sponges to inhibit Ddx58 degradation, which can affect SARS-CoV-1 replication. Additionally, as observed by Zhang et al.,¹²⁴ host circRNAs mainly served as the miRNA sponges to influence MERS-CoV replication. From Figure 3B, hsa_circ_0067985 originates in FNDC3B gene, which sponges hsa-miR-1275, while hsa_circ_0006275 originates in CNOT1 gene and sponges hsa-miR-2392. These two circRNAs show significant upregulation during MERS-CoV infection, thereby regulating the expression of key downstream target genes including MYO15B, MAP3K9, MEF2C, SPOCK1, ZBTB11, and USP15. Collectively, the findings provide novel insights into the circRNA regulation and associated signaling pathways, presenting potential host-targeted antiviral measures to resist SARS-CoV-2 infection.

CircRNAs affect cytokines to regulate the immunity against SARS-CoV-2 infection. CircRNAs exert a vital role in preventing viral damage by regulating the immunity in host–virus interaction.¹²⁵ According to Li et al.,¹²⁶ NF90/NF110, originating in human interleukin-enhanced binding factor 3 (ILF3), made a direct impact on governing back-splicing and collaboration with circRNA generation during viral infection. In the case of viral invasion, NF90/NF110 translocates into nucleus, which partly contributes to a reduction in circRNA production. The accumulation of NF90/NF110–circRNP complexes in the cytoplasm probably affects the host immunity. Based on the obtained results, circRNAs compete against viral mRNAs to bind to NF90/NF110, which potentially serve as the NF90/NF110 molecular reservoir to achieve rapid immunity during viral infection.^{126, 127} According to Chen et al.,¹²⁸ self-splicing introns-derived circRNAs could bind to retinoic acid-inducible gene I (RIG-I) receptor to efficiently activate immune signaling during viral infection. It has been recently suggested that hsa_circ_0000479 upregulation among COVID-19 patients modulates RIG-I and IL-6 levels through sequestering hsa-miR-149-5p.¹²⁹ Furthermore, circRNAs regulate immunity by mediating signaling pathways and cytokine production. Recently, an integrative analysis of protein transcription demonstrates significant modulation of hypoxia-inducible factor-1 (HIF-1), epidermal growth factor receptor (ErbB), tumor necrosis factor (TNF), and mammalian target of rapamycin (mTOR) signaling pathways through SARS-CoV-2 infection.¹³⁰ For DEcircRNAs, their parental genes associated with the above pathways are involved in various antiviral signaling pathways, including chemokines, interferon (IFN), RIG-I-like receptors, and mitogen-activated protein kinases (MAPKs).¹¹⁷ Therefore, circRNAs can regulate cellular signaling and immuno-inflammatory responses in SARS-CoV-2 infection.

Small molecular polypeptides, cytokines and glycoproteins produced via various cells, exert vital effects on numerous physiological events, including immuno-inflammatory response regulation. Acting as immunomodulators, cytokines are related to endocrine, paracrine, and autocrine signaling, contributing to the immune response against viral infections.^{131, 132} Viral infections can induce cytokine production, which is of great importance for immune response control, antiviral defense, and target cell capacity to support viral replication.¹³³ The infection with SARS-CoV-2 induces auto-immunity by activating specific immune factors, including Th1 chemokines CXCL9/10/11, interferon-inducible proteins, and 2′−5′ oligoadenylate synthase (OAS1-3).¹³⁴ Additionally, it decreases ribosomal protein transcription.¹³⁴ OAS activation necessitates the participation of viral genome dsRNA, leading to 2′−5′ oligoadenylate (2′−5′A) generation. Then, 2′−5′A markedly enhances RNase L activity to execute the antiviral activity, thereby degrading viral RNA whereas impeding viral protein production.¹³⁵ As discovered by Liu et al.,¹⁰⁰ dsRNA-activated protein kinase R (PKR) inhibitors are generated by endogenous circRNAs (16–26 bp) duplexes. In addition, circRNAs undergo global degradation via endonuclease RNase L, activating PKR antiviral pathway.

CircRNAs and airway smooth muscle cells

In general, mild asthma can be efficiently treated, while severe asthma will bring significant burdens on the economy and society. It usually entails notable alterations of airway smooth muscle cells (ASMCs), including hypertrophy, hyperplasia, as well as changes of fibrosis- and inflammation-associated proteins.¹³⁶ These aberrations in ASMCs make significant contributions to airway remodeling in severe asthma.¹³⁷ Regrettably, there still lack such strategies for addressing such changes in ASMCs. Hence, it is imperative to thoroughly understand the asthma pathogenesis to develop efficient ways for mitigating the influence on ASMCs caused by severe asthma and improve patient prognosis.

The development of next-generation sequencing has made it possible to identify many circRNAs in different types of cells and species.⁸⁸ In asthma, there are few studies investigating the effect of circRNAs on ASMC dysfunction. The role of circRNAs in regulating the biological and cellular processes of ASMCs during asthma has recently been revealed in numerous studies. These processes, such as growth, invasion, apoptosis, extracellular matrix (ECM) generation, and inflammatory factor production, are primarily mediated through their role as ceRNAs for miRNAs, thereby regulating target gene mRNAs.

CircERBB2, originating at ERBB2 locus, is a notable circRNA in this context and is correlated with tumor genesis. CircERBB2 upregulation can be detected within bronchial biopsy tissues in patients with asthma and ASMCs treated with platelet-derived growth factor BB (PDGF-BB), the in vitro model of asthma.¹³⁸ PDGF-BB has been the factor implicated in inflammation and airway remodeling during asthma,¹³⁹ which can stimulate ASMCs growth and invasion.^{140, 141} Noteworthily, PDGF-BB upregulation is closely related to asthma severity, with increased ASMC proliferation and invasion being observed in asthma patients.¹⁴² CircERBB2 silencing with small hairpin RNA (shRNA) significantly alleviates the proliferation, invasion and inflammation of ASMCs caused by PDGF-BB.¹³⁸ From the mechanism perspective, circERBB2 can sponge miR-98-5p to regulate the expression of insulin-like growth factor 1 receptor (IGF1R) and facilitate ASMC growth and invasion after PDGF-BB administration.¹³⁸

Another circRNA of interest, circHIPK3, also called circ_0000284, shows upregulation in ASMCs upon PDGF-BB stimulation.^{143, 144} CircHIPK3 can enhance the growth and invasion of PDGF-BB-treated ASMCs and inhibit their apoptosis by interacting with miR-326 for regulating stromal interaction molecule 1 (STIM1).¹⁴³ STIM1 is an endoplasmic reticulum membrane protein exerting a role of the calcium sensor as well as the stimulatory factor for airway hyperresponsiveness (AHR) and ASMC remodeling during asthma.^{145, 146} In addition, circHIPK3 enhances ASMC growth, migration, and invasion after PDGF administration, which is achieved by regulating miR-375/matrix metalloproteinase-16 (MMP-16) axis.¹⁴⁴

Circ_0002594 and circ_CSNK1E are upregulated in PDGF-stimulated ASMCs, and circ_0000029 exhibits decreased expression.^147-149 Both circ_0002594 and circ_CSNK1E are significantly upregulated in samples collected from asthma patients.^{147, 148} Functional research shows that shRNA-induced circ_CSNK1E silencing significantly decreases ASMC proliferation and migration after PDGF treatment.¹⁴⁷ Based on computational research and experimental verification, circ_CSNK1E can regulate vesicle-associated membrane protein 2 through binding to miR-34a-5p, therefore promoting the PDGF-mediated ASMC growth and invasion.¹⁴⁷ Besides, circ_0002594 can promote ASMC growth, invasion, and inflammation after PDGF administration. Moreover, circ_0002594 can sponge miR-139-5p to regulate tripartite motif 8 (TRIM8) expression, finally triggering cellular injury.¹⁴⁸ In some studies, miR-139-5p administration and TRIM8 suppression decrease ASMC growth through suppressing the chromatin remodeling factor BRG1 and inhibiting NF-κB pathway.¹⁵⁰ Circ_0000029 can modulate ASMC growth and invasion by regulating KCNA1 and miR-576-5p expression.¹⁴⁹ Combined with the above-mentioned results, circRNAs are vital for regulating crucial activities of asthmatic ASMCs by interacting with miRNAs and target genes. Nevertheless, the complete circRNA activity scope and their interplay with additional mechanisms within ASMCs during asthma remain unclear. More investigations are needed to comprehensively investigate the impacts of circRNAs on asthmatic ASMCs.

CircRNAs and airway and bronchial epithelial cells

Airway epithelial cells play a crucial role as the first-line protecting barrier, safeguarding both airway and lungs from antigens and inflammatory stimuli. Airway and bronchial epithelial cell impairments have been identified as a common feature of asthma. Abnormal circRNAs expression can significantly impact the functions of airway and bronchial epithelial cells.

In a comprehensive profiling study, Jia et al.¹⁵¹ utilized microarrays to investigate circRNAs within PM2.5-exposed human bronchial epithelial cells (BEAS-2B), and PM2.5 was the respiratory risk associated with pulmonary disorders like asthma. Circ_406961 is one of the dysregulated circRNAs, exhibiting the notable difference in fold change. Experiments indicate that circ_406961 is dose-dependently downregulated within PM2.5-exposed BEAS-2B cells. Circ_406961 knockdown with siRNA aggravates inflammation and decreases cell viability, while its upregulation exerts the opposite effects. Mechanism research suggests that circ_406961 is vital for regulating inflammation through interaction with ILF2 and activation of STAT3/JNK pathways.

Moreover, circVPS33A is upregulated in plasma samples collected from plasma patients and in BEAS-2B cells stimulated by house dust mites (HDM) protein (Der p1).¹⁵² Der p1 stimulation reduces cell viability, growth, invasion, migration, and promotes autophagy, inflammation, and apoptosis of BEAS-2B cells. CircVPS33A knockdown with siRNA can mitigate the Der p1-induced cellular injuries. Mechanically, circVPS33A can improve the above cellular injuries through sponging miR-192-5p, thus regulating HMGB1 expression. CircARRDC3 can promote mucus generation and aggravate inflammation of IL-13-treated nasal epithelial cells during allergic rhinitis, serving as the prominent risk factor related to asthma progression.¹⁵³ Its regulation can be achieved via miR-375/KLF4 axis. Collectively, the above results suggest the contribution of dysregulated circRNAs to asthma-related dysfunction in bronchial epithelial cells. Nevertheless, the above results need to be validated by clinical studies and in vivo animal models.

CircRNAs and airway goblet cells

Airway goblet cells are crucial for keeping airway homeostasis by producing mucin.¹⁵⁴ Goblet cells can differentiate from airway epithelial cells called club (Clara) cells upon stimulation with growth factors, genetic factors, environmental insults, and inflammation.^{154, 155} The out-of-control proliferation of goblet cells promote mucin generation in the airway.¹⁵⁶ Excess mucin accumulation can be frequently observed in different airway disorders, like asthma, cystic fibrosis, and chronic obstructive pulmonary disease (COPD).¹⁵⁶ Hyperplasia of goblet cells (the higher goblet cell count) in asthma is an important airway remodeling feature, which is correlated with chronic excessive mucus production and the resulting airway obstruction.¹⁵⁷ CircRNAs can induce hyperplasia of goblet cells. For example, Wang et al.¹⁵⁸ suggested that circZNF652, also called circ_0000782, enhanced goblet cell hyperplasia in the allergic epithelium. CircZNF652 is specifically and significantly expressed within airway epithelium in asthmatic children and OVA-mediated experimental mouse models.¹⁵⁸ CircZNF652 upregulation promotes metaplasia of bronchial goblet cells and result in excess mucus production. CircZNF652 can sponge miR-452-5p and activate JAK2/STAT6 axis to promote goblet cell metaplasia.¹⁵⁸ ESRP1, which is a splicing factor, can facilitate circZNF652 expression, leading to goblet cell metaplasia.¹⁵⁸ The above results indicate that ESRP1/circZNF652/miR-452-5p/JAK2/STAT6 axis is vital for modulating mucus hypersecretion, goblet cell metaplasia, and AHR within the experimental asthmatic mouse models. Therefore, it is the possible strategy to target circZNF652 and activate miR-452-5p to intervene with epithelial remodeling in experimental asthma. However, the precise mechanism related to ESRP1 upregulation in allergic airway epithelium is still unknown. CircZNF652 can regulate RhoA, EIF5, JAK2, and SRSF3. JAK2 is found to be related to goblet cell metaplasia during asthma, while the effects of additional factors on asthma need to be further explored.

