Role of long noncoding RNAs in diabetes-associated peripheral arterial disease

Abstract

Diabetes mellitus (DM) is a metabolic disease that heightens the risks of many vascular complications, including peripheral arterial disease (PAD). Various types of cells, including but not limited to endothelial cells (ECs), vascular smooth muscle cells (VSMCs), and macrophages (MΦs), play crucial roles in the pathogenesis of DM-PAD. Long non-coding RNAs (lncRNAs) are epigenetic regulators that play important roles in cellular function, and their dysregulation in DM can contribute to PAD. This review focuses on the developing field of lncRNAs and their emerging roles in linking DM and PAD. We review the studies investigating the role of lncRNAs in crucial cellular processes contributing to DM-PAD, including those in ECs, VSMCs, and MΦ. By examining the intricate molecular landscape governed by lncRNAs in these relevant cell types, we hope to shed light on the roles of lncRNAs in EC dysfunction, inflammatory responses, and vascular remodeling contributing to DM-PAD. Additionally, we provide an overview of the research approach and methodologies, from identifying disease-relevant lncRNAs to characterizing their molecular and cellular functions in the context of DM-PAD. We also discuss the potential of leveraging lncRNAs in the diagnosis and therapeutics for DM-PAD. Collectively, this review provides a summary of lncRNA-regulated cell functions contributing to DM-PAD and highlights the translational potential of leveraging lncRNA biology to tackle this increasingly prevalent and complex disease.

Graphical Abstract

Introduction

Diabetes or diabetes mellitus (DM) is a disease characterized by the poor control of blood glucose levels and is broadly divided into Type 1 (T1D), Type 2 (T2D), and gestational diabetes, all resulting in hyperglycemia. In 2021, the CDC estimated global diabetes prevalence in individuals ranging from 20 to 70 years old was 10.5% (half a billion people), which is predicted to rise to 12.2% (783 million people) by 2045 [1]. The financial burden attributed to diabetes-related expenditures is estimated to be over a trillion USD by 2045 [2], posing a daunting socio-economic burden. Patients with DM are at high risk for various cardiovascular diseases (CVD), e.g., coronary artery disease (CAD), stroke, and peripheral artery disease (PAD), which are the leading causes of DM-related mortality and morbidity. Specifically, PAD occurs in patients with DM at a two- to four-fold higher incidence than in individuals without DM [3]. While sharing some etiology with CAD and stroke, PAD, defined by reduced blood flow as a result of the narrowing of arteries to the limbs, has its unique pathogenesis. Although PAD can also affect the upper limbs, it is much more prevalent in lower limbs.

Concomitant DM-PAD represents a significant threat to both life and limb. There are 185,000 amputations in the US each year, and more than half are attributable to DM-PAD [4, 5]. DM-PAD carries a risk of limb loss four times higher than the national average [6, 7]. The primary therapy for limb preservation in this patient cohort is revascularization [8]. However, the efficacy of revascularization in this patient cohort is severely hampered by two biological hurdles that are closely associated with DM: (1) severe vessel calcification and (2) microvascular disease. Significant medial artery calcification is a common histologic feature among people with diabetes and is associated with adverse outcomes [9,10,11]. A myriad of surgical techniques, including atherectomy, lithotripsy, and covered stents, have been developed to treat calcium-laden atherosclerotic lesions. However, none of these have improved clinical outcomes [12]. Microvascular disease, defined as occlusive lesions in blood vessels ≤ 100 μm in diameter, is strongly associated with diabetes [13] and increases the risk of amputation four-fold [14]. Currently, there are no surgical or medical therapies for microvascular disease. Given the aging population and growing incidence of diabetes, there is a critical, unmet need to deepen our understanding of the pathobiology of both DM and PAD and to leverage that knowledge to develop novel therapies to improve PAD outcomes.

Many risk factors associated with DM, e.g., hyperglycemia, hyperlipidemia, and hypertension, can contribute to atherosclerosis, characterized by a buildup of plaque in the inner arterial wall resulting in restricted blood flow [15, 16]. Atherosclerosis begins with dysfunctional endothelial cells (ECs), which, in a classical model, recruit monocytes to attach, migrate, and differentiate into macrophages (MΦs). After engulfing low-density lipoprotein (LDL) cholesterol, MΦs become foam cells in the vascular wall, a hallmark of atherosclerotic lesions [16]. Other MΦs polarize to an inflammatory state, which, along with ECs, secrete platelet-derived growth factors (PDGFs), vascular endothelial growth factors (VEGF), and other inflammatory cytokines, resulting in the increased proliferation of vascular smooth muscle cells (VSMCs), which can undergo dedifferentiation and phenotypic switching towards myofibroblast- or an osteoblast-like phenotype [16, 17]. These cells secrete collagen and other extracellular matrix (ECM) components, leading to lesion progression, calcification, and fibrosis of the arterial wall. At the later stage, prolonged exposure to an inflammatory milieu results in apoptosis of ECs, VSMCs, and MΦ, forming a necrotic core in the atherosclerotic lesion. In a diabetic setting, where patients are atheroprone, the capability of MΦs to participate in efferocytosis, the process of removing diseased and dying cells through phagocytosis, is often hampered [18]. Without inflammation resolution, the plaque can destabilize and rupture, causing thromboembolism, inadequate oxygenation, and nutrient delivery in the periphery.

To aggravate these detrimental alterations, the microvascular ECs in the diabetic vessels can have impaired angiogenic and wound healing capacity, failing to form adequate collateral vessels to alleviate the tissue ischemia. Compensatory angiogenesis is supported by tissue-resident MΦ recruitment, which secretes pro-angiogenic factors such as VEGF and recruits other support cells, including monocytes and neutrophils. However, given the dominance of pro-inflammatory MΦ due to the DM milieu, this restorative mechanism is often insufficient to overcome the ischemia [19, 20]. The compromised tissue perfusion and ischemia in the limbs and extremities can lead to intermittent claudication (IC), manifesting in cramping and pain, a common symptom of PAD. When these conditions continue to worsen, PAD can progress to critical limb tissue ischemia (CLTI), the most severe form of PAD [21]. CLTI presents as resting pain, severe tissue ischemia, and impaired wound healing and can be coupled with peripheral sensitive neuropathy, a condition also more common in patients with diabetes [22]. Concomitant DM-PAD can cause the formation of blisters/ulcers from imbalanced weight support, numbness even in the presence of injury, and insufficient wound healing to resolve skin wounds, all of which place patients at higher risk of infections [23]. These detrimental factors work synergistically to further drive tissue death and potential limb amputation as a consequence of diabetic foot.

