Role of flow-sensitive microRNAs in endothelial dysfunction and atherosclerosis – “Mechanosensitive Athero-miRs”

Abstract

Atherosclerosis preferentially occurs in arterial regions exposed to disturbed flow, in part, due to alterations in gene expression. MicroRNAs (miRNAs) are small, noncoding genes that post-transcriptionally regulate gene expression by targeting messenger RNA transcripts. Emerging evidence indicates that alteration of flow conditions regulate expression of miRNAs in endothelial cells both in vitro and in vivo. These flow-sensitive microRNAs, known as “mechano-miRs”, regulate endothelial gene expression, and can regulate endothelial dysfunction and atherosclerosis. MiRNAs such as, miR-10a, -19a, -23b, -17~92, -21, -663, -92a, -143/145, -101, -126, -712, 205, and -155, have been identified as mechano-miRs. Many of these miRNAs were initially identified as flow-sensitive in vitro and were later found to play a critical role in endothelial function and/or atherosclerosis in vivo through either gain-of-function or loss-of-function approaches. The key signaling pathways that are targeted by these mechano-miRs include the endothelial cell cycle, inflammation, apoptosis, and nitric oxide signaling. Furthermore, we have recently shown that the miR-712/205 family, which is upregulated by disturbed flow, contributes to endothelial inflammation and vascular hyper-permeability by targeting tissue inhibitor of metalloproteinase-3 (TIMP3), which regulates metalloproteinases (MMPs) and a disintegrin and metalloproteinases (ADAMs). The mechano-miRs that are implicated in atherosclerosis are termed as “mechanosensitive athero-miRs” and are potential therapeutic targets to prevent or treat atherosclerosis. This review summarizes the current knowledge of mechanosensitive athero-miRs and their role in vascular biology and atherosclerosis.

Introduction

Atherosclerosis is a chronic inflammatory disease of the vascular wall which leads to cardiovascular pathologies such as myocardial infarction, ischemic stroke, and peripheral arterial disease. It is a leading cause of morbidity and mortality in developed countries^1–3. Despite the presence of systemic risk factors such as hypercholesterolemia, hypertension and diabetes, atherosclerosis preferentially occurs in arterial regions exposed to disturbed flow (d-flow), whereas the arterial regions exposed to stable flow (s-flow) remain healthy. Recent studies have demonstrated that d-flow induces atherosclerosis in hypercholesterolemic conditions in mouse models⁴. D-flow, which typically occurs in branched or curved arteries, is characterized by complex flow patterns with low-magnitude and oscillatory shear stress (OS), whereas stable (s-flow) is characterized by high-magnitude, unidirectional laminar shear stress (LS) in straight sections of the vasculature.

The mechanisms by which flow affects endothelial function and atherosclerosis have been reviewed previously^5–9 and are also the subject of the accompanying reports in this ATVB Focus issue. Briefly, d-flow induces and s-flow prevents endothelial dysfunction and atherosclerosis, respectively, in part due to alterations in gene expression and the epigenetic landscape^10–14. Vascular endothelial cells respond to blood flow through mechanosensors, which transduce the mechanical force associated with flow (also known as shear stress) into cell signaling events and ultimately, changes in gene expression^{5, 8, 15}. D-flow promotes and s-flow suppresses atherogenesis, respectively, in part, through differential regulation of pro-atherogenic and atheroprotective genes^{10, 16, 17}. S-flow upregulates atheroprotective genes such as Klf2¹⁸, Klf4¹⁹, and eNOS²⁰, whereas d-flow induces a number of pro-atherogenic genes. These include vascular cell adhesion molecule-1 (VCAM-1)^{4, 10}, bone morphogenic protein-4 (BMP4), and matrix metalloproteinases (MMPs)²¹, which mediate pro- angiogenic, -inflammatory, -thrombogenic, and -apoptotic responses^{5, 22, 23}.

Small non-coding RNAs, microRNAs (miRNAs), have been implicated in the regulation of gene expression primarily as post-transcriptional repressors²⁴. miRNAs interact with the 3’untranslated region (UTR) of specific target mRNAs in a sequence-specific manner, resulting in mRNA degradation or translational inhibition²⁵. It has been previously demonstrated that LS and OS differentially regulate the expression of miRNAs in endothelial cells. The majority of flow-sensitive miRNAs (mechano-miRs) has been identified and characterized using cultured endothelial cells that were subjected to LS or OS conditions. Of these, a select few were subsequently validated in vivo. Therefore, only a limited number of direct linkages to atherogenesis have been established. Here, we review and briefly summarize our current knowledge of mechano-miRs and their role in endothelial dysfunction and atherosclerosis.

