Emerging Roles for MicroRNAs in Diabetic Microvascular Disease: Novel Targets for Therapy

ESSENTIAL POINTS

• Impaired microvascular function is a hallmark during the development of insulin resistance and in diabetic subjects

• MicroRNAs, small noncoding RNAs that fine-tune gene expression, have emerged as critical regulators in insulin-responsive tissues and cell types important for maintaining vascular homeostasis and preventing the sequelae of diabetes-induced end organ injury

• Current pathophysiological paradigms of microRNAs and their target genes demonstrate novel mechanisms and therapeutic targets in regulating the diabetic microvasculature in diabetes-associated complications such as retinopathy, nephropathy, wound healing, and myocardial injury

• Circulating microRNAs in exosomes or microvesicles may be diagnostically useful in type I or type II diabetes

• Emerging delivery platforms for manipulating microRNA expression or function may serve as the next frontier in therapeutic intervention to improve diabetes-associated microvascular dysfunction and its attendant clinical consequences

Abstract

Chronic, low-grade systemic inflammation and impaired microvascular function are critical hallmarks in the development of insulin resistance. Accordingly, insulin resistance is a major risk factor for type 2 diabetes and cardiovascular disease. Accumulating studies demonstrate that restoration of impaired function of the diabetic macro- and microvasculature may ameliorate a range of cardiovascular disease states and diabetes-associated complications. In this review, we focus on the emerging role of microRNAs (miRNAs), noncoding RNAs that fine-tune target gene expression and signaling pathways, in insulin-responsive tissues and cell types important for maintaining optimal vascular homeostasis and preventing the sequelae of diabetes-induced end organ injury. We highlight current pathophysiological paradigms of miRNAs and their targets involved in regulating the diabetic microvasculature in a range of diabetes-associated complications such as retinopathy, nephropathy, wound healing, and myocardial injury. We provide an update of the potential use of circulating miRNAs diagnostically in type I or type II diabetes. Finally, we discuss emerging delivery platforms for manipulating miRNA expression or function as the next frontier in therapeutic intervention to improve diabetes-associated microvascular dysfunction and its attendant clinical consequences.

I. Introduction

Diabetic subjects exhibit both accelerated macro- and microvasculature disease, an effect that significantly contributes to cardiovascular disease and diabetic-associated complications such as retinopathy, nephropathy, and neuropathy. Insulin resistance, a major risk factor for both type 2 diabetes and ischemic cardiovascular disease, is a common pathologic state caused by decreased sensitivity to insulin, a decrease in maximal response to insulin, or a combination of both (1, 2). It is central to the pathogenesis of type 2 diabetes and helps to maintain the diabetic state (3–7).

The recent recognition of microRNAs (miRNAs) as important mediators in the diabetic macro- and microvasculature has provided unique molecular and cellular insights into their impact on pathophysiological pathways in insulin resistance, type 2 diabetes, and obesity. Moreover, the appreciation that miRNAs can be detected extracellularly, such as in circulating blood or urine, raises the potential for their use as biomarkers for diagnosis and prognosis or in response to diabetic or cardiovascular therapeutics.

Herein, we focus on the emerging pathophysiological roles of miRNAs, evolutionarily conserved small noncoding RNAs of ∼22 nucleotides that fine-tune target gene expression and signaling pathways in insulin-responsive tissues and cell types important for maintaining optimal vascular homeostasis and preventing the complications of diabetes-induced end organ injury. We highlight current paradigms of miRNAs and their target genes involved in regulating the diabetic microvasculature in a range of diabetes-associated sequelae such as retinopathy, nephropathy, wound healing, and myocardial injury. We provide an update of the potential use of circulating miRNAs diagnostically in type 1 or type 2 diabetes. Finally, we discuss emerging delivery platforms for manipulation of miRNA expression or function as the next therapeutic frontier to improve diabetes-associated microvascular dysfunction and its attendant clinical consequences. We also review the potential challenges and opportunities for miRNAs as therapeutics in diabetic vascular disease.

II. Biogenesis and Function of miRNAs

MiRNAs are one class of small, evolutionarily conserved noncoding RNAs of ∼22 nucleotides in length that are involved in nearly all developmental and pathophysiological processes in animals as they target over one-half of protein-coding transcripts (8, 9). The first miRNA, Caenorhabditis elegans lineage-4 (lin-4), was identified in 1981 (10) and was cloned in 1993 (11), followed by lethal-7 (let-7) (12). miRNAs are initially transcribed by RNA polymerase II in the nucleus to form long RNA transcripts, termed primary-miRNA (pri-miRNA), with the stem-loop structure in which the mature miRNA sequences are embedded (13, 14) (Fig. 1). Pri-miRNAs contain the cap structure and poly(A) tails, which are the unique properties of class II gene transcripts (15). The nuclear ribonuclease (RNase)-III enzymes, Drosha, cleaves pri-miRNA into hairpin-shaped RNA of approximately 70 base pairs precursor-miRNAs (premiRNA), which consist of an imperfect stem-loop structure (16). Drosha is a nonspecific RNase; to mediate the genesis of miRNAs from the pri-miRNA, it forms a protein complex of approximately 600 kDa, termed microprocessor, with the essential cofactor DGCR8, also known as Pasha; a protein containing two double-stranded RNA (dsRNA)-binding domains (17–19). The specific cleavage at the stem-loop of the hairpin RNA structure of pri-miRNA is directed by DGCR8 (19). The microprocessor complex is also crucial for the posttranscriptional cross-regulation between Drosha and DGCR8 (20). DGCR8 stabilizes Drosha through an interaction between the middle domain of Drosha and the conserved C-terminal domain of DGCR8 (20, 21), whereas Drosha decreases DGCR8 mRNA level by cleaving it at a hairpin in the second exon (20, 22) (Fig. 1).

Figure 1.

The biogenesis of miRNA begins with the transcription of miRNA to pri-miRNA in the nucleus. The pri-miRNA is cleaved by the Drosha/DGCR8 complex to form premiRNA, which can then be exported from the nucleus to the cytoplasm by exportin5. In the cytoplasm, premiRNA is further processed into the mature miRNA duplex by Dicer. One strand (usually the guide strand) of the mature miRNA forms a complex with Dicer and the Argonaute protein, known as the miRNA-containing RISC, where miRNA binds to the 3′ UTR of its target mRNA, resulting in the degradation (if the miRNA:mRNA duplex complementarity is perfect) or suppression of the translation of the target mRNA (if the complementarity is not perfect).

Following the initial cleavage by Drosha, the premiRNAs are exported by a Ran-guanosine triphosphate–dependent nucleo/cytoplasmic cargo transporter named exportin 5 (Exp5) from the nucleus to the cytoplasm, where the maturation of miRNA takes place (23). Exp5 was originally known as a minor export transporter for transfer RNAs (tRNAs) (24). Recent evidence indicated that the affinity of Exp5 for premiRNA is much higher than that for tRNA (23). As individual miRNAs are abundantly present at up to 50,000 copies per cell in some organisms (25), premiRNAs are probably the main cargo for this transporter compared with other RNAs, including short hairpin RNAs and tRNAs (23, 26). Exp5 also protects nuclear premiRNA from degradation by forming a complex with premiRNA and Ran-guanosine triphosphate together (27). Once in the cytoplasm, the premiRNA is cleaved by another RNase-III enzyme, Dicer, into a small, imperfect dsRNA duplex that contains both the mature miRNA strand and its complementary strand (28–30). Dicer is an RNase III–type endonuclease of ∼200 kDa (31); it is highly conserved in almost all eukaryotic organisms such as plants and animals (14). It cleaves at a site that is about 21 to 28 nucleotides (nt), depending on the species and the type of Dicer (32), from the 3′ end of the terminus of dsRNAs (the 3′-counting rule) (31, 33, 34). In mammals, Dicer also binds to the 5′ phosphorylated end of the premiRNA and cleaves it at about 22 nt from the 5′ end (the 5′-counting rule) (35). After cleavage by Dicer, the resulted 21 to 24 nt miRNA duplex needs to be unwound to release one of the strands for entering Argonaute for function (36, 37) (Fig. 1). Usually, one strand (“passenger” strand, also known as the star strand or “-5-p”) of this short-lived duplex disappears, whereas the other strand (“guide” strand or “-3p”) becomes the mature miRNA (38). It was originally thought that the passenger strands are subject to rapid degradation (37, 39, 40). However, recently, studies found that these strands could also be loaded on Argonaute and directed repression of target mRNAs (41, 42).