In the past decades, the DM-PAD field has primarily focused on dissecting the role of proteome and the related cellular pathways in regulating various cellular processes contributing to this ischemic disease. Recent technological advances in human transcriptome profiling have enabled the thorough investigation of the non-coding genome, including long non-coding RNAs (lncRNAs). Like messenger RNAs (mRNAs), most lncRNAs are transcribed by RNA polymerase II, 7-methyl guanosine (m⁷G)-capped at the 5′ ends and polyadenylated at the 3′ end [24, 25]. Moreover, lncRNAs display distinct features in their transcription, processing, and subcellular localizations, which dictate their modes of action and cellular functions. Relative to protein-coding mRNAs, lncRNAs display higher cell-type specificity and exert varied functions in different cell types. While there are multiple ways to classify lncRNAs, one is by their genomic locus and orientation to nearby protein-coding genes [26,27,28]. This results in four classes: intergenic, intronic, sense, and antisense lncRNA. Early studies focused on long intergenic non-coding RNAs (lincRNAs), whose transcriptional units do not overlap with nearby protein-coding genes. Intronic lncRNAs initiate from the intron of a protein-coding gene, while sense lncRNAs are transcribed from the sense strand of a protein-coding gene and its exons, overlapping with part of protein-coding genes. In contrast, antisense lncRNAs are transcribed from the antisense strand of protein-coding genes, usually overlapping with protein-coding exons.

LncRNAs can also be classified based on their mechanism of action, i.e., cis- and trans-acting [29,30,31]. While cis-acting lncRNAs exert regulatory control over the transcription within the same or adjacent loci, trans-acting lncRNAs exhibit their regulatory roles on distal targets, the mechanism of which appears much more diverse and less characterized. However, some lncRNAs can follow both cis- and trans-acting modes of action. To add to the complex biology of lncRNAs, promoter regions and pseudogenes can also give rise to lncRNAs. A significant proportion of lncRNAs are transcribed from enhancers, the DNA-regulatory elements that activate transcription to higher levels than those in their absence [32,33,34,35]. The genomic loci of these lncRNAs are typically associated with enhancer histone hallmarks or bound by transcription factors (TFs), also termed super-enhancers (SEs) [36, 37]. Previous studies, including ours, reported that lncRNAs derived from SE regions show strong cell type specificity and participate in transcription regulation of cell identity, fate, and stress responses (i.e., angiogenesis and proliferation) [38,39,40,41]. Understanding the classification of lncRNAs can aid in a systematic, comprehensive approach to their diverse roles and regulatory mechanisms in diseases like DM-PAD, which can be leveraged to design new avenues for disease management, treatment, and intervention.

In this review, we summarize the existing literature on lncRNA regulation of the dominant vascular cell types contributing to DM-PAD, using selective lncRNAs that have been extensively studied as examples. While atherosclerosis is the main cause of PAD and impaired angiogenesis is a primary contributor to PAD, several excellent review articles have discussed the role of lncRNA in atherosclerosis and angiogenesis [42,43,44,45,46,47,48,49,50]. Herein, we place an emphasis on lncRNAs that can serve as missing links between DM and PAD. We then provide suggestions for experimental approaches to study lncRNAs in DM-PAD. Finally, we discuss the potential of exploiting lncRNAs as diagnostic and therapeutic targets for treating DM-PAD. Collectively, we hope to provide an overview of the emerging role of lncRNA in DM-PAD that can spur future exploration in this research area.

LncRNAs in endothelial cells (ECs)

The endothelium constitutes the critical interface between tissues and circulation. In physiological conditions, ECs are quiescent, exerting anti-inflammatory and anti-thrombotic properties. In the diabetic milieu, ECs become activated and upregulate adhesion molecules such as vascular cell adhesion molecule-1 (VCAM-1) and monocyte chemoattractant protein (MCP1, coded by CCL2), recruiting leukocytes, a key step of vascular inflammation. Diabetic stress also leads to decreased nitric oxide (NO) bioavailability and impaired NO-mediated endothelium-dependent vasodilation, the hallmark of EC dysfunction. Increased oxidative stress is another detrimental factor due to diabetic stress, resulting from activation of NADPH oxidases (NOX), mitochondrial dysfunction, and endothelial NO synthase (eNOS) uncoupling [51,52,53]. The chronic elevation of glucose, pro-inflammatory cytokines (e.g., tumor necrosis alpha (TNF-α), interleukin-1 beta (IL-1β)), and oxidative stress can induce endothelial-mesenchymal transition (EndoMT), characterized by downregulation of the endothelial lineage markers (e.g., cadherin-5, and eNOS) and upregulation of the mesenchymal markers (e.g., alpha-smooth muscle actin (α-SMA), type I collagen, and fibronectin 1 (FN1) [54, 55]. EC apoptosis can occur at the late stage of PAD, resulting in plaque erosion and thrombosis. While these pathogenic processes can be shared by both large and small vessels, microvascular EC dysfunction can manifest as reduced angiogenic capacity, a crucial aspect of the pathogenesis of PAD [56]. An increasing list of lncRNAs has been shown to regulate EC function, as summarized by several excellent reviews [42, 50, 57,58,59,60,61,62,63,64,65]. Below, we highlight the reported roles of several lncRNAs that may serve as molecular links connecting DM and PAD.

Metastasis Associated Lung Adenocarcinoma Transcript 1 (MALAT1) is one of the first identified and most extensively studied lncRNAs implicated in numerous diseases, including DM, CAD, and PAD [66,67,68]. It is ubiquitously expressed, highly conserved across mammalian species [69], and primarily localized to the nucleus [70, 71]. MALAT1 is encoded within Chromosome 11q13 in humans and Chromosome 19qA in mice. In ECs, MALAT1 regulation seems context-dependent with different modes of action. For example, MALAT1 is increased in streptozotocin (STZ)-treated diabetic mice, and high-glucose (HG) treated retinal ECs [72,73,74]. Increased levels of circulating MALAT1 have also been reported in patients with T2D [75]. Molecularly, in HG-treated ECs, MALAT1 promotes serum amyloid antigen 3, an inflammatory ligand, to increase TNF-α and interleukin-6 (IL-6) [76, 77]. On the other hand, hypoxia also increases MALAT1 expression, while suppression of MALAT1 leads to impaired EC proliferation and blood flow recovery in a mouse model with hindlimb ischemia (HLI) [67, 78]. Interestingly, MALAT1 has been reported to be transcriptionally regulated by Krϋpple-like Factor 4 (KLF4) [79], a key TF responsible for EC homeostasis [80], suggesting a potential of TF in mediating the context-dependent induction of MALAT1. Given these reported roles of MALAT1 in ECs, it will be important to directly assess the role of MALAT1 in the context of DM-PAD, where it may be induced by both HG to promote a pro-inflammatory response, as well as by hypoxia/ischemia to confer a pro-angiogenic response.

