Atherosclerosis

For decades, various genetic and environmental factors have been studied to elucidate the onset, progression, and causes of diseases [1-4]. Cardiovascular disease (CVD) as a representative disease with high mortality rates progresses through complex pathophysiological processes through communication between the many involved cells [5,6]. Atherosclerosis, a representative CVD, has traditionally focused on evaluating immune cells, especially macrophages that phagocytose lipids to become foam cells and secrete various inflammatory substances [7-9]. Recently, many studies have been published showing that endothelial cells (ECs) and smooth muscle cells (SMCs) constituting the intima and media of arteries play major roles in the formation of atherosclerotic lesions [10-12]. Blood factors such as cholesterol and low-density lipoprotein (LDL) cause endothelial dysfunction in the vascular intima and allow primary immune cells such as monocytes and macrophages to infiltrate the intima and media [13,14]. These macrophages are transformed into foam cells and inflammatory phenotypes that secrete inflammatory substances to stimulate vascular smooth muscle cells (VSMCs). VSMCs, which are capable of phenotypic switching, are converted to proliferative and migratory phenotypes by inflammatory stimuli and play a major role in plaque formation [15-17]. Interestingly, new cell types with characteristic transcriptome patterns within atherosclerotic lesions have been discovered with the recently developed single-cell RNA sequencing technology suggesting that there are still unknown cell types involved in atherosclerosis [18-21]. Therefore, evaluation of the crosstalk between the cell types involved in the lesion is important for understanding these diseases.

Non-coding RNAs

Studies on the pathophysiological mechanisms of CVDs have been conducted for numerous genes [22-25]. Furthermore, several regulatory non-coding RNAs (ncRNAs) involved in this mechanism have been discovered [26]. Based on 200 nucleotides, it can be divided into small ncRNAs and long ncRNAs (lncRNAs) [27]. First, microRNA (miRNA), a typical regulatory small ncRNA, forms a complex with proteins including argonaute (AGO), and binds to the 3ʹ-untranslated region (3ʹ-UTR) of the target messenger RNA (mRNA) in a specific base pair, thereby interfering with mRNA translation [27]. Many lncRNAs have also been identified, and they regulate the expression of various genes [28]. The types and roles of lncRNAs are further subdivided according to their location on the genome (intronic, intergenic, bidirectional), the direction of transcription relative to the nearby protein-coding gene (sense and antisense), and the regulation manner for other genes (cis and trans) [27]. Finally, circular RNA (circRNA), the most recently identified regulatory ncRNA, is being actively studied for its biogenesis and function [29]. Early on, it was discovered in eukaryotes that circRNAs have closed and single-stranded structures made from the back-splicing of host mRNAs [30,31]. The analysis from high-throughput RNA sequencing has confirmed that many circRNAs are widely expressed in cells and tissues [29,32].

Co-culture systems

Although many researchers are conducting experiments using animal models and analyzing patient-derived clinical samples to elucidate the molecular mechanisms of disease, there are some limitations [33]. In vivo experiments using animal models such as mice are of great help in predicting disease mechanisms but have limitations in that the molecular mechanisms do not completely match those of humans [34]. In addition, research using clinical samples is difficult to expand to molecular study and the supply of samples is also limited. Therefore, a plan must be designed to reduce the excessive time and cost of in vivo experiments and to efficiently utilize clinical samples after confirming the precise mechanisms in cells through in vitro experiments. Elucidation of the action mechanism of a specific gene in a single cell type has continuously progressed. However, to further explore the complex process in a particular disease, it is also important to examine the intercellular crosstalk response according to gene expression changes by co-cultivating two or more cells. Furthermore, cell co-culture systems have been well established, and they are often used in in vitro experiments in disease mechanism studies [17].