In animals, most miRNAs are incorporated into the RNA-induced silencing complex (RISC) (43) to form miRNA-containing RISC, which is also known as miRNA-containing ribonucleoprotein complex or mirgonaute (14, 38). The miRNA directs the RISC to the 3′ untranslated region (UTR) of the target mRNA to negatively regulate the expression of the latter at the posttranscriptional level by promoting mRNA cleavage (44, 45) and at the translational level by serving as a translational repressor to inhibit the translation of protein from mRNA (46), or a combination of the two (47). The initial evidence for miRNA targeting 3′ UTR of RNA came from the observation that a part of lin-4 sequence was complementary with the 3′ UTR of lin-14 mRNA (11,48) and is required for the repression of lin-14 by lin-4 (49). Recent studies have reported that, in animals, the 7 nt “seed region,” mapping to positions 2 to 8 at the 5′ end, of miRNA (50, 51) plays a key role on its target recognition and is highly conserved both within and between species (47, 52). When there is a perfectly complementary sequence between the miRNA and its target, the target mRNA is cleaved (53–55). However, in mammals, this phenomenon is rarely observed; the sequences of miRNA and its target are only partially complementary (53). In this case, mRNA is deadenylated and then decapped (56–59). An in vivo study showed that the decapped mRNAs were degraded (60). However, in cultured cell extracts, deadenylated mRNAs are not further degraded but remain in the deadenylated state, followed by translational repression (56).

III. miRNAs Regulating Insulin-Signaling Pathways and the Microvasculature

Under homeostatic conditions, cellular signaling by insulin is initiated by binding to insulin receptors (61, 62). The insulin receptor is a cell surface integral membrane protein expressed on most mammalian cells such as liver, adipose, and skeletal muscle cells (63–65). Insulin exerts its effect on three main target tissues: fat, liver, and skeletal muscle (66). Binding of insulin triggers oligomerization and receptor autophosphorylation on tyrosine residues (67–69) and tyrosine phosphorylation of insulin receptor substrates (IRSs) IRS1, IRS2, IRS3, IRS4, IRS5/DOK4, and IRS6/DOK5 (70–74). This phosphorylation provides the basis for the subsequent association with downstream signal proteins that diverge into three different pathways: the phosphoinositide-3-kinase (PI3K)/Akt pathway, the CAP/Cbl/TC10 pathway, and the mitogen-activated protein-kinase (MAPK)–dependent pathway (75–77).

PI3K interacts with phosphorylated tyrosines on IRS molecules (78, 79), resulting in increased levels of phosphatidyl inositol-4,5-diphosphate to phosphatidyl inositol-3,4,5-triphosphate (80, 81) followed by activation of phosphoinositide-dependent kinase 1 (82), which further phosphorylates protein kinase B (PKB) and atypical protein kinase C (83–85). Phosphorylation of PKB inactivates glycogen synthase kinase 3 (GSK-3), which results in glycogen synthesis (86). PKB also mediates stimulation of the fatty acid synthase promotor (87), inhibition of phosphoenolpyruvate carboxykinase gene transcription (88), and translocation of the main glucose transporter GLUT4 to the plasma membrane (89). In addition, PKB regulates downstream protein synthesis and degradation, such as 4E-BP1 and eIF-4G, which is mediated by phosphorylation and activation of mammalian target of rapamycin (mTOR) (90, 91).

For glucose uptake to be fully manifested, proto-oncogene Cbl in the second pathway is recruited to the phosphorylated insulin receptor by interaction with the adapter protein CAP (92, 93), which finally results in reinforced GLUT4 translocation in parallel with but not affecting PI3K signaling in response to insulin (76, 93). The third pathway after IRS activation includes GRB2, SHP2, and several other proteins, which lead to mitogen-activated protein-kinase activation, cellular proliferation, and differentiation via gene transcription regulation (77, 94–96).

A. Visceral fat

It is well recognized that human subjects with generalized obesity suffer from a high risk of insulin resistance and its metabolic complications, such as type 2 diabetes mellitus and hypertriglyceridemia (97). Inflammation in adipose tissue, as suggested by the presence of macrophages in the form of crown-like structures, has been identified as a mediator of systemic insulin resistance (98). Obese human subjects with excess visceral adipose tissue, or abdominal obesity, are at higher risk for type 2 diabetes components than those whose fat is located predominantly in the lower body, subcutaneously (99, 100). Only a few miRNAs have been reported to directly regulate adipocytes and visceral adipose tissue during insulin resistance (Table 1).

Table 1.

List of miRNAs Involved in Insulin-Relevant Tissues

miRNA	Target	Function	Reference
Liver
miR-103	–	Promotes adipogenesis in hepatocytes	(108)
miR-122	PTP1B	Dephosphorylates tyrosine residues on the IR and IRS	(290)
miR-143	–	Impairs insulin-stimulated Akt activation and glucose homeostasis	(291)
miR-145	–	Inhibits glucose uptake in HepG2 cells, diminishes the phosphorylation of Akt and IRS1, and induces insulin resistance in hepatocytes	(292)
miR-190b	IGF-1	Induces insulin resistance	(293)
miR-200a	–	Increases the expression level of leptin receptor and IRS2, reduces body weight gain, and restores liver insulin responsiveness	(294)
miR-26a	–	Improves insulin sensitivity, decreases hepatic glucose production, and decreases fatty acid synthesis	(295)
miR-379	–	Increases VLDL-associated triglyceride	(296)
miR-802	Hnf1b	Impairs glucose tolerance and attenuates insulin sensitivity	(297)
VAT
miR-222	ER-α	Reduces GLUT4	(298)
miR-26b	PTEN	Promotes insulin-stimulated glucose uptake and increases insulin-stimulated GLUT4 translocation	(299)
miR-181b	PHLPP2	Reduces visceral fat inflammation and improves glucose homeostasis and insulin sensitivity by increasing adipose tissue endothelial insulin signaling and Akt/eNOS activation	(101)
Skeletal muscle
miR125b	IGF2	Regulates PI3K/Akt or MAPK pathway	(300, 301)
miR-135a	IRS-2	Inhibits PI3K/Akt pathway and glucose uptake	(302)
miR-182	FoxO3	Attenuates atrophy-related gene expression	(303)
miR-199a-3p	IGF, PIK3r1, and mTOR	Activates AKT/mTOR signal pathway	(304)
miR-21	TGFβI	Promotes PI3K/Akt/mTOR signaling	(305)
miR-24	p38	Inhibits activation of p38 MAPK pathway	(306)
miR-29a	IRS-1	Impairs Akt/GSK-3 pathway and glucose uptake	(307)