H19 is another lncRNA extensively studied in a number of cell types and biological processes [81, 82]. H19 is imprinted in a gene cluster containing insulin-like growth factor gene 2 (IGF2) on Chromosome 11 in humans, which is conserved in Chromosome 7 in mice [83, 84]. H19 is primarily localized in the cytoplasm and plays pleiotropic and cell type dependent roles in ECs, VSMCs, and MΦ during the progression of cardiovascular diseases [85]. Regarding diabetic retinopathy, exposure to high glucose decreases H19 levels, and this results in the activation of the mitogen-activated protein kinase-extracellular signal-regulated kinase 1/2 (MAPK-ERK1/2) pathway downstream of transforming growth factor beta (TGFβ) signaling, which in turn promotes EndoMT [86]. In line with the requirement of H19 for EC homeostasis, loss of H19 in ECs also upregulates inflammatory markers, such as VCAM-1 and intercellular adhesion molecule-1 (ICAM-1), likely through induction of signal transducers and activators of transcription 3 (STAT3) [87]. H19 is also decreased in the endothelium of aged mice and in the atherosclerotic plaques in humans compared to healthy carotid artery biopsies, suggesting loss of H19 as part of vascular pathologies. Mechanistically, decreased levels of H19 led to an upregulation of senescence markers p16 and p21 [88], attendant with a reduction of EC proliferation and inhibition of sprouting capacity of mouse aortic rings [87]. The importance of H19 is further demonstrated in vivo, with EC-specific inducible H19 knockout (KO) mice exhibiting increased systolic blood pressure and reduced capillary density in an HLI model [87]. While these studies report a decreased H19 in ECs in various disease settings, increased circulating H19 has been reported in diabetic patients versus healthy controls [89,90,91]. These divergent patterns of H19 in ECs versus circulation are likely due to the levels of H19 in different cell types and the potential release mechanisms of H19 into circulating blood, which warrants further investigation. The therapeutic potential of H19 has also been tested, leveraging extracellular vesicle-mimetic nanovesicles (EMNVs) to deliver H19 to ECs affected by hyperglycemia [92]. The EMNV-restored H19 promoted the recovery of diabetic wounds using an in vivo STZ-induced impaired wound closure assay [92]. Given the suppression of H19 in DM and its protective role in ECs, restoration of H19 may be a promising approach to attenuate DM-PAD.

Encoded in the DLK1-MEG3 locus on Chromosome 14q32, maternally expressed gene 3 (MEG3) has been extensively studied in cancer [93], diabetes [94], renal ischemia [95], and CVDs [96, 97]. Increased levels of MEG3 have been reported in cardiomyocytes treated with HG, as well as in serum and villous samples from patients with gestational diabetes, compared to non-diabetic samples [98]. Inhibition of MEG3 was found to promote the EC angiogenic function in vitro and enhance ischemic recovery in aged mice with HLI [99]. The role of MEG3 in angiogenesis may be through epigenetic regulation by which MEG3 guides enhancer of zeste homolog 2 (EZH2) to repress integrin subunit alpha 4 (ITGA4) [100]. In line with these findings supporting a negative role of MEG3 in EC homeostasis, overexpression of human MEG3 promotes aortic cellular senescence and aggravates atherosclerotic lesions in mice [101]. However, MEG3 has also been shown to protect EC function by regulating the DNA damage response through its interaction with an RNA-binding protein (RBP) polypyrimidine tract binding protein 3 (PTBP3), which restrains the p53-induced EC apoptosis [102]. It is also worthwhile to note that in the context of diabetic retinopathy, MEG3 is decreased in the retinas of STZ-treated mice and HG-treated retinal ECs [103]. MEG3 knockdown (KD) aggravates retinal vessel dysfunction in vivo, evidenced by increased vascular leakage, increased acellular capillaries, and exacerbated inflammatory damage [102, 104]. Furthermore, MEG3 has been reported to be increased in plasma samples from patients with diabetes [105, 106]. Taken together, given the complexity of MEG3-regulated EC function, the role of MEG3 in DM-PAD remains to be directly evaluated and identified.

We have identified a lncRNA that enhances eNOS (encoded by NOS3) expression (LEENE) and regulates EC function [107]. Encoded by LINC00520 in Chromosome 14 in humans, the transcription of LEENE is increased by physiological flow and hypoxia and decreased by pathophysiological conditions, including disturbed flow, HG, and TNF-α [107, 108]. The levels of LEENE also decreased in the mesenteric arteries isolated from donors with diabetes. Inhibition of LEENE in human ECs decreased the expression of eNOS, Kinase insert Domain Receptor (KDR), and a set of genes promoting angiogenesis. Interestingly, the mouse homolog of LEENE, also encoded in Chromosome 14, is increased in the atheroprotective region of the aorta as compared to the atheroprone region, as well as in the ischemic limb as compared to the non-ischemic limb in the non-diabetic wild type mice. However, the ischemia induction of mouse leene was ablated in the diabetic mice. Furthermore, genetic deletion of leene in mice impaired flow perfusion and tissue recovery in the HLI model, especially under diabetic conditions. Importantly, supplementation of human LEENE RNA (despite the limited sequence conservation between human and mouse transcripts) rescued the impaired ischemic recovery in the KO mice. Mechanistically, LEENE promotes the transcription of pro-angiogenic genes, such as KDR and NOS3, by interacting with their promoters and RNA binding proteins (e.g., LEO1, a key component of the RNA polymerase II–associated factor complex) [108]. Leene-KO mice also developed worsened hypertension in an angiotensin II-infused model [109]. These data support the functional conservation of lncRNAs in humans and mice and highlight the promise of targeting LEENE to ameliorate DM-PAD [110]. Of note, LINC00520 also gives rise to LncRNA Activated by Sheer Stress In the Endothelium (LASSIE), which can interact with intermediate filament protein nestin and the adherens junction components in the cytoplasm to regulate EC permeability [111], suggesting different splicing variants of LINC00520 may have diverse impacts on EC functions.

LINC00607, encoded near FN1 in Chromosome 2 in the human genome, is another EC-enriched lncRNA found to be unregulated by DM and modulate EC function. KD of LINC00607 under HG and TNF-α treatment (HT) attenuated pro-inflammatory and pro-EndoMT phenotypes of ECs [112]. Intriguingly, at a baseline condition, LINC00607 KD caused a profound inhibition of EC angiogenic functions, supporting an essential role of LINC00607 in maintaining normal EC function [113], which was confirmed by a separate study [114]. Mechanistically, LINC00607 can interact with a chromatin remodeling protein Brahma-Related Gene-1 (BRG1) to regulate the chromatin state at ERG (ETS Transcription Factor)-binding loci and hence facilitates the transcription of angiogenic genes involved in VEGF-signaling pathway, including KDR, Tetraspanin 12 (TSPAN12), and Von Willebrand Factor (VWF) [114]. These findings suggest a context-dependent mode of action of LINC00607 in ECs, which may involve complex molecular mechanisms engaging different TFs, co-regulators, and gene targets. For example, under HT, LINC00607 is upregulated in the nucleus and interacts with the chromatin to increase FN1 and Serpin Family Member 1 (SERPINE1) expression, which may be in concert with the function of MYC [113]. Like LINC00520 and many other lncRNAs, LINC00607 produces 13 transcripts that appear to localize to different subcellular compartments. It is possible that different transcripts respond to distinct upstream signals and exert shared or transcript-specific functions, which warrants future investigation. Intriguingly, we found that only the locked nucleic acid (LNA)-GapmeR that inhibited the LINC00607 level efficiently in the nucleus led to significant SERPINE1 and FN1 reduction. In contrast, inhibition of cytoplasmic-localized LINC00607 did not exert the same effect, suggesting a cellular localization-dependent role of LINC00607 in EC regulation [112]. The functional importance of LINC00607 and its disease relevance, in particular, to DM-PAD remains to be explored.