1. Mechano-miRs and their effect on endothelium and atherosclerosis

Emerging evidence indicates that mechano-miRs are responsible for the overall flow-dependent control of endothelial dysfunction and atherosclerosis. Current literature divides mechano-miRs into three classes according to their putative pro- and/or anti-atherogenic effects: (1) anti-atherogenic mechano-miRs, (2) pro-atherogenic mechano-miRs and (3) dual-role mechano-miRs (Table 1). Anti-atherogenic mechano-miRs include miR-10a, 19a, 23b, 101, 143 and 145, all of which are either upregulated by s-flow/LS or are downregulated by d-flow/OS. Pro-atherogenic mechano-miRs include miR-17~92, 92a, 663, 712, and 205, all of which are either upregulated by d-flow/OS or are downregulated by s-flow/LS. Dual-role mechano-miRs include miR-21, 155, and 126. Dual-role mechano-miRs have been implicated in either pro- or anti-atherogenic events, but with inconsistent flow-sensitivity. Table 1 also shows the flow/shear dependent responses of each mechano-miR, the experimental system in which the miRNA was studied, its validated target genes, and its role in endothelial dysfunction and atherosclerosis. Figure 1 shows a short summary of the role of mechanosensitive miRNAs tested in atherosclerosis in vivo.

Table 1.

Role of mechano-miRs in endothelial (EC) dysfunction and atherosclerosis

Atherogenic effect		miRNA	Shear	miRNA Targets	Indirect targets & Affected signaling pathways	EC dysfunction and Atherosclerosis	Exp. System	Ref.
Anti-atherogenic Mechano-miR		10a	d-flow ↓	MAP3K7 β-TRC	VCAM-1, E-selectin and NFκB signaling ↓	Anti-inflammatory; Inhibit EC NFκB activation	in vitro shear pig AA	27
		19a	LS ↑	cyclin D1	Cell cycle arrest (G1/S)	Inhibits EC proliferation	in vitro shear	28, 29
		23b	LS ↑	E2F1	Rb hypo-phosphorylation ↑	Inhibits EC proliferation	in vitro shear	29
		101	LS ↑	mTOR	Cell cycle arrest (G1/S)	Inhibits EC proliferation	in vitro shear	30
		143 145	LS ↑	ELK1, KLF4 CAMK2d SSH2 PHACTR4 FL1	Atheroprotective SMC phenotype ↑	miR-143/145 inhibits atherosclerosis in ApoE−/− mice.	in vitro shear in vivo ApoE−/− mouse model	32
		145	NT	KLF4	myocardin ↑, VSMC contractile phenotype ↑	SMC-targeted miR-145 inhibits atherosclerosis	in vivo ApoE−/− mouse model	87
Pro- atherogenic Mechano-miR		17–92	LS ↓	NT	NT	NT	In vitro shear	29
		92a	OS ↑ d-flow ↑	KLF2 KLF4 ITGA5 SIRT1	EC inflammation	Anti-miR-92a inhibits EC inflammation and atherosclerosis	In vitro shear in vivo ApoE−/− mouse model	33, 34, 88
		663	OS ↑	NT		EC inflammation	in vitro flow system	35
		712 205	d-flow ↑ OS ↑	TIMP3	MMPs activity ↑	Anti-miR-712 prevents EC inflammation and atherosclerosis Anti-miR-205 inhibits EC inflammation	in vitro shear in vivo partial carotid ligation model	26, 86
Dual role Mechano-miR	Pro- atherogenic	21	OS ↑	PPARα	Activates AP-1, monocyte adhesion	EC inflammation	in vitro shear	89
	Pro- atherogenic	21	NT	PTEN Bcl2		miR-21−/− inhibits neointimal formation in vein graft model	In vivo vein graft model (mouse, porcine)	90
	Anti- atherogenic	21	LS ↑	PTEN	eNOS phosphorylation and NO production ↑	Increases NO bioavailability and decreases EC apoptosis	in vitro shear	53
	Pro- atherogenic	155	NT	Bcl6	NFκB activation	Promotes atherosclerosis	In vivo partial carotid ligation model	62
	Pro- atherogenic	155	NT	SOCS-1	Cholesterol efflux ↓, pro-inflammatory cytokine (TNFα, IL-6, IL-1β) ↑	miR-155−/− inhibits inflammation and atherosclerosis	in vivo ApoE−/− mouse model	63
	Anti- atherogenic	155	LS ↑	MYLK	RhoA ↑	Inhibits EC migration and proliferation	in vitro shear mouse AA	60
	Anti- atherogenic	155	NT	NT	Regulates monocyte subset	BMT with miR-155 knockout destabilizes plaque and stimulates atherosclerosis	in vivo ApoE−/− mouse model	61
	Pro- atherogenic	Secretion of 126-3p	LS ↓	FOXO3 BCL2 IRS1	SMC turnover	miR-126−/− inhibits neointimal lesion formation	In vitro shear in vivo ligation-induced neointimal model	57
	Anti- atherogenic	126-5p	d-flow ↑	Dlk1	EC repair and proliferation (G2/M) ↑	Induces EC proliferation and inhibits atherosclerosis	in vivo partial carotid ligation model	58