For co-culture, donor cells that generate signals upon stimulation by external factors are required, as in diseases [35-37]. A recipient cell is also needed to receive these signals. There are several co-culture techniques for CVD research [38]. First, indirect co-culture confirms the effect only through secreted factors (cytokines, growth factors, and extracellular vesicles [EVs]) produced by donor cells [38]. Indirect co-culture is divided into a method that involves direct treatment of the donor cell-cultured medium and a method that involves exchanging only secreted factors in a separate space through a transwell chamber (Fig. 1A). Conversely, direct co-culture induces cell-to-cell contact by placing two cells in a culture dish or attaching each cell to the opposite sides of a transwell chamber (Fig. 1B) [38,39]. In addition to secreted factors, the effects of adhesion molecules such as intercellular adhesion molecule 1 (ICAM-1) and vascular cell adhesion molecule 1 (VCAM-1) can also be tested [38]. Finally, there is another method that involves creating a three-dimensional environment such as tissue using two or more types of cells and a gel scaffold platform (Fig. 1C) [38]. This is a typical atherosclerosis-simulating co-culture model in which the intima (ECs), media (VSMCs) with the extracellular matrix (gel scaffold), and immune cells (monocyte/macrophage) infiltrating these layers of the blood vessel model are all structured together [38].

Figure 1. Co-culture systems. (A) Indirect co-culture model. Intercellular crosstalk is mediated through factors secreted from cells. In a model, each cell type is cultured separately, and the supernatant cultivated with one cell type (conditioned medium) is applied to the other cell type. Otherwise, each cell type is cultured separative at the same culture well through the transwell chamber and the secreted factors (cytokines, chemokines, growth factors, and extracellular vesicles) are exchanged. (B) Direct co-culture model. Both cells are cultured together in the same culture well for interaction through secreted factors and adhesion molecules (ICAM-1, VCAM-1, E-selectin, and receptors) allowing cell-to-cell contact. Two different types of adherent cells (vascular smooth muscle cells or endothelial cells), or both adherent and suspension cells (immune cells), are cultured together. Otherwise, each cell type attached to the opposite sides of the permeable membrane of the transwell chamber is allowed to interact through contact and to exchange secreted factor. (C) Gel scaffold 3D co-culture model. An extracellular matrix (ECM) is used to construct a gel scaffold in a culture well or transwell chamber. Through the ECM layer, compartments are divided between different cell types, and intercellular crosstalk and cell migration can be observed through this layer. ICAM-1, intercellular adhesion molecule 1; VCAM-1, vascular cell adhesion molecule 1; 3D, three-dimensional.

Intercellular crosstalk studies that utilized the abovementioned co-culture systems to create various disease-mimic environments have been performed widely [17,40,41]. However, intercellular crosstalk studies of ncRNAs are relatively scarce, and few reviews have summarized the utilization of co-culture systems. Therefore, this review summarizes intercellular crosstalk studies of various ncRNAs and the co-culture systems used therein. In particular, among complex diseases, research on CVDs including atherosclerosis is intensively discussed herein. The studies involving the direct delivery of ncRNAs through EVs and those evaluating the indirect influence of ncRNAs on other cells through the regulation of secreted factors and cell adhesion molecules are mainly described (Fig. 2).

Figure 2. The intercellular crosstalk and non-coding RNAs (ncRNAs) in diseases. ncRNAs are differentially expressed by various cellular stresses that cause disease, and they transmit signals to other cells in two major ways. First, ncRNAs are directly delivered from donor cells to recipient cells by extracellular vesicles such as exosomes and microvesicles (red arrow). These delivered ncRNAs interfere with the protein translation of target mRNAs (miRNA) or regulate the expression of other genes by binding to RNA-binding proteins or miRNAs (lncRNA and circRNA). In another way, ncRNAs indirectly affect recipient cells by regulating the expression of secreted factors (cytokines, chemokines, and growth factors) or cell adhesion molecules (ICAM-1, VCAM-1, E-selectin, and receptors) in donor cells (blue arrow). miRNA, microRNA; lncRNA, long non-coding RNA; circRNA, circular RNA; MVB, multivesicular body; Ang II, angiotensin II; I/R injury, ischemia-reperfusion injury; oxLDL, oxidized low density lipoprotein; TNF-α, tumor necrosis factor-α; ILs, interleukins; MMPs, matrix metalloproteinases; GFs, growth factors; sICAM-1, soluble intercellular adhesion molecule-1; sVCAM-1, soluble vascular cell adhesion molecule-1.