LncRNAs in macrophages (MΦs)

MΦs are involved in both the progression and resolution of PAD. At the steady state, the primary functions of MΦ are tissue surveillance, where cells are poised to fight against infection, remove toxic or dead/dying cells, and restore homeostasis. Under diabetic conditions, such as HT, monocytes, the MΦ precursors, are increasingly recruited to the artery wall. Here, they differentiate into MΦs and secrete factors propagating the chronic low-grade arterial inflammation, which accelerates atherosclerosis, the cause of PAD. Apoptosis of MΦs can also occur during advanced atherosclerosis, contributing to plaque instability and rupture. In the context of PAD, MΦs can be recruited to the ischemic muscle, where they have a continuum of subtypes and polarization states to regulate the inflammatory and immune response [115]. In these contexts, MΦs are dominantly derived from monocytes, and their numbers, particularly pro-inflammatory MΦs, expand greatly in ischemic muscle [116]. Simplistically, MΦs can switch between two phenotypes: M1-like (pro-inflammatory) and M2-like (anti-inflammatory or tissue-repairing), although several other functional and disease-relevant MΦ subtypes have been identified [117,118,119]. Typically, M1-like MΦs promote inflammatory processes and tissue damage, whereas M2-like MΦs promote tissue repair and resolution of inflammation [120]. These MΦs are involved in the clearance of dead cells and debris from the plaque and support the production of anti-inflammatory cytokines, contributing to tissue healing and stabilization of plaques. In addition, there are resident MΦs seeded during embryonic development, which reside in skeletal muscle. When polarized by M2-ligands, such as interleukin-4 (IL-4), these resident MΦs can promote wound healing and angiogenic repair [121]. Recent studies have revealed unexpected heterogeneity of MΦ and identified new functional subtypes involved in ischemic response. For instance, perivascular MΦs, which accumulate within the ischemic site of human PAD patients and mouse PAD models, produce NO via inducible NOS (iNOS) to regulate blood flow after tissue injury [19].

Although no MΦ-expressing lncRNAs have been directly tested to regulate the development of PAD, lncRNAs have been shown to regulate many functional aspects of MΦs, including chemotaxis, phagocytosis, lipid metabolism, foam cell formation, and efferocytosis [122,123,124], all of which are highly relevant to DM-PAD. Below, we discuss several lncRNAs that have been shown to regulate MΦ function in DM and atherosclerosis and their potential involvement in DM-PAD.

LncRNA Dynamin 3 opposite strand (Dnm3os) is encoded on the opposite strand of Dynamin 3 (Dnm3) on Chromosome 1 [125]. Mouse Dnm3os and its human ortholog DNM3OS share 83% homology. Nuclear-localized Dnm3os/DNM3OS are increased by diabetic conditions in monocytes and MΦ. This phenotype was observed in cultured monocytes and MΦ under HG and palmitic acid (PA) treatment, db/db mice, high-fat diet-fed mice, and STZ-treated ApoE^−/− mice. DNM3OS also increases in CD14⁺ monocytes from patients with DM as compared to healthy controls. Downstream of nuclear factor-κB (NF-κB), Dnm3os interacts with nucleolin (NCL), an RBP responsible for regulating chromatin structure, by negatively regulating the expression of pro-inflammatory gene expression through promoting a repressive chromatin state [126]. In diabetic conditions, NCL is decreased while Dmn3os is increased, leading to chromatin opening at inflammatory genes. As a result of this dysregulated Dmn3os-NCL interaction, inflammation and phagocytosis in MΦ are activated in DM. Given the identified role of Dnm3os, Dnm3os is also likely to be elevated in DM-PAD and contributes to the unresolved vascular inflammation.

Palmitic Acid-Regulated Anti-Inflammatory lncRNA (PARAIL) is another DM-regulated lncRNA found to be altered by DM conditions that regulate MΦ function [127]. PARAIL (A530072M111Rik, in mouse) is encoded on Chromosome 8q21.3 and conserved between humans and mice [127]. It is divergently transcribed from its proximal receptor-interacting protein kinase 2 (RIPK2), a kinase important for innate and adaptive immunity. Like DNM3OS, PARAIL is also enriched in the nucleus and increased by PA in CD14⁺ MΦ from healthy donors and THP1-derived MΦs in an NF-κB-dependent fashion. PARAIL is highly enriched in the nuclear fractions of MΦs, particularly in the chromatin fraction. PARAIL is induced in MΦ during the inflammation resolution phase. Mechanistically, PARAIL interacts with an RBP, human antigen R (HuR, gene name ELAV like RNA binding protein 1 (ELAVL1)), which is known to bind AU-rich elements (AREs). When PARAIL is overexpressed in THP1-derived MΦs, inflammatory markers are decreased. In contrast, when PARAIL is suppressed, its interaction with HuR/ELAVL1 increases, promoting the stability and expression of ARE-containing inflammatory genes. In the diabetic db/db and STZ-treated ApoE^−/− mice, PARAIL expression was significantly reduced, supporting that loss of PARAIL contributes to enhanced inflammation in a diabetic setting. Interestingly, PARAIL was also found to be increased by a pro-inflammatory cytokine IL-1β in human ECs, suggesting its functional role in vasculature in addition to MΦs. However, whether PARAIL partakes in the pathogenesis of DM-PAD remains to be determined.

First identified to be upregulated after the occurrence of renal ischemia-reperfusion [128], the Macrophage-Enriched lncRNA Regulates Inflammation, Chemotaxis, and Atherosclerosis LncRNA (MERRICAL, previously named AI662270) has been found to play a pro-atherogenic role in mouse models, including ApoE^−/− mice fed high-fat diet and Ldlr^−/− mice fed a high-fat, high-sucrose diet [128, 129]. MERRICAL is encoded in Chromosome 11, proximal to a family of chemokine genes, including C–C Motif Chemokine Ligand 3 (CCL3) and CCL4. MERRICAL, together with CCL3 and CCL4, are co-induced in intimal atherosclerotic lesions of mice. It is also exponentially upregulated with the differentiation of bone marrow-derived macrophages (BMDMs), especially in pro-inflammatory M1-like MΦs. MERRICAL is enriched in the nucleus, where it interacts with mixed-lineage leukemia 1 (MLL1), a histone methyltransferase, and an RNA-binding adapter protein WD repeat-containing protein 5 (WDR5) to promote activation of promoter regions of CCL3 and CCL4 genes, leading to their increased expression. Consistently, inhibition of MERRICAL in vivo decreased leukocyte recruitment to the vascular wall and markedly reduced DM-accelerated atherosclerotic lesion progression. While these findings strongly support a pro-inflammatory and pro-atherosclerotic role of MERRICAL, the effect of DM alone on MERRICAL and its role in DM-PAD remains to be defined.

