Long Noncoding RNA/Circular RNA-miRNA-mRNA Axes in Ischemia-Reperfusion Injury

Ischemia-reperfusion injury (IRI) elicits tissue injury involved in a wide range of pathologies. Multiple studies have demonstrated that noncoding RNAs (ncRNAs), including long noncoding RNAs (lncRNAs), circular RNAs (circRNAs), and microRNAs (miRNAs), participate in the pathological development of IRI, and they may act as biomarkers, therapeutic targets, or prognostic indicators. Nonetheless, the specific molecular mechanisms of ncRNAs in IRI have not been completely elucidated. Regulatory networks among lncRNAs/circRNAs, miRNAs, and mRNAs have been the focus of attention in recent years. Studies on the underlying molecular mechanisms have contributed to the discovery of therapeutic targets or strategies in IRI. In this review, we comprehensively summarize the current research on the lncRNA/circRNA-miRNA-mRNA axes and highlight the important role of these axes in IRI.


Introduction
Ischemia-reperfusion injury (IRI) occurs after an initial restriction of blood supply to an organ followed by restoration of perfusion [1]. The mechanisms contributing to the pathogenesis of IRI include oxidative/nitrosative stress, mitochondrial dysfunction, calcium overload, inflammation, and activation of apoptotic and autophagic pathways, among other mechanisms [2]. Studies have reported single-target interventions for these pathogeneses of IRI. Nitric oxide (NO) reduces mitochondrial damage and reactive oxygen species (ROS) derived from reperfusion by mimicking the protective effect of kinase pathways that decrease apoptosis and tissue damage. However, it has been difficult to determine the optimal NO dose, and excess NO levels have been determined to be harmful [3]. To reduce the calcium overload, the inhibition of proteins, of which sustained activation causes excessive cation influx, is believed to have a protective effect in ischemia models [4]. When preventative strategies against IRI cannot be used, suppression of the inflammatory response is beneficial for IRI. However, the inflammatory pathways are so complex that blocking any medium in the system may not provide definitive and effective treatment [5]. Compared with single-target interventions, multitarget interventions may have better efficacy in the treatment of IRI.
For decades, research has focused on the 2% of the human genome that codes for proteins [6]. In recent years, researchers have found that the remaining 98% of the genome that was once considered as nonfunctional "junk" includes noncoding RNAs (ncRNAs) that play important roles in a wide range of biological processes such as growth, development, and organ function. Furthermore, ncRNAs have been found to function in all kinds of human diseases and conditions, including IRI [7][8][9]. MicroRNAs (miRNAs) are a family of ncDNAs comprising 21-25 nucleotides and are the most commonly researched class of ncRNAs. miR-NAs play essential regulatory roles in the expression of proteins by binding specific target mRNAs for cleavage or translational repression [10]. Long noncoding RNAs (lncRNAs), the class of ncRNA making up the largest portion of the mammalian ncRNAs, are a heterogeneous group of ncRNAs more than 200 nucleotides long that regulate gene expression through a diverse range of mechanisms [11]. Circular RNAs (circRNAs), characterized by their covalently closed-loop structures without 5 ′ caps and 3 ′ poly tails, comprise a large class of ncRNAs that are produced by a noncanonical splicing event called back-splicing [12]. Recent studies have also revealed a role of lncRNAs/circRNAs as competing endogenous RNAs (ceRNA) that sponge specific miRNAs to indirectly regulate the expression of many genes. Increasing evidence has identified the abnormal expression of ncRNAs in IRI of multiple organs, especially the heart, brain, liver, and kidney [13][14][15][16]. Furthermore, several studies have identified lncRNAs/circRNAs that function as ceRNAs in regulating the expression of many genes as vital to the development and progression of IRI, which may provide multitarget interventions for the treatment of IRI.
In this review, we provide an overview of the roles of the lncRNA/circRNA-miRNA-mRNA axis as potential biomarkers and therapeutic targets for the detection and treatment of IRI in different organs, including the heart, brain, liver, and kidney ( Figures 1 and 2) and as mediators and effectors of organ protection. In addition, we discuss prospective tactics for targeting ncRNAs as potential novel therapies for IRI to reduce tissue injury of important organs.

lncRNA/circRNA-miRNA-mRNA Axis in IRI
Tissue injury elicited by ischemia and reperfusion (I/R) occurs in a wide range of pathologies, especially in myocardial infarction (MI), ischemic stroke of the brain, acute kidney injury (AKI), intestinal ischemia, retinal vascular occlusion, and organ transplantation [1]. Previous studies have shown that ncRNAs play an important role in I/R. Multiple pathological processes that contribute to I/R are associated with cell dysfunction, including apoptosis and necrosis, or autophagy dysfunction, cell proliferation, and sterile inflammation [1,2,17].