Direct delivery of ncRNAs via EVs

EV-derived miRNAs: The pathophysiological role and mechanism of action of ncRNAs identified in each disease have been extensively studied and summarized in other reviews [28,42]. Several EV-derived ncRNAs that play important roles in CVDs, including atherosclerosis, are reviewed here. In studies of ncRNAs delivered directly through EVs (exosomes and microvesicles), many miRNAs, which are relatively small in their size compared with lncRNAs and circRNAs, have been discovered and studied [43,44].

Representatively, the miRNA produced from the miR-143/145 cluster is well known as an EV-derived miRNA that regulates the phenotype of VSMCs [45-47]. Krüppel-like factor 2 (KLF2), a transcription factor associated with various diseases including atherosclerosis, induces upregulation of the miR-143/145 cluster in ECs [43]. The miR-143/145 cluster is enriched in the EVs released from human umbilical vein ECs (HUVECs) in which KLF2 is overexpressed or stimulated by shear stress. These are delivered to human aortic SMCs (HASMCs) and reduce the expression of miR-143/145 targets to prevent dedifferentiation of VSMCs [43]. This result was confirmed by the isolation of EVs and the co-culture of these two cell types. In this study, when the phospholipid membrane of the EV was disrupted by phospholipase (Triton X-100 or cyclodextrin), miR-143/145 in the EV was mostly degraded by RNase [43]. This suggests that EVs such as exosomes and microvesicles play an important role in stably delivering ncRNAs to a target.

MiR-155 is another EV-derived miRNA that is highly expressed in human carotid plaque samples [48]. It plays an important role in vascular inflammation and atherogenesis by suppressing the expression of B-cell lymphoma 6 (BCL6), a transcription factor that attenuates the pro-inflammatory nuclear factor-κB (NF-κB) signal. In addition, miR-155, which is strongly expressed in VSMCs by overexpression of KLF5, is transferred to ECs through exosomes [49]. This transferred miR-155 facilitates endothelial dysfunction and atherosclerosis progression, suggesting that exosomal miR-155 could be a target for atherosclerosis treatment.

Exosomal miR-155 has not only been studied in atherosclerosis but also in acute myocardial infarction (AMI) [50]. Mouse macrophages in the heart injured by AMI were found to secrete a large amount of miR-155 through exosomes. When mouse cardiac fibroblasts were treated with the injured macrophage-conditioned medium, the expression of son of sevenless 1 (SOS1) and suppressor of cytokine signaling 1 (SOCS1) was suppressed, and inflammation of cardiac fibroblasts was induced [50].

EV-derived lncRNAs: EV-mediated functions of lncRNAs and circRNAs are also being elucidated [28,51]. Metastasis-associated lung adenocarcinoma transcript 1 (MALAT1) is a representative multifunctional lncRNA that regulates various diseases and molecular mechanisms, and many studies have been conducted on CVDs [52-55]. MALAT1 expression is greatly increased in ECs by hypoxia, and silencing of MALAT1 enhances the migratory response of ECs and suppresses their proliferation [56]. Interestingly, plaque size and CD45⁺ cell infiltration were found to be increased in a MALAT1-deficient atherosclerosis mouse model (ApoE^-/- Malat1^-/-), and significantly lower expression of MALAT1 was observed in human plaques than in normal arteries [52]. These lines of evidence indicate that reduced levels of MALAT1 are associated with atherosclerotic lesion formation in mice and humans.

In addition, it was recently verified that a large amount of MALAT1 is distributed in the exosomes of HUVEC treated with oxidized LDL (oxLDL), a key risk factor for CVDs [57]. When monocytes were co-cultured with oxLDL-treated HUVECs, exosomal MALAT1 reduced the expression of pro-inflammatory markers (IL-12) and increased the expression of anti-inflammatory markers (CD206, Arg-1, IL-10) in macrophages. This was demonstrated by the reversal of the progression of M2 macrophage polarization in MALAT1-inhibited monocytes [57]. In summary, MALAT1 is a lncRNA that mediates the communication among several related vascular cells through exosomes in the pathophysiology of atherosclerosis.