Macrophage-Associated Atherosclerosis lncRNA Sequence (MAARS), encoded by Gm14461 on Chromosome 2, is a MΦ-specific lncRNA found to localize primarily in the nucleus. MAARS is a critical regulator of MΦ apoptosis and efferocytosis, but it has only been studied in mice so far [130]. Its expression dramatically increases in the mouse aortic intima during atherosclerotic progression and decreases with atherosclerotic regression. Inhibition of MAARS decreased MΦ apoptosis in vitro and reduced atherosclerotic lesion formation independent of lipid profile and inflammation changes. The anti-atherosclerotic effect of MAARS inhibition was attributed to the decrease in MΦ apoptosis and the increase in efferocytosis in the vessel wall. Like PARAIL, MAARS also interacts with HuR/ELAVL1, and this interaction prevents cytosolic shuttling of HuR, where it regulates genes involved in apoptosis, such as p53 and caspase-9. The function of MAARS in the diabetic and ischemic context remains to be resolved. However, given its relevance in key MΦ functions and possible conservation between humans and mice, it may also play a role in DM-PAD disease progression.

Suppressor of Inflammatory Macrophage Apoptosis LncRNA (SIMALR) is a non-conserved, nuclear-localized human lncRNA encoded in Chromosome 6q23.3 [131]. SIMALR directs MΦ polarization, as it is highly expressed in inflammatory/activated M1-like, but not in non-activated anti-inflammatory M2-like MΦ and is increased in the MΦ of human atherosclerotic plaques. In M1-like MΦ, depletion of SIMALR leads to elevated expression of apoptotic markers such as Poly (ADP-ribose) Polymerase (PARP), caspase-9, and caspase-3, suggesting a protective role of SIMALR in MΦs. Mechanistically, SIMALR acts in trans by interacting with hypoxia-inducible factor 1α (HIF1α) and promoting the recruitment of this transcription factor to the promoter of Netrin-1 (NTN1) a laminin-related protein, which promotes MΦ survival. The SIMALR KD-induced apoptosis was attenuated by treating MΦ with recombinant NTN1 protein [131]. Given the anti-inflammatory and anti-apoptotic effects of SIMALR, it can potentially play a positive role in DM-PAD. Therefore, SIMALR could be a promising target in targeting diabetes-affected MΦs. In line with the study, SIMALR was increased in the MΦs of human atherosclerotic plaques, suggesting an increase of M1-like MΦs, where they can contribute to necrotic core formation [132]. In the context of DM-PAD, SIMALR has not yet been extensively studied. Based on the current data, it is a promising target for mitigating MΦ accumulation and apoptosis in DM-PAD.

Cholesterol homeostasis regulator of microRNA expression (CHROMR, previously named CHROME [133, 134]) is encoded by AC009948.5 on Chromosome 2q31.2. Like SIMALR, it is not conserved and is only present in the genomes of primates [133]. Expression of CHROMR in both the plasma and neointima was increased in patients with atherosclerosis, with the latter being localized to the infiltrating inflammatory cells. Furthermore, its expression was positively correlated with dietary cholesterol levels in MΦs as well as hepatocytes, where it was found to regulate cholesterol efflux. The molecular function of CHROMR has been linked to sterol-activated liver X receptors (LXR), key transcription factors known to regulate the response to cholesterol homeostasis. The mode of action of CHROMR differs from the aforementioned MΦ lncRNAs. Specifically, CHROMR interacts with Argonaute 2 protein (key protein in miRNA-induced silencing complex), and its cellular action on MΦ cholesterol efflux is mediated by sequestering miR-27b, miR-33a, miR-33b, and miR-128 in the cytoplasm [133]. Among these CHROMR-associated miRNAs, circulating miR-27b was found to be downregulated in smokers predisposed to PAD [135], while circulating miR-33a was found to be elevated in T2D patient serum [136], and inhibition of miR-33a/b reduced atherosclerosis [137]. Interestingly, CHROMR is also increased by viral infection, e.g., Influenza A and SARS-COV-2, for which DM poses significant risks to developing severe thrombotic and inflammatory vasculopathies [133, 138]. Given the crucial role of CHROMR in MΦ metabolism, it would be interesting to investigate its regulatory role in DM-PAD.

LncRNAs in vascular smooth muscle cells (VMSCs)

As the immediate adjacent cell types to ECs, VSMCs and pericytes are mural cells found in the walls of blood vessels. VSMCs receive EC-derived signals, such as NO, cyclic guanosine monophosphate (cGMP), and endothelin-1, and are responsible for vaso-constriction/relaxation to regulate blood flow and vascular tone [139]. While VSMCs constitute the media of the blood vessel wall, pericytes are VSMC counterparts in the capillaries and small venules. Under homeostatic conditions, VSMCs primarily regulate blood vessel diameter and structural integrity in larger vessels, while pericytes are key to the stability and regulation of microvasculature. In a hyperglycemic state, VSMC can undergo proliferation, dedifferentiation, and phenotypic switching, leading to abnormal vascular tone and remodeling, contributing to atherosclerosis and arterial stiffening [16, 140]. On the other hand, the contribution of altered pericyte function in the pathogenesis of PAD is still unclear. However, pericyte dysfunction generally leads to microvascular instability, increased permeability, and impaired angiogenesis, contributing to ischemia and damage in other tissues such as the eye and kidney.

LncRNAs have been shown to regulate various processes contributing to the dysfunction of VSMC and pericytes, as reviewed elsewhere [141, 142]. Some of the lncRNAs discussed earlier in the EC and MΦ sections, e.g., MALAT1, H19, and LINC00607, have also been shown to regulate VSMC function [113, 143, 144]. For example, the EC-enriched LINC00607 is increased by Angiotensin II (Ang II) (which can be elevated in DM) in VSMCs and likely regulates VSMC proliferation [113]. Thus, the regulatory function of these lncRNAs can act in concert to modulate the pathogenesis of DM-PAD. Below, we summarize several well-characterized VSMC-enriched lncRNAs that regulate VSMC proliferation, phenotypic switching, and migration and discuss their potential contribution to DM-PAD.