CircRNAs and immune cells

The airways are infiltrated and activated by different immune cells during asthma, including dendritic cells (DCs), innate lymphoid cells (ILCs), T cells (Th1/Th2/Th9/Th17/Th22 cells), neutrophils, eosinophils, mast cells, and B cells.¹⁵⁹ The two major types of asthma are type 2 (dominant type) and non-type 2. Type 2 asthma is generally caused by allergies and nonallergic factors. In contrast, non-type 2 asthma usually results from infections and contaminants. Once ILCs and Th2 cells, particularly ILC2 cells, detect an allergen, they are capable of secreting cytokines including IL-4, IL-5, and IL-13. Subsequently, these cytokines trigger the recruitment of eosinophils, ultimately causing inflammation in the airways.¹⁵⁹ Mast cells can produce histamine, thereby facilitating bronchoconstriction, while neutrophils are associated with severe asthma. To initiate immune responses, DCs present antigens to T cells, whereas macrophages accelerate or suppress inflammation according to their M1 or M2 type. Allergies can be triggered by IgE antibodies produced by B cells. It is possible to experience typical asthma symptoms, like bronchoconstriction or acute hypotension, caused by the above-mentioned immune mechanisms.

Recently, circRNAs have been suggested as important ceRNAs for miRNAs in the function and development of T cells. m_circRasGEF1B, the LPS-mediated circRNA, is an example suggesting that circRNAs are involved in immunoregulation. It can modulate the ICAM-1 mRNA stability to promote immunoregulation in human body.¹⁶⁰ As the cell surface glycoprotein, ICAM1 can be found in certain immune cells and endothelial cells, and it has been identified as the inflammation biomarker. Through the analysis of circRNA expression patterns in diverse cells, different expression patterns are revealed, indicating that they can regulate some cellular processes. For instance, hsa_circ_0012919 downregulation can result in DNA methylation of CD11a and CD70 in CD4 T cells.^161-164 Hsa_circ_0045272 may negatively regulate IL2 production and T cell apoptosis through interaction with hsa-miR-6127.¹⁵⁸ CircIKZF1, circTNIK, circTXK, and circFBXW7 are circRNAs with specific expression in T cells.¹⁶³

Recently, circRNAs have been increasingly suggested to be associated with the asthma pathophysiology via regulating immune responses.^{165, 166} CD4 T cells are substantially demonstrated to drive disease occurrence by modulating IgE generation upon allergic conditions, generating proinflammatory factors, and recruiting and activating inflammatory cells including neutrophils, eosinophils, and macrophages.^167-176 CD4⁺ T cells can produce cytokines. Specifically, Th1 cells produce IFN-γ, while Th2 cells secrete IL-4, IL-5, IL-6, IL-9 and IL-13, and Th17 cells can release IL-17A and IL-17F.¹⁷⁷ In asthma, the complicated immune response can be modulated through the complicated cytokine network.^{168, 173-175, 178} To analyze the effect of circRNAs on CD4 T cells during asthma, Huan et al.¹⁷⁹ examined the circRNAs expression patterns within CD4 T cells collected in five asthma patients and five normal subjects by micro-array analysis. Totally 597 circRNAs were abnormally expressed between asthma patients and normal subjects. Among them, hsa_circ_0005519 was most significantly upregulated, as subsequently verified through qRT-PCR assay in the current cohort samples and in another cohort including 65 asthma individuals and 30 normal controls. Through bioinformatic analysis, literature search and reporter assays, the authors suggested the role of hsa_circ_0005519 in regulating IL13 and IL6 expression, probably by interaction with let-7a-5p. IL-13 and IL-6 are extensively suggested to be significantly related to the asthma pathogenesis.^{167, 178, 180, 181} Hsa_circ_0005519 expression within CD4 T cells is negatively related to let-7a-5p, whereas positively related to IL13 and IL6 mRNA expression, FeNO and peripheral blood eosinophil ratio. As a result, hsa_circ_0005519 is the candidate biomarker for asthma.

In one article, hsa_circ_0002594 was found to be upregulated within CD4⁺ T cells from a present microarray dataset containing five asthma patients and five normal subjects.¹⁷⁹ For validation, another cohort including 83 asthma patients and 54 normal controls was analyzed, revealing that hsa_circ_0002594 expression was significantly upregulated in CD4⁺ T cells from asthma patients.¹⁷⁹ Hsa_circ_0002594 expression in CD4+ T cells of asthma patients was positively related to FeNO, but negatively correlated with PD20 (the methacholine dosage necessary for inducing FEV1 decline by 20%). The correlation between hsa_circ_0002594 expression of asthma group and specific clinical factors was analyzed, showing that the hsa_circ_0002594 high-expression group had increased FeNO expression but decreased PD20 expression in relative to hsa_circ_0002594 low-expression group. The high-expression subgroup had more skin prick test (SPT)-positive and Th2-high patients than SPT-negative and non-Th2 inflammation counterparts. Based on the above results, asthma patients who showed hsa_circ_0002594 high-expression exhibited clinical indicators consistent with a Th2-mediated allergic response. Inhaled corticosteroids (ICS) administration significantly reduces hsa_circ_0002594 expression, suggesting that it may be the therapeutic target. Furthermore, hsa_circ_0002594 is sensitive and specific in asthma diagnosis, no matter whether ICS is applied or not. Molecular mechanisms related to hsa_circ_0002594 underlying asthma occurrence were explored by computational research; as a result, possible target miRNAs were identified, proposing the role of hsa_circ_0002594 in the competitive sequestering of the activities of hsa-miR-503-5p, hsa-let-7e-5p, hsa-miR-587, hsa-miR-16-5p, and hsa-miR-514a-3p. However, this study still has the following limitations: (1) The sample size was small, and thus large-scale and more studies are warranted, (2) experimental verification for miRNA-hsa_circ_0002594 interaction was not available, and (3) the hsa_circ_0002594 level in circulation was not evaluated for its potential as a biomarker.

Mmu_circ_0001359 has been recently found to be downregulated within lungs in the murine model of OVA-mediated asthma.¹⁸² Delivering adipose-derived stem cells (ADSCs)-derived exosomes after mmu_circ_0001359 modification can alleviate airway remodeling by activating M2-like macrophages via miR-183-5p/FoxO1 pathway.¹⁸²

CircRNAs expression within lungs in mice with HDM-induced experimental asthma was analyzed.¹⁸³ The circRNA–miRNA axis was explored, which identified two circRNAs with upregulated expression, including circ_0000629 and circ_0000455 targeting miR-29b and miR-15a separately. Previous research reported that such miRNAs were inversely related to allergic reactions.^{134, 184} By contrast, two circRNAs with downregulated expression, namely, circ_0000723 and circ_0001454, target miR-214 and miR-146b, respectively. MiR-214 and miR-146b show a positive relationship to asthma.^{135, 185} Therefore, the aforementioned four circRNAs are candidates deserving subsequent analysis concerning the association with asthma.

Type 1 diabetes

Type 1 diabetes (T1D), also called insulin-dependent diabetes, is the chronic autoimmune disorder with the representative characteristic of β-cell destruction and dysfunction, absolute insulin deficiency, and increased blood glucose content.²¹⁸ Substantial evidence demonstrates the effects of circRNAs on autoimmune disorders, partially by evaluating autoimmune disorders in T1D research.

There were 61 circRNAs with upregulation while 7 with downregulation detected within plasma from new-onset T1D patients, with four circRNAs (hsa_circRNA_100332, hsa_circRNA_101062, hsa_circRNA_085129, and hsa_circRNA_103845) being demonstrated in succession using quantitative real-time PCR.²¹⁹ There were 93 DEcircular transcripts detected from peripheral blood in T1D patients relative to normal subjects, with 30 being upregulated while 63 being downregulated. Based on GO and KEGG analyses, such circRNAs contributed to T1D occurrence through diverse pathways. Hsa_circ_0072697 was the possible endogenous miRNA sponge to regulate β-cell activity.²²⁰ The above dysregulated circRNAs serve as potential biomarkers for T1D, which should be further validated based on longitudinal studies. CircRNAs are related to the inflammatory response during T1D, which may offer the new therapeutic target for the disease. CircPPM1F is increasingly suggested to be mainly expressed within and promotes M1 macrophage activation through circPPM1F–HuR–PPM1F–NF-κB pathway. CircPPM1F upregulation may aggravate injuries to pancreatic islet.²²¹ Macrophages, which are the immune response inflammatory mediators, are essential for insulitis. Besides, they can support autoimmune T cells during T1D through promoting the infiltration of inflammatory cells.²²² As discovered by Yang et al.,²²³ hsa_circ_0060450 expression increased in T1D patients, and it sponged miR-199a-5p for suppressing macrophage-induced inflammation through JAK-STAT pathway. Moreover, 1922 significantly expressed circRNAs were detected in mouse islet β-cells after induction of different cytokines, including interleukin-1β (IL-1β), TNF-α, and INF-γ.²²⁴ To analyze the impacts of circRNAs on T1D, Yi et al. constructed a circRNA–lncRNA–miRNA–mRNA ceRNA network of T1D by computational and bioinformatic analyses.²²⁵ Nonetheless, there is limited research on therapeutic circRNAs for T1D, and a larger sample size is needed for further validation.

Type 2 diabetes

Type 2 diabetes (T2D) has an increasing incidence rate due to the growing living standards and the changed diet structures, bringing significant health burdens on the patients and society.^{226, 227} The function of circRNAs in T2D pathogenic mechanisms and development has recently been explored, suggesting that circRNAs may serve as diagnostic biomarkers and therapeutic targets in diabetics.