Another lncRNA, growth arrest-specific 5 (GAS5), is also involved in intercellular mechanisms in vascular cells [58]. GAS5 inhibits SMC differentiation by competitively binding with SMAD3 protein and preventing SMAD3 from binding to the Smad-binding elements of the transforming growth factor (TGF)-β-responsive contractile gene promoter [58]. Interestingly, the expression of GAS5 was greatly increased in plaques of human patients with atherosclerosis and animal models, and its inhibition reduced apoptosis of oxLDL-treated THP-1 cells [59]. GAS5 was also found in the exosomes of oxLDL-treated THP-1 cells. When these exosomes were isolated and directly treated with HUVECs, apoptosis was enhanced [59]. These findings suggest that GAS5 regulates the apoptosis of macrophages and ECs through exosomes and can be an effective therapeutic target for atherosclerosis.

EV-derived circRNAs: Among the aforementioned regulatory ncRNAs, circRNAs have been identified most recently, and their functions are less known than those of miRNAs and lncRNAs. One of the well-known roles of circRNAs is as competing endogenous RNAs [60]. In particular, many circRNAs act as sponges for miRNAs [61,62]. CircNPHP4, a circRNA that is produced from the nephronophthisis 4 (NPHP4) gene locus and abundant in small extracellular vesicles (SEVs) of monocyte, is significantly upregulated in patients with coronary heart atherosclerotic disease [63]. Treatment of human coronary artery endothelial cells (HCAECs) with these monocyte-derived SEVs significantly enhanced heterogeneous adhesion between monocytes and HCAECs. CircNPHP4 acts as a sponge for miR-1231, which suppresses the protein expression of epidermal growth factor receptor in HCAECs [63].

Another circRNA, circUbe3a, has been studied in myocardial fibrosis caused by AMI [64]. CircUbe3a was increased and loaded into SEVs during increased M2 macrophage infiltration after AMI. This circRNA sponges miR-138-5p and then regulates the translation of the Ras homolog gene family member C (RhoC), the target of miR-138-5p [64]. Co-culture with M2 macrophages differentiated from bone marrow-derived macrophages through transwell chambers increased fibrosis markers (collagen I, collagen III, and α-SMA) and proliferation and migration of mouse primary cardiac fibroblasts [64].

A recent study revealed the role and mechanism of circHIPK3 in the regulation of cardiac senescence [65]. This study showed an interesting mechanism in which circHIPK3 regulates p21 expression by serving as a scaffold for HuR and E3 ubiquitin ligase β-TrCP. In circHIPK3-knockout mice, cardiac function is poor, and the telomere length in the heart is shortened [65]. However, aging and weakening of heart function were recovered in cardiomyocytes (mouse primary cardiomyocytes and H9C2 cells) and mice injected with exosomes derived from human umbilical cord mesenchymal stem cells, which have a high amount of circHIPK3 [65]. Another report has shown that circR-284 (another name for circHIPK3) contains a binding site for miR-221 and that the ratio of circHIPK3 to miR-221 in serum can be a diagnostic biomarker for carotid plaque rupture and stroke [66,67].

The studies of EV-derived ncRNAs associated with the CVDs described here and others that have been further investigated are summarized in Table 1. In the section below, several representative studies of the indirect effects of ncRNAs through the regulation of secreted factors are reviewed (summarized in Table 2).

Table 1 . Direct delivery of ncRNAs via extracellular vesicles.

ncRNA	Disease	Donor cell	Recipient cell	Co-culture system	Target in the recipient cell	Reference
miRNA
miR-143/145	AS	HUVEC	HASMC	Transwell co-culture	ELK1, KLF4, CAMK2d, CFL1, PHACTR4, SSH2, MMP3	[43]
miR-155	AS	HASMC	HCMEC	Transwell co-culture, exosome treatment	ZO-1	[49]
	AMI	Mouse BMDM	Mouse CF	Conditioned medium, exosome treatment	Sos1, Socs1	[50]
miR-221/222	AMI	ADSC	H9C2 cell	Conditioned medium	PUMA, ETS-1	[84]
				Exosome treatment		[85]
miR-21-3p	AS	RAW264.7 cell	Mouse VSMC	Transwell co-culture, exosome treatment	PTEN	[86]
miR-106a-3p	AS	THP-1 cell	Human VSMC	Transwell co-culture, exosome treatment	CASP9	[87]
lncRNA
MALAT1	AS	HUVEC /TERT2	THP-1 cell	Exosome treatment	Unknown	[57]
GAS5	AS	THP-1 cell	HUVEC	Exosome treatment	Unknown	[59]
circRNA
circNPHP4	CAD	PBMC-derived monocyte	HCAEC	Co-culture (monocyte adhesion assay), conditioned media, SEV treatment	miR-1231/EGFR	[63]
circUbe3a	AMI	Mouse BMDM	Mouse CF	Transwell co-culture, SEV treatment	miR-138-5p/RhoC	[64]
circHIPK3	Cardiac senescence	UMSC	Mouse CM, H9C2 cell	Exosome treatment	β-TrCP, HuR	[65,66]