Although alterations in adipocytes and macrophage accumulation have been clearly associated with insulin resistance (98), the role of the microvasculature (i.e., endothelial cells [ECs]) in visceral fat in regulating insulin resistance is poorly understood. A recent study by our group identified that in response to high-fat-diet–induced insulin resistance, the anti-inflammatory miRNA miR-181b (101, 103) is significantly reduced in adipose tissue ECs, but not adipocytes, from visceral fat (epididymal white adipose tissue [eWAT]) after just one week. This finding raised the possibility that “miR-181b replacement therapy” may ameliorate inflammation in visceral fat and potentially insulin resistance. Indeed, high-fat-diet–induced, insulin-resistant mice that were tail vein injected with liposomally encapsulated miR-181b exhibited markedly improved glucose homeostasis and insulin sensitivity compared with controls (101). Mechanistically, miR-181b overexpression in adipose ECs enhanced insulin-mediated Akt phosphorylation at Ser473, and reduced endothelial dysfunction, an effect that shifted macrophage polarization toward an M2 anti-inflammatory phenotype in eWAT. These effects were associated with induction of endothelial nitric oxide synthase (eNOS), nitric oxide (NO) activity, and FoxO1 phosphorylation specifically in eWAT, but not in liver or skeletal muscle. Importantly, miR-181b overexpression in peripheral blood mononuclear cells had no effect on macrophage activation, proliferation, or recruitment to visceral fat. Similarly, adipocytes overexpressing miR-181b had no direct effect on glucose uptake; however, when conditioned medium from ECs overexpressing miR-181b was added to adipocyte cultures, these adipocytes were subsequently rendered less insulin resistant with improved glucose uptake. Bioinformatics and gene profiling studies revealed that the target of miR-181b was plexkstrin homology domain leucine-rich repeat protein (PHLPP2), a phosphatase known to dephosphorylate Akt at Ser473. Indeed, small interfering RNA (siRNA)-mediated knockdown of PHLPP2 recapitulated miR-181b’s favorable effects on glucose homeostasis, insulin sensitivity, and inflammation of eWAT. Interestingly, PHLPP2 is expressed robustly in eWAT, whereas it is barely detectable in liver and skeletal muscle, which may underlie the miR-181b tissue-specific effect in eWAT. Finally, these findings have translational significance, as ECs from diabetic subjects expressed higher PHLPP2 expression compared with healthy controls. Collectively, these findings highlight that miR-181b plays a critical homeostatic role in the microvasculature of visceral fat to control EC inflammation and insulin resistance (Fig. 2).

Figure 2.

miR-181b improves endothelial dysfunction and consequently improves glucose homeostasis and insulin sensitivity in white adipose tissue by targeting the phosphatase PHLPP2 and reducing its expression. Reduced PHLPP2 enhances insulin signaling and AKT and eNOS expression.

B. Liver and skeletal muscle

Insulin resistance is one of the key components of the metabolic syndrome, and it eventually leads to the development of type 2 diabetes (103, 104). During the development of insulin resistance, hepatic lipid accumulation and hepatic inflammation may be observed in liver (104). In addition, skeletal muscle is quantitatively an important site of insulin resistance in type 2 diabetic patients and a direct risk factor for the development of cardiovascular disease (105, 106). More recently, several miRNAs have been reported to regulate hepatic and skeletal muscle function in type 2 diabetes (107–110) (Table 1; Fig. 3). Although some of these miRNAs, such as miR-143, miR-26, miR-29, and miR-103, have been implicated in regulating cell types important in the vessel wall such as ECs, a definitive role for these miRNAs in the microvasculature of these relevant tissues has not been elucidated, but may provide future pathophysiological insights and targets for therapy.

Figure 3.

miRNA regulation of insulin signaling and glucose homeostasis. Many miRNAs have been identified to regulate insulin signaling in adipose tissue, liver, and skeletal muscle. miR-181b targets the phosphatase PHLPP2 in ECs of adipose tissue. miR-122, miR-103, miR-802, miR-143, and miR-26a regulate insulin signaling in the liver by regulating the expression of PTP1B, Cav-1, Hnf1b, ORP8, ACSL3, and PKCδ. miR-503, miR-125b, and miR-135a act in skeletal muscle on targets, including EFNB2 and VEGFA, IGF2, and IRS2, respectively. Endothelial-enriched miRNAs are highlighted in red.

IV. Role of miRNAs Involved in Diabetes-Induced Vascular Dysfunction

It is known that type 2 diabetes is associated with an increased risk of both micro- and macrovascular complications (111). In the macrovasculature, hyperglycemia can regulate multiple proinflammatory and proatherosclerotic target genes in ECs, vascular smooth muscle cells (VSMCs), and macrophages (112, 113). The morbidity associated with diabetic microvasculature disease, including retinopathy (114), neuropathy (115), nephropathy (116), and limb ischemia, is staggering (117). miRNAs have been implicated in the epigenetic regulation of key metabolic, inflammatory, and antiangiogenic pathways in type 2 diabetes and may contribute to common disease complications (118) (Table 2).

Table 2.

List of miRNAs Involved in Diabetes-Associated Vascular Dysfunction

miRNA	Target	Function	Reference
Macrovasculature
let-7a/b	–	Prevents atherosclerosis	(138)
miR-1	–	Prevents endothelial dysfunction	(141)
miR-125b	Suv39h1	Promotes the inflammatory reaction	(156)
miR-126	Spred-1	Promotes EPC proliferation and migration	(136)
miR-143/145	–	Increases VSMC cell area and reduces cell proliferation	(153)
miR-144-3p	ABCA1	Promotes THP-1 macrophage differentiation into foam cells	(308)
miR-146a	IRAK1, IL-6	Alters insulin sensitivity	(131)
miR-155	p65	Contributes to the progression of atherosclerosis	(134)
	–	Promotes macrophage-derived foam cell formation	(164)
miR-181b	Importin-α3 PHLPP2 Card10	Confers anti-inflammatory effects in the macro- and microvasculature by regulating NF-κB and insulin signaling	(101, 102, 143, 144)
miR-195	SIRT1	Increases expression of FN	(309)
miR-200	Zeb1	Induces MCP-1 and COX-2 expression and promotes monocyte binding	(157)
miR-206	–	Enhances cholesterol efflux	(164)
miR-21	SPI	Stimulates VSMC proliferation and reduces VSMC differentiation marker gene expression	(155)
miR-24	–	Reduces intraplaque macrophage proliferation	(165)
miR-492	resistin	Reduces insulin resistance and endothelial dysfunction	(310)
miR-9	ACAT1	Decreases THP-1 macrophage-derived foam cell formation	(311)
miR-92a	MKK4 and JNK	Inhibits oxidative stress–induced VSMC apoptosis	(312)
Microvasculature
miR-130a	Runx3	Maintains the normal EPC function	(172)
miR-135a	TRPC1	Improves renal function and ameliorates progression	(313)
miR-146	CARD10 HuR	Inhibits NF-κB activation	(176)
miR-146a	IRAK1	Reduces IL-1β–induced ICAM-1 expression	(198)
miR-146a	FN	Inhibits thrombosis	(127)
miR-146a	IRAK1 TRAF6	Promotes the complications of diabetic sensory neuropathy	(198)
miR-195	–	Prevents SIRT1-mediated tissue damage	(309)
miR-200b	VEGF	Prevents the glucose-induced increased vascular permeability and angiogenesis	(169)
miR-21	DAXX	Prevents endothelial apoptosis	(171)
miR-29b	–	Abolishes TGF-β/Smad3 pathway activation	(314)
miR-29c	Spry1	Inhibition of miR-29c reduces mesangial matrix accumulation and albuminuria	(186)
miR-320	–	Inhibits EC migration and tube formation	(174)
miR-34a	–	Alleviates glomerular hypertrophy	(315)
miR-93	Msk2/VEGF-A	MiR-93 overexpression reduces features of diabetic nephropahy	(187, 188)

Hyperglycemia and insulin resistance have been identified as key players in the development of diabetic atherosclerosis, with metabolic insulin signaling being an important contributor to normal vascular function and homeostasis (119). Abnormalities in vascular EC and VSMC function and macrophage activation are important contributors to vascular complications and propensity to both macro- and microvascular dysfunction (120).