EC: endothelial cell; LS, laminar shear stress; OS, oscillatory shear stress; NT, not tested; d-flow, disturbed flow; KLF2 and 4, Krȕppel-like factor 2 and 4; ITGA5, integrin alpha 5; SIRT1, Sirtuin 1 (Silent Mating Type Information Regulation 2); FOSB, Finkel–Biskis–Jinkins (FBJ) murine osteosarcoma viral oncogene homolog B; SLC7A5, Solute carrier family 7 member 5; NAV2, neuron navigator; TIMP3, tissue inhibitor of metalloproteinase 3; PPARα, Peroxisome proliferator-activated receptor alpha; PTEN, Phosphatase and tensin homolog; BCL6, B-cell lymphoma 6; SOCS-1, Suppressor of cytokine signaling 1, MYLK, Myosin light chain kinase; FOXO3, Forkhead box O3, BCL2, B-cell lymphoma 2; IRS1, insulin receptor substrate 1; Dlk1, delta-like 1 homolog; MAP3K7, mitogen-activated protein kinase kinase kinase 7; β-TRC, β -transducin repeat-containing gene; AA : aortic arch; mTOR, Mammalian target of rapamycin; CAMK2d, calcium/calmodulin-dependent protein kinase type II delta chain; SSH2, slingshot homolog 2; PHACTR4, phosphatase and actin regulator 4; FL1, Follicular lymphoma, susceptibility to 1; MMP, matrix metalloproteinase; Rb, retinoblastoma protein; SMC, smooth muscle cell;

Figure 1. Role of mechanosensitive miRNAs in atherosclerosis.

The role of the following mechano-miRs in atherosclerosis has been tested in vivo. Expression of miR143/145 is in increased under LS condition in a KLF2-dependent pathway. Endothelial-derived miR-143/145 is transported to the medial SMCs in extracellular vesicles. This vesicle-mediated delivery of miR-143/145 had an antiatherogenic effect by preventing SMC de-differentiation. Low shear stress upregulates miR-92a and induces endothelial inflammation and atherosclerosis by co-regulating KLF2 and KLF4. miR-126-5p is downregulated by d-flow and inhibits EC proliferation and atherosclerosis through upregulation of Dlk1. Expression of miR-10a was lower in the athero-susceptible regions of the inner aortic arch as compared to other regions in the porcine arterial system. miR-10a induces a proinflammatory endothelial phenotype by targeting MAP3K7 and β-TRC. D-flow induces expression of miR-712 in an XRN1-dependent manner. D-flow also induces expression of miR-205. Both miR-712 and miR-205 target TIMP3. Loss of TIMP3 leads to the activation of MMPs and ADAMs, which result in endothelial inflammation and hyper-permeability and ultimately, atherosclerosis. miR-712 and miR-205 were also increased in circulation and were also increased in the medial smooth muscle cells and circulating immune cells where they affected SMC migration and leukocyte-EC interactions, respectively, further contributing to atherosclerosis. Systemic treatment with anti-miR-712 prevented atherosclerosis in murine models of atherosclerosis. miR, microRNA; EV, extracellular vesicle; d-flow, disturbed blood flow; s-flow, stable flow; EC, endothelial cell; ECM, extracellular matrix; VSMC, vascular smooth muscle cell; MØ, macrophage; SSH2, Slingshot Protein Phosphatase 2; PHACTR4, Phosphatase And Actin Regulator 4; CAMK2d, Calcium/Calmodulin-Dependent Protein Kinase II Delta; XRN1, exoribonuclease-1.

Shu Chien and colleagues were the first to report mechano-miRs (miR-19a and 23b) in cultured endothelial cells in vitro, whereas Peter Davies and colleagues reported miR-10a as the first mechano-miR identified directly from the endothelium in vivo. Although the majority of mechano-miRs have been identified using endothelial cells in vitro, it is important to validate them in vivo since numerous mechanosensitive genes identified in vivo are known to be either dysregulated or lost during endothelial cell culture^{10, 26}. Currently, only a few studies have examined miRNA expression profiles in vivo, in part, due to technical difficulties collecting endothelial-specific RNAs in these studies^27,32.

While most of the mechanosensitive miRNAs have been identified from in vitro studies using cone-and-plate viscometers or flow chambers, miR-712 and miR-10a are two mechanosensitive miRNAs that have been identified directly from in vivo studies. The gene expression of cells in vivo is intricately regulated by intracellular signaling components, paracrine factors from neighboring cells, and circulating humoral factors. Although cells may continue to survive and expand in vitro, the gene and miRNA expression profiles are altered during the process of adaptation from the in vivo environment to the culture dish. This becomes evident upon careful review of various reports comparing the gene expression of newly extracted cells to cells expanded ex vivo, including hematopoietic progenitor cells, adipocytes, neutrophils, NK Cells, and other hematopoietic cells^43–52. This phenomenon has been observed in endothelial cells as well, however, there are limited studies comparing the gene and/or microRNA expression changes between newly isolated endothelial cells and established endothelial cell lines^{10, 27}. To study the effect of shear stress on gene expression changes in endothelial cells, it is extremely important to correlate the in vitro and in vivo gene and miRNA expression changes. Therefore, methodologies to directly isolate ECs from arterial regions exposed to d-flow or s-flow provide direct insight into the mechanosensitivity of a particular gene or miRNA. Whether these mechanosensitive genes or miRNAs retain or lose their response to shear stress after adapting to in vitro culture and how the sensitivity is lost or dysregulated is important to validate whether in vitro findings are relevant to that of in vivo.