ncRNA, non-coding RNA; miRNA, microRNA; lncRNA, long ncRNA; circRNA, circular RNA; AS, atherosclerosis; AMI, acute myocardial infarction; CAD, coronary heart atherosclerotic disease; HUVEC, human umbilical vein endothelial cell; BMDM, bone marrow-derived macrophage; ADSC, adipose-derived stem cell; HUVEC/TERT2, hTERT-immortalized HUVEC; PBMC, peripheral blood mononuclear cell; UMSC, umbilical cord mesenchymal stem cell; HASMC, human aortic smooth muscle cell; HCMEC, human cardiac microvascular endothelial cell; CF, cardiac fibroblast; VSMC, vascular smooth muscle cell; HCAEC, human coronary artery endothelial cell; CM, cardiomyocyte; SEV, small extracellular vesicle; ZO-1, zonula occludens-1; Sos1, son of sevenless homolog 1; Socs1, suppressor of cytokine signaling 1; PUMA, p53 upregulated modulator of apoptosis; ETS-1, ETS proto-oncogene 1; PTEN, phosphatase and tension homologue; CASP9, caspase-9; EGFR, epidermal growth factor receptor; RhoC, ras homolog gene family member C; β-TrCP, E3 ubiquitin ligase; HuR, human antigen R.

Table 2 . Indirect functions of ncRNAs via secreted factors.

ncRNA	Disease	Donor cell	Recipient cell	Co-culture system	Target in the donor cell	Secreted factors	Reference
miRNA
miR-146a, -708, -451,-98	AS	HAEC	HASMC	Transwell co-culture (flow chamber)	IRAK1, IKBKG, IL6R, IKBKA	NF-κB-related genes	[68]
let-7c	AS	Monocyte-derived DC	CD4+ T cell	Co-culture(96-well plate, FACS)	Unknown	TNF-α, IL-1β/6/10, TGF-β	[70]
miR-30-5p	AS	THP-1 cell	HUVEC	Transwell co-culture	TCF21	TNF-α	[88]
lncRNA
NEXN-AS1	AS	HUVEC	THP-1 cell	Co-culture (monocyte adhesion assay)	BAZ1A/ NEXN	MCP-1, ICAM-1, VCAM-1, TNF-α, IL-6, MMP-1/9	[72]
LINC00305	AS	THP-1 cell	HASMC	Co-culture	LIMR/ AHRR	IL-1β/8, TNF-α, MMP-9, CD14	[73]
VINAS (DEPDC4 in human)	AS	b.End.3 cell, HUVEC	Mouse PBMCs, THP-1 cell	Co-culture (monocyte adhesion assay)	Unknown	MCP-1, TNF-α, IL-1β, COX-2, VCAM-1, ICAM-1, E-selectin	[77]
lncRNA-MAP3K4	AS	b.End.3 cell	Mouse PBMCs	Co-culture (monocyte adhesion assay)	MAP3K4	MCP-1, TNF-α, IL-1β, COX-2, ICAM-1, E-selectin	[78]
circRNA
Hsa_circ_ 0087352	AAA	THP-1 cell	HA-VSMC	Transwell co-culture	miR-149-5p/IL6	IL-6/1B, TNFa	[79]