2.1. Heart. Cell death is a cardinal contributor to most cardiac diseases such as MI, IRI, and heart failure [18]. The high morbidity and mortality of cardiac diseases is mainly caused by myocardial cell death due to I/R [19]. See Table 1 for a summary of the studies of the lncRNA-miRNA-mRNA axes in myocardial IRI.
2.1.1. lncRNA/circRNA-miRNA-mRNA Axis Regulates Apoptosis of Cardiomyocytes in IRI. Apoptosis is the major form of programmed cell death. Accumulating evidence has demonstrated that the lncRNA/circRNA-miRNA-mRNA axis plays an important role in IRI by mediating cell apoptosis. lncRNA metastasis-associated lung adenocarcinoma transcript 1 (MALAT1) was reported to promote cardiomyocyte apoptosis in an IRI-induced MI mouse model via PDCD4 (Programmed cell death 4) upregulation by sponging miR-200a-3p [20]. Another study suggested that MALAT1 may function as a ceRNA to upregulate NLRC5 (nucleotide-binding and oligomerization domain-like receptor C5) by binding to miR-125b-5p in an IRI-induced acute myocardial infarction (AMI) mouse model, leading to the apoptosis of myocardial cells [21]. Another report showed that the cardiac protective effect of fentanyl was abrogated by MALAT1 through its negative regulation of the miR-145/Bnip3 (Bcl2 19 kDa Protein-Interacting Protein 3) pathway [22].
The lncRNA H19 is transcribed from the imprinted H19insulin growth factor 2 locus [23]. The exact functions of H19 in cancer have been controversial because it has been identified not only as an oncogene but also a tumor suppressor. Similarly, H19 seems to have a contradictory effect in IRI. Some studies showed that the overexpression of H19 reduced cell apoptosis and alleviated myocardial IRI of mice and cardiomyocyte injury induced by H 2 O 2 or hypoxiareoxygenation (H/R) through interacting with different miR-NAs and mRNAs [24][25][26]. Conversely, Luo et al. [27] showed that knockdown of H19 promoted cell viability, inhibited cell apoptosis, reduced inflammatory cytokines, suppressed oxidative stress, and decreased infarct size in a myocardial I/R mouse model through the miR-675/PPARα (Peroxisome proliferator-activated receptor α) pathway. Further research is needed to clarify the definitive mechanism and function of H19 in IRI of the heart.
The lncRNA nuclear-enriched abundant transcript 1 (NEAT1), transcribed from a common promoter by RNA polymerase II, is commonly expressed in mammalian cells and acts as a scaffold for the nucleus [28][29][30]. Multiple studies that investigated the effects of NEAT1 in myocardial IRI demonstrated that NEAT1 was abnormally upregulated in vitro and in vivo [31][32][33][34]. NEAT1 was also significantly upregulated in peripheral blood of patients with unstable angina and patients with ischemic cardiomyopathy/MI in comparison with healthy controls [34]. Furthermore, the overexpression of NEAT1 promoted the apoptosis of cardiomyocytes and enhanced myocardial IRI via different axes, such as the NEAT-miR-495-3p-MAPK6 and NEAT-miR-27b-PINK1 axes [31,35]. However, Yan et al. reported the opposite results and found that NEAT1 was downregulated in cardiomyocytes following IRI in vivo and hydrogen peroxide treatment in vitro and acted as a miRNA sponge to target miR-125a-5p, leading to the upregulation of Bcl12l12 (B-cell lymphoma-12-like 12) and the inhibition of cardiomyocyte apoptosis [36].
Previous studies showed that suppression of the lncRNA HOX transcript antisense RNA (HOTAIR) exasperated cell viability and migration potential and increased apoptosis induced by oxidative stress in H9C2 cells, which may be partly attributed to the HOTAIR/miR-125/MMP-2 (Matrix metalloproteinase-2) axis [37]. Correspondingly, HOTAIR prevented oxidative stress and cardiac myocyte apoptosis in myocardial IRI, which involves AMPKα activation via the EZH2/miR-451/Cab39 axis [38]. Yu and Chen [39] speculated that circulating HOTAIR/miR-126 may be a potential biomarker and risk factor predictor for myocardial IRI.