Smooth muscle and Endothelial cell–enriched migration/differentiation-associated long NonCoding RNA (SENCR) is one of the first lncRNAs identified to be enriched in both VSMCs and ECs encoded by Chromosome 11q24.3 [145]. It is transcribed as an antisense lncRNA from the 5′ end of the Friend leukemia virus integration 1 (FLI1) gene, predominantly cytoplasmic, and poorly conserved between human and mouse [145]. SENCR is downregulated in diabetic conditions, such as in db/db mice and HG and Ang-II-treated VSMCs [146, 147]. In VSMCs, SENCR KD decreases the expression of myocardin (MYOCD), a master regulator of SMCs contractile genes [148], and increases pro-migratory genes, leading to a hyper-motile SMC phenotype. On the other hand, SENCR overexpression in mice inhibits aortic dissection-associated VSMC proliferation, migration, phenotypic switching, and AngII-induced VSMC apoptosis and ECM degradation [147]. Mechanistically, SENCR regulates VSMC function through miRNAs and Forkhead box O (FoxO) proteins [146, 149, 150]. Interestingly, SENCR is also expressed in ECs and induced by laminar shear stress. By interacting with cytoskeletal-associated protein 4 (CKAP4, a noncanonical RBP) and cadherin 5 (CDH5) (encoding vascular endothelial-cadherin (VE-cadherin)), SENCR has been shown to maintain membrane integrity and reduce endothelial permeability [151]. Of note is that the levels of SENCR were found to be markedly lower in ECs isolated from patients with premature CAD [152, 153] and in muscles with CTLI [152]. Collectively, these studies indicate SENCR as an essential regulator for vascular homeostasis, whose gain of function may be beneficial in DM-PAD.

As previously mentioned, MYOCD is a key transcriptional cofactor for serum response factor (SRF), which both form a master switch complex controlling VSMC differentiation and lineage [154]. MYOcardin-induced smooth muscle lncRNA inducer of Differentiation (MYOSLID) [155], previously annotated AC007879.7, is a VSMC-specific antisense lncRNA located on Chromosome 2q33.3 in humans [156]. Despite the proximity to Krϋppel Like Factor 7 (KLF7) and cAMP responsive element binding protein 1 (CREB1), MYOSLID does not appear to regulate these neighboring genes. Interestingly, the human MYOSLID shares conservation with primates but not lower species, including mice, suggesting an evolutionarily acquired function of MYOSLID. MYOSLID is transcriptionally regulated by MYOCD/SRF and localized in the cytosol, where it participates in nuclear localization of megakaryoblastic leukemia 1 (MKL1, another coactivator of SRF and key player in maintaining VSMC differentiation) [157] to promote contractile gene expression governed by MYOCD/SRF, e.g. α-SMA, calponin 1 (CNN1), and myosin heavy chain 11 (MYH11). Thus, MYOSLID acts as a lncRNA amplifier for the maintenance of VSMC differentiation through a feed-forward mechanism.

Cardiac mesoderm enhancer-associated non-coding RNA (CARMN), encoded immediately upstream of miR-143 and − 145, two miRNAs regulating VSMC function, is a cardiomyocyte and VSMC-enriched lncRNA conserved in mice and humans located on Chromosome 18 and 5q32, respectively. Originally identified as a regulator of cardiac differentiation [158], the function of CARMN in VSMC has been characterized by three independent studies [159,160,161]. In high-cholesterol diet-fed atherosclerotic mice, CARMN expression was significantly decreased during the disease progression but restored after returning to a normal diet [160]. Likewise, CARMN expression decreased in human atherosclerotic plaques compared to non-atherosclerotic arteries [160]. In primary human coronary arterial SMCs, CARMN and miR-143 and − 145 were downregulated by proliferative stimuli PDGF-subunit B (PDGF-BB) and oxidized LDL (ox-LDL) [159]. CARMN inhibition, similar to the effect of these stimuli, decreased the contractile markers in VSMC, which appears to be independent of miR-143 and − 145 [159]. The functional importance of CARMN in vivo has been demonstrated using constitutive CARMN-KO mice, which exhibited increased lesion area [159], while VSMC-specific and inducible Carmn-KO mice had exacerbated neointima formation after vascular injury [161]. Intriguingly, the KD of CARMN in atherogenic Ldlr^−/−mice showed reduced VSMC proliferation and atherosclerotic lesions [160]. These seemingly paradoxical findings suggest a potential dose-dependent effect of CARMN, which may also vary between in vitro and in vivo settings. At the molecular level, CARMN seems to bind to both MYOCD [161] and SRF [160, 162], which are crucial for VSMC differentiation and contractile gene program [160, 161]. Given the decreased MYOCD in many vasculopathies, it is possible that the MYOCD-interacting lncRNAs, e.g., MYOSLID and CARMN, and their impaired actions also play a role in the development of DM-PAD.

While SENCR, MYOSLID, and CARMN are downregulated in conditions that induce VSMC dysfunction, the expression of smooth muscle-induced lncRNA enhancer of replication (SMILR) is increased in VSMCs stimulated with IL1-α, and PDGF, two cytokines promoting VSMC-related pathologies [163]. SMILR (previously annotated RP11-94a24.1) is encoded downstream of the proximal gene hyaluronic acid synthase 2 (HAS2) in Chromosome 8q24.13 in humans. SMILR expression increases in both the cytoplasm and nucleus of VSMCs during phenotypic switching, subsequently leading to its release into the extracellular space, which allows for the detection of SMILR in plasma, making it a potential biomarker for patients at risk of atherosclerosis-related PAD. Mechanistically, SMILR forms RNA-RNA hybrids with centromere protein F, which may subsequently recruit the Staufen1 protein to regulate VSMC proliferation and migration [163]. Given its role as a key driver of VSMC proliferation and its response to proliferation and pro-inflammatory cytokines, which are often increased in DM, one may speculate that SMILR may also partake in the progression of DM-PAD. However, like several lncRNAs discussed earlier, SMILR is poorly conserved, making it challenging to interrogate its function in mouse models.

Inflammatory MLK1 Interacting Long Noncoding RNA (INKILN) is another lncRNA recently found to be exclusively expressed in VSMC [164]. INKILN is an intergenic cytoplasmic lncRNA encoded on Chromosome 4q13.3, 20 kb upstream of IL-8 [164]. INKILN and its interacting protein MKL1 [165] are induced in VSMCs undergoing phenotypic switching, treatment with TNF-α and IL-1β, and in injured carotid arteries. INKILN activates pro-inflammatory gene expression in cultured human VSMCs and ex vivo cultured mouse vessels. However, studying INKILN in vivo is not straightforward because, like MYOSLID, SMILR, and SENCR, INKILN is poorly conserved between humans and mice. To tackle this challenge, a humanized INKILN transgenic (Tg) mouse model was created. After ligation injury, the INKILN-Tg mice developed exacerbated neointimal formation in the carotid artery. To explain the underlying molecular mechanism, INKILN was found to interact with and stabilize MKL1 through Ubiquitin Specific Peptidase 10 (UPS10), a deubiquitinating enzyme, leading to MKL1 and p65 transactivation of VSMCs’ pro-inflammatory gene profile [164]. Given the reported pro-inflammatory effect of INKILN in vascular inflammation and VSMC dysfunction, it likely also plays a role in DM-PAD.

In addition to the lncRNAs described above (summarized in Fig. 1), a growing list of lncRNAs have been shown to be altered in DM conditions and affect cellular functions relevant to PAD. While some of these lncRNAs are exclusively or preferentially expressed in one or a few cell types, others tend to regulate the functions in multiple cell types with shared or unique modes of action. We have provided a table listing these lncRNAs and their relevant alteration and function to DM-PAD (Table 1).  