1) Anti-atherogenic mechano-miRs

These miRNAs were either increased by s-flow/LS or decreased by d-flow/OS in endothelial cells and were shown to be anti-atherogenic (Table 1).

miR-10a

MiR-10a expression in the endothelium is decreased by d-flow in the athero-prone, lesser curvature region of the porcine aortic arch as compared to that of the athero-protected, thoracic aorta²⁷. Mechanistically, miR-10a is an anti-inflammatory miRNA that inhibits NFκB activation by targeting MAP3K7 and βTRC, both of which promote IκB degradation and p65 translocation²⁷.

miR-19a and miR-23b

Initial miRNA expression analysis studies used cultured human umbilical vein endothelial cells (HUVECs) to determine the flow-sensitivity of miRNAs²⁸. Following 12 h of LS (12 dyn/cm²), 35 miRs were upregulated and 26 miRs were downregulated as compared to the static control cells. Among these, LS increased expression of miR-19a, which targets cyclin D1, thereby inducing endothelial quiescence²⁸. Later studies identified 8 upregulated and 13 downregulated miRNAs in response to 24 h of pulsatile LS (12±4 dyn/cm²) as compared to the static no-flow condition²⁹. One of the upregulated miRNAs was miR-23b, which suppressed endothelial proliferation by reducing E2F1 expression and Rb phosphorylation²⁹.

miR-101 and miR-143/145

MiR-101 is upregulated in response to LS and was reported to target the mTOR gene, thus inducing cell cycle arrest³⁰. MiR-143/145 levels were increased by LS in an AMPKα2-dependent manner³¹. A subsequent study showed that LS increased the expression of miR-143/145 by a KLF2-dependent pathway³². Furthermore, it was demonstrated that endothelial-derived miR-143/145 can be transferred to medial smooth muscle cells (SMCs) via extracellular vesicles and prevent atherogenesis by preventing SMC de-differentiation³².

2) Pro-atherogenic mechano-miRs

These miRNAs were either increased by d-flow/OS or decreased by s-flow/LS in endothelial cells and were shown to induce endothelial dysfunction and pro-atherogenic responses.

miR-17~92 and miR-92a

The miR-17~92a cluster comprises several miRs, including miR-17, 18a, 19a 19b, 20a and 92a. The miR-17~92 cluster is regulated by shear stress in that some members (miR-17, miR-19b, miR-20a, miR-92a) were downregulated by 24 h pulsatile LS²⁹. Subsequent studies showed that miR-92a was downregulated by LS and was upregulated by OS³³. These in vitro findings are consistent with in vivo studies showing that ECs in the athero-susceptible porcine aortic arch have increased miR-92a levels as compared to those of the athero-resistant thoracic aorta³³. Further studies demonstrated that miR-92a induced endothelial inflammation by targeting KLF2 and KLF4³⁴.

miR-663

From a miRNA array study using HUVECs exposed to 24 h LS or OS, miR-663 was identified as one of the most highly upregulated mechano-miR in OS conditions³⁵. Furthermore, miR-663 overexpression induced endothelial inflammation, thus suggesting its potential pro-atherogenic role³⁵. Consistent with this report, miR-663 was upregulated in HUVECs exposed to pro-atherogenic oxidized phospholipids, and was found to play a permissive role in the induction of VEGF and activation of ATF4 branch of unfolded protein response in ECs^{36, 37}. Interestingly, miR-663 in vascular smooth muscle cells induced a contractile phenotype and forced overexpression of this human-specific mechano-miR with an adenoviral construct reduced neointimal formation in mice with complete carotid ligation³⁸. This shows that miR-663 may have cell type-dependent responses which ultimately determine its role in atherosclerosis.

miR-712 and miR-205 family

Recently, two mechano-miRs, miR-712 and miR-205, were identified from a mouse model of flow-induced atherosclerosis, known as the partial carotid ligation (PCL) model⁴. In this model, 3 out of 4 of the caudal branches of the left common carotid artery (LCA) are ligated, thus inducing d-flow along the entire length of the straight section of the artery. As the contralateral right common carotid artery (RCA) is left undisturbed, it continues to be exposed to s-flow^{4, 39}. This model directly demonstrates the causal relationship between d-flow and atherosclerosis, as the LCA develops atherosclerosis within two weeks of surgery, while the RCA remains plaque-free. Additionally, by carefully extracting endothelial-enriched RNA from mouse carotid arteries following ligation, we were able to identify more than 500 mechanosensitive genes^{10, 40–42}. Using the same mouse model, two mechano-miRs, miR-712 and 205, were identified and their function was established in vivo in two different animal models of atherosclerosis²⁶. We found that d-flow induced expression of miR-712, and that it contributed to endothelial dysfunction, and thus atherosclerosis. Additionally, treatment with anti-miR-712 prevented plaque development. The pro-inflammatory and pro-atherogenic effects of miR-712 were mediated by its target, tissue inhibitor of metalloproteinase-3 (TIMP3). Loss of TIMP3 by miR-712 resulted in activation of matrix metalloproteinases (MMPs) and a disintegrin and metalloproteinases (ADAMs), ultimately leading to endothelial inflammation, hyper-permeability, and atherosclerosis²⁶. Interestingly, miR-712 is murine-specific, and miR-205 is a homolog in human and other vertebrates²⁶. MiR-205 (1) shares the same “seed sequence” with miR-712, (2) its expression is upregulated by d-flow, and (3) it also targets TIMP3 in a flow-dependent manner²⁶.