Serum hsa_circ_0054633 was used to diagnose prediabetes and T2D among 63 prediabetics, 64 T2D patients and 60 normal controls.²²⁸ Serum hsa_circ_0054633 expression among T2D patients was related to low-density lipoprotein, hemoglobin A1c, and fasting blood glucose levels.²²⁹ Hsa_circ_0063425 combined with hsa_circ_0056891 could differentiate T2D patients and those with fasting glucose impairment from normal subjects among 313 subjects, and the AUC values were 0.837 and 0.719, respectively.²³⁰ Hsa_circ_0063425 and hsa_circ_0056891 are found to sponge miR-19a-3p and miR-1-3p, respectively, aiming to exert crucial effects on insulin phosphatidylinositol 3-kinase/protein kinase B pathway.²³⁰ Furthermore, circHIPK3 and CDR1as have increased expression among T2D patients.²³¹ CircHIPK3, also known as hsa_circ_0000284, is a highly abundant circRNA within pancreatic islets and is obtained based on exon 2 in HIPK3 that is conserved between mouse and human cells.²³² CircHIPK3 facilitates insulin resistance and hyperglycemia by increasing FOXO1 expression and sponging miR-192-5p.²³³ CircHIPK3 expression declines within DB/DB islets. CircHIPK3 knockdown promotes β-cell apoptosis while decreasing glucose-stimulated insulin production. CircHIPK3 significantly affects islet cells partially by sponging miR-29-3p, miR-30, miR-124-3p, and miR-338-3p, which is achieved through regulating critical β-cell genes like Slc2a2, Mtpn, and Akt1.²³⁴ The above miRNAs are widely suggested to significantly affect β-cell proliferation, development, function, and survival.^235-238 CircHIPK3 can differentiate T2D patients from normal subjects, and the specificity and sensitivity are 88.4 and 50.0%, respectively.²³¹ CDR1, first discovered in brain tissue²³⁹ is widely expressed in pancreatic β-cells. CDR1as is reported to modulate insulin production and transcription in islet cells by inhibiting miR-7 through regulating endogenous target genes Pax6 and Myrip.²⁴⁰ CDR1as expression is downregulated within OB/OB and DB/DB mice, conforming to its effect on diabetes occurrence. Furthermore, CDR1as overexpression increases islet cell insulin production and insulin level,²⁴⁰ while CDR1as knockdown has opposite effects and reduces the prolactin-mediated MIN6B cell proliferation in isolated rat islets.²³⁴ CDR1as is shown to differentiate prediabetics from normal controls (AUC = 0.605).²³¹

Circ-Tulp4 is the candidate treatment against T2D.²⁴¹ Based on in vitro experimental results, circTulp4 modulated MIN6 cell proliferation. Its overexpression upregulated SOAT1, which then increased cyclin D1 expression, finally promoting cell cycle progression while alleviating β-cell dysfunction by inhibiting miR-7222-3p.²⁴¹ CircGlis3, the exocircRNA obtained in β-cells, exhibits higher expression within serum samples collected from T2D patients and diabetic mice. It is translocated in islet endothelial cells and serves as the therapeutic target.²⁴² CircGlis3 is related to lipotoxicity-mediated β-cell disease and diabetes through inhibiting cell growth and insulin production. From the mechanism perspective, circGlis3 can decrease islet endothelial cell invasion and viability and is transported to target cells by exosomes. Intronic circRNAs produced by insulin genes exert a vital impact on T2D occurrence. For instance, circular intronic-Ins2 (ci-Ins2) is found to modulate β-cell activity.²⁴³ CircRNA containing the lariat sequence of intron 2 in insulin gene is reported to modulate insulin production in T2D patients and diabetes rodent models, which may clarify the impairment of secretion ability.²⁴⁴ CircANKRD36 is also associated with chronic inflammatory factors in T2D. According to subsequent functional analysis, circANKRD36 expression was positively correlated with IL-b and blood glucose levels.²⁴⁵ Furthermore, dysregulated circRNAs is correlated with pancreatic β-cell autophagy in T2D rats, which may result from rno_circRNA_008565 after interacting with miRNAs.²⁴⁶ Collectively, circRNAs are potentially related to inflammatory response, insulin resistance, and β-cell failure within T2D, and may be the candidate therapeutic target for T2D.

Gestational diabetes mellitus

CircRNAs are crucial for the prediction of several adverse reactions among gestational diabetes mellitus (GDM) patients. CircVEGFC, the new regulatory factor for glucose metabolism, has elevated expression among GDM pregnant women. Its upregulation is related to hypertension and fetal malformations.²⁴⁷ Pregnant women with circACTR2 upregulation may experience GDM. According to Zhu et al.²⁴⁸ aberrant circACTR2 expression was related to premature delivery, fetal malformations, miscarriage and intrauterine infection. Therefore, circRNAs are the candidate biomarkers that can be used for predicting the complication risk of GDM patients. However, their clinical application and the associated mechanisms should be further explored.

GDM is the frequently seen pregnancy complication, which has seriously threatened the maternal and fetal health. CircRNAs are found to be of therapeutic significance. For example, placental villous circRNAs are explored for their expression patterns in GDM women, suggesting that the dysregulated circRNAs are probably associated with GDM.^{249, 250} Chen et al.²⁵¹ discovered the high circ_0008285 expression among GDM patients, and circ_0008285 depletion inhibited cell growth, migration, and invasion. Hsa_circ_0005243 is related to suppressing cell growth and invasion and promoting IL-6 and TNF-α expression within GDM via β-catenin/NF-κB pathway.²⁵² CircPNPT1 expression increases within placental tissues in GDM patients and high glucose (HG)-stimulated trophoblast cells.²⁵³ At the mechanism level, circPNPT1 sponges miR-889-3p and activates p21-activated kinase 1 to modulate trophoblast cell dysfunction. CircPNPT1 is packaged in exosomes and subsequently delivered to additional cells.²⁵³

Effects of circRNAs on diabetic complications

Diabetes complications can result from persistent hyperglycemia, including diabetic retinopathy (DR), diabetic nephropathy (DN), diabetic neuropathy (DNP), diabetic cardiomyopathy (DCM), and HG-mediated endothelial dysfunction. The complications have seriously threatened human life. CircRNAs are widely suggested to play a key role in the occurrence of diabetic complications. Figure 4 displays circRNAs that are verified to be related to diabetes-related disorders.

CircRNAs in DR

DR can be classified as proliferative or nonproliferative DR based on its severity.²⁵⁴ DR, the frequent diabetes-related microvascular complication, is the major factor inducing blindness among working-aged individuals.²⁵⁵ It is correlated with alterations of endothelial cells and pericytes and shows the typical features of vascular leakage and capillary occlusion.²⁵⁵ CircRNA dysregulation is linked with DR progression; therefore, they may be the key biomarkers and reasonable therapeutic targets.

With regard to proliferative DR, circRNAs are the diagnostic biomarkers. According to Zhang et al.,²⁵⁶ the blood hsa_circ_0001953 expression increased in proliferative DR individuals in relative to nonproliferative DR counterparts and normal subjects. Nevertheless, this should be verified with a larger population. Totally 529 dysregulated circRNAs were detected from two diabetic human retinas. Among them, hsa_circ_0005015 had increased expression within plasma, fibrovascular membrane, and vitreous samples in DR individuals. It promoted retinal endothelial angiogenesis through modulating endothelial cell growth, tube formation, and migration. Hsa_circ_0005015 was the candidate diagnostic biomarker for DR. Subsequent functional experimental analysis showed that hsa_circ_0005015 sponged miR-519d-3p to inhibit miR-519d-3p activity, thus upregulating XIAP, MMP-2, and STAT3.²⁵⁷ Hsa_circ_0001883 and hsa_circ_0095008 have been recently identified as early diagnostic biomarkers for DR patients, and the AUCs for diagnosis are 0.607 and 0.671, respectively.²⁵⁸

Circhipk3 is upregulated in retinal endothelial cells and diabetic retinas, which blocks mir-30a-3p to increase the expression of vascular endothelial growth factor-C (VEGF-C), WNT2 and FZD4, resulting in vascular dysfunction and endothelial proliferation.²⁵⁹ circfto, circdnmt3b, circcol1a2, circ_001209, circslc16a12, circ-UBAP2, and exosomal circfndc3 can modulate the diabetic retinal vascular function.^260-264 They are the candidate targets for controlling DR. CircDNMT3B knockdown aggravated visual injury, while its upregulation alleviated retinal vascular dysfunction. By contrast, upregulation of circ_001209 sponges miR-15b-5p to exacerbate retinal injury of diabetic rats.^{265, 266} CircRNAs affect the pericyte-endothelial cell crosstalk to modulate DR.^{267, 268} cPWWP2A, the new DR-related circRNA, is the ceRNA interacting with miR-579 for promoting retinal vascular dysfunction by upregulating angiopoietin 1, SIRT1, and occludin.²⁶⁷ Interestingly, cPWWP2A is expressed on pericytes rather than on endothelial cells, while it has indirect regulation on endothelial cell biology via exosomes. Pericytes-derived exosomal circEhmt1 is transported into endothelial cells, where it prevents from HG-induced injury.²⁶⁸ CircRNAs are essential for human retinal pigment epithelial (ARPE-19) cells. CircRNA_0084043 and hsa_circ_0041795 are significantly upregulated within HG-mediated ARPE-19 cells, which can decrease their growth and survival. At the mechanism level, hsa_circ_0041795 sponges miR-646 to facilitate HG-induced apoptosis in ARPE-19 cells by activating VEGFC miR-646.²⁶⁹ CircRNA_0084043 deficiency sponges miR-140-3p and induces TGFA to significantly increase HG-exposed ARPE-19 cell apoptosis. CircRNA_0084043 deficiency efficiently suppresses HG-induced inflammation by suppressing the inflammatory factors TNF-α, Cox-2, and IL-6 within ARPE-19 cells.²⁷⁰ Circ-PSEN1 and circ-ADAM9 are upregulated in DR. Circ-ADAM9 promotes injuries to ARPE-19 cells, while circ-PSEN1 knockdown alleviates ferroptosis.^{271, 272} CircZNF532 modulates pericyte biology through sponging miR-29a-3p, thereby upregulating NG2, CDK2, and LOXL2 expression. Besides, it regulates pyroptosis and apoptosis of ARPE-19 cells to facilitate inflammation and angiogenesis in DR.^273-275 Circ-ITCH is related to the DR pathology and physiology in rat retinal pigment epithelial cells.²⁷⁶ Recent research has shown that silencing circMET can reduce retinal angiogenesis.²⁷⁷ CircRNAs are involved in DR progression and may be used for the treatment of the disease.

CircRNAs in DN

DN is a major factor inducing mortality among diabetics.²⁷⁸ Its pathological features include progressive ECM deposition, basement membrane thickening, persistent mesangial cells (MCs) proliferation, and renal fibrosis.²⁷⁹ MC and ECM dysfunction play a vital role and serve as early warning signals for DN pathogenesis. It is of great importance to unravel molecular mechanism underlying DN and identify candidate therapeutic targets, aiming to improve patient survival.

Currently, circRNAs are identified as biomarkers for DN. Hsa_circ_0000867 and hsa_circ_0001831 show increased expression within peripheral blood in early type 2 DN cases.²⁸⁰ Hsa_circ_0001831 can be used to diagnose early type 2 DN, and the sensitivity, specificity, and AUC values are 0.85, 0.85, and 0.95, respectively.²⁸⁰ Hsa_circRNA_102682 is identified to independently predict the risk of DR in one article involving 73 patients), with the AUC for prediction being 0.97.²⁸¹

CircRNAs are widely used to treat DN, including circEIF4G2, circTAOK1, circACTR2, and hsa_circ_0125310.^282-285 CircFBXW12 has elevated expression within the serum of DN cases and HG-induced human MCs. Based on in vitro functional experimental results, circFBXW12 knockdown inhibited human MC proliferation and decreased ECM generation and oxidative stress via miR-31-5p/LIN28B pathway.²⁸⁶ CircSMAD4 can mitigate ECM accumulation, while circRNA_15698 can promote ECM accumulation.^{287, 288} Liu et al.²⁸⁹ suggested that circ_0080425 expression was related to DN development, which also positively affected MCs fibrosis and growth via the miR-24-3p/FGF11 axis. Circ_0000712 enhances the HG-induced apoptosis, fibrosis, oxidative stress and inflammation in DN through regulating SOX6 and miR-879-5p expression.²⁹⁰ Similarly, circ_0000491 sponges miR-455-3p to regulate Hmgb1 expression, therefore promoting HG-mediated apoptosis, fibrosis, oxidative stress and inflammation in DN.²⁹¹ In vivo, circRNA_010383 upregulation suppresses miR-135a expression, increases transient receptor potential cation channel, subfamily C, and member 1 expression, and suppresses kidney fibrosis and proteinuria of DN mice.²⁹² Circ_0000285 and circ_0037128 are upregulated within the DN mouse models. Circ_0000285 sponges miR-654-3p to activate MAPK6 expression and induce podocyte injury.²⁹³ Circ_0037128 has been recently suggested to modulate DN occurrence by miR-17-3p/AKT3 pathway.²⁹⁴ Circ_0037128 is reported to suppress HK-2 cell growth and enhance their fibrosis and inflammation by miR-497-5p/NFAT5 pathway. By contrast, circAKT3, circLARP4, circITCH, and mmu_circRNA_0000309 expression decreases within DN rats and mice, as well as in DN cell models and tissues.^{224, 295-297} Upregulation of circAKT3 sponges miR-296-3p to suppress the fibrosis-associated protein production, therefore modulating E-cadherin, and it may be the significant therapeutic target for patients with DN.²⁹⁵ CircHIPK3 can modulate insulin production and islet proliferation in order to facilitate insulin resistance. In two articles regarding DN, circHIPK3 significantly influenced rat MCs and human renal tubular epithelial HK-2 cells. From the mechanism perspective, circHIPK3 expression significantly decreased within HG-treated HK-2 cells. CircHIPK3 overexpression reduced the accumulation of miR-326 and miR-487a-3p, thereby upregulating SIRT1 while reducing HG-mediated suppression on HK-2 proliferation.²⁹⁸ Furthermore, circHIPK3 expression was found to be upregulated within rat MCs. CircHIPK3 knockdown suppressed cell growth and inhibited the mRNA abundance of PCNA, cyclin D1, TGF-β1, FN, and Col. I within MCs.²⁹⁹ However, more investigations are warranted to demonstrate the effect of circHIPK3 on DN.