In addition to lncRNA-related axes, increasing evidence has suggested that circRNA-miRNA-mRNA axes regulate the apoptosis of cardiomyocytes in IRI of heart disease (see Table 2 for a summary of the studies of circRNAs-miRNAs-mRNAs in IRI). A letter published in 2019 speculated that the circDLGAP4/miR-143 pathway may be a potential regulator of cardiomyocyte apoptosis in myocardial IRI [60]. Intriguingly, a study by Chen et al. confirmed this speculation [61]. The authors found that the overexpression of circDLGAP4 effectively restored the decreased expression of the circRNA HECT domain E3 ubiquitin protein ligase 1 (HECTD1) resulting from miR-143 inhibition in human umbilical vein endothelial cells, which contributed to the attenuation of endothelial cell dysfunction induced by IRI by increasing cell viability and decreasing cell apoptosis and migration [61]. The circRNA sodium/calcium exchanger 1 (circNCX1) was upregulated in both H9C2 cells and neonatal rat myocardial cells after treatment of H 2 O 2 or H/R and promoted the production of ROS and myocardial cell apoptosis induced by IRI by targeting miR-133a-3p and leading to overexpression of proapoptotic gene cell death-inducing protein (CDIP1) in a myocardial I/R mouse model [62]. In addition, mitochondrial fission and apoptosis-related circRNA (MFACR) regulated mitochondrial fission and apoptosis in the heart by directly targeting and downregulating miR-652-3p, which blocked mitochondrial fission and reduced cardiomyocyte cell death by suppressing MTP18 translation. Consequently, the knockdown of MFACR attenuated the I/R-induced upregulation of mitochondrial fission, apoptosis, and MI size [63]. Furthermore, the circRNA serine/threonine-protein kinase tousled-like 1 (circTLK1) was prominently upregulated in a myocardial IRI mouse model, leading to significantly increased cardiomyocyte apoptosis by its activity as a sponge of miR-214. miR-214 abolished the negative effects of receptor-interacting serine/threonineprotein kinase 1 (RIPK1) in myocardial IRI, including an impaired cardiac function index, distensible infarct area, and cell apoptosis. These results indicate that the cir-cTLK1/miR-214/RIPK1 axis plays a crucial role in myocardial IRI, which may provide therapeutic targets for treatment [64]. Chang et al. [65] reported that circ_100338 regulates angiogenesis and metastasis of myocardial I/R through miRNA-200a-3p/FUS.
As described above, many lncRNA/circRNA-miRNA-mRNA axes mediate cell apoptosis of cardiomyocytes in IRI. Apoptosis is induced by the "extrinsic" and "intrinsic" pathways, and there are multiple biochemical and functional linkages between the two pathways. I/R, as cytotoxic stimuli, induces the translocation and integration of prodeath members of the Bcl2 protein family (e.g., Bax and Bak) into the outer mitochondrial membrane [18]. However, ischemia per se is not sufficient for activation of Bcl2 proteins because many are redox sensitive, requiring the oxidative stress that is evoked by reperfusion. I/R-induced cell death is reduced in animals treated with pan-caspase inhibitors, providing additional support for the notion that apoptosis contributes to the death of cardiac myocytes [66,67]. While such observations might lead to the proposal that targeting caspases may be an important therapeutic means to reduce I/R injury, caspase inhibition may not be ideal because other aspects of mitochondrial function will be adversely affected. Therefore, these lncRNA/circRNA-miRNA-mRNA axes may be a suitable alternative in the regulation of cardiomyocyte apoptosis and represent potential therapeutic targets of cell apoptosis during cardiac I/R injury.
2.1.2. lncRNA-miRNA-mRNA Axis Regulates Autophagy of Cardiomyocytes in IRI. Autophagy is a highly conserved catabolic process that provides organelle quality control and generates intracellular nutrients from lysosomal processing of cellular structures [68]. Yu et al. first speculated that MALAT1 may negatively regulate the expression of miR-204, which could increase the autophagy of cardiomyocytes and myocardial IRI [69]. Furthermore, MALAT1 promoted OGD/R-induced H9C2 cell injury by sponging miR-20b to enhance Beclin1-mediated autophagy [70]. miR-204 was found to regulate autophagy through LC3-II during myocardial I/R [71]. Based on the above conclusions, Wang et al. [72] ultimately speculated that the MALAT1/miR-204/LC3-II pathway may be an important regulatory axis of autophagy in myocardial IRI. Nonetheless, further experimental evidence is needed to confirm this possibility. The lncRNA AK139128 is involved in the regulation of autophagy and apoptosis in myocardial IRI by targeting the miR-499/FOXO4 axis [73]. In addition, silencing AK139328 by siRNA significantly enhanced miR-204-3p expression and suppressed cardiomyocyte autophagy, thereby attenuating myocardial IRI in diabetic mice [74]. The lncRNA AK088388 competitively binds to miR-30a, which promotes the expression of autophagy-related proteins, Beclin1 and LC3-II, and eventually leads to cell damage in myocardial    IRI [75]. The lncRNA autophagy-promoting factor (APF) mediates the conduction of autophagic-related signals in cardiomyocytes and competitively binds to miR-188-3p, thus indirectly upregulating the expression of ATG7 (Autophagy-related gene 7) and affecting autophagic cell death and MI [76]. Furthermore, Chen et al. [77] found that I/R induced a significant increase in miR-128 associated with a decrease in UCA1 and HSP70, which was reversed by morphine postconditioning treatment that also ameliorated infarct size and cell autophagy. This result suggested that morphine postconditioning treatment preserved myocardium from injury by mediating the UCA1/miR-128/HSP70 pathway.