ncRNA, non-coding RNA; miRNA, microRNA; lncRNA, long ncRNA; circRNA, circular RNA; AS, atherosclerosis; AAA, abdominal aortic aneurysms; HAEC, human aortic endothelial cell; DC, dendritic cell; HUVEC, human umbilical vein endothelial cell; HASMC, human aortic smooth muscle cell; PBMC, peripheral blood mononuclear cell; HA-VSMC, human aortic-vascular smooth muscle cell; IRAK1, IL-1 receptor-associated kinase; IKBKG, NF-κB kinase subunit-γ; IL6R, IL-6 receptor; IKBKA, NF-κB kinase subunit-α; TCF21, transcription factor 21; BAZ1A, bromodomain adjacent to zinc finger domain protein 1A; NEXN, nexilin F-actin-binding protein; LIMR, lipocalin-1 interacting membrane receptor; AHRR, aryl-hydrocarbon receptor repressor; MAP3K4, mitogen-activated protein kinase kinase kinase 4; IL-6, interleukin-6; NF-κB, nuclear factor-κB; TNF-α, tumor necrosis factor-α; TGF-β, transforming growth factor-β; MCP-1, monocyte chemoattractant protein 1; ICAM-1, intercellular adhesion molecule 1; VCAM-1, vascular cell adhesion molecule 1; MMP-1, matrix metalloproteinase-1; COX-2, cyclooxygenase-2.

Indirect functions of ncRNAs via secreted factors

It has been reported that the synthetic SMCs converted in the stagnant state of the blood flow caused by atherosclerotic lesions lead to abnormal miRNA profiles of ECs [68]. Anti-inflammatory miRNAs (miR-146a, miR-708, miR-451, and miR-98) derived by shear stresses target NF-κB-related genes (IRAK1, IKBKG, IL6R, and IKBKA) in EC, which results in the relieve of inflammation of adjacent synthetic SMCs [68]. In this study, the co-culture of human aortic ECs and SMCs was established by plating each cell on the opposite side of a porous membrane and incorporating it into a parallel-plate flow chamber [39,68]. Furthermore, treatment with GW4869, a neutral sphingomyelinase inhibitor that suppresses the secretion of exosomal miRNAs, demonstrated that miRNAs in ECs were not directly delivered to regulate SMC inflammation [68,69].

Activated immune cells, including monocytes, macrophages, dendritic cells (DCs), and T cells, which actively produce factors such as cytokines, are the main cause of atherosclerotic inflammation [20]. Interestingly, the miRNA let-7c was increased in oxLDL-treated DCs, and its suppression reversed the oxLDL-induced increase in pro-inflammatory cytokines (TNF-α, IL-1β, and IL-6) and decrease in anti-inflammatory cytokines (TGF-β and IL-10) [70]. In addition, inhibition of let-7c limited the activation of T cells co-cultured with mature DCs and promoted the induction of T regulatory cells, which was confirmed by flow cytometry following the co-culture of these two cells [70].

A few years ago, a large-scale meta-analysis using the genome-wide association studies datasets showed the possibility that the lncRNA nexilin antisense RNA 1 (NEXN-AS1) may contribute to lung cancer susceptibility [71]. However, its detailed mechanism of action was reported for the first time in a study on arteriosclerosis [72]. NEXN-AS1 was decreased in human atherosclerotic plaques, and its increase in ECs inhibited Toll-like receptor 4 oligomerization and NF-κB activity and downregulated the expression of adhesion molecules (ICAM-1 and VCAM-1), inflammatory cytokines (MCP-1, TNF-α, and IL-6), and extracellular matrix-degrading enzymes (MMP-1 and MMP-9) [72]. These findings were confirmed by the co-culture of ECs and monocytes. Interestingly, NEXN-AS1 interacts with chromatin remodeler BAZ1A (bromodomain adjacent to zinc finger domain protein 1A) at the promoter region of NEXN to upregulate its expression [72]. This function of regulating the expression of nearby protein-coding genes on the same chromatin is a typical cis-acting mode of action of lncRNAs [28].

LINC00305 was identified as an atherosclerosis-related lncRNA from the genome-wide association studies database (GWASdb) [73]. Indeed, LINC00305 was more abundant in human peripheral blood mononuclear cell (PBMC)-derived CD14⁺ cells, THP-1 cells, and atherosclerotic plaques than in HASMCs and HUVECs [73]. Its overexpression promotes the expression of inflammatory genes (IL1B, IL8, TNF, MMP9, and CD14) in THP-1 cells and reduces the expression of contractile markers (CNN1, MYH11, and SMTN) in co-cultured HASMCs [73]. In this study, lipocalin-interacting membrane receptor (LIMR), a binding partner of LINC00305, was identified, and the mechanism of inflammation regulation by LIMR was reported for the first time [73,74]. LINC00305-LIMR binding enhances the interaction between LIMR and aryl hydrocarbon receptor repressor (AHRR), which cooperates with NF-κB in inflammation, promoting protein expression and nuclear localization of AHRR [73,75,76].