A. Role of miRNAs in endothelial dysfunction and insulin resistance/type 2 diabetes

Both insulin resistance and endothelial dysfunction are associated with increased circulating markers of inflammation (121). In physiological conditions, insulin promotes endothelium-dependent relaxation by a mechanism that involves an increase of NO production via activation of PI3K and Akt kinase pathways (110, 122). However, in pathological conditions, insulin resistance reduced bioavailability of NO production via inhibition of eNOS activation (123) and eNOS expression (124), an effect that may be mediated by TNF-α-induced NF-κB signaling pathway (123, 125). Indeed, genetic inhibition of NF-κB signaling in ECs protects mice from the development of obesity- and aging-associated insulin resistance (126). This is accompanied by decreased macrophage infiltration into adipose tissue and oxidative stress in plasma, and increased blood flow, mitochondrial content, and insulin delivery in muscle (126). These favorable metabolic phenotypes are partially explained by increased eNOS expression. Emerging studies highlight important roles for miRNA-mediated regulation of endothelial dysfunction and diabetes (Table 2).

For example, miR-146a expression was reduced in both high glucose–treated ECs and streptozotocin (STZ)-induced diabetic rats (127, 128). Further studies demonstrated that miR-146a lowered ICAM-1 expression by targeting IL-1 receptor-associated kinase (IRAK1) (129) and by inhibiting the inflammatory signaling via binding to IL-6 (130), which alters insulin sensitivity and contributes to the progression of atherosclerosis (131). miR-155 expression was increased in type 2 diabetic mice (132) and impaired endothelium-dependent vasorelaxation by directly targeting the 3′ UTR of eNOS (133). However, Wu et al. found that TNF-α–induced miR-155 serves as a negative feedback regulator in endothelial inflammation and atherosclerosis by targeting the nuclear transcription factor p65 (134). Moreover, clinical studies showed that expression of miR-126 was reduced in type 2 diabetic patients with coronary artery disease, and type 1 diabetic subjects with reduced miR-126 expression was associated with higher micro- and macro-vascular complications (118, 135–137). miR-126 promoted endothelial progenitor cell (EPC) proliferation and migration and inhibited apoptosis via targeting Spred-1, an intracellular inhibitor of the Ras/extracellular signal-regulated kinase (ERK) cascade, which may influence the PI3K/Akt/eNOS signaling pathway (136), an important insulin-signaling pathway. In addition, overexpression of let-7a and let-7b inhibited oxidized-low-density-lipoprotein–induced EC apoptosis, NO deficiency, and eNOS downregulation and attenuated oxidized-low-density-lipoprotein–induced p38MAPK phosphorylation, NF-κB nuclear translocation, IκB degradation, and PKB dephosphorylation. These findings may provide new insights into the protective properties of let-7a and let-7b in preventing atherosclerosis and potentially insulin resistance (138).

Overexpression of miR-492 reduced insulin resistance and endothelial dysfunction induced by high glucose both in vitro and in vivo by directly reducing the expression of resistin, which also plays an important role in the development of both atherosclerosis and insulin resistance (139, 140). Feng et al. (127) reported that miR-1 was reduced in STZ-induced diabetic rats and, at the same time, endothelin-1 (ET-1) expression, a key marker indicating endothelial dysfunction was increased. Furthermore, transfection of miR-1 mimics inhibited the high glucose­–mediated induction of ET-1 in human umbilical vein endothelial cells (HUVECs), suggesting that miR-1 prevents endothelial dysfunction induced by hyperglycemia (141). In addition, induction of miR-195 after high glucose treatment increased expression of fibronectin (FN) via targeting SIRT1, a regulator of aging in ECs, indicating that miR-195 has the potential to promote thrombosis via induction of FN (142).

Finally, the miRNA miR-181b not only regulates vascular inflammation in the microvasculature as highlighted above in adipose tissue ECs, but also suppresses EC inflammation in the macrovascualture (102, 143–146). Indeed, systemic delivery of miR-181b markedly reduced atherosclerotic lesion formation independent of any effects on lipid profiles and reduced arterial thrombosis in mice (143, 144) (Table 2).

B. Role of miRNAs in VSMC function and insulin resistance/type 2 diabetes

VSMCs are also targets of insulin, an effect that bears clinical importance for vasculoproliferative states, both macrovascular (e.g., coronary artery disease and intimal hyperplastic restenosis to microvascular disease (e.g., myocardial arteriogenesis, retinopathy) (147, 148). Indeed, compared with nondiabetic VSMCs, human diabetic VSMCs exhibit abnormal morphology in culture, increased proliferation, and greater adhesion and migration rates (147). Through the PI3-K pathway, insulin stimulates glucose transport, induces the well-differentiated contractile state, antagonizes the effects of platelet-derived growth factor, and increases NO production (149). When the PI3-K pathway is inhibited, insulin increases chemotaxis (150) and VSMC proliferation via the MAPK pathway (151). Moreover, by the cooperation of the two pathways, insulin activates the HIF/VEGF pathway (152). Accumulating studies highlight several VSMC-enriched miRNAs that contribute to vascular dysfunction in diabetes (Table 2).

For example, miR-143 and miR-145, two upregulated miRNAs in VSMCs isolated from type 2 diabetic patients, increased VSMC area and reduced cell proliferation. Transfection of anti–miR-143/145 conferred type 2 diabetes-VSMC characteristics into a nondiabetic phenotype, implying that miR-143/145 impaired VSMC phenotype and function, suggesting a new therapeutic intervention of vascular function in type 2 diabetic patients (153). Another miRNA, miR-21, may also be important in regulating VSMC proliferation. Insulin resistance led to a significant reduction of miR-21 expression (154), and overexpression of miR-21 stimulated VSMC proliferation and reduced VSMC differentiation marker gene expression by targeting specificity protein-1 (155) (Table 2).

Finally, miR-125b and miR-200 have been implicated in modulating VSMC inflammatory effects. In VSMCs from diabetic db/db mice, both miR-125b and miR-200 were upregulated compared with nondiabetics cells (156, 157). Via targeting histone methyltransferase Suv39h1, miR-125b increased the expression of MCP-1 and IL-6 and enhanced monocyte binding to promote inflammation in db/db mice (156). Moreover, overexpression of miR-200 mimics also induced MCP-1 and COX-2 expression and promoted monocyte adhesion via binding Zeb1, an E-box binding transcriptional repressor. Interestingly, silencing Zeb1 increased miR-200 expression, indicating a feedback regulatory loop during this process (157). Collectively, therapeutic silencing of miR-143/miR-145, miR-21, miR-125b, or miR-200 may limit VSMC-associated proliferative states or VSMC-derived inflammation important to a range of diabetic macro- or microvascular complications.

C. Role of miRNAs in macrophage function and insulin resistance/type 2 diabetes

Abundant studies indicate that monocytes are recruited to peripheral tissues (i.e., pancreas, liver, and adipose tissue) to become resident macrophages and contribute to local inflammation and development of insulin resistance (158). In the state of insulin resistance, there is a distinct switch, termed macrophage polarization (159), in the macrophage populations present from an anti-inflammatory (M2) population to an inflammatory (M1) population, which releases cytokines and chemotactic factors that exacerbate and amplify the local inflammatory environment, thereby creating a vicious cycle (160). miRNAs figure prominently in regulating macrophage activation and polarization dynamics (Table 2).

For example, through targeting the adenosine triphosphate–binding cassette transporter A1 (ABCA1), miR-144-3p mimics increased the secretion of inflammatory cytokines such as IL-6, IL-1β, and TNF-α to promote THP-1 macrophage differentiation into proinflammatory foam cells. Activation of Liver X receptors (LXRs) in macrophages promotes cholesterol efflux (161, 162) and protects against the development of insulin resistance and atherosclerosis (163). Overexpression of miR-206 in THP-1 macrophages increased LXRα expression and enhanced cholesterol efflux. Activation of LXRα also reduced miR-206 expression, indicating that there is a feedback loop between miR-206 and LXRα in regulating cholesterol efflux, vascular inflammation, insulin resistance, and atherosclerosis (164). Finally, miR-24 reduced intraplaque macrophage proliferation rates by 34% by targeting the matrix metalloproteinases-14 (MMP-14), highlighting its therapeutic potential in macrophage proinflammatory functions (165).