3) Mechano-miRs with dual role in atherosclerosis

Some mechano-miRs, such as miR-21, 126, and 155, have been implicated in both anti- and pro-atherogenic responses. This may reflect the fact that a single miRNA can target numerous target mRNAs, some of which mediate pro-atherogenic responses while others act in an opposite manner. The overall response of the cell depends on cellular context, cell type, or environment³².

miR-21

miR-21 was upregulated by LS in HUVECs and prevented apoptosis by targeting PTEN⁵³. However, another study showed that OS upregulated miR-21 in a time-dependent manner inhibited peroxisome-proliferator-activated receptor-α (PPARα), thus leading to the enhanced expression of the pro-inflammatory VCAM-1⁵⁴. It is interesting that miR-21 was upregulated by both LS and OS when compared to static conditions, suggesting that the mechano-sensitivity of miR-21 is a transient adaptive response to imposed shear. Interestingly, miR-21 was increased in arterial endothelium exposed to d-flow as compared to that of s-flow in the mouse PCL model^{4, 39}. Whether miR-21 plays a pro- or anti-atherogenic role still remains to be determined.

miR-126

The effect of flow on the expression of miR-126 and its role in atherosclerosis is also currently confusing. MiR-126 (also referred to as miR-126-3p) and miR-126* (also known as miR-126-5p) are highly expressed in endothelial cells (ECs), and have been shown to regulate vascular integrity and angiogenesis⁵⁵, as well as inflammation⁵⁶. The secretion of miR-126-3p into conditioned media, but not its intracellular expression per se, was decreased by LS and increased by OS, respectively, in HUVEC⁵⁷. Endothelial-derived miR-126-3p regulated smooth muscle cell (SMC) turnover in an EC-SMC co-culture system. Furthermore, genetic knockout of miR-126 inhibited neointimal formation in a complete carotid ligation model, while local reintroduction of miR-126 in the knockout mice enhanced neointimal formation. This suggests that miR-126 is a pro-atherogenic miRNA. In contrast, a recent study showed that d-flow decreased expression of both miR-126-5p and 126-3p, however treatment with miR-126-5p mimic, and not miR-126-3p, reduced atherosclerotic lesion formation. The anti-atherogenic effect of miR-126-5p was mediated by targeting the Notch1 inhibitor, delta-like 1 homolog (Dlk1), which promoted the proliferative potential of endothelial cells⁵⁸. Given the well-known effect of d-flow on increased endothelial proliferation in athero-prone regions, however, the underlying mechanisms by which miR-126-5p inhibited atherosclerosis remains to be seen. Another interesting report from the same group showed that miR-126-3p present in apoptotic bodies had an atheroprotective effect in a murine carotid cuff model of atherosclerosis⁵⁹. Together, these results suggest that the role of miR-126-5p and miR-126-3p still needs further clarification.

miR-155

MiR-155 expression is increased by LS in HUVECs and is abundantly expressed in the intima of the thoracic aorta, which is naturally exposed to s-flow in vivo, thus suggesting it is an anti-atherogenic mechano-miR⁶⁰. Consistent with this idea, hematopoietic deficiency of miR-155 induced atherosclerosis and decreased plaque stability by increasing myeloid inflammatory cell recruitment to the plaque regions⁶¹. However, other studies showed evidence suggesting that miR-155 mediates pro-atherogenic responses^{62, 63}. MiR-155 was shown to directly target eNOS mRNA in HUVECs and impair endothelium-dependent vascular relaxation in human arteries⁶⁴. Leukocyte-specific miR-155 deficiency reduced plaque size and lesional macrophage count in the PCL model of atherosclerosis⁶². Another study showed that genetic knockdown of miR-155 ameliorated atherogenesis in ApoE^−/− mice by reducing inflammatory responses of macrophages and increasing macrophage cholesterol efflux⁶³. Also, tissue specific genetic knockdown of miR-155 in bone marrow-derived cells suppressed atherogenesis in ApoE^−/− mice⁶³, demonstrating its role as pro-atherogenic miRNA.

2. Role of non-mechanosensitive athero-miRs

From early endothelial dysfunction to fully developed atherosclerosis, there are multiple cell types in the vessel wall that play a critical role in the development and progression of atherosclerosis. These include endothelial cells, medial smooth muscle cells, adventitial fibroblasts and adipocytes, platelets, and immune cells. Additionally, lipid metabolism, which regulates the circulating levels of triglycerides, cholesterol and oxidized-low-density-lipoprotein (LDL), also affects the process of atherogenesis. MicroRNAs have been shown to regulate vascular inflammation and atherosclerosis by affecting various cell types in the vasculature. Several studies have recently shown the importance of specific miRNAs in atherosclerosis that are not necessarily flow-sensitive (non-mechanosensitive athero-miRs). These non-mechanosensitive miRNAs can be divided into two classes according to their putative pro- or anti-atherogenic roles: 1) anti-athero-miRs and (2) pro-athero-miRs (Table 2).