CircRNAs in DNP

DNP, a common diabetic complication, can substantially decline the life quality of diabetic patients. DNP involves the sensory functional loss, which initiates at the distal lower extremities and is featured by stimulus-evoked and spontaneous pain, along with many illnesses.³⁰⁰ Nonetheless, the DNP etiology is still unknown. CircRNAs are found to regulate DNP progression.

NcRNAs, including miRNAs, circRNAs, and long ncRNAs, were analyzed for their expression in streptozotocin-caused DNP mice. There were totally 148 miRNAs, 135 circRNAs, 30 mRNAs, and nine long ncRNAs with significant dysregulation among DNP mice. They were enriched in Rap1, MAPK, protein digestion, and absorption pathways, and human T-cell lymphotropic virus-I infection pathways.³⁰¹ In addition, 133 mRNAs and 15 circRNAs with dysregulated expression were identified within dorsal root ganglia in diabetic and wild-type mice through chip scanning and high-throughput RNA sequencing. Meanwhile, there were 14 mRNAs and seven circRNAs closely associated with diabetic peripheral neuropathy.³⁰² They are the candidate diagnostic and prognostic biomarkers for DNP.

CircHIPK3 is abundantly expressed in DNP rats and patients. CircHIPK3 overexpression is positively related to neuropathic pain among T2D patients. CircHIPK3 knockdown mitigates neuropathic pain of diabetic rats, which may be probably associated with the reduced neuroinflammation. Mechanistically, circHIPK3 interacts with miR-124 to downregulate miR-124. The study provides the first evidence supporting the use of intrathecal circHIPK3 shRNA in the treatment of rat DNP.³⁰³

circRNAs in DCM

DCM is a main factor inducing morbidity and mortality among diabetics and brings significant burdens on the economy and society globally.^{304, 305} DCM exhibits the typical features of cardiomyocyte apoptosis, hypertrophy, myocardial interstitial fibrosis, as well as subsequent HF without dyslipidemia, hypertension and coronary artery disease (CAD).^{304, 306}

CircRNAs are substantially related to DCM development. According to Zhou et al., circRNA_010567 was significantly upregulated within Ang-II-induced cardiac fibroblasts (CFs) and the myocardium of diabetic mice. It modulated fibrosis-associated protein expression within CFs, including Col I, α-SMA, and Col III by miR-141/TGF-β1 pathway.³⁰⁷ CircRNA_000203 expression significantly increased within CFs and DCM mouse myocardium. Functional experimental analysis showed that circRNA_000203 upregulation inhibited miR-26b-5p and activates CTGF and Col1a2, finally upregulating α-SMA and Col3a1 within CFs.³⁰⁸ Subsequently, CircHIPK3 was found to promote myocardial fibrosis while reducing cardiac function of DCM mice. CircHIPK3 deficiency suppressed CFs proliferation by the Col3a1 and miR-29b-3p/Col1a1 regulatory network.³⁰⁹ Therefore, circRNA_010567, circHIPK3, and circRNA_000203 are related to the myocardial fibrosis pathogenesis, providing the therapeutic target and molecular mechanism for DCM.

CircHIPK3 is also discovered with downregulated expression in HG-exposed human cardiomyocytes and in DCM. CircHIPK3 upregulation inhibits the expression of PTEN, the tumor suppressor, and suppresses the HG-induced apoptosis in human cardiac myocytes.³¹⁰ Circ_0071269 is markedly upregulated in HG-exposed rat cardiomyocytes. Circ_0071269 silencing can sponge miR-145 to enhance cell viability and inhibit inflammatory response, pyroptosis, and cytotoxicity of rat cardiomyocytes in vitro, thereby upregulating GSDMA.³¹¹ Hsa_circ_0076631, called caspase-1-associated circRNA (CACR) as well, has elevated expression in HG-exposed cardiomyocytes and patients with diabetes. CACR knockdown activates caspase-1 by targeting miR-214-3p within cardiomyocytes. On the contrary, miR-214-3p deficiency partly reverses the benefits of CACR knockdown for the pyroptosis of cardiomyocytes. Consequently, CACR is the candidate new therapeutic target for DCM through regulating CACR/miR-214-3p/caspase-1 pathway.³¹²

CircRNAs in hyperglycemia-associated wound healing and endothelial dysfunction

Endothelial cells constitute the inner vascular lining, which are the anticoagulant barrier between blood and vessel wall.³¹³ These unique multifunctional cells regulate vasomotor tone, metabolic homeostasis, vascular smooth muscle cell (VSMC) growth, and immune/inflammatory responses.³¹⁴ Hyperglycemia has been previously suggested to induce endothelial cell dysfunction among diabetics.^{315, 316} Hyperglycemia is widely suggested to change circRNA expression, while circRNAs are related to different hyperglycemia-mediated endothelial dysfunction processes during diabetes.³¹⁷

Human umbilical vein endothelial cells (HUVECs) can be easily obtained and extensively applied for in vitro research on endothelial cells. Based on Cheng et al.,³¹⁸ hsa_circ_0068087 promoted the inflammation and dysfunction of endothelial cells via miR-197/TLR4/NF-κB/NLR family pyrin domain containing 3 (NLRP3) pathway. CircBPTF is under close regulation in hyperglycemia-mediated HUVECs. CircBPTF deficiency targets miR-384 and downregulates LIN28B to suppress oxidative stress and inflammatory injuries.³¹⁹ Comparatively, circ_CLASP2 expression declines in hyperglycemia-mediated HUVECs. Circ_CLASP2 regulates the miR-140-5p/FBXW7 pathway to prevent hyperglycemia-induced dysfunction of HUVECs.³²⁰

Based on significant evidence, circRNAs like circWDR77, circLRP6, and circYTHDC2 show upregulation within hyperglycemia-exposed VSMCs.^321-323 The above circRNAs enhance VSMCs growth and invasion via diverse mechanisms. Suppressing circWDR77 inhibits VSMC growth and invasion via miR-124/fibroblast growth factor 2 pathway.³²¹ CircYTHDC2 is found to modulate VSMC dysfunction through YTHDC2-induced m6A modification and regulate ten-eleven translocation 2 expression.³²³ CircLRP6, which sponges miR-545-3p, contributes to the growth and invasion of hyperglycemia-exposed VSMCs by downregulating HMGA1.³²²

Various circRNAs are related to hyperglycemia-mediated wound healing. Circ-Klhl8 expression is enhanced in endothelial progenitor cells under hypoxic pretreatment conditions, whereas circ-Gcap14 expression is enhanced in ADSCs.^{324, 325} Upregulation of circ-Gcap14 enhances angiopoiesis and accelerates hypoxic ADSC wound closure in diabetics.³²⁴ Based on Shang et al.,³²⁵ circ-Klhl8 promoted autophagy, suppressed hyperglycemia-mediated endothelial cell injury, and enhanced diabetic wound closure via the miR-212-3p/SIRT5 pathway. Recently, exosomal circARHGAP12 and mmu_circ_0000250 are indicated to promote diabetic wound closure by regulating autophagy.^{326, 327} From the mechanism perspective, circARHGAP12 can sponge miR-301b-3p to enhance diabetic wound healing, thereby regulating target genes ATG16L1 and ULK2 expression.³²⁷ CircARHGAP12 is the candidate novel therapeutic target for diabetic wound closure.

Aberrant adipocyte metabolism

Adipocyte metabolism refers to a process to digest, absorb, synthesize and decompose body adipose tissues under the action of different enzymes. Aberrant adipocyte metabolism usually results in further disorders, including metabolic syndrome, hyperlipidemia, AS, and nonalcoholic fatty liver disease (NAFLD).^328-330 Adipocyte metabolism is a complicated physiological process that involves a variety of regulatory factors and genes. It becomes crucial to maintain the adipocyte metabolic homeostasis to prevent and treat adipogenesis-associated diseases. CircRNAs display specific expression profiles in diverse tissues, cells, and developmental stages, impacting the metabolic processes in vivo.³³¹ CircRNAs are increasingly suggested to be associated with adipocyte metabolism and the associated disorders (Table 2).

CircRNAs modulates adipocyte metabolism within adipose tissue

CircRNAs can impact adipose function in numerous species, including mice, humans, cattle and pigs. As reported by Zhu et al.,³³² hsa_circH19 exhibited upregulation within the blood in patients experiencing metabolic syndrome. Hsa_circH19 silencing enhanced adipose stem cell differentiation via PTBP1. By conducting sequencing on human subcutaneous and visceral adipose tissues, Arcinas et al.³³³ identified circTshz2-1 and circArhgap5-2 as the essential regulatory factors for adipogenesis in vitro. Based on the obtained results, circRNAs expression exerts a vital role in keeping adipocyte metabolism. CircSAMD4A also sponges miR-138-5p and modulates EZH2 expression in human obese disorders to control the adipogenesis of obese people.³³⁴ CircARF3 can sponge miR-103 to inhibit miR-103 activity and upregulate TRAF3, miR-103's downstream target. Then, it alleviates the high-fat-diet-induced inflammation in mouse adipose tissue.³⁴⁶

CircRNAs can sponge miRNA within adipose tissue and directly bind to RBPs, thus participating in expression. Circ_0075932 expression exerts a vital role in human normal adipose tissue. Exosomes derived from circ_0075932 overexpression adipose cells significantly enhance cell apoptosis and inflammation. circ_0075932 accelerates cell apoptosis and inflammation through direct binding to RBP PUM2 and enhancing PUM2-induced AuroraA/NF-κB pathway activation.³³⁵

Mesenchymal stem cells (MSCs) are stem cells with high plasticity, which include bone marrow stem cells (BMSCs) and adipose-derived mesenchymal stem cells (ADSCs). They are capable of self-renewal and differentiation into bone cells, adipose cells and chondrocytes.³⁴⁷ As recently discovered, many circRNAs are related to MSCs adipogenesis.^{336, 337, 348} For instance, circFOXP1 sponges miR-17-3p/miR-127-5p and shows indirect regulation of EGFR and noncanonical Wnt pathways, thus maintaining the MSC stemness.³⁴⁹ CircRNA CDR1as silencing can promote osteogenic differentiation of BMSCs and suppress their adipogenic differentiation, whereas CDR1as overexpression has opposite effects. Mechanistically, CDR1as competitively binds miR-7-5p to WNT5B for enhancing adipogenic differentiation.³³⁷ CircRNA–vgll3, through the direct sequestering of cytoplasmic miR-326-5p and inhibition of miR-326-5p activity, promotes osteogenic differentiation of ADSCs.³³⁶ miR-338-3p can upregulate circRNA_013422 and circRNA_22566 in osteogenic differentiation of mouse ADSCs.³³⁸ miR-338-3p targets Fgfr2 and Runx2 to modulate osteogenic differentiation of mouse bone marrow stromal stem cells.³³⁹ CircRNA_23525 targets miR-30a-3p to modulate Runx2 expression, which has been detected as the active regulatory factor for the osteogenic differentiation of ADSCs.³⁴⁰