Autophagy is actually a cell survival mechanism rather than a cell death process and can be activated by I/R-related conditions (e.g., energy deprivation, oxidative stress, and ER stress) [78]. However, uncontrolled autophagy ultimately leads to cell death and may contribute to I/R injury. Autophagy is involved in myocardial IRI through a dual regulation: protection of myocardial cell death during the myocardial ischemia stage and prevention of myocardial cell death during the myocardial reperfusion stage. Inhibition of autophagy has been shown to amplify I/R-induced damage [78,79], while pharmacologic stimulation of autophagy confers protection against I/R [80, 81].

lncRNA-miRNA-mRNA Axis Regulates Necrosis of Cardiomyocytes in IRI.
Distinct from the programmed property of apoptosis and autophagy, necrosis is an uncontrolled process that occurs randomly under the condition of overwhelming stress and contributes to the "accidental" death of the cell [2]. The lncRNA necrosis-related factor (NRF) functions by directly binding to miR-873 and regulates RIPK1/RIPK3 expression and necrosis. Necrosis in cardiomyocytes and MI induced by IRI is attenuated when the expression of NRF is knocked down. Furthermore, p53 regulates cell necrosis in the heart by targeting NRF, miR-873, and the RIPK1/RIPK3 axis in the necrotic cascades [82]. The death program of cytokine-induced necrosis in myocardial IRI was further aggravated when H19 was downregulated by RNA interference. Further research found that H19 decreased the necrotic cell death of cardiomyocytes by interfering with the expression of miR-103/107 that promoted cell necrosis in a cellular model treated with H 2 O 2 and in a myocardial IR mouse model by inhibiting the expression of FADD (Fas-associated protein with death domain) [83]. Further studies should explore how the H19-miR-103/107-FADD pathway is involved in the intricate necrotic cascade.
Necrosis is one of the main forms of cell death that is most prominent in the I/R heart. Cells can be driven to necrosis by I/R via the activation of at least three separate signaling pathways: necroptosis, mitochondrial permeability transition-dependent regulated necrosis, and parthanatos [84,85]. Although these lncRNA-miRNA-mRNA axes mediate necrosis of cardiomyocytes in IRI, how these axes integrate into the complex necrotic cascade and the relationships with other necrotic-related factors remain to be studied.

lncRNA-miRNA-mRNA Axis Regulates Inflammation of Cardiomyocytes in IRI.
The I/R-induced inflammatory response in most organs has been termed sterile inflammation because of the absence of microorganisms. However, similar to the response to all kinds of microorganism pathogens, sterile inflammation derived from IRI is characterized by the recruitment of peripheral immune cells to the injured tissue sites, accompanied with the production and release of cytokines and chemokines [86]. As discussed above, MALAT1 plays vital roles in I/R pathogenesis by mediating cell death. Some studies speculated that MALAT1 regulates the inflammatory response in myocardial IRI via targeting different targets. One of the first studies showed that MALAT1 upregulates NLRP3 inflammasome expression potentially by sponging miR-133 in the I/R-injured heart [87]. Moreover, Ruan et al. [88] speculated that MALAT1 may aggravate the inflammation response through regulating PTGS2 (Prostaglandin-endoperoxide synthase 2) by targeting miR-26b in myocardial IRI. MALAT1/miR-203 was also considered important in I/R by increasing cardiomyocyte inflammation and myocardial injury [89].
Inflammation plays a prominent role in the reperfusion component of total tissue injury in I/R, which is characterized by leukocyte trafficking to ischemic sites that occur primarily during reperfusion, and I/R-induced leukocyte infiltration contributes to a large number of pathologic processes [86]. Furthermore, leukocyte endothelial cell adhesive interactions, which precipitate the microvascular complications and tissue injury induced by reperfusion, are one of the earliest signs of tissue dysfunction and injury elicited by I/R [90,91]. Multiple factors are involved in the dynamic regulation of inflammation in myocardial IRI upon AMI, and thus inhibition of the inflammatory response may be a potential therapeutic strategy [92]. Therefore, it may be possible to reduce or prevent the production of IRI by interfering with the inflammatory response produced by these axes.

Brain.