V. Diabetic Complications

A. miRNAs involved in aberrant angiogenic growth of ECs and diabetic retinopathy

Diabetic retinopathy is one of the leading causes of blindness (139). Nearly 60% of type 2 diabetic patients are expected to have some form of retinopathy by the first decade after diagnosis (166). Aberrant angiogenic growth of ECs is now recognized to be an important component of diabetic retinopathy (167, 168). Several miRNAs have been implicated in regulating aberrant EC growth (Fig. 4). In the retina of STZ-induced diabetic rats, miR-200b expression decreased, whereas its direct target, VEGF, increased on both the mRNA and protein level. Overexpression of miR-200b normalized VEGF expression in the diabetic rat retina and prevented the glucose-induced increased permeability and angiogenesis in HUVECs, highlighting the importance of miR-200b in the regulation of VEGF-mediated diabetic retinopathy (169). Under hypoxic conditions, miR-126 expression is reduced in diabetic retinas. Restoration of miR-126 expression halted the hypoxia-induced neovascularization by suspending cell cycle progression and inhibiting the expression of VEGF and MMP-9 (170). Inhibition of a different miRNA, miR-21, significantly enhanced the high glucose–induced endothelial cytotoxicity. Furthermore, overexpression of miR-21 inhibited the expression of death domain associated protein (DAXX), a proapoptotic mediator, whereas silencing DAXX mRNA reversed the inhibition of miR-21 on endothelial apoptosis induced by high glucose, implying that miR-21 could protect ECs from apoptosis by inhibiting DAXX expression (171). In EPCs derived from patients with diabetes, transfection of miR-130a increased the protein level of ERK/VEGF and Akt via repression of Runx3, suggesting that the miR-130a can maintain normal EPC function through ERK/VEGF and Akt signaling pathways (172). In diabetic rats, miR-320 negatively regulated glucose-induced increase of VEGF and FN (173) and inhibited EC migration and tube formation (174). Finally, in response to high glucose conditions in ECs, several miRNAs such as miR-181c, miR-20b, and miR-15a, were reduced; however, their functional effects have not been explored (175).

Figure 4.

miRNA regulation of diabetic wound healing and myocardial angiogenesis. These miRNAs are involved in regulating diabetic wound healing and myocardial angiogenesis through effects on their targets that control cell proliferation, migration, apoptosis, and endothelial microvascular function. In response to diabetic conditions, several miRNAs are dysregulated (marked with an arrow).

NF-κB is a ubiquitous inducible transcription factor and is a key proinflammatory mediator of cellular pathways in the early phases of diabetic retinopathy development. In human retinal ECs, thrombin increased miR-146 expression in an NF-κB–dependent manner, and at the same time, miR-146 promoted a negative feedback to thrombin-induced NF-κB activation via targeting CARD10 (176). Human antigen R (HuR) has been reviewed as a promising therapeutic target for angiogenesis (177). Through binding to HuR, miR-146 also repressed endothelial activation by inhibiting the proinflammatory NF-κB pathway (178). miR-146a reduced the expression of the NF-κB–responsive adhesion molecule ICAM-1 by binding to the 3′ UTR of IRAK1, an upstream adapter protein activated by IL-1β in both human retinal ECs and diabetic rat retinas (179). Moreover, high glucose increased miR-195 and decreased SIRT1 expression in diabetic rat retinal ECs. Neutralization of miR-195 rescued SIRT1 expression and decreased tissue damage in diabetic retinopathy (180), highlighting an important role for miR-195-SIRT1 signaling axis in modulating retinal EC function.

B. miRNAs involved in the pathogenesis of diabetic nephropathy and diabetic microvascular kidney disease

Diabetic nephropathy is the major cause of end-stage renal disease and high mortality in diabetic patients (181, 182). Accumulating evidence has demonstrated that miRNAs regulate signaling pathways involved in the pathogenesis of diabetic nephropathy, including the microvasculature that figures prominently in this process (182) (Fig. 4). Renal hypertrophy and abnormal extracellular matrix deposition are hallmarks of diabetic nephropathy (183). FN is widely expressed in embryos and adults, especially in regions of active morphogenesis, cell migration, and inflammation (184). The main function of FN is to serve as a scaffold for cell adhesion and migration, thereby also mediating cell proliferation and differentiation (185). FN also contributes to micro- and macrovascular thrombosis (142). In kidneys of STZ-induced diabetic mice, both endothelin-1 and FN was increased with reduced expression of miR-1; in addition, miR-1 overexpression in ECs prevented the high glucose–induced ET-1 upregulation, suggesting a potential link in diabetic microvascular kidney disease (141). In HUVECs and STZ-induced diabetic rats, high glucose decreased miR-146a and increased FN expression. Interestingly, overexpression of miR-146a decreased FN expression in ECs, raising the possibility that efforts to maintain miR-146a expression in the microvasculature may be therapeutically relevant for diabetic nephropathy (127).

Finally, two other miRNAs, miR-29c and miR-93, have been implicated in regulating diabetic nephropathy through unique mechanisms directly or indirectly related to microvascular function. miR-29c inhibition reduced albuminuria and kidney mesangial matrix accumulation in a diabetic mouse model, an effect thought to be mediated by miR-29c targeting Sprouty homolog 1 (Spry1) and, in turn, regulation of Rho kinase (186). However, Spry1 has also been implicated as a MAPK antiangiogenic mediator, raising the possibility that miR-29c’s effects could also be related to angiogenic effects. miR-93, on the other hand, has been more directly associated with alterations of angiogenic signaling and diabetic nephropathy. Mice with inducible overexpression of miR-93 in podocytes showed marked improvements in diabetic nephropathy through modulating chromatin reorganization and targeting Msk2, a histone kinase, and its substrate H3S10 (187). Interestingly, miR-93 is localized within the intron of MCM7 gene, and its expression is reduced in response to high glucose. The authors showed that hyperglycemia decreased MCM7 promoter activity, which, in turn, decreased miR-93 and increased its target gene VEGF-A and VEGF signaling in kidneys, an effect that increased collagen and FN expression, paving the way to diabetic nephropathy (188). Taken together, overexpression of miR-146a and miR-93, or neutralization of miR-93, may provide potential therapeutic targets for regulating endothelial function and key steps in diabetic nephropathy.

C. miRNA regulation of diabetic neuropathy

Diabetic neuropathy is a nerve disorder that occurs in nearly 50% of patients with diabetes (189–191). It is found in patients who have had type 1 diabetes for more than five years and in the early stages of type 2 diabetes (192–194). The diabetic neuropathies are comprised of several different syndromes: distal symmetric polyneuropathy (typically presents with the gradual onset of numbness, paresthesias, or dysesthesias in the feet), median mononeuropathy (selective injury of individual nerves), visceral autonomic neuropathy (such as genitourinary dysfunction and orthostatic hypotension), small fiber and painful neuropathy, and other diabetic nerve injury (195). Several miRNAs have been reported to be involved in diabetic neuropathy (196–198) (Fig. 4).