CircRNAs and liver lipid metabolism

The liver exerts an essential effect on various metabolic processes, including lipid digestion, absorption, production, transportation, and decomposition. Drink abuse, obesity, and diabetes can induce the liver lipid metabolic disturbance. Excess liver lipid accumulation produces massive reactive oxygen species, which causes mitochondrial dysfunction and endoplasmic reticulum stress within hepatocytes, ultimately inducing NAFLD.^350-352 Numerous circRNAs are suggested to regulate lipid metabolism and influence lipid disorder occurrence.^{345, 353}

A high-fat diet in the mother can change liver miRNAs expression in her offspring and impair the metabolic health.³⁵⁴ Current research has clarified a novel mechanism underlying the role of maternal obesity in liver lipid metabolism of offspring. Through liver RNA sequencing on offspring of high-fat-diet-fed C57BL/6J mice, Chen et al.³⁴¹ detected 231 DEcircRNAs, including 121 with upregulation whereas 110 with downregulation. circRNA_0000660 expression is significantly related to Igfbp1 expression, while circRNA_0000660 silencing decreases liver lipid accumulation. Maternal obesity induces metabolic disorders in offspring by changing liver circRNA expression. CircHIPK3 modulates lipid metabolism within hepatocytes. It promotes triglyceride level and adipose deposition within HepG2 cells, and induces diabetes-associated metabolic diseases like insulin resistance and hyperglycemia by decreasing miR-192-5p and increasing its downstream TF FOXO1.²³³

CircRNAs are widely suggested with vital effects on regulating hepatic steatosis^{355, 356} (Figure 5). miR-34a can bind to PPARα in rodents, which suppresses lipid metabolism and enhances hepatic steatosis, while circRNA_0046366 and circRNA_0046367 compete against PPARα for binding to miR-34a. These circRNAs act as endogenous regulators of miR-34a, reducing hepatic steatosis by suppressing the interaction of miR-34a with MRE within PPARα mRNA. Aberrant circRNA_0046366 and circRNA_0046367/miR-34a/PPARα signaling is the potential novel target for treating hepatic steatosis.^{342, 343} In high-fat-diet-mediated hepatic steatosis, circRNA_021412 expression significantly decreases.³⁴⁴ circRNA_021412 downregulation suppresses the competitive inhibition on miR-1972, therefore repressing Lpin1, the target gene of miR-1972. Lpin1 downregulation can downregulate long-chain acyl CoA synthase, causing hepatic steatosis.

NAFLD is the condition of excess fat accumulation within hepatocytes not associated with alcohol abuse. Many NAFLD patients experience nonalcoholic steatohepatitis (NASH).³⁵⁷ Aberrant circSCD1 expression influences hepatocyte lipidation and enhances fatty liver disease via JAK2/STAT5 pathway.³⁴⁵ Jin et al.³⁵³ established four circRNA–miRNA–mRNA pathways using the NASH mouse model. Their results showed that circRNA expression patterns might be employed to diagnose NASH.

Role as biomarkers

Huang et al.³⁸³ found that hsa_circ_0116795 was significantly upregulated in the peripheral blood of AMI patients, and its expression increased in cardiomyocytes under hypoxic conditions, suggesting it may be used as a biomarker for early diagnosis of AMI. Liu et al.’s study³⁸⁴ indicated that the expression of circPRDM5 in the serum of AMI patients was significantly reduced. The sensitivity of AMI diagnosis could be significantly improved when circPRDM5 was combined with cardiac troponin T and creatine kinase-MB.³⁸⁴ Tian et al.,³⁸⁵ while exploring biomarkers of sudden cardiac death (SCD) caused by acute ischemic heart disease (IHD), found time-dependent expression patterns of circSLC8A1 and circNFIX in animal and cellular models that matched findings from autopsy cases. This suggests their potential value in distinguishing IHD-related SCD. In particular, the high sensitivity and specificity of circSLC8A1, as well as its positive correlation with creatine kinase-MB in pericardial fluid, and downregulated circNFIX levels may reflect ischemic myocardial injury, which is negatively correlated with the degree of coronary artery stenosis.³⁸⁵ The study by Tong et al.³⁸⁶ analyzed microarray data of exocircRNA and mRNA in AMI from CHD patients. The constructed ceRNA network analysis revealed specific circRNAs (such as circRNA0001785, circRNA0000973, circRNA0001741, and circRNA0003922) as promising candidates for effective prediction of CHD.³⁸⁶

CircRNA promotes the progress of ACS

MI and I/R injury are the most common and fatal manifestations of ACS. Increasing evidence shows that circRNAs play an important role in the occurrence and development of ACS.

First, circRNAs affect cardiac fibrosis after MI by regulating the proliferation and migration of CFs. Pang et al.³⁸⁷ showed that circHelz can bind to Yes-associated protein (YAP) and promote its localization to the nucleus, thereby promoting cell proliferation and growth. Reducing circHelz expression alleviates cardiac fibrosis and the activation of CFs.³⁸⁷ Wang et al.³⁸⁸ pointed out that exosome circUbe3a released by M2-type macrophages promotes the proliferation and migration of CFs by adsorbing miR-138-5p and inhibiting the translation of Rho family GTPase 3 (RhoC). Deng et al.³⁸⁹ revealed that circ-HIPK3 interacts with miR-17-3p and adenylate cyclase type 6 (ADCY6) to regulate calcium ion concentration in cardiomyocytes and affect cardiac function. Reducing its level can reduce cardiac fibrosis and maintain cardiac function after MI.³⁸⁹ These studies suggest that regulating the expression of specific circRNAs may effectively reduce cardiac fibrosis after MI, providing a new strategy for the treatment of CVDs.

Second, circRNAs affect MI and I/R injury by regulating inflammatory responses and immunomodulatory processes. For example, Huang et al. found that circ_SMG6 binds to miR-138-5p, upregulates early growth response 1 (EGR1) expression, promotes neutrophil recruitment, and exacerbates myocardial injury.³⁹⁰ Additionally, Zhu et al.³⁹¹ showed that the expression of circITGB1 was increased in the plasma of AMI patients, which increased the expression of NFAT activating molecule 1 (NFAM1) by competitively binding to miR-342-3p and affected the maturation of DCs, highlighting the role of circRNA in inflammatory response and immune regulation. Ye et al.³⁹² pointed out that the serum concentrations of IL-1β and IL-18 were increased, circ-NNT and ubiquitin specific peptidase 46 (USP46) were upregulated, while miR-33a-5p was downregulated in MI patients. By regulating USP46, circ-NNT promoted pyroptosis in myocardial I/R injury.³⁹²

Notably, circRNAs also act by affecting specific signaling pathways. Zhang et al.³⁹³ found that circ_0030235 activated the PI3K/AKT and MEK/ERK signaling pathways by regulating miR-526b to promote injury in MI models using an oxygen glucose deprivation/reperfusion (OGD/R) model. Zhu et al.³⁹⁴ found that circMAT2B increased OGD-induced cell damage and inflammatory damage by positively regulating miR-133 and activating the PI3K/AKT and Raf/MEK/ERK signaling pathways in MI. Li et al.³⁹⁵ found that inhibition of circARPA1 could effectively reduce myocardial cell fibrosis and apoptosis, emphasizing the importance of the Wnt/β-catenin signaling pathway in this process.

In addition, numerous studies have revealed that circRNAs aggravate myocardial injury. For example, circROBO2 can act as a sponge for miR-1184, promoting cardiomyocyte apoptosis by upregulating the expression of TNF receptor associated death domain.³⁹⁶ Circ_0049271 aggravated myocardial cell damage through the miR-17-3p/FZD4 signaling axis in hypoxia-reoxygenation induced myocardial cell damage.³⁹⁷ The expression of circHSPG2 and MAPK kinase kinase 2 (MAP3K2) is increased in the serum of AMI patients, while the expression of miR-1184 is decreased. Downregulation of circHSPG2 can reduce myocardial cell injury.³⁹⁸ Circ_0023461 is upregulated under hypoxic conditions and upregulates the expression of phosphodiesterase 4D (PDE4D) by acting on miR-370-3p, thereby regulating the viability and glycolytic ability of AC16 cardiomyocytes.³⁹⁹ CircRbms1 promotes myocardial cell injury by regulating the miR-742-3p/forkhead box O1 (FOXO1) axis in the progression of MI.⁴⁰⁰ Circ_0124644 increases myocardial cell injury in AMI by targeting miR-590-3p and affecting the expression of SRY-box TF 4.⁴⁰¹ Circ_0007059 is upregulated during MI and promotes inflammation and cardiomyocyte apoptosis by acting as a sponge for miR-378 and miR-383.⁴⁰² Under hypoxia/reoxygenation (H/R) conditions, the expression of circ-CBFB in cardiomyocytes increases, promoting apoptosis by adsorbing miR-495-3p, revealing the role of circ-CBFB in H/R injury.⁴⁰³

In summary, circRNAs play a key role in cardiac fibrosis, myocardial apoptosis, inflammatory response, and myocardial I/R injury after MI by regulating the proliferation, migration, myocardial apoptosis, inflammatory response, and immunomodulatory processes of CFs, as well as specific signaling pathways. These studies not only deepen our understanding of the pathogenesis of CVDs but also provide important clues for the development of new therapeutic strategies.

CircRNA inhibits the progress of ACS

By acting as a “molecular sponge” for miRNA, circRNAs can affect the activity of miRNA and regulate the expression of downstream target genes. This mechanism also plays a key inhibitory role in the occurrence and development of ACS.

CircRNAs play an important role in alleviating myocardial hypoxia. For example, Zhao et al.⁴⁰⁴ revealed the protective effect of circHSPG2 on cardiomyocytes under hypoxic conditions. By acting as a molecular sponge for miR-25-3p, circHSPG2 positively regulates the miR-25-3p/proapoptotic WT1 regulator axis and protects cardiomyocytes from hypoxia-induced damage.⁴⁰⁴ Similarly, Zhou et al.⁴⁰⁵ found that circ_0001747 in ADSC exosomes alleviated I/R injury in cardiomyocytes by targeting miR-199b-3p. Gao et al.⁴⁰⁶ found that H/R-induced cardiomyocyte death could be alleviated by enhancing the expression of mmu_circ_000338 in H/R-treated and I/R-injured mouse hearts. By directly binding to histone deacetylase 7 (HDAC7) and interfering with its entry into the nucleus, mmu_circ_000338 alleviated the inhibition of HDAC7 on forkhead box protein A2 (FOXA2) transcription, thereby upregulating FOXA2 and inhibiting receptor-interacting protein kinase 3-dependent necrosis.⁴⁰⁶ Zhao et al.⁴⁰⁷ investigated hypoxic conditions in vitro and found that circRNA_010567 enhanced mitochondrial activity by adsorbing miR-141 and increasing death-associated protein kinase 1 expression, thus playing a protective role in hypoxia-induced myocardial cell damage. Liu et al.⁴⁰⁸ revealed the protective role of circ_0002612 in I/R injury. Circ_0002612 releases the expression of its target peroxisome proliferator-activated receptor gamma coactivator 1 alpha (Ppargc1a) by competitively binding to miR-30a-5p, thereby inhibiting the expression of NLRP3 and contributing to the survival of cardiomyocytes and reducing apoptosis.⁴⁰⁸