Acute ischemic stroke (AIS) is a pathological process that starts with local cerebral vascular occlusion and is accompanied by a series of changes in cellular behaviors, leading to sudden local brain dysfunction [93]. The effective treatment for AIS is to restore blood flow, which can lead to reperfusion injury. Cerebral IRI often occurs in stroke and cardiac arrest and induces neuronal damage. Increasing evidence demonstrates that ischemia is often associated with a series of neurological disorders, such as hypoxia, oxidative stress, and inflammatory responses, which eventually lead to acute necrosis, apoptosis, and autophagy of ischemic brain cells [94]. In recent years, ncRNAs were found to play important roles in physiopathological processes related to stroke (see Table 3 for a summary of the studies on lncRNAs-miRNAs-mRNAs in cerebral IRI).
2.2.1. lncRNA-miRNA-mRNA Axis Regulates Apoptosis of Nerve Cells in IRI. In cerebral IRI, the main mechanism of brain injury mainly involves apoptosis of nerve cells. MALAT1 and TUG1 may exhibit similar roles in cerebral IRI. MALAT1 and TUG1 were significantly upregulated in both a middle cerebral artery occlusion/reperfusion 18 BioMed Research International
2.2.2. lncRNA-miRNA-mRNA Axis Regulates Autophagy of Nerve Cells in IRI. By regulating the miR-26b/ULK2 and miR-30a/Beclin1 pathways, MALAT1 promoted OGD/Rinduced nerve cell autophagy and ischemic injury [107,108]. Yu et al. [109] showed that the lncRNA KCNQ1OT1 promotes I/R-induced autophagy and reduces cell viability by regulating the miR-200a/FOXO3/ATG7 axis. Furthermore, KCNQ1OT1 was significantly increased in the plasma of patients with AIS, and its expression was positively correlated with the severity of stroke, which implied that KCNQ1OT1 may be a diagnostic biomarker or severity evaluation indicator.
2.2.4. circRNA-miRNA-mRNA Axis in Brain IRI. Lin et al. [115] explored the potential function of circRNAs in the etiopathogenesis of cerebral IRI. The authors investigated the expression profiles of circRNAs between HT22 cells with OGD/R and controls using a circRNA microarray. The results showed that 15 circRNAs were markedly altered in the OGD/R model group. The authors selected mmu-circRNA-015947 for further verification by qRT-PCR. Bioinformatics analysis showed that mmu-circRNA-015947 could bind with miRNAs (mmu-miR-188-3p, mmu-miR-329-5p, mmu-miR-3057-3p, mmu-miR-5098, and mmu-miR-683) and thereby elevate the expression of target genes. This research indicates that increasing the level of mmu-circRNA-015947 might contribute to the process of cerebral IRI and provides a potential strategy for clinical treatment. In addition, three circRNAs (cZNF292, TLK1, and circ_ 008018) were significantly increased in the MCAO/R mouse model and mouse neurons under OGD/R and knockdown of the circRNAs attenuated neuronal injury during cerebral I/R. These effects were related to cell apoptosis mediated by targeting miR-22/Wnt/b-catenin (under OGD/R), miR-335-3p/TIPARP (under OGD/R and cerebral I/R), and miR-99a/Pi3k/Akt/GSK3b (cerebral I/R) [116][117][118]. Furthermore, the expression levels of circTLK1 in patients with AIS were notably increased in comparison with healthy controls. Interestingly, the expression levels of circTLK1 in plasma from patients with stroke were associated with lesion localization and infarct volumes [117]. Han et al. [119] found that the levels of circHECTD1 were elevated in tMCAO mouse stroke models and in plasma of AIS patients. Knockdown of cir-cHECTD1 contributed to reduction in infarct areas, attenuation of neuronal deficits, and amelioration of astrocyte activation via mediating autophagy in tMCAO mice through miR-142/TIPARP. Thus, circHECTD1 was considered a new biomarker and therapeutic target for stroke. Jiang et al. [120] found that an upregulated circRNA, antisense noncoding RNA in the INK4A locus (circANRIL), inhibited OGD/Rinduced HBMEC proliferation and promoted cell apoptosis and phosphorylation of p65 and IκBα, which were abrogated by miR-622. These results demonstrated that circANRIL aggravated OGD/R-induced injury in HBMECs by mediating the NF-κB pathway through sponging miR-622.