However, only one miRNA, miR-146a, uniquely suppresses the same proinflammatory NF-κB targets in neurons and ECs relevant to the pathogenesis of diabetic neuropathy and the diabetic microvasculature. Specifically, overexpression of miR-146a in dorsal root ganglion (DRG) neurons inhibited the upregulation of the targets IRAK1 and TRAF6 (198–200). In db/db mice, which is a mouse model of diabetic peripheral neuropathy, hyperglycemia suppressed miR-146a expression and increased the protein levels of the proinflammatory adaptor proteins IRAK1 and TRAF6; similar results were also observed in cultured DRG neurons (198), which comprise thermoceptors, mechanoceptors, and itch sensors (201–203) thought to be responsible for the complication of diabetic sensory neuropathy (204–206). As such, miR-146a plays an important role in protecting DRG neurons from apoptosis under hyperglycemic conditions, suggesting that miR-146a might also prevent the development of diabetic neuropathy. Because miR-146a also confers favorable anti-inflammatory, protective effects in ECs in the microvasculature by regulating IRAK1, TRAF6, and HuR/eNOS, it may serve as a particularly important miRNA in the microvasculature and in peripheral neuropathy. Collectively, given the emerging role of EC plasticity governed by sympathetic nerve innervation (207), future studies will be necessary to explore the importance of miRNAs in peripheral nerves of patients with diabetes and the potential dual role of miRNAs regulating the microvasculature and peripheral nerves, as highlighted by miR-146a.

D. miRNAs involved in diabetic wound healing and angiogenesis

Accumulating studies suggest a number of miRNAs may be dysregulated and participate in different pathophysiological phases of diabetic wound healing (208) (Fig. 4). For example, miR-26a was shown to be induced in diabetic dermal wounds in mice. Local inhibition of miR-26a expression increased granulation tissue, induced angiogenesis, and promoted wound healing through an ID1/SMAD1 signaling pathway in ECs. In addition, local inhibition of miR-26a had no effect on leukocyte accumulation and enhanced the accumulation of myofibroblasts, which are important for wound contraction (209). Another miRNA, miR-155, known to functionally regulate diverse aspects of the immune response was found to be induced in diabetic wounds in mice (210). Deficiency of miR-155 led to a reduced inflammatory response and improved wound closure, an effect associated with increased expression of miR-155 target genes BCL6, RhoA, SHIP1, and FIZZ1 (210). In addition, expression of the members of the miR-99 family were found to be reduced in diabetic wounds, and overexpression of this miR family reduced PI3K/Akt signaling and migration and proliferation of keratinocytes, implicating their potential role in the later phase of re-epithelialization (211). The expression of a different miRNA, miR-21, was induced late (day 8) in dermal wounds of diabetic mice. Although miR-21 overexpression increased fibroblast migration in vitro, the role of altering miR-21 expression in diabetic wound healing remains unknown (208). In addition, the proinflammatory cytokine TNF-α increased expression of miR-200b in diabetic wounds, an effect that decreased GATA2 and VEGFR2 and altered the angiogenic response; conversely, miR-200b deficiency attenuated TNF-α’s effects and promoted angiogenesis (212). In diabetic limbal corneas, miR-146a was shown to be significantly increased. Overexpression of miR-146a in limbal epithelial cells decreased epithelial cell migration and wound closure in addition to phosphorylating p38 and EGFR. Ex vivo studies showed that inhibition of miR-146a in cultured human diabetic corneas promoted wound healing (213). In diabetes, several cell types exhibit impaired function including bone marrow–derived angiogenic cells (BMACs). Overexpression of miR-27b increased cell proliferation, tube formation, adhesion, and delayed apoptosis by directly targeting TSP1, TSP2, semaphorin 6A, and p66shc (Src homology 2 domain containing transforming protein 1) in BMACs. Topical transplantation of BMACs overexpressing miR-27b in wounds of diabetic mice resulted in improved wound healing and wound perfusion. Interestingly, direct delivery of miR-27b to the wounds of diabetic mice only partially improved wound healing, suggesting that additional miRNAs are likely to be involved (214). Finally, examination of human subjects with type 2 diabetes revealed distinct circulating miRNA patterns, including increased expression of miR-191, in the presence of chronic wounds (215). Interestingly, endothelial-derived miR-191 was taken up in dermal cells, leading to a reduction in the miR-191 target zonula occludens-1. Moreover, increased miR-191 expression impaired angiogenesis by suppressing migration of both diabetic dermal ECs and fibroblasts, in part, by targeting zonula occludens-1 (215). Collectively, these studies reveal important paracrine or “miracrine”-mediated effects from the microvasculature of wounds on dermal fibroblast function, an effect that may impact diabetic wound healing.

E. miRNAs involved in myocardial injury and angiogenesis

A number of miRNAs are implicated in myocardial injury and angiogenesis (Figure 4). miR-26a was shown to be important in both physiological and pathological angiogenesis by suppressing endothelial SMAD1 expression, an effect leading to reduced Id1 and increased p21^WAF/CIP1 and p27 expression and EC cycle arrest (209). miR-26a expression is increased in response to acute myocardial infarction in mice and in human subjects with acute coronary syndromes. miR-26a overexpression impairs physiological angiogenic responses such as exercise-induced angiogenesis in mice and developmental angiogenic responses such as the formation of the caudal venous plexus formation in zebrafish (209). In contrast, in vivo neutralization of miR-26a reduced myocardial infarct size by rapidly inducing robust angiogenesis by two days with improved left ventricular function, suggesting a new therapeutic approach for diseases associated with impaired pathological angiogenesis, including diabetic myocardial or peripheral artery disease (209). Inhibition of miR-92a, a part of the miR-17-to-miR-92 cluster, also reduced infarct size and promoted neovascularization in response to myocardial infarction by targeting integrin-α5 and eNOS (216). The role of miR-92a in physiological angiogenesis and diabetes remains unknown. miR-126 targets sprouty-related EVH-1 domain-containing 1 and regulated developmental angiogenesis in zebrafish and neovascularization after ischemic myocardial injury in mice (217). Members of the miR-23∼miR-27∼miR-24 cluster target several known angiogenic factors, including semaphorin 6A, Sprouty2, GATA2, and p21-activated kinase PAK4, to promote myocardial and retinal neovascularization (218–221). Finally, overexpression of miR-342-5p inhibited EC proliferation and angiogenesis in vitro and in vivo through direct targeting of endoglin. In addition, miR-342-5p inhibited TGF-β signaling by repressing SMAD1/5 phosphorylation in ECs. In a pathological model of angiogenesis using lased-induced choroidal neovascularization (CNV) in mice, miR-342-5p expression was significantly induced. Overexpression of miR-342-5p in this model system inhibited neovascularization and decreased the CNV area (222). Because CNV may coexist with chronic diabetic retinopathy (223), miRNAs may figure prominently in disease pathogenesis or response to therapy.

Although there are a number of studies that focus on the role of miRs in the context of diabetic angiogenesis, only a few studies examined the role of miRNAs specifically in type 1 diabetes. For example, circulating miR-27b and miR-320a are implicated in the progression of retinopathy in type 1 diabetes. miR-27b directly targets TSP-1, an extracellular matrix protein that is implicated in EC proliferation, migration, and angiogenesis, whereas miR-320a indirectly targets TSP-1 by reducing its secretion by ECs (224). In vitro inhibition of miR-195 protects against palmitate-induced EC apoptosis, suggesting that it may play a role in EC angiogenesis and metabolic-induced ER stress. In a mouse model of STZ-induced type 1 diabetes, miR-195 expression was increased in mouse hearts, and inhibition of miR-195 increased myocardial capillary density and coronary blood flow and reduced myocardial hypertrophy, in part, by attenuating oxidative damage and caspase 3 activity. Mechanistically, silencing of miR-195 increased BCL-2 and Sirt-1 protein expression (225). Because reduced coronary flow reserve is a strong indicator of coronary microvascular dysfunction in patients with diabetes and predicts cardiac mortality (226), these findings suggest that neutralization of miR-195 may improve diabetic coronary microvascular dysfunction, an emerging therapeutic interest to reduce cardiovascular burden in this predisposed population (227).