In addition, circRNAs are also involved in the process of cardiac repair and remodeling. These studies reveal the important role of circRNA in the pathological process of the heart and provide a new target for the treatment of heart diseases. For example, Wang et al.⁴⁰⁹ found that circHIPK3 could promote cardiac endothelial cell cycle progression and proliferation by reducing miR-29a activity and increasing vascular endothelial growth factor A (VEGFA) expression, thus alleviating damage after MI. Garikipati et al.⁴¹⁰ showed that overexpression of circFndc3b, which was downregulated after MI, promoted angiogenesis, reduced cardiomyocyte apoptosis, and improved cardiac function. Ju et al.⁴¹¹ found that overexpression of circRNA FEACR could inhibit MI and improve cardiac function by directly binding to nicotinamide phosphoribosyltransferase (NAMPT), improving NAMPT protein stability, enhancing SIRT1 expression, and promoting forkhead box protein O1 (FOXO1) transcriptional activity, thereby upregulating the transcription of the ferroptosis inhibitor Fth1. Li et al.⁴¹² revealed that regulating the circCELF1/miR-636/Dickkopf WNT signaling pathway inhibitor 2 (DKK2) pathway can reduce myocardial fibrosis and improve cardiac function. Zhu et al.⁴¹³ showed that circNFIB expression was reduced after MI, and its overexpression could reduce the proliferation of CFs by adsorbing miR-433. Zhou et al.⁴¹⁴ found that circRNA ACR activated (Pink1) expression and reduced myocardial infarct size by blocking DNA methyltransferase 3B (DNMT3B)-mediated DNA methylation of the Pink1 promoter. Sun et al.⁴¹⁵ explored the role of circFoxo3 in MI and found that overexpression of circFoxo3 improved cardiac dysfunction caused by MI. Wang et al.⁴¹⁶ focused on the role of mitochondrial fission and apoptosis-related circRNA (MFACR) in cardiac cell death and pointed out that MFACR prevented mitochondrial protein 18 kDa (MTP18) translation by inhibiting miR-652-3p and reducing cardiac cell death. Zhao et al.⁴¹⁷ found that circMACF1 could act as a “sponge” of miR-500b-5p to upregulate the expression of EMP1 and slow down the progression of AMI. Notably, Wei et al.⁴¹⁸ found that circWhsc1 in EVs produced by endothelial cells under hypoxic conditions promoted cardiomyocyte proliferation and cardiac regeneration. This process enhances tripartite motif containing 59 (TRIM59) phosphorylation, thereby increasing TRIM59 binding to signal transducer and activator of transcription 3 (STAT3) and promoting STAT3 phosphorylation, which ultimately leads to cardiomyocyte proliferation, potentially providing a therapeutic target for cardiac repair after MI.⁴¹⁸ Yu et al.⁴¹⁹ showed that circCEBPZOS levels were reduced in the serum of patients with poor cardiac remodeling. Overexpression of circCEBPZOS can improve cardiac function by regulating the miR-1178-3p/phosphoinositide-dependent kinase-1 (PDPK1) axis, thereby reducing adverse remodeling and promoting angiogenesis after MI.

In terms of enhancing the antioxidant capacity of cardiomyocytes, studies have shown that circRNAs are involved in regulation through different mechanisms. For example, Fan et al.⁴²⁰ found that H₂O₂-induced cardiomyocyte injury could be alleviated by increasing circ_HIPK3 expression. The mechanism is that circ_HIPK3 upregulates the expression of its target IRS1 by adsorbing miR-33a-5p, thereby improving cell survival.⁴²⁰ Chen et al.⁴²¹ found that elevated circ-SWT1 expression attenuated H₂O₂-induced cardiomyocyte injury in an AMI model. Circ-SWT1 interacts with miR-192-5p to reduce the inhibition of SOD2 by miR-192-5p, thereby enhancing the antioxidant capacity of cells.⁴²¹

These research results demonstrate the multifaceted role of circRNAs in MI and its subsequent cardiac pathological processes, involving not only the protection and repair of cardiomyocytes but also the regulation of cardiomyocyte death. These findings advance our understanding of the molecular mechanisms of CVDs and provide possibilities for developing new therapeutic strategies.

Cell proliferation

Several reports indicate that exocircRNAs enhance cellular proliferation. Human hepatic epithelial cells exposed to arsenite release circRNA 100284, which is then conveyed via exosomes to surrounding normal cells. By acting as a miR-271 sponge and upregulating EZH2 expression, exocircRNA_100284 accelerates the cell cycle and enhances cellular proliferation.⁶²³ A circRNA called exosomal CDR1-AS promotes the onset and progression of HCC. By acting as a sponge for miR-1270 and increasing alpha-fetoprotein expression in HCC cells, the upregulation of CDR1-AS enhances cell proliferation and motility. Exosomes carrying CDR1-AS are then delivered to nearby cells, stimulating their migration and proliferation.⁶²⁴ Circ0000284 is found in plasma exosomes, malignant tissues, and cholangiocarcinoma cell lines. Circ0000284 upregulates LY6E expression and encourages cholangiocarcinoma cell proliferation, migration, and invasion by acting as a sponge for miR-637 in cells. Circ0000284, released into exosomes, is then delivered to surrounding cells, promoting their proliferation and motility.⁶²⁵

ExocircRNAs produced by healthy cells inhibit proliferation in recipient cells, in contrast to those produced by tumors, which increase it. For example, circ0051443, downregulated in HCC patients’ plasma exosomes, is released by healthy hepatic cells and transported to adjacent HCC cells, where it increases BAK1 expression and sponges miR331-3p, resulting in cell cycle arrest and apoptosis in tumor cells.⁶²⁶

EMT, invasion, and metastasis

Intercellular communication mediated by exosomes has been shown to facilitate the formation of prometastatic niches, induce EMT, and alter the TME, all of which contribute to tumor invasion and metastasis.^627-629 Recent research implicates circRNAs in exosomes in the invasion and metastasis of tumors. Pancreatic cancer cells release circPDE8A, and overexpression of exosomal circPDE8A in patient plasma is linked to a poorer prognosis. In pancreatic cancer cells, circPDE8A acts as a miR-338 sponge, activating the metastasis-associated in colon cancer (MACC)/MET/ERK or AKT pathways and promoting invasive development. Additionally, by stimulating the MACC/MET/ERK or AKT pathway, exosomal circPDE8A promotes EMT in recipient cells.⁶³⁰ CircRanGAP1 is upregulated in plasma exosomes from preoperative GC patients compared with those from healthy controls and postoperative GC patients. More significantly, the plasma exosomes from these individuals increased the migratory and invasive rate of GC cells, suggesting that exosomal circRanGAP1 may perform a role in the onset and advancement of GC.⁶³¹ As opposed to nonmetastatic or low-metastatic HCC cells, exosomes from highly metastatic HCC cells have a higher concentration of circPTGR1. The miR-449a/MET pathway, activated by coculture with exosomal circPTGR1, promotes HCC cell invasion and motility. Additionally, patients with HCC have a poorer prognosis and more advanced clinical stages when circPTGR1 expression is high in serum exosomes.⁶³²

Exosoma CircRNA may facilitate the invasion of cancerous cells not only by increasing their invasiveness but also by changing the TME. HUVECs receive circIARS via exosomes secreted from pancreatic cancer cells. Pancreatic cancer cells may more readily invade HUVECs due to the presence of exosomal circIARS, which increases their permeability by upregulating RhoA and sponging miR-122.⁶³³ ExocircRNA-100338, which is increased in an HCC cell line with a high level of metastasis, also enhances the ability of HUVECs to permeate, proliferate, and undergo angiogenesis. CircRNA-100338 may bind RBPs such as NOVA2, which is known to regulate vascular growth and lumen development, as shown by RNA pull-down analysis. These findings suggest that circRNA-100338 could regulate angiogenesis by interacting with NOVA2, but further research is needed to confirm these hypotheses.⁶³

Multiple reports indicate that various circRNAs are implicated in metastasis. CircPRMT5, commonly upregulated in urothelial carcinoma of the bladder (UCB) cells, acts as a miR-30c sponge and modulates the SNAIL1/E-cadherin pathway, stimulating EMT in UCB cells. Additionally, the presence of circPRMT5 in serum and urine exosomes has been positively associated with lymph node metastasis (LNM) and advanced tumor progression.⁵⁴⁸ OC tissues exhibit elevated levels of circWHSC1, which upregulates hTERT and MUC1, acts as a miR-1182 and miR-145 sponge, and enhances cell proliferation, motility, and invasiveness. Exosomal circWHSC1 upregulates N-cadherin and MUC1, causing peritoneal mesothelial cells to adopt a fibroblast-like morphology. Additionally, OC cells, when injected intraperitoneally into mice along with exosomes from cells overexpressing circWHSC1, enhance peritoneal dissemination. This suggests that exosomal circWHSC1 may be transported to peritoneal mesothelial cells, facilitating peritoneal dissemination in vivo.⁶³⁴

Chemoresistance

CircRNAs in exosomes are associated with chemoresistance. In a microarray analysis of exosomal RNAs from chemosensitive and chemoresistant CRC cells, the expression levels of 105 circRNAs were significantly upregulated, while those of 34 circRNAs were downregulated in exosomes from chemoresistant CRC cells. Research suggests that hsa_circ_0032883 could be a potential biomarker for chemotherapy response, as it was found to be significantly elevated in the serum exosomes of chemosensitive patients compared with those who are chemoresistant.⁶³⁵ Another study identified a positive correlation between ciRS-122 and chemoresistance in CRC. ciRS-122 enhances the expression of PKM2 and acts as a miR-122 sponge. ciRS-122, transferred from chemoresistant cells to recipient cells, promotes glycolysis, and reduces drug sensitivity.⁶³⁶

Adipocyte browning

CircRNAs may be transported into adipose tissues by exosomes, potentially causing cancer cachexia in addition to facilitating tumor cells' intracellular crosstalk. Unlike cancer without cachexia, which typically has a better prognosis, cancer cachexia is characterized by an uncontrollable decrease in total body weight. Reports indicate that cancer cachexia is driven by abnormally increased thermogenesis due to adipocyte browning.^{637, 638} These findings suggest that exocircRNAs from tumors may contribute to the development of cancer cachexia.

ExocircRNA's potential in clinical settings

CircRNAs can be found in serum,²⁶ urine,⁶⁶ and saliva,¹⁰⁸ with their tertiary structure contributing to their stability. Additionally, various circRNAs are dysregulated in different malignancies.⁶³⁹ To date, circRNAs have shown promise as noninvasive biomarkers for tumor detection. Recent evidence using high-throughput sequencing indicates that various exocircRNAs have potential as diagnostic biomarkers for cancers. Circ-0051443 was shown to be considerably reduced in exosomes generated from the plasma of patients with HCC compared with healthy controls utilizing microarray analysis. The diagnostic potential of exosomal circ-0051443 was evaluated based on plasma exosomes from 60 HCC patients and 60 healthy individuals. An AUC value of 0.8089 for exosomal circ-0051443 indicated moderate diagnostic accuracy.⁶²⁶ Hsa_circ_0000419, circ-KIAA, and hsa_circ_0065149 exhibited lower levels in plasma exosomes from patients with GC than in those from healthy individuals. AUC values of 0.84, 0.75, and 0.64, respectively, for these circRNAs indicated reasonable diagnostic accuracy.^{78, 640, 641} In a circRNA microarray analysis of three pairs of esophageal squamous cell carcinoma (ESCC) and nontumor tissues, 1045 circRNAs were found to be upregulated and 1032 downregulated in ESCC. Both hsa_circ_0043603 and hsa_circ_0001946 were found to be significantly downregulated in plasma from 50 ESCC patients compared with 50 controls. When used together as a diagnostic tool, these two circRNAs achieved a specificity of 98% and a sensitivity of 84%.⁶⁴² Exosomal hsa-circ-0004771 levels in the circulation were shown to be considerably elevated in the serum samples of 110 CRC patients in contrast to 35 patients with benign intestinal diseases and 35 healthy controls. Exosomal hsa-circ-0004771 in the circulation demonstrated a moderate ability to distinguish between stage I/II CRC patients and healthy controls, with AUC values of 0.86 and 0.88, respectively.⁶⁴³ Poor prognosis and clinical response to treatment in small-cell lung cancer were linked to serum exosomal FECR1.⁶⁴⁴ CircPDAC was elevated and detectable in the sera of pancreatic cancer patients (specificity 0.90, sensitivity 0.45). Additionally, circPDAC was also detected in the sera of patients with intraductal papillary mucinous neoplasm, a known precancerous lesion.⁶⁴⁵