Table 2.

Non-mechanosensitive athero-miRs that induce inflammation and atherosclerosis

Atherogenic Effect	miRNA	miR- mimic or inhibitor	Major Targets	Indirect targets & Affected signaling pathways	Pathophysiological Effects	Exp. System	Ref.
Anti-atherogenic Non-mechano-miR	1	anti-miR	MLCK	ERK phosphorylation	Antago-miR-1 enhances endothelial permeability	in vivo high-fat diet-induced EC permeability	65
	30c	Lentiviral miR-30c	MTP	Lipid synthesis and ApoB production	anti-miR-30c increases hyperlipidemia and atherosclerosis. miR-30c reduces atherosclerosis	ApoE−/− mouse model	70
	144	mimic	ABCA1	HDL	Anti-miR-144 increases hepatic ABCA1 and HDL cholesterol	ApoE−/− mouse model	72
	146a	mimic	HuR	NFκB and MAP kinase	Represses NFκB pathway	in vitro EC inflammation	66
	181b	mimic	importin-α3	EC NFκB signaling	miR-181b inhibits endothelial inflammation and atherosclerosis	ApoE−/− mouse model	67, 68
	195	Adenoviral miR-195	CDC42, CCND1, FGF1	IL-1β, IL-6 and IL-8	miR-195 prevents neointimal formation	Balloon-injury rat neointimal model	69
	467b	mimic	LPL	Lipid accumulation and cytokine secretion (IL-6, IL-1β, TNF-α, MCP-1)	miR-467b reduces plasma cholesterol and atherosclerosis.AntagomiR-467b increases atherosclerosis.	ApoE−/− mouse model	71
Pro-atherogenic Non-mechano-miR	342-5p	anti-miR	Akt1	Nos2	Macrophage-miR-342-5p stimulates inflammation and atherosclerosis	in vivo partial carotid ligation model	73
	33	miR-33−/− ApoE−/−	ABCA1 ABCG1	HDL-C levels (cholesterol efflux)	miR-33−/− knockout inhibits atherosclerosis	ApoE−/− mouse model	76
	33	anti-miR	ABCA1	HDL, cholesterol transport	Anti-miR-33 inhibits inflammation and atherosclerosis	in vivo atherosclerosis regression model	74
	33	anti-miR	ABCA1	HDL, cholesterol efflux	Anti-miR-33 inhibits atherosclerosis	ApoE−/− mouse model	77
	33	anti-miR	ABCA1, CROT,CPT1A, HADHB	HDL, VLDL triglyceride	anti-miR-33 increases hepatic ABCA1	Non-human primate model(African green monkey)	75
No effect on atherosclerosis (in this study)	33	anti-miR	ABCA1	Triglyceride, HDL level was not maintained for long-term experiment	Long-term anti-miR-33 treatment has no effect on atherosclerosis	ApoE−/− mouse model	78

NT, not tested; ABCA1, ATP-binding cassette transporter A1; RIP140, ABCG1, ATP-Binding Cassette Sub-Family G Member 1; CROT, carnitine O-octanoyltransferase; CPT1A, Carnitine Palmitoyltransferase 1A; HADHB, hydroxyacyl-CoA dehydrogenase/3-ketoacyl-CoA thiolase/enoyl-CoA hydratase (trifunctional protein), beta subunit; MLCK, Myosin light-chain kinase; MTP, Microsomal triaglyceride transfer protein; HuR, human antigen R; CDC42, Cell division control protein 42 homolog; CCND1, Cyclin D1; FGF1, Fibroblast growth factor 1; LPL, Lipoprotein lipase; HDL, High-density lipoprotein; LDL, low-density lipoprotein; VLDL, very low-density lipoprotein; NOS2, Nitric oxide synthase 2.

1) Anti-athero-miRs

The following miRNAs have been experimentally validated as anti-atherogenic miRNAs through in vitro and in vivo studies.

miR-1, 146a, 181b, and 195

Inhibition of miR-1 using antago-miR enhanced the endothelial hyper-permeability effect seen under hyper-cholesterolemic conditions and overexpression using miR-1 mimic prevented endothelial barrier dysfunction⁶⁵. MiR-146 negatively regulate inflammation in that overexpression of miR-146a blunts endothelial activation, whereas in vitro inhibition of miR-146a/b or deletion of miR-146a in mice aggravates endothelial inflammation⁶⁶. Mechanistically, miR-146 represses the NFκB and MAP kinase pathways through downregulation of HuR, which promotes endothelial activation in the vessel wall⁶⁶. Similarly, systemic delivery of miR-181b reduced atherosclerosis through inhibition its target gene, importin-α3, which mediates NFκB nuclear translocation specifically in the vascular endothelium^{67, 68}. Also, adenoviral delivery of miR-195 reduced neointimal formation in a balloon-injury carotid artery model by inhibiting VSMC proliferation, migration, and IL-1β, IL-6, and IL-8 synthesis⁶⁹.