Fig. 1

Role of lncRNAs in vascular cell functions related to DM-PAD. A schematic showing the roles of lncRNAs in the functional regulation of endothelial cells (ECs), macrophages (MΦ) and vascular smooth muscle cells (VSMCs) under conditions related to DM-PAD. The reported mechanism of action and their consequent functional role in each cell type is indicated in the boxes. Arrows placed next to each lncRNA and their processes indicate reported or putative changes of expression or activity under DM-PAD.

Table 1 Long noncoding RNAs in diabetes-associated peripheral arterial disease (place before experimental methods and approaches to studying lncRNAs)

Experimental methods and approaches to studying lncRNAs

As reviewed above, numerous lncRNAs have been implicated in a myriad of cellular processes relevant to DM-PAD. To profile lncRNAs, RNA-sequencing (RNA-seq) is widely used to provide quantitative measurements with high sensitivity. In addition, RNA-seq can also be used to identify alternatively spliced transcripts and different isoforms of lncRNAs, which may confer different biological functions. Two commonly used RNA-seq library preparation workflows are polyadenylic acid (polyA) enrichment and ribosomal RNA (rRNA) depletion. PolyA enrichment employs the oligo(dT)-tagged magnetic beads to enrich mRNAs and polyA-containing lncRNAs, thus excluding the non-polyadenylated RNAs [211]. Alternatively, to retain non-polyadenylated RNAs, rRNA can be removed, e.g., by using magnetic beads coupled to oligonucleotides that specifically hybridize with rRNAs [212, 213]. The remaining pool of transcripts preserves the protein-coding mRNAs, lncRNAs, and other non-coding RNAs, providing a favorable tool for de novo lncRNA profiling and detecting lncRNA transcripts of low abundance [214, 215]. Given the relatively lower abundance of lncRNAs to mRNAs, a higher sequencing depth is generally recommended.

To allow for more effective and accurate detection of lowly expressed lncRNAs, target enrichment techniques utilizing probe-based hybridization, e.g., RNA CaptureSeq, can be used [216,217,218]. Target enrichment is critical for full-length or long-read sequencing of individual lncRNA transcripts, e.g., using Oxford Nanopore for direct RNA sequencing [219, 220] or Pacific Biosciences’ Single Molecule, Real-Time sequencing [221, 222]. Following the initial RNA-seq profiling or RNA-seq data mining, it is highly recommended to validate findings from these high-throughput technologies using traditional approaches, e.g., northern blotting and reverse transcription-PCR (RT-PCR), to avoid false positives. In addition, to validate the 5′ and 3′ terminal sequences of lncRNA transcripts, rapid amplification of cDNA ends (RACE) is a valuable assay. Combined with sequencing, also termed RACE-seq, this method effectively annotates the transcription start site and obtains the full-length sequences of lncRNAs [223].

Because LncRNAs are a class of immensely adaptable molecules that function through RNA-DNA, RNA-RNA, or RNA-protein interaction, their biological functions are closely related to their cellular localization. To determine the subcellular localization of lncRNAs, single-molecule fluorescent in-situ hybridization (smFISH) and fractionation are two standard and complementary methods. smFISH utilizes the amplification of multiple tandem probes that can bind to their RNA targets where their fluorescent signal is enhanced, which allows visualization of low-abundance RNA targets [224, 225]. smFISH enables the quantification of RNA transcripts by molecular counts and the detection of co-localization of multiple lncRNAs. As expected, the key to reliable smFISH detection is the rigorous probe design, which can be facilitated by bioinformatic tools such as ProbeDealer [226]. Due to the low abundance and presence of repetitive elements in some lncRNAs, off-target effects, and microscopy interference such as autofluorescence, lncRNA detection by smFISH could be challenging or infeasible in some cases. Cell fractionation, which physically isolates subcellular compartments based on intact organelle purification or partition along sucrose gradients, can be applied to overcome the challenges of smFISH [227, 228], which can also provide validation for smFISH. RNA extracted from different subcellular fractions can be quantified by RT-PCR for individual lncRNAs or RNA-seq for the entire transcriptome. With this method, lncRNAs have been mapped to chromatin-associated fraction, nuclear soluble fraction, and cytoplasmic fraction.

Recent developments of new techniques such as engineered ascorbate peroxidase (APEX)-catalyzed proximity biotinylation and multiplexed error robust fluorescence in situ hybridization (MERIFISH) have provided more powerful tools for improved lncRNA profiling. Originally utilized as a genetic tag for protein mapping, APEX catalyzes the formation of biotin-phenoxyl radicals between its substrates, biotin-phenol, and hydrogen peroxide in different compartments of live cells and covalently biotinylated nearby endogenous proteins. By cross-linking the protein-RNA in situ, APEX-tagged subcellular proteomes also provide localization information of their binding RNA partners [229, 230]. When combined with RNA Immunoprecipitation, these RNAs can be identified by qPCR or profiled by RNA-seq. APEX-RNA-seq has been shown to identify hundreds of cytoplasmic and nuclear lncRNAs in HEK-293T cells. As a high-throughput version of smFISH, MERFISH was developed to simultaneously detect hundreds of RNA transcripts using combinatorial FISH labeling with encoding schemes in successive rounds of hybridization and imaging [231, 232]. However, extensive optimization is needed to apply MERFISH to detect multiple lncRNAs within different subcellular compartments, esp. in the nucleus. As each of these techniques has strengths and limitations, one should consider using at least two independent methods to confirm the subcellular localization of a lncRNA candidate.

As many lncRNAs exert their activity via RNA-protein interactions, various modern approaches have increased the capacity to identify their protein partners for investigating lncRNA molecular mechanisms and pathways. Cross-linking and immunoprecipitation (CLIP) or RBP immunoprecipitation (RIP) can be employed to examine if the given lncRNA is binding to a protein of interest [233, 234]. Multiple techniques have also been developed to capture lncRNA-chromatin interaction in either one vs. the whole genome or all RNAs vs. the whole genome manner [235], among which the chromatin isolation by RNA purification (ChIRP) is the most cited method and has been used to identify genomic sites and RBPs bound by many lncRNAs [236, 237].

To investigate the cellular functions of lncRNAs, experiments that alter RNA function are commonly used to isolate one variable. The gain-of-function of lncRNA function typically uses viral vectors, such as adenovirus, lentivirus, and adeno-associated virus. For example, we have used adenovirus to overexpress human LEENE RNA in the mouse ablated with the leene homolog, which improved ischemic recovery [108]. To study the loss-of-function of lncRNAs, the most common techniques used are RNA interference (RNAi) and antisense oligos (ASOs). While RNAi is mainly active in the cytoplasm and thus more suitable to target cytoplasmic lncRNAs, ASO, esp. the LNA-GapmeRs, can induce the RNase H-dependent RNA degradation and therefore inhibit the nuclear-localized lncRNAs [238,239,240,241], as demonstrated by many studies including the ones cited in this article. Besides these perturbations at the RNA levels, various clustered, regularly interspaced short palindromic repeats (CRISPR) systems are powerful tools for both gain and loss-of-function studies. CRISPR interference (CRISPRi) and activation (CRISPRa) facilitate efficient control over lncRNA transcription inhibition or activation by coupling nuclease-deficient Cas9 (dCas9) with a transcriptional repressor or activator, such as Krüppel associated box (KRAB) or VP64, respectively [242]. In addition, to tether a specific lncRNA to specific DNA loci, CRISPR-DISPLAY (CRISP-DISP) can be employed, which uses dCas9 to deploy RNA cargos to DNA loci as directed by guide RNA [243, 244].