Liver
2.3.1. lncRNA-miRNA-mRNA Axis Regulates Apoptosis of Hepatocytes in IRI. Liver IRI, which occurs in hemorrhagic shock, resection, and transplantation, starts with local ischemic insult, followed by inflammation-mediated reperfusion injury [121] (see Table 4 for a summary of the studies of lncRNAs-miRNAs-mRNAs in other diseases with IRI). A recent study investigated the role of the lncRNA Gm4419 in hepatic I/R [122]. The authors found that Gm4419 was upregulated in hepatic IRI rats, and knockdown of Gm4419 aggravated I/R-induced liver damage in hepatic IRI rats. 22 BioMed Research International

BioMed Research International
Gm4419 promoted H/R-induced apoptosis by sponging miR-455 and regulating SOX6 in BRL-3A cells. Therefore, Gm4419 accelerated hepatic IRI by interacting with the miR-455/SOX6 axis. Huang et al. [123] reported that the expression levels of MEG3 and Nrf2 were decreased in hepatic I/R mice and in HL7702 cells with H/R treatment, while miR-34a was increased. Overexpression of MEG3 inhibited apoptosis and affected the production of ROS in vitro, which was abrogated by miR-34a inhibitor treatment. MEG3 overexpression ameliorated the hepatic function of hepatic I/R mice and significantly reduced the level of serum ALT and AST. These results indicated that MEG3 protected hepatocytes from hepatic IRI through miR-34a/Nrf2. Furthermore, Dai et al. [124] found that the levels of AK054386, miR-199, and CHOP (C/EBP homologous protein) were elevated, decreased, and elevated in hepatic IRI models, respectively. Overexpression of AK054386 promoted cell apoptosis in the BNL-CL2 IRI cell model and CHOP expression, which were rescued by miR-199 overexpression. These results suggest that AK054386 plays a crucial role in hepatic IRI via miR-199 by mediating the ERS pathway.
2.3.2. lncRNA-miRNA-mRNA Axis Regulates Autophagy of Hepatocytes in IRI. Autophagy has been proven to be involved in hepatic IRI [125,126]. Liver IRI induces elevated levels of HOTAIR and ATG7 and increases autophagy, which is attenuated by the knockdown of HOTAIR. In addition, HOTAIR acts as a ceRNA for miR-20b-5p and increases the expression of ATG7. These results indicated that the HOTAIR/miR-20b-5p/ATG7 axis plays a crucial role in hepatic IRI via autophagy [127].
2.3.3. circRNA-miRNA-mRNA Axis in Liver IRI. Zhang et al. [128] first examined circRNA expression profiles during hepatic IRI by microarray hybridization analysis and found that circRNAs are closely associated with hepatic IRI and ischemic postconditioning (IPO). The analyses revealed that the expression of 1599 circRNAs was altered, including 213 upregulated and 493 downregulated circRNAs, between the I/R group and the control group. In a comparison of the IPO group with the I/R group, the results revealed that 641 circRNAs were upregulated and 252 circRNAs were downregulated. Moreover, the ceRNA network, including 6 cir-cRNAs, 47 miRNAs, and 90 mRNAs, illustrated that the "housekeeping" function of circRNAs is abnormally regulated in hepatic IRI. The mmu_circRNA_005186/miR-124-3p/Epha2 axis was chosen for further study after qRT-PCR validation. Silencing of mmu_circRNA_005186 moderated lipopolysaccharide-induced inflammation by elevating miR-124-3p and reducing Epha2, which suggests that the mmu_ circRNA_005186/miR-124-3p/Epha2 axis might play an important role in hepatic IRI.

Kidney
2.4.1. lncRNA-miRNA-mRNA Axis in Renal IRI. Renal IRI, which contributes greatly to AKI, is one of the most critical issues for many clinical situations, including renal transplantation, nephrectomy, sepsis, and repair of suprarenal aneu-rism [129][130][131]. Hu et al. [132] used a microarray assay to find that 2218 genes are differentially expressed in renal IRI, including 1103 upregulated genes and 1115 downregulated genes. The authors further reported that both the dysregulated lncRNA MALAT1 and miR-139-5p participated in IRI and were closely associated with cell proliferation [133]. In addition, Geng et al. [134] revealed that GAS5, which is prominently elevated in renal IRI, was inhibited by delayed ischemic preconditioning (IPC), and GAS5 levels were increased by knocking down miR-21 before IPC in vivo. A negative regulatory relationship was detected between miR-21 and TSP-1 (Thrombospondin 1) both in vitro and in vivo. The role of GAS5 in H/R-induced cell apoptosis was already illuminated in the authors' previous work [135]. Together, these results indicated that GAS5 promoted apoptosis by functioning as a miR-21 sponge and downregulating TSP-1 in renal IRI.
NEAT1 was reported to facilitate renal tubular epithelial apoptosis induced by H/R via binding to miR-27a-3p, and reciprocal inhibition was detected between NEAT1 and miR-27a-3p. Furthermore, miR-27a-3p was expressed at low levels while NEAT1 was overexpressed in AKI patients compared with healthy controls [136]. Xu et al. [137] confirmed that TUG1 silencing attenuates inflammation and apoptosis via binding to miR-449b-5p and downregulating HMGB1 and MMP-2 expression in renal IRI. In addition, Tang et al. [138] reported that loss of XIST and PDCD4 enhanced CoCl2-processed HK-2 cell proliferation and repressed cellular apoptosis, which was reversed by miR-142-5p. Furthermore, knockdown of LINC00520 protected against AKI both in vitro and in vivo by mediating PI3K/AKT through miR-27b-3p/OSMR [139].