Biological properties and activity of exocircRNAs in NSCLC

Continuous proliferation and evasion of apoptosis are characteristic changes in cancer cells.⁶⁴⁷ The specific biological implications of numerous circRNAs on NSCLC are yet unknown. ExocircRNAs are found to be abnormally expressed in body fluids, tissues, plasma, and serum in cases of NSCLC.⁶⁴⁸ This review outlines the research conducted to understand how exocircRNAs are related to NSCLC (Table 3). In recent years, exocircRNAs have been linked to the progression of NSCLC, specifically to proliferation, apoptosis, invasion, metastasis, angiogenesis, glycolysis in NSCLC cells, recurrence, mechanisms of treatment resistance, EMT in NSCLC cells, and mortality, by regulating their target miRNAs and downstream tumor-associated signaling pathways^{649, 650} (Figure 8). Table 3 presents the positive correlation between several upregulated circRNAs and NSCLC progression. Additionally, the downregulation of hsa_circ_0102537 and circ_0007761 inhibits the progression of NSCLC.⁶⁵¹

ExocircRNAs target specific organs and cells by transporting exosomes, which subsequently play a role in controlling the proliferative, apoptotic, invasive, and metastatic processes of NSCLC cells. CircRNA_102481 promotes the proliferation of cells sensitive to EGFR-TKIs while inhibiting their apoptosis by acting as a sponge for miR-30a-5p and downregulating ROR1 expression.⁶⁵²

Another study revealed an increase in hsa_circ_0008717 expression in serum exosomes from lung adenocarcinoma (LUAD) cases. Suppression of hsa_circ_0008717 decreased malignant behaviors in LUAD cells, arrested the cell cycle, and induced apoptosis, while repressing their proliferation, migration, and invasiveness, by modulating the miR-1287-5p/PAK2 axis.⁶⁵³

Similarly, circDNER expression is increased in exosomes from LUAD cells and cases. CircDNER acts as a sponge for miR-139-5p, while ITGB8 is the target of miR-139-5p. Consequently, circDNER enhances the ability of LUAD cells to proliferate and migrate by modulating the miR-139-5p/ITGB8 axis.⁶⁵⁹ Circ-CPA4 is also involved in transport within LUAD cells. As discovered by Hong and colleagues, let-7 is sponged by circ-CPA4, while PD-L1 is a target of let-7. Mechanistically, circ-CPA4 knockdown inhibited in vitro LUAD cell growth and suppressed in vivo LUAD tumor growth via the let-7/PD-L1 axis.⁶⁵⁰

Hsa_circ_0001346, hsa_circ_0000690, and hsa_circ_0001439 were found to be increased in lung cancer, presenting potential novel diagnostic and treatment biomarkers.⁶⁵⁸ Ning et al.⁶⁶² discovered that overexpression of hsa_circ_0007385 enhances NSCLC cell proliferation and stemness via the miR-1253/FAM83A signaling pathway. Circ_0048856 levels are increased in NSCLC cells and tissues, promoting proliferation, migration, and invasion, while inhibiting apoptosis in NSCLC cells.⁶⁶⁰ Additionally, hsa_circ_0102537 was predominantly packaged in exosomes, but less commonly found in LUAD tissues and plasma compared with normal subjects.⁶⁵¹

ExocircRNAs target signaling pathways in NSCLC

An expanding volume of studies demonstrates that exocircRNAs contribute to the progression of NSCLC by influencing cancer-related signaling pathways, specifically the PI3K/AKT and MAPK pathways. The PI3K/AKT signaling pathway holds significant importance in controlling diverse cellular biological processes such as cell cycle, glucose transport, and the development of cancer.⁶⁸² Aberrations in the PI3K/AKT pathway are tightly linked to the advancement of NSCLC.⁶⁸³ Multiple exocircRNAs have been pinpointed as modulators of NSCLC tumorigenesis through their interaction with the PI3K/AKT signaling cascade. As an illustration, Chen and colleagues⁶⁴⁹ uncovered that exosomal circFARSA stimulates the PI3K/AKT pathway, bolstering the growth of NSCLC cell lines, specifically A549 and PC9. The MAPK signaling pathway is renowned for its role in cancer development. Disturbances in the MAPK/ERK cascade contribute significantly to various facets of cancer progression.⁶⁸⁴ Wang et al.²⁵ revealed that circ-ZKSCAN1 hinders the proliferation of NSCLC cells. This mechanism involves the upregulation of FAM83A expression by sequestering miR-330-5p, thereby suppressing the MAPK signaling pathway.²⁵ Furthermore, reports indicate that hsa_circ_0004050 downregulates DUSP9 expression by sponging miR-1233-3p in the NSCLC cell line A549, thus impeding the ERK/JNK signaling cascade.⁶⁸⁵ Our investigations have demonstrated elevated levels of circ-ZKSCAN1 in both NSCLC tissues and cell lines. This upregulation leads to the suppression of the MAPK signaling pathway by absorbing miR-330-5p, resulting in increased FAM83A expression and consequently, augmented NSCLC cell growth.²⁵ Taken together, these findings underscore the pivotal role of circRNAs in regulating NSCLC signaling cascades (Figure 9).

Relations of exocircRNAs with NSCLC clinicopathological features

Presently, due to the intricate pathological traits of NSCLC, detecting its early symptoms is challenging. Consequently, many patients receive a diagnosis only when the disease has progressed to an advanced stage, thereby missing the optimal window for surgical intervention. Furthermore, the inadequate prognosis evaluation significantly impedes the customization of treatment plans for individual patients and efforts to extend their survival. In the past few years, several traditional biomarkers, including AFP, PD-L1, and ALK, have been utilized for both the diagnosis and prognosis assessment of various cancers, NSCLC being one of them. Table 4 presents the relations of exocircRNAs with clinicopathological features in NSCLC cases. Wang and colleagues explored the relation of circ_0008717 level with clinicopathological features in 24 LUAD cases, revealing that LUAD cases who had circ_0008717 upregulation showed an advanced TNM and a large tumor size compared with those showing circ_0008717 downregulation.⁶⁵³ According to Hong et al.,⁶⁵⁰ exosomal circ-CAP4 upregulation was tightly associated with TNM and extra-lung metastasis (EM). As discovered by Xian et al.,⁶⁵⁶ upregulation of the serum markers hsa_circ_0025580 and hsa_circ_0014235 in LUAD patients was strongly linked to tumor sizes and TNM stages, which was not the case for sex, age, or smoking status. Moreover, exosomal circ_0047921 and circ_0056285 expression significantly increased among LUAD cases except circ_0007761 was decreased, which showed a positive relation to family history of cancer, but not with other clinical features.⁶⁵⁶ Furthermore, hsa_circ_0026403, hsa_circ_0025580, and hsa_circ_0014235 expression was also related to TNM stage and tumor sizes, rather than to sex, age, and smoking status.⁶⁵⁷ Moreover, hsa_circ_0069313 and Circ-MEMO1 upregulation showed a positive relation to TNM, LNM as well as EM.^{661, 664} The expression of hsa_circ_0056616 showed a positive association with tumor size, TNM stages, and EM, which was not the case for age and sex.⁶⁶⁹ Based on the above evidence, tumor size, and TNM stage exhibited the highest association with aberrant exocircRNA expression whereas age, gender, number of tumors, smoking status, and LNM presented no significant relationship.

ExocircRNAs’ significance in diagnosing and predicting the prognosis of NSCLC

Although great progresses have been made in surgical, locoregional, and medical treatments, NSCLC still has a high mortality rate due to disease metastasis and relapse postoperatively.⁶⁸⁶ NSCLC-derived exosomes are found to redirect cancer cell metastasis in cells with no specific organ metastatic capacity.⁶⁵⁰ Exosomes can shuttle circRNAs between cells while regulating tissue genesis and cell differentiation. ExocircRNAs have been implicated in recent research as candidate prognostic and diagnostic biological markers for NSCLC.^{651, 655, 661} Table 5 presents the latest research on the involvement of exocircRNAs in diagnosing and predicting the prognosis of NSCLC.

In NSCLC individuals, there is an increase in the levels of hsa_circ_0025580 and hsa_circ_0014235, which may have an impact on NSCLC progression⁶⁵⁷ According to Wang and colleagues,⁶⁵⁷ hsa_circ_0014235 was a new marker to diagnose NSCLC, and its AUC value of the ROC curve value was 0.8254 (95 CI%: 0.762–0.889) to discriminate NSCLC.⁶⁵⁷ Moreover, hsa_circ_0025580 had 0.8003 (95 CI%: 0.741–0.862).⁶⁵⁷ According to He and coworkers,⁶⁶⁰ circ_0048856 showed upregulation in plasma-derived exosomes with better diagnostic ability in NSCLC (AUC = 0.943, and the sensitivity and specificity were 88 and 80%, respectively). In addition, hsa_circ_0069313, circRNA-002178, and circSHKBP1 have the ROC values of 0.724, 0.9967, and 0.853, respectively, to distinguish NSCLC from health, respectively.^{655, 660, 661} Moreover, Zhang et al.⁶⁷¹ reported that circSATB2 was mostly released via exosomes in NSCLC cells, with significant upregulation in NSCLC cells and tissues compared with normal lung cells and para-carcinoma tissues. According to the ROC curve, circSATB2 displayed great significance in diagnosing NSCLC, and the AUC was 0.66 (95 CI%: 0.581–0.74).⁶⁷¹ As suggested by Ding and colleagues,⁶⁶⁴ circMEMO1 upregulation in nearby tissues predicted dismal survival for NSCLC cases.

Nevertheless, further study is required before circRNAs can be used in clinical settings since their function in NSCLC is still relatively unknown. Moreover, a large number of samples, multiple centers, and independent external cohort validation are required for circRNA to become a widely accepted biomarker. Besides, there are only a very few cirRNAs that have been found to have an impact on tumors. In addition, only a limited number of cirRNAs have been identified in NSCLC, and further research is required.

Exosome-derived circRNA promotes HCC

Numerous studies have provided compelling evidence of the significant involvement of exosome-derived circRNAs in advancing the progression of HCC. These specific circRNAs exert a substantial influence on the biological characteristics of HCC cells, including proliferation, metastasis, and drug resistance, thus actively fostering the advancement of HCC.