F. miRNAs and hyperglycemic memory

Hyperglycemic memory describes a phenomenon that the detrimental effects of hyperglycemia may persist even after restoration of normal glucose levels in diabetic patients (228, 229). Large population studies revealed that the persistent risk for vascular complications is observed for both type 1 and 2 diabetic patients (230, 231). Emerging evidence demonstrate that miRNAs contribute to hyperglycemic memory. For example, it has been shown that hyperglycemia significantly affects miRNA expression in the heart of diabetic mice, and many of them remained significantly altered after three weeks of intensive glycemic control with insulin (232). In miR-125b, miR-146a-5p, and miR-29a-3p were associated with persistent impaired endothelial function and altered proinflammatory gene expression after transient high-glucose treatment (233). The role of miRNAs in the pathogenesis of diabetic complications and metabolic memory has also been recently reviewed by Prattichizzo et al. (234). The causal role of miRNAs in hyperglycemic memory awaits further investigation. Targeting these miRNAs associated with hyperglycemic memory could open new therapeutic avenues to reduce cardiovascular events or residual cardiovascular risk in patients with diabetes.

VI. Circulating miRNAs

Exosomes are spherical-to-cup-shaped, 30- to 100-nm microvesicles (235) and can be released by all cell types (236). The composition of exosomes is very complicated, and they are enriched in various lipids, proteins, miRNAs, mRNAs, and DNAs (237). After they are released, exosomes enter the circulatory system and bodily fluids, including plasma, urine, bile, and breast milk, among others (238, 239). Therefore, the content of exosomes can be transferred from host cells to recipient cells carrying their content, which can affect the function of recipient cells, thereby facilitating communication between host and recipient cells (235, 240, 241).

miRNAs can also be detected in the extracellular environment, circulating in various biological fluids such as plasma, sera, urine, cerebrospinal fluid, and tears (242). The majority of detectable miRNAs are primarily concentrated in exosomes in serum and saliva (243). Accumulating studies demonstrate that circulating miRNAs carried in exosomes may influence target cell function in an analogous manner as some circulating hormones (244, 245). In this review, we summarize the effects of exosomal miRNAs on diabetic mellitus and their potential as biomarkers.

A. Exosomal miRNAs in type 1 diabetes

β-cell dysfunction and impaired insulin production are hallmarks of diabetes. Thioredoxin-interacting protein, a cellular redox regulator, has been reported to induce the expression of miR-204 in β-cells. By targeting MAFA, a known insulin transcription factor, miR-204 can block insulin production (246). Moreover, miR-204 can also target the protein kinase R-like ER kinase (PERK) to regulate β-cell apoptosis through the unfolded protein response (247). Other studies revealed that miR-204 was abundantly expressed in exosomes of human aqueous humor (248), is regulated by the gut microbiome, and impairs endothelium-dependent vasorelaxation by targeting the Sirtuin1 lysine deacetylase (Sirt1) (249). These findings highlight the important role of circulating miR-204 in both diabetes and endothelial function.

A cross-sectional cohort study revealed that circulating miRNA levels of miR-21 and miR-210 were significantly upregulated in the plasma and urine of type 1 diabetic patients, whereas urinary miR-126 levels were significantly lower in diabetic patients. These results suggest that miRNAs may participate in the early onset of type 1 diabetes (250). Because miR-21 is also induced by TGF-β1, an essential mediator in the pathogenesis of diabetic nephropathy, considerable attention has focused on miR-21 as a potential mediator in this condition (251). For example, renal cortices from type 1 diabetic mice demonstrated significantly elevated levels of miR-21, which mediates critical pathologic features of diabetic kidney disease through inhibiting the expression of its target PTEN and increasing Akt phosphorylation (252). Collectively, these findings provide a new potential diagnostic biomarker for diabetic nephropathy in type 1 diabetic patients.

In new-onset type 1 diabetes in children, miR-25 was negatively associated with residual β-cell function and positively associated with glycemic control, suggesting a predictive biomarker for type 1 diabetes diagnosis and a new target for therapeutic intervention (251). Inflammatory cytokines contribute to the progressive deterioration of pancreatic β-cell function and development of type 1 diabetes (253, 254). In response to IL-1β, for example, miR-101a and miR-30b participate in β-cell dysfunction by decreasing insulin content and gene expression and increasing β-cell death (255).

B. Exosomal miRNAs in type 2 diabetes

Accumulating studies demonstrate that circulating miR-126 levels are decreased in the blood of patients with type 2 diabetes (256–258). miR-126 is also highly expressed in ECs (38) to maintain endothelial homeostasis and vascular integrity in part through regulating the VEGF pathway (217, 259). miR-126 was identified as a key miRNA that confers the proangiogenic properties of circulating CD34⁺ cells (260). The CD34⁺ cell subset secretes miR-126 largely in microvesicles and exosomes, which are taken-up by ECs, and promotes angiogenesis (260). However, this proangiogenic capacity of CD34⁺ cells from diabetic patients was significantly attenuated due to the reduction of miR-126 expression (260). Recent clinical studies also demonstrated that, in peripheral whole blood, circulating miR-126 levels in type 2 diabetic patients are lower than in healthy controls and increased significantly after improved glycemic control. These findings suggest that circulating levels of miR-126 may be considered as a potential blood-based biomarker, perhaps useful for monitoring therapeutic responses in type 2 diabetic patients, including diabetic angiogenic responses important to diabetic coronary or peripheral artery disease (258, 261).

Serum miR-130b levels were significantly decreased in type 2 diabetic patients with diabetic nephropathy (262), and increased expression of miR-130b improves renal tubulointerstitial fibrosis by binding to the 3′ UTR of Snail mRNA both in vitro and in vivo (263). Furthermore, miR-130b can also be shuttled by microvesicles and transferred into target cells to regulate recipient cell function (264). Therefore, microvesicle-shuttled miR-130b has the potential to be a circulating therapeutic tool in diabetic nephropathy.

Several other miRNAs may serve as circulating biomarkers or mediators in type 2 diabetes. For instance, miR-1 has the potential to prevent endothelial dysfunction induced by hyperglycemia (141). Several studies have also found that plasma miR-1 levels were significantly higher in patients with myocardial infarction (265). In addition, microarray profiling studies identified the miRNAs miR-144, miR-146a, miR-150, and miR-182 as four diabetes-related miRNAs in rat peripheral blood (128). Among these miRNAs, circulating miR-144 levels were correlated with a reduction in the expression of its target, IRS1, at both the mRNA and protein levels (128), implying a new potential diagnostic and therapeutic target of type 2 diabetes. High-glucose treatment also reduced the miR-126 content of endothelial apoptotic bodies, which was similar to circulating microvesicles in plasma of patients with type 2 diabetes (256). miR-375 was recently characterized to be a pancreatic islet–specific miRNA, and knockdown of endogenous pri-mir-375 expression almost completely protected NIT-1 cells from palmitate-induced lipoapoptosis by repressing its target V1 expression, suggesting a potential therapeutic target for β-cell dysfunction and β-mass loss in type 2 diabetes (266). Interestingly, miR-375 is significantly increased in type 2 diabetic patients and has the potential to serve as a new biomarker for type 2 diabetes (267, 268). Taken together, emerging studies indicate that these exosomal miRNAs in the circulation of type 2 diabetic subjects may not only serve as potential biomarkers, but they may also regulate key processes in the pathology of a range of type 2 diabetes–associated conditions.