To investigate the in vivo function of lncRNAs, one can employ a similar approach as used for protein-coding genes, most typically genetic deletion in mice. However, one needs to be cautioned that deletion of lncRNAs derived from regulatory genomic regions, e.g., enhancers, can ablate the function of DNA rather than RNA. One approach to address this is to re-introduce the lncRNA in conditions where DNA is removed, e.g., as demonstrated for LEENE [108]. Another approach to address this and the low lncRNA homology [245] is to create a knock-in (KI) of human lncRNA expression in mice to study the exogenous expression of human lncRNAs under the in vivo setting, as demonstrated for the VSMC-specific INKILN [164]. A KI lncRNA mouse model is generated by targeted insertion of the lncRNA gene or cDNA at a selected locus, most commonly at the genomic locus homologous to humans. DNA without any functional gene or regulatory elements flanks the insert, and homologous recombination allows the target of the trans-gene to that specific integration site. These genetic engineering strategies can be combined with tissue-specific targeting for protein-coding gene studies. Alternatively, gain- or loss-of-function of lncRNA may be achieved by using viral vectors or ASOs as introduced for cell-based studies, using careful consideration of the administration route based on the lncRNA features and research aims. Interested readers are recommended to read several reviews discussing the utilization of mouse models to study lncRNAs [246,247,248,249]. Upon obtaining the relevant mouse models, one can subject them to relevant models of diabetes(e.g., either STZ-induced T1D or special diet-induced T2D) or atherosclerosis (e.g., through induction of proprotein convertase subtilisin/kexin type 9 (PCSK9) under a high-cholesterol diet), which can be further combined with the femoral artery ligation to induce HLI, the most commonly used model for PAD [68, 250,251,252,253]. We have provided a graphical summary of methods to discover, validate, and functionally characterize lncRNAs above (Fig. 2).

Fig. 2

Experimental methods and approaches for lncRNA study. A Discovery and validation of lncRNAs using a RNA-seq and CaptureSeq and b using Northern blotting, RT-qPCR, and smFISH. B Characterization of lncRNA function. a Depending on their subcellular localization, lncRNAs mediate miRNA sponging, protein-protein interactions, or mRNA stability in the cytoplasm or epigenetic regulation, chromatin state modulation, or RNA splicing in the nucleus. b RNA interactions with proteins or DNA can be identified through eCLIP- and ChIRP-seq. c Investigation of the role of lncRNAs in cells through gain-of-function using of Ad, AAV, and lentivirus or loss-of-function via siRNA or ASO. CRISPRa, CRISPRi, and CRISPR-Display can be used to reveal chromatin-associated lncRNA functions. d In vivo studies using mouse models with CRISPR-KO, Tg, and CRISPR-KI of lncRNA of interest. smFISH single-molecule fluorescence in situ hybridization, eCLIP: enhanced crosslinking and immunoprecipitation, ChIRP chromatin isolation by RNA pulldown, Ad adenovirus, AAV adeno-associated virus, ASO antisense oligonucleotides, RISC RNA-induced silencing complex, CRISPRa/i CRISPR activation/interference, gRNA guide RNA, CRISPR-KO CRISPR-knockout, Tg transgenic, CRISPR-KI CRISPR-Knockin

Conclusion and perspectives

LncRNAs have emerged as crucial regulators of various cellular processes, including those in several vascular cell types discussed herein. These RNA molecules, which do not encode proteins, influence gene expression and cellular function through diverse mechanisms such as chromatin remodeling, transcriptional control, and post-transcriptional regulation. In the context of vascular cells, lncRNAs play significant roles in endothelial function, smooth muscle cell behavior, and the inflammatory responses critical to vascular health. Importantly, in DM-PAD, a condition characterized by reduced blood flow and vascular complications due to DM, lncRNAs have been implicated in disease progression and pathogenesis. Altered lncRNA expression in diabetic conditions can exacerbate vascular dysfunction, contributing to the development and progression of PAD.

While most of the lncRNA studies focused on their intracellular function, recent studies have revealed the presence of lncRNAs in circulation, either in a cell-free format or encapsulated in extracellular vesicles [254,255,256], hinting at a potential role of lncRNA in cell communication and its potential as biomarkers for a disease state. At the time of this review, no publication has reported any systemic lncRNA profiling in circulating peripheral blood monocytes, serum, or plasma samples of patients with PAD. However, as discussed earlier, several lncRNAs, e.g., H19, have shown diagnostic potential in CAD [89, 257,258,259]. It will be of interest to explore the diagnostic value of lncRNAs in PAD and DM-PAD.

Targeting lncRNAs represents an innovative therapeutic strategy for DM-PAD. Compared to protein-coding genes, their non-coding feature, relatively low expression levels, and high tissue/cell type-specificity may provide higher specificity and lower off-target effect with a lower necessary dosage [27]. Emerging evidence supports the idea that the impaired angiogenic signaling pathways downstream of the angiogenic growth factor can be governed by lncRNAs [47]. Therefore, modulating the levels of these lncRNAs can likely yield additive therapeutic effects through cooperative mechanisms. The most developed approach to targeting lncRNAs is ASO-based therapeutics [260, 261]. Several ASOs targeting oncogenic lncRNAs are currently being tested clinically [262, 263], which will yield valuable information for the therapeutic potential of targeting lncRNAs for non-cancerous diseases such as DM-PAD. Complementary to inhibiting the disease-driving lncRNA, another approach would be to restore the DM-impaired lncRNAs that would otherwise confer anti-inflammatory and pro-repair effects. This may be achieved by viral vectors or modified RNAs, which have been tested in clinical trials for PAD therapeutics [264] and proven successful in the vaccines against SARS-CoV2 [265,266,267,268].

In conclusion, the accumulating evidence supports a highly complex and integrative lncRNA-mediated regulation of various cell types driving DM-PAD. The dysregulated lncRNAs in DM can impair vascular homeostasis and regenerative capacity and provoke prolonged vascular inflammation, remodeling, and damage. Beyond what has been discussed in this review, lncRNA dysregulation in other cell types, such as pericytes and fibro/adipogenic progenitors, may also exacerbate vascular dysfunction [269,270,271]. Improved understanding of the intricate molecular mechanisms governing context-dependent lncRNA functionality across various cell types and careful evaluation of their diagnostic and therapeutic potential in DM-PAD presents a new opportunity to address this challenging medical condition.