2.4.2. circRNA-miRNA-mRNA Axis in Renal IRI. Huang et al. [140] reported that the circRNA yes-associated protein 1 (circYAP1) expression was downregulated in AKI patients as well as in I/R-treated HK-2 cells. circYAP1 overexpression promoted cell growth and attenuated the secretion of inflammatory factors and ROS generation in HK-2 cells under I/R treatment. In addition, the authors found that circYAP1 was a functional sponge for miR-21-5p that reversed the inhibitory effects of circYAP1 on cell injury. Furthermore, circYAP1 inhibits miR-21-5p to activate the PI3K/AKT/m-TOR pathway. These results revealed that circYAP1 activates the PI3K/AKT/mTOR signal pathway and protects HK-2 cells against renal IRI via binding to miR-21-5p.
2.5. Other IRI 2.5.1. lncRNA-miRNA-mRNA Axis in Retinal IRI. Retinal IRI, a cause of irreversible visual damage, occurs with glaucoma, diabetic retinopathy, and retinal vascular occlusive disorders [141]. One study [142] reported that the Mbd2-AL1 (Methyl-CpG-binding domain protein 2-associated long noncoding RNA 1)/miR-188-3p/Traf3 axis plays an important role in mediating retinal ganglion cell apoptosis and visual function in Mbd2-KO mice. The results implied that the lncRNA Mbd2 may be a novel therapeutic target for retinal ischemic diseases. In addition, there may be many 26 BioMed Research International similarities between the roles of miR-21/PDCD4 in retinal IRI and cerebral IRI. Wan et al. [143] reported that I/R-mediated H19 overexpression facilitates NLRP3/6 inflammasome imbalance and leads to retinal microglial pyroptosis, excess cytokine secretion, and neuronal death. MEG3 aggravates ischemic damage and deteriorates overall neurological functions by binding to miR-21 and downregulating PDCD4, which mediates sterile inflammation and neuronal lesions in cerebral IRI [144].  [146] found that inhibition of the lncRNA cancer susceptibility candidate 7 (CasC7) promoted cell apoptosis in SH5Y-SY cells under OGD/R treatment and increased infarct size in spinal cord IRI rats through miR-30c/Beclin1, which was reversed by NaSH preprocessing. The study concluded that hydrogen sulfide saves the spinal cord from IRI by the CasC7/miR-30c/Beclin1 axis. The expression of MALAT1 and Bcl2 was suppressed while miR-204 was upregulated in a rat spinal cord IRI model and hypoxia-induced neurocyte lines [147]. Furthermore, knockdown of MALAT1 promoted cell apoptosis, which was associated with downregulation of Bcl2 and upregulation of miR-204. MALAT1-treated spinal cord IRI rats also showed lower motor deficit index scores. Therefore, these results indicated that MALAT1 plays a neuroprotective role in spinal cord IRI rats by binding miR-204/Bcl2.
2.5.4. lncRNA/circRNA-miRNA-mRNA Axis in Intestinal IRI. Zou et al. [148] revealed that H19 overexpression increased the level of miR-675, which in turn inhibited the expression of ZO-1 (Zonula occludens 1) and E-cadherin, leading to dysfunction of the epithelial barrier. These effects were reversed by upregulation of the RNA-binding protein HuR in H19overexpressing cells. These results revealed that H19 and HuR act upon each other, and H19 mediates the intestinal epithelial barrier function via the miR-675/ZO-1/E-cadherin axis. Feng et al. [149] explored the expression profiles of circRNAs after intestinal I/R with or without IPO and investigated the underlying mechanisms of IPO associated with the altered cir-cRNAs. The authors identified 62 circRNAs and 521 mRNAs differentially expressed in the intestinal I/R group compared with the sham group, as well as 33 circRNAs and 303 mRNAs that were altered between the IPO group and I/R group. Two circRNAs, circRNA_012412 and circRNA_016863, were identified as closely related to the protective mechanisms of IPO. Ultimately, four pathways, circRNA_012412/miR-7649-3p/Sertad1 (SERTA domain-containing protein 1), cir-cRNA_012412/miR-3473c/Sertad1, circRNA_012412/miR-6991-3p/Nudcd1 (NudC domain-containing protein 1), and circRNA_012412/miR-6991-3p/Jam2 (Junctional adhesion molecule B), were constructed based on the TargetScan, miRanda, and miRDB databases. These circRNA regulatory pathways may be closely associated with endogenous protective signaling in IPO during intestinal I/R and warrant further investigation. The study was the first to fully describe the cir-cRNA expression profiles during intestinal I/R, and the results showed that IPO was associated with altered circRNAs, which provides a new perspective to clarify how IPO protects against intestinal IRI.