miR-30c, 144, and 467b

These three miRs play an important role in regulation of plasma cholesterol level, a critical risk factor of atherosclerosis. Overexpression of miR-30c in the liver reduced the hyperlipidemia seen in Western diet-fed mice by decreasing lipid synthesis and the secretion of triglyceride-rich ApoB-containing lipoproteins, which in turn prevented atherosclerosis in ApoE^−/− mice⁷⁰. Also, miR-467b reduces plasma cholesterol and atherosclerosis. Specifically, miR-467b inhibition by using antago-miR-467 enhanced the progression of atherosclerosis by increasing lipid accumulation and inflammatory cytokine secretion⁷¹. Additionally, miR-144 regulates hepatic expression of the ABCA1 protein, thereby high density lipoprotien (HDL)-cholesterol as well, but its contribution to the development of atherosclerosis is unknown⁷².

2) Pro-athero-miRs

miR-342

Macrophage-derived miR-342-5p increases atherosclerosis and inflammatory stimulation of macrophages through inhibition of Akt1⁷³. Suppression of Akt1 by miR-342-5p indirectly induces Nos2 and IL-6 in macrophages through miR-155. Therefore, systemic miR-342-5p inhibition reduced atherosclerotic plaque development in the aorta of ApoE^−/− mice.

miR-33

MiR-33 plays a critical role in the suppression of inflammation and atherosclerosis. Multiple studies show that silencing of miR-33 prevents inflammation and atherosclerosis^74–77. Anti-miR-33 treatment or genetic knockout of miR-33 in mice promotes reverse cholesterol transport and inhibition of atherosclerosis through regulation of ABCA1^{74, 76, 78}. Also, anti-miR-33 increased hepatic expression of ABCA1 and induces a sustained increase in plasma HDL in a non-human primate model⁷⁵, thus validating its anti-atherogenic role. However, a recent report showed that long-term silencing of miR-33 with anti-miR-33 failed to demonstrate the anti-atherogenic effects in LDL receptor knockout mice⁷⁸. The inconsistent responses reported in these studies may reflect a different target genes and roles of miR-33 in various cells and tissue types. While miR-33 may control its major gene target ABCA1 in some cell types such as hepatocytes, it may regulate other less well-characterized target genes or off-targets in other cell types leading to complex responses in vivo. Thus, these observations raise an important concern regarding the off-target effects of systemically delivered miRNAs, anti-miRs, or antago-miRs in animal studies and clinical use.

3. Circulating miRNAs and their transport via vesicle-dependent and independent pathways

miRNAs serve as messengers between the ECs and SMCs and are critical for homeostasis and maintenance of the vasculature. Alterations in EC–SMC communication have been implicated in the pathogenesis of atherosclerosis as well as in the formation of aneurysms. While knowledge of circulating miRNAs in cardiovascular disease is emerging, fundamental questions still remain as to which cells produced them by which mechanisms and how they act on the ECs and SMCs as well as other cell types. In a previous report, extracellular vesicles (EVs) enriched in miR-143/145 are secreted in response to overexpression of the mechanosensitive transcription factor KLF2 in HUVECs. It was also shown that endothelial-derived miR-143/145 can be transferred to SMCs via extracellular vesicles and prevent atherogenesis by preventing SMC de-differentiation³². In another study, miR-126 was shown to be secreted by ECs, and transported via RNA and RNA-protein complexes, independently of vesicular or DNA components, This was followed by uptake by co-cultured SMCs, which promoted SMC proliferation, cell cycle progression, and apoptosis⁵⁷. Exposure of endothelial cells to laminar shear stress inhibits miR-126 secretion, which in turn abrogates the atherogenic actions of miR-126 on the co-cultured SMCs. However, the mechanisms that control the specific loading of miRNA species into extracellular vesicles (exosomes) remain unknown. Recently, the heterogeneous nuclear ribonucleoprotein A2B1 (hnRNPA2B1) was identified as a miRNA sorting protein that specifically binds exosomal miRNAs through the recognition of specific motifs and controls their loading into the vesicles⁷⁹. We recently reported that miR-712 and miR-205 are increased in medial SMCs and circulating immune cells, which serve to regulate SMC migration and leukocyte-EC interaction, respectively, but whether the miRNAs are transported via a vesicle-dependent or -independent mechanism, is not yet clear. Nevertheless, these reports suggest that circulating miRNAs, regardless of the mechanism of cellular export, can regulate communication between endothelial cells and SMCs, which in turn can control the overall outcome on endothelial inflammation and atherosclerosis.

4. Biogenesis of mechano-miRs: Atypical pathway

Typically, miRNAs are generated as primary miRNA (pri-miRNAs) transcripts by RNA polymerase II from genes⁸⁰. These pri-miRNA transcripts are initially processed within the nucleus by the RNase III enzyme Drosha and its cofactor DGCR8 into small 60- to 100-nt-long hairpin pre-miRs^81–83. Pre-miRNAs are then exported into the cytosol where they are further processed into mature miRNAs by Dicer enzyme. Mature miRNAs are then incorporated into the miRNA-induced silencing complex (miR-RISC) where they bind to the complementary 3’-untranslated region of the respective target gene(s), resulting in mRNA decrease or translation inhibition⁸⁴. In addition to the aforementioned canonical biogenesis pathway, some atypical miRNAs, such as mirtrons and simtrons can be generated by non-canonical pathways that do not require the microprocessor components Drosha and DGCR8. Mirtrons represent very short introns of genes which form pre-miRNAs directly after splicing, thus bypassing Drosha. whereas simtrons are splicing machinery-independent mirtron-like miRNAs⁸⁵.