VII. Conclusions and Future Therapeutic Opportunities

A. Rationale for targeting miRNAs in the diabetic microvasculature

Therapeutic targeting of noncoding RNAs such as miRNAs represent an emerging frontier in the treatment of a variety of chronic disease states, including diabetes and a range of diabetes-associated complications that manifest with microvasculature dysfunction. Although several classes of RNA therapeutics are currently under development, including miRNA mimetics and inhibitors, siRNAs, and antisense oligonucleotides, only one antisense oligonucleotide drug, mipomersen, which targets apoB, is currently approved by the U.S. Food and Drug Administration for patients with homozygous familial hypercholesterolemia (269). In contrast to single-gene targeting using siRNA, therapeutic targeting of miRNA offers the ability to regulate gene networks important to the pathogenesis of complex diseases such as diabetes. Effective delivery of miRNA mimics or inhibitors to the microvasculature may facilitate improved cellular signaling and maladaptive biological processes in the local environmental of injured diabetic tissues.

The ability of miRNAs to circulate in exosomes and microvesicles provides a unique opportunity to use these “hormone-like” noncoding RNAs as biomarkers not only for the diagnosis and prognosis in a range of diabetes-associated conditions with altered microvascular function, but also in response to current or emerging therapies. In addition, because miRNAs may be expressed in specific cell types or in response to certain stimuli, careful interpretation is required for determining which miRNAs may best serve as potential biomarkers and therapeutic targets. Furthermore, miRNA target genes may be differentially expressed in a tissue- or cell-specific manner, which may be important for distinct diabetic microvascular environments.

B. Obstacles and challenges for effective miRNA delivery and targeting

Modulation of miRNAs may be facilitated locally or systemically using a range of chemically modified antisense microRNA inhibitors (antimiRs) (to inhibit miRNA) or miRNA mimics (to overexpress miRNA). Naked antimiRs are readily degraded by endogenous circulating or tissue RNases. However, several strategies have been used to chemically modify antimiR oligonucleotides to enhance target stability, affinity, or tissue uptake (270–272). For example, locked nucleic acids exhibit high binding efficiencies and improved stability with the addition of a methylene link between the 2′-oxygen and the 4′-carbon resulting in a locked position and reducing the flexibility of the ribose ring (273). Other modifications include ribose 2’-OH group modifications, such as 2′4′–constrained 2′O′-ethyl and 2’methoxyethyl, and phosphorothioate modification. Because a recent study highlighted that phosphorothioate modifications may facilitate nucleotide-based drugs to bind and activate platelets eliciting thrombus formation in response to carotid injury, pulmonary thromboembolism, and mesenteric artery injury in mice (274), careful scrutiny of chemical modifications will be required to minimize potential side effects and unanticipated toxicities. In contrast to their double-stranded counterparts (miRNA mimics), single-stranded antimiR oligonucleotides can be formulated in saline or phosphate-buffered saline for subcutaneous or intravenous delivery and do not require lipid- or nanoparticle-based delivery systems. Upon systemic delivery, these antimiRs rapidly leave the plasma and are taken up by multiple tissues, most notably the liver, the spleen, the kidney, adipose tissue, and bone marrow (275). Upon cellular uptake, the antimiR generates a high-affinity, stable bond, with the miRNA reducing the availability of the endogenous miRNA for binding to the 3′ UTR of the mRNA target(s). For a general view of miRNA targets, delivery of antimiRs, and therapeutic developments, please refer to the seminal review by van Rooij et al. (276).

C. Future prospects and unique platforms for miRNAs

Preclinical studies in nonhuman primates using naked antimiR oligonucleotides have shown promise for targeting miRNAs expressed highly in the liver, such as miR-122 or miR-33 (277). Cholesterol analogs have been added to antimiRs to facilitate cellular uptake, particularly in the liver, and this enhances their incorporation into high-density lipoprotein and low-density lipoprotein (278). Alternative approaches for miRNA inhibition include the use of competitive miRNA inhibitors, such as miRNA sponges or decoy transcripts that contain miRNA binding sites complementary to the seed sequence of the miRNA of interest (279).

Several approaches have been established to reconstitute miRNAs in disease states in which miRNA deficiency occurs. For instance, reduced miR-181b expression in the microvasculature of visceral fat is associated with the onset of insulin resistance, and delivery of miR-181b mimics was used to restore its expression and ameliorate insulin resistance in mice (101). Because of their double-stranded nature, the delivery of therapeutic miRNA molecules in vivo has many of the same challenges as siRNAs. As such, drug delivery vehicles such as liposomes, polymeric micelles, and nanoparticle-based drug carriers are being developed to deliver these oligonucleotides to cells and tissues. One of the major challenges associated with miRNA mimics replacement technologies include the ability to target miRNA to a specific cell type and the potential requirement of multiple doses to achieve sustained target repression. Viral vectors can be used to direct sustained expression, including short hairpin RNAs that can be processed in the target cell to the mature miRNA, an approach that has been used in multiple preclinical studies (280, 281). However, viral delivery systems will require careful scrutiny for clinical use. Cell-type–specific ligands, peptides, and nanoparticles are likely to provide unique delivery platforms that enable sustained and prolonged miRNA expression or knockdown.

In addition to chemical modification of miRs or antimiRs, nanotechnology is also an effective tool for systemic delivery of miRNAs and may provide enhanced specificity for targeting the diabetic vasculature. Typically, nanotechnology includes nanosized particles composed of different entities such as lipids, polymers, and inorganic materials (282, 283). These nanoparticles can also be modified to selectively deliver miRNAs. For example, to selectively deliver antimiR-712 to aortic ECs, cationic lipopartcles incorporating the peptide (VHPK) to target VCAM-1 were used, an effect resulting in inhibition of atherosclerosis by targeting TIMP3 (284). In a similar approach, loading miR-146 and miR-181b into E-selectin–targeting multistage nanoparticles effectively transported these two miRNAs to ECs, an effect that ameliorated endothelial inflammation and atherosclerosis (285). These modified nanocarriers may improve therapeutic effects and also reduce “off-target” effects, thereby providing a potential framework for targeting, for example, the diabetic microvasculature in pathologies such as diabetic wound healing, retinopathy, or impaired myocardial angiogenesis, among others. Finally, other potential vehicles for modulating miRNA expression in the vasculature include exosomes, a class of microvesicles released by cells that contain cell-specific proteins (286) that can recognize target cells (287). Therefore, engineering cell-specific exosomes enriched with a cell surface signature relevant to the diabetic microvasculature may hold promise for serving as an ideal miRNA carrier for a range of diabetic pathologies.

D. Examples of miRNA mimics and inhibitors in clinical trials

Several miRNA-based therapeutics are currently in clinical trials. The first is a locked nucleic acid directed against miR-122 (Miravirsen), which targets hepatitis C virus RNA (288). In a phase 2 study, Miraversen demonstrated dose-dependent antiviral activity maintained over four weeks (289). Another antimiR in clinical development is an inhibitor to miR-155 for cutaneous T cell lymphoma (ClinicalTrials.gov no. NCT02580552). miRNA mimics in earlier-stage clinical trials include intradermal delivery of miR-29b mimics for patients with cutaneous scleroderma (ClinicalTrials.gov no. NCT02603224) and MRX34, a miR-34 mimic encapsulated in a liposomal nanoparticle formulation, which was to be examined in patients with a range of advanced solid tumors and hematologic malignancies. Due to multiple immune-related severe adverse events, the phase 1 clinical trial investigating MRX34 was halted, raising concerns about both dosing and targets for this miR-34 mimic (ClinicalTrials.gov no. NCT01829971). Nevertheless, the results of clinical trials involving miRNA mimics or inhibitors provide new therapeutic opportunities for “fine-tuning” disease management involving the diabetic microvasculture.

In conclusion, accumulating studies highlight critical roles for miRNAs in diverse aspects of diabetic microvascular disease, including insulin resistance, diabetic wound healing, myocardial angiogenesis, diabetic retinopathy, and diabetic nephropathy, and a range of important cell signaling pathways in disease pathogenesis. miRNAs may offer new targets for early detection and therapeutic intervention, paving the way for unique means to identify and target the devastating sequelae of diabetic microvasculature disease.