Conclusions and Future Perspectives
Over the past few decades, ncRNAs have been found to play complex roles in the development and gene regulatory processes of many diseases. However, understanding of the mechanisms of lncRNAs and circRNAs in different organs and dysfunctional states and their potential as therapeutic targets or diagnostic markers of IRI is still in its infancy. Furthermore, the interactions of the lncRNAs/circRNAs-miR-NAs-mRNAs are complex and dynamic. First, a specific lncRNA or circRNA plays a variety of roles in IRI of different organs by targeting different miRNAs/mRNAs. The expression levels of lncRNAs and circRNAs in animal and cell models seem to be similar as in patient serum and were associated with the severity of the disease, providing evidence for clinical diagnosis and prognosis. Second, the same lncRNA or circRNA can play a contradictory role in IRI of the same organ, potentially contributing to other molecules in response to stress conditions. For example, H19 showed opposite expression levels and effects in myocardial IRI, and future studies should investigate the various regulatory mechanisms. Third, some lncRNAs and circRNAs can target the same miRNAs/mRNAs in different organs. For instance, MEG3 and H19 both performed their functions in cerebral IRI and in retinal IRI, respectively, by sponging miR-21 and targeting PDCD4 to mediate cell apoptosis [143,144]. Finally, some drugs and treatments play an important role in IRI through the functions of these axes. For example, metformin protects against oxidative stress injury in cerebral IRI by affecting the H19/miR-148a-3p/Rock2 axis [99], while the cardioprotective effects of fentanyl in cardiac IR appeared to be abrogated by the MALAT1/miR-145/Bnip3 axis [22]. Furthermore, IPO can attenuate liver IRI and exhibit significant intestinal protection through circRNAs-miRNAs-mRNAs [128,149], and the renal protection of delayed IPC involves preconditioning-induced upregulation of miR-21 and downregulated expression of GAS5 and TSP-1 [134]. Although these studies have shown some benefit, they have not been able to identify a specific effective protocol of IPC/IPO for a clinical study, perhaps because it has been very difficult to systematically illuminate the underlying complex mechanisms of ncRNAs. Therefore, further investigations focusing on revealing the specific molecular mechanisms of ncRNAs in the development of IRI and abnormal conditions are needed.

BioMed Research International
Although cardiac magnetic resonance imaging remains the gold standard to assess the consequences of acute IRI in AMI patients in terms of MI size and adverse left ventricular remodeling, this technique is limited in terms of ease of access and availability of skilled personnel [150]. However, the presence of ncRNAs in the serum plasma has suggested that these molecules may serve as biomarkers in AMI patients or patients with myocardial IRI [34,41,117,119,151]. Although plasma contains RNases, circulating ncRNAs, especially lncRNAs, have been shown to be stable in this environment, which indicates that they are relatively resistant to nucleolytic degradation and making them potentially useful as circulating biomarkers for AMI. Furthermore, ncRNA or the antisense molecules may be delivered to the ischemic heart using several approaches including intravenous or intramyocardial injections or carriers such as viruses, nanoparticles, or exosomes. Zhang et al. [152] found that hypoxia modified the expression of several miRNAs in exosomes secreted by H9c2 cells, and these exosomal miRNAs protected H9c2 cells against simulated IRI and prevented apoptosis through HIF-1, TNF, MAPK, and mTOR signaling pathways. Some miRNAs or lncRNAs that are downregulated following AMI are known to be beneficial for cardioprotection, and one therapeutic strategy is to deliver ncRNA mimics targeting these miRNAs or lncRNAs to the ischemic heart. Furthermore, an alternative approach to upregulate cardioprotective miRNAs may involve a smallmolecule therapeutic strategy. Therefore, further research on these ncRNAs involved in IRI of different organs is urgently needed, and these studies may lead to the identification of other unknown signaling pathways or reactions in IRI. This research will also help elucidate the contribution of ncRNAs to the pathophysiology of IRI and subsequent clinical outcomes and provide support for them as potential markers or therapeutic targets for IRI.
In summary, the lncRNA/circRNA-miRNA-mRNA axes have multiple roles in IRI, and factors in these axes may function as diagnostic markers and therapeutic targets. Several studies involving the lncRNA/circRNA-miRNA-mRNA axis are imperative and continuous efforts will be necessary for us to integrate these axes with clinical practice.

Conflicts of Interest
The authors declare that there are no conflicts of interest.