Although most mechano-miRs are regulated by the canonical pathway, some miRNAs, such as miR-712 and miR-663, are regulated by atypical pathways. Recently, we found that the murine-specific miR-712 and the human specific miR-663 are the most highly upregulated miRNAs in conditions of d-flow, as determined from two completely independent array studies^{26, 35}. These intriguing results suggested that there may be a common mechanism between the two mechanosensitive miRNAs. We have found that both miRNAs are synthesized from an unexpected source, pre-ribosomal RNA, derived from the RN45S genes. Human and mouse RN45s genes are present in at least 5 different chromosomes, each containing 30 to 40 repeats, and they are not officially annotated yet. A RN45S unit is polycistronic, coding for 3 rRNAs with two intervening internal transcribed spacer (ITS) sequences called ITS1 and ITS2 (18S rRNA - ITS1 - 5.8S rRNA - ITS2 - 28S rRNA). Through sequence alignments, we found that miR-712 is derived from the ITS2 region of the murine RN45s gene, while miR-663 is derived from the ITS1 region of the human RN45s gene²⁶. This is the first report which demonstrates that miRNAs can be derived from the spacer regions of the rRNAs. Interestingly, these spacer regions of ribosomal RNA contain species-specific nucleotide sequences that are used as phylogenetic markers. Typically under normal conditions, these spacer regions are rapidly degraded by exoribonuclease-1 (XRN1), following the transcription of rRNAs. However, d-flow decreases the expression of XRN1, but not the canonical miRNA processors DGCR8/DROSHA, in both human and mouse endothelial cells²⁶. This indicates that there may be a common role of XRN1 in the expression of human miR-663 and murine miR-712. Knockdown of XRN1 significantly increased both miR-712 and miR-663 expression, while knockdown of DGCR8 did not affect their expression. These results indicate that miR-712 induction by d-flow is dependent on XRN1 but independent of the canonical miRNA processors. At present, it is not known whether XRN1 is directly or indirectly responsible for biogenesis of these atypical miRNAs and this need to be determined. In addition to flow, miR-712 expression is also regulated by Angiotensin-II (AngII), suggesting that miR-712 and possibly other ITS-derived miRNAs such as miR-663 can be regulated by nonmechanical, humoral stimuli as well⁸⁶. Since, two of the most flow-sensitive mechano-miRs are produced from the ITS regions of the ribosomal RNA gene in a similar atypical mechanism, it would be interesting to determine whether there are more miRNAs generated in this manner and the general role of miRNAs processed in such a manner.

5. Summary and perspectives

Here, we summarized the role of flow-sensitive miRNAs (mechano-miRs) in the regulation of endothelial function and atherosclerosis (athero-miRs). Taken together, these results demonstrate that mechano-miRs are crucial mediators of endothelial function, and atherosclerosis. However, our knowledge of miRNAs, especially mechano-miRs and athero-miRs, their regulation, and their functions is still in its infancy. Further investigation is required to identify other RNAs and explore their biological roles, mechanisms, and potential therapeutic applications. There have been some promising results from recent Phase II clinical trials indicating the safety, feasibility, and therapeutic potential of anti-miRs for the treatment of hepatitis. In animal studies, some anti-miRs targeting mechano-miRs such as miR-712, miR-205, or miR-155, and athero-miRs such as miR-33 have demonstrated their potential as anti-atherogenic therapies. This encourages the development of miRNA therapeutics for the treatment of atherosclerosis in humans. Despite these encouraging results, the development of miRNA therapeutics must overcome several major challenges. First, the current methods to inhibit miRNAs include anti-miR/antagomiR and miR-sponge, both of which directly bind to miRNAs, thereby affecting all of their target genes (~ hundreds of genes) indiscriminately, potentially causing undesirable effects. Therefore, to minimize the potentially undesirable effects of miRNA inhibition, better strategies should be developed to deliver these inhibitors either specifically to cells of interest or to design more specific inhibitors that can specifically block a unique but desired mRNA-miRNA interaction without affecting the expression of off-target genes. Second, targeting ubiquitously expressed miRNAs will lead to unintended side effects in other cells or tissues. Therefore, the use of cell type–specific delivery strategies such as nanotechnology-based platforms should be explored. Finally, the biology of miRNAs is still emerging and identification of specific target genes of miRNAs in each cell type and organs continues to be a major challenge in understanding the role of miRNAs. In summary, expression of miRNAs is robustly regulated by different flow conditions in endothelial cells. These mechano-miRs are critical regulators of vascular function and atherosclerosis, thereby serving as diagnostic and therapeutic targets of the vascular disease.