Another lncRNA, vascular inflammation and atherosclerosis lncRNA (VINAS), that modulates vascular inflammation was discovered through RNA sequencing profiling of atherosclerotic intimal lesions [77]. Interestingly, VINAS knockdown inhibited vascular wall inflammation and reduced atherosclerotic lesion formation by more than half in LDL receptor-deficient (LDLR^-/-) mice. This lncRNA regulated the expression of key inflammatory markers (MCP-1, TNF-α, IL-1β, and COX-2) and adhesion molecules (VCAM-1, ICAM-1, and E-selectin) by regulating NF-κB and MAPK signaling in ECs [77]. This effect was confirmed through the co-culture of mouse ECs (b.End.3 cells) and PBMC-derived monocytes, and this was also verified for a conserved human ortholog of VINAS, DEP domain containing 4 (DEPDC4), in the experiment using HUVECs and THP-1 cells [77]. The group that conducted this study recently discovered a new endothelial lncRNA called lncRNA-MAP3K4 that regulates vascular inflammation, which has similar functions to VINAS [78]. In addition, lncRNA-MAP3K4 shares a bidirectional promoter with the neighboring protein-coding gene MAP3K4 and, similar to NEXN-AS1, cis-regulates the expression of this neighboring gene. Interestingly, the three abovementioned lncRNAs NEXN-AS1, VINAS, and lncRNA-MAP3K4 are expressed not only in ECs but also in other atherosclerosis-related cells (SMCs and macrophages) and have some regulatory functions in inflammation, indicating their broad role in CVDs [72,77,78].

In the human specimen of abdominal aortic aneurysms (AAAs), a common CVD, the circRNA hsa_circ_0087352, produced from ubiquilin 1 (UBQLN1) gene locus, is highly expressed [79]. The progression of AAA is significantly influenced by the focal distribution of macrophages to the arterial wall. Recently, hsa_circ_0087352 has been found to be upregulated in LPS-induced THP-1 inflammatory macrophages [79]. Hsa_circ_0087352 sponges miR-149-5p promoting IL6 mRNA expression and increasing the secretion of inflammatory cytokine. In addition, overexpression of hsa_circ_0087352 in macrophages was found to induce apoptosis of human aortic-VSMCs in transwell co-cultures through induction of ERK/NF-κB signaling and release of apoptotic inflammatory cytokines (IL-6 and TNF-α) [79].

Recently, the diverse roles of many regulatory ncRNAs in CVDs have been extensively studied [26,80]. However, compared with studies of direct delivery through EVs, there are relatively fewer reports on indirect intercellular crosstalk mediated by ncRNAs [81,82]. miRNAs and circRNAs are found in large numbers in serum and secreted vesicles due to their structural stability [67,83]. This point shows the potential of extracellular ncRNAs as a predictive and diagnostic biomarker for various diseases. Measurement of ncRNA expression in blood, which is easier to obtain than lesional tissue samples, can enhance clinical accessibility limited by ethical issues. In addition, the regulation of intercellular crosstalk of ncRNAs demonstrates their broad role in vascular disease. Along with the ncRNAs discussed in detail above, we briefly describe additional ncRNAs in Tables 1 and 2, which we hope will be helpful to researchers [84-88]. The study of the pathophysiological mechanisms of ncRNA utilizing co-culture systems discussed in this review will play an important role in obtaining reliable results in future experiments using animal models and clinical samples.

Figs. 1 and 2 were created using BioRender (https://biorender.com/).

This study was supported by grants from the Basic Science Research Program (NRF-2021R1A2B5B02001501, NRF-2022R1A4A2000767, NRF-2022M3A9E4017151, NRF-2021R1A6A3A13044540, and NRF-2022R1I1A1A01070056) of the National Research Foundation of Korea funded by the Korean government (MSIT). The funders had no role in the literature collection and analysis, decision to publish, or preparation of the manuscript.

The authors declare no conflicts of interest.