First, exocircRNAs derived from HCC cells can function as miRNA “sponges,” adsorbing miRNAs and thereby contributing to HCC progression. For example, circWHSC1 absorbs miR-142-3p and promotes proliferation and metastasis of HCC cells by directly targeting homeobox A1 (HOXA1), suggesting that circWHSC1 may become a diagnostic indicator of HCC.⁶⁹³ Circ_MMP2 interacts with miR-136-5p and MMP2, leading to the upregulation of MMP2 and the promotion of HCC cell progression. Elevated MMP2 levels are associated with HCC metastasis and poor overall survival in HCC patients.⁶⁹⁴ Similarly, circPTGR1 promotes cancer cell metastasis by competing for miR449a binding to MET.⁶³² CircRNA Cdr1as promotes AFP expression by adsorbing miR-1270.⁶²⁴ Circ_002136 augments Ras-related expression by blocking miR-19a-3p, increasing Ras-related protein Rab-1A (RAB1A) expression and promotes HCC cell progression.⁶⁹⁰ CircTGFBR2 enhances ATG5-mediated cytoprotective autophagy by interfering with the interaction between miR-205-5p and autophagy-related protein 5 (ATG5), thereby promoting HCC progression.⁶⁸⁹ Notably, in arsenide-induced HCC, circRNA_100284 can be transferred into normal hepatocytes and induce cell cycle acceleration, cell proliferation, and malignant transformation of normal hepatocytes by acting as a sponge for miRNA-217, suggesting that circRNAs have the potential to induce malignant transformation in normal cells.⁶²³

Second, exocircRNAs may mediate drug resistance in HCC cells. For example, exosome-derived circZFR from cancer-associated fibroblasts (CAFs) is highly expressed in cisplatin (DDP)-resistant HCC cell lines and can inhibit the STAT3/NF-κB pathway, promote tumor growth and enhance DDP resistance. Targeting circZFR and related pathways may help improve the efficacy and survival of patients treated with chemotherapy for HCC.⁶⁹⁵ Exosome-derived circRNA-SORE from HCC cells is frequently upregulated in sorafenib-resistant HCC cells, and further studies have shown that circRNA-SORE stabilizes the cytoplasmic oncogenic protein YBX1, which is a key component in the development of HCC. Silencing of circRNA-SORE by siRNA could overcome sorafenib resistance in HCC patients.⁶⁹⁶ CircPAK1 derived from HCC cell exosomes promotes HCC cell proliferation, invasion, and metastasis, while inhibiting apoptosis and angiogenesis. Further studies showed that exosomes from ravastatinib-resistant HCC cells mediate the transport of circPAK1, contributing to ravastatinib resistance.⁶⁹⁷

In addition, HCC-derived exocircRNAs can inhibit immune cells in the TME and promote angiogenesis to exert procancer effects. For example, circUHRF1 promotes tumor lung metastasis and reduces the number of natural killer group 2, member D (NKG2D)-positive cells in metastatic tumor nodules and contributes to NK cell dysfunction and poor clinical prognosis in HCC patients.⁶⁹⁸ Hsa_circ_0074854 regulates the expression of HCC cells by interacting with HuR protein and regulating its protein expression level, thereby affecting the EMT process in HCC cells. Conversely, exosomes carrying downregulated hsa_circ_0074854 can translocate to macrophages and inhibit their polarization, which may ultimately impede the migration, invasion, and EMT of HCC cells.⁶⁹¹ CircFUT8, whose expression is upregulated in HCC cells, can be modified by m6A, thereby promoting HCC progression. However, M1-type macrophage-derived exosomes can inhibit HCC progression by transfecting miR-628-5p into HCC cells and downregulating m6A modification, suggesting an anticancer role for M1-type macrophage-derived exosomes.⁶⁹⁹ CircGSE1 can maintain the expression of transforming growth factor-beta receptor 1 (TGFBR1) and Smad3, which promotes the expansion of Treg cells and thereby inhibits the activity of effector T cells, such as CD8⁺ T cells, resulting in immune escape from HCC cells.⁷⁰⁰ In addition, circTMEM181 increases CD39 expression on macrophages and inhibits CD8⁺ T cells, leading to resistance to PD-1 treatment in HCC.⁷⁰¹ Similarly, circCCAR1 is able to promote CD8⁺ T cell exhaustion, leading to impaired antitumor immunity in TME. The mechanism is that circCCAR1 prevents the ubiquitinated degradation of the PD1 protein by binding directly to PD1, thereby impairing the tumor killing capacity of CD8⁺ T cells⁷⁰² Notably, in vitro and animal studies have shown that circRNA-100338 promotes the invasion and metastasis of HCC cells and promotes vascular endothelial cell proliferation, angiogenesis, and which correlates with lung metastasis and reduced survival in HCC patients.⁶³

Exosome-derived circRNA inhibits HCC

Conversely, it has also been shown that exosome-derived circRNAs can inhibit HCC progression by adsorbing miRNAs. For example, circ_0051443 can act as a sponge for miR-331-3p, thereby attenuating the regulatory effect of miR-331-3p on BCL2 antagonist/killer 1 (BAK1), which inhibit HCC cell progression.⁶²⁶ Macrophage exosomes overexpressing recombination signal binding proteinJκ (RBPJ) overexpress hsa_circ_0004658, which led to the upregulation of junctional adhesion molecule 3 through the adsorption of miR-499b-5p, thereby inhibiting HCC cell proliferation and promoting apoptosis, suggesting the diagnostic and therapeutic potential of hsa_circ_0004658 as a biomarker and target for HCC cell therapy.⁶⁹²

Taken together, these data demonstrate the important role of exosome-derived circRNAs in promoting/inhibiting HCC development, mediating drug resistance, promoting proliferation, metastasis, and regulating immune escape. These findings provide important clues for understanding the mechanism of exosome-derived circRNAs in the pathophysiology of HCC cells and offer new opportunities for their clinical application as diagnostic and prognostic markers as well as therapeutic targets in HCC cells.

Exosome-derived circRNA promotes CRC

CircRNAs in CRC tissues can block downstream signaling pathways by adsorbing miRNAs, thereby affecting cancer cell growth, metastasis and drug resistance.

First, exosome-derived circRNAs in CRC cells can promote tumor invasion, and metastasis by promoting endothelial cell migration and proangiogenesis. For example, circ_001422 was able to inhibit the expression of miR-195-5p, which in turn promotes the kinase insert domain receptor (KDR)/mTOR signaling pathway and facilitates endothelial cell migration and angiogenesis, suggesting that circ_001422 has a role as a biomarker in the diagnosis of CRC.⁷⁰⁶ CircTUBGCP4 activates AKT through regulation of the miR-146b-3p pathway, thereby promoting vascular endothelial cell migration, lumen formation and tumor metastasis, providing important clues to understand the molecular mechanisms of CRC development.⁷⁰⁷ Second, cancer cell exosome-derived circRNAs can attenuate the sensitivity of CRC cells to radiation, resulting in a reduced effect of radiotherapy. For example, circ_0067835 interacts with miR-296-5p and exerts a procancer effect by regulating the miR-296-5p/insulin-like growth factor 1 receptor (IGF1R) pathway in cancer cells. When downregulated in vivo, circ_0067835 inhibited CRC growth and increased cellular sensitivity to radiation.⁷⁰⁸ In addition, circ_IFT80 can target miR-296-5p and further inhibit musashi1 (MSI1), a gene associated with tumor progression and radiosensitivity, thereby promoting tumorigenesis and reducing cellular sensitivity to radiation.⁷⁰⁹ In addition, exosome-derived circRNAs can also promote tumorigenesis under hypoxic conditions. For example, under hypoxic conditions, CAFs exosomes can secrete circEIF3K, which in turn promotes CRC cell proliferation, invasion and metastasis. The miR-214/PD-L1 pathway plays an important role in this process.⁷¹⁰ The hypoxic CRC cell-derived exosome circ-133 is able to translocate to normal tumor cells and promote migration via the miR-133a/guanine nucleotide exchange factor-H1 (GEF-H1)/RhoA axis, suggesting a difference in metastatic potential between hypoxic and relatively normoxic cancer cells.⁷¹¹

In terms of drug resistance, circRNAs may act as miRNA sponges to mediate CRC resistance to chemotherapeutic agents. For example, circ_0006174-enriched exosomes from doxorubicin (DOX)-resistant CRC cells promoted DOX resistance by binding miR-1205 and upregulating cyclin D2 (CCND2). This finding has important clinical implications for the treatment of CRC patients and is expected to be a future target for diagnosis and therapy.⁷¹² The exosome-derived hsa_circ_0005963 of oxaliplatin-resistant CRC cells enhances cellular resistance to chemotherapeutic agents by adsorbing miR-122, leading to overexpression of the M2 isoform of pyruvate kinase (PKM2), which subsequently increases cellular glycolytic metabolism. Targeting circRNA with siRNA (si-ciRS-122) can inhibit glycolysis and reverse resistance to oxaliplatin.⁶³⁶ In addition, increased expression of circ_0000338 in CRC cells correlates with increased 5-FU resistance. It has been shown that circ_0000338 negatively regulates the expression of miRNAs miR-217 and miR-485-3p by interacting with these two miRNAs, leading to 5-FU resistance in CRC patients.⁷¹³ Notably, some exosome-derived circRNAs are also capable of encoding proteins, thereby mediating drug resistance. For example, circATG4B is expressed at elevated levels in exosomes released from drug-resistant CRC cells and is capable of encoding the protein circATG4B-222aa. CircATG4B-222aa competitively interacts with the membrane vesicle transport protein transmembrane p24-trafficking protein 10 (TMED10) and acts as a drug resistance blocker. This competitive interaction with transmembrane p24-trafficking protein 10 (TMED10), acting as a decoy to prevent TMED10 from binding to autophagy-related gene 4B (ATG4B), an interaction that leads to increased autophagy in CRC cells, thereby inducing the development of chemotherapeutic resistance.⁷¹⁴

Exosome-derived circRNAs inhibit CRC

In contrast, the effect of circRNAs on CRC also has therapeutic potential, and certain exosome-derived circRNAs are also able to inhibit CRC, providing an important basis for clinical practice and the development of therapeutic strategies.

It has been suggested that adsorption of miRNAs by circRNAs may also inhibit CRC. For example, circFNDC3B in CRC cell lines and exosomal samples is able to adsorb miR-937-5p and thus inhibit its function. In contrast, miR-937-5p is known to promote EMT. In animal studies, overexpression or exosomal treatment of circFNDC3B was shown to inhibit the growth, angiogenesis and liver metastasis of CRC transplant tumors.⁷¹⁵ CircEPB41L2 acts as a sponge for miR-215p and miR-9425p in exosomes, and overexpression of circEPB41L2 inhibits CRC cell proliferation, enhances apoptosis, and inhibits migration and invasion, which subsequently blocks the regulation of PTEN/AKT signaling pathway by these two miRNAs, ultimately presenting an inhibitory effect on CRC progression in cells and animal models.⁷¹⁶

In addition, some circRNAs act as protein decoys capable of binding proteins to exert inhibitory effects. For example, the plasma exosome circLPAR1 can be internalized by CRC cells and inhibits METTL3–eukaryotic translation initiation factor 3 subunit h (eIF3h) through binding to eIF3h, thereby inhibiting mRNA translation, which in turn inhibits the expression of the oncogene bromodomain-containing protein 4 (BRD4), thereby suppressing CRC cell proliferation, invasion and migration. Clinical studies have shown that downregulation of circLPAR1 is associated with poor survival in CRC patients. Therefore, circLPAR1 may serve as a potential diagnostic marker for CRC.⁶²

Some circRNAs have oncogenic effects in cancer cells, but tumor cell exosomes excrete them, thus escaping their inhibitory effects and maintaining cancer cell adaptation. An example is the exosome-derived circRHOBTB3 in CRC cells. in response to this situation, researchers have proposed to intervene in the secretion process of circRHOBTB3 through antisense oligonucleotides, thereby increasing its intracellular expression and consequently enhancing its inhibitory effect on CRC.⁷¹⁷

Notably, it has been shown that some circRNAs not only inhibit tumors but also promote resistance to chemotherapeutic agents. For example, hsa_circ_0000338 has a tumor growth inhibitory effect in CRC cells, but its delivery to other CRC cells via exosomes leads to a role in promoting drug resistance. When knockdown of hsa_circ_0000338 was targeted by siRNA, it was shown to enhance drug resistance, suggesting that hsa_circ_0000338 may have a tumor growth inhibitory effect in CRC cells. These findings provide important clues to further understand the function of circRNAs.⁶³⁵ More studies on exosome-derived circRNAs in CRC have been presented in Table 6.

In summary, exosome-derived circRNAs have important roles in the development and treatment of CRC. These studies provide potential therapeutic targets and theoretical bases by deeply exploring the functions of circRNAs in exosomes, which provide important clues for future clinical practice and the development of therapeutic strategies. However, the different effects of circRNAs further suggest that researchers need to consider their mechanisms of action and regulatory networks more comprehensively in order to better understand their roles in disease development and provide more accurate targets and pathways for future therapeutic strategies.