Skip main navigation

Deregulation of microRNA-503 Contributes to Diabetes Mellitus–Induced Impairment of Endothelial Function and Reparative Angiogenesis After Limb Ischemia

Originally published 2011;123:282–291



Diabetes mellitus impairs endothelial cell (EC) function and postischemic reparative neovascularization by molecular mechanisms that are not fully understood. microRNAs negatively regulate the expression of target genes mainly by interaction in their 3′ untranslated region.

Methods and Results—

We found that microRNA-503 (miR-503) expression in ECs is upregulated in culture conditions mimicking diabetes mellitus (high D-glucose) and ischemia-associated starvation (low growth factors). Under normal culture conditions, lentivirus-mediated miR-503–forced expression inhibited EC proliferation, migration, and network formation on Matrigel (comparisons versus lentivirus.GFP control). Conversely, blocking miR-503 activity by either adenovirus-mediated transfer of a miR-503 decoy (Ad.decoymiR-503) or by antimiR-503 (antisense oligonucleotide) improved the functional capacities of ECs cultured under high D-glucose/low growth factors. We identified CCNE1 and cdc25A as direct miR-503 targets which are downregulated by high glucose/low growth factors in ECs. Next, we obtained evidence that miR-503 expression is increased in ischemic limb muscles of streptozotocin-diabetic mice and in ECs enriched from these muscles. Moreover, Ad.decoymiR-503 delivery to the ischemic adductor of diabetic mice corrected diabetes mellitus–induced impairment of postischemic angiogenesis and blood flow recovery. We finally investigated miR-503 and target gene expression in muscular specimens from the amputated ischemic legs of diabetic patients. As controls, calf biopsies of nondiabetic and nonischemic patients undergoing saphenous vein stripping were used. In diabetic muscles, miR-503 expression was remarkably higher, and it inversely correlated with cdc25 protein expression. Plasma miR-503 levels were also elevated in the diabetic individuals.


Our data suggest miR-503 as a possible therapeutic target in diabetic patients with critical limb ischemia.

Ischemic complications represent the leading cause of morbidity and mortality in diabetic patients.1 Early in the course of diabetes, intracellular hyperglycemia causes endothelial dysfunction and microvascular rarefaction.2 The overall result is tissue hypoperfusion, which, in limbs, results in the formation of not-healing ulcers. Moreover, diabetes mellitus impairs endogenous reperfusion accomplished by reparative angiogenesis, thereby worsening the recovery from an ischemic insult.2 In diabetic patients, the ischemic disease follows an inexorable course, and limb amputation is too often the ultimate remedy.2 A better understanding of the molecular mechanisms underpinning diabetes mellitus–associated vascular complications is urgently needed to improve therapeutic options.2

Editorial see p 236

Clinical Perspective on p 291

Because of their incapacity to regulate glucose influx, endothelial cells (ECs) represent an important target for diabetes mellitus–induced damage. In particular, it is well established that ECs cultured in high glucose show delayed replication,3,4 abnormal cell cycling,5 and increased apoptosis.6 Progression through the cell cycle is a tightly regulated process that includes multiple checkpoints. An orderly expression of cyclin-dependent serine/threonine kinases (cdks), their binding partners, cyclins (CCNs), and associated regulatory proteins modulates entry into each of the cell-cycle phases. In particular, CCNDs activate Cdk4 and/or Cdk6 in early G1 phase, whereas CCNEs activate Cdk2 in the late G1 phase, leading to the passage into the S-phase of the cell cycle. CCNA/Cdk2 complexes assure S phase progression, and CCNB/Cdk1 complexes complete mitosis (M).7 However, among Cdks and CCNs, widespread compensations exist.7 Phosphatases of the cell division cycle 25 family (cdc25) are critical for timely Cdk1 and Cdk2 activation, thus regulating G1-S and G2-M transitions.8 Regulation of cell-cycle progression through modulation of cdc25A, CCNE1, CCND1 and CCND2 has been reported in ECs.9,10

microRNA (miRNAs) are small noncoding RNAs of ≈22-nucleotides. miRNAs are transcribed as precursors, processed by the RNase III enzymes Drosha and Dicer, and ultimately assembled into cytoplasmic protein-RNA complexes called RISCs (RNA-Induced Silencing Complexes), which mediate degradation of targeted messenger RNA (mRNA) transcripts and/or translational arrest, mainly through imperfect base pairing with the 3′ untranslated region (3′-UTR). miRNAs are important regulators of gene expression.11 Initially described for having pathogenic roles in cancer, miRNAs have become a hot-topic as biomarkers and therapeutic targets in cardiovascular diseases.12,13 Because a single miRNA can control multiple genes, modulation of 1 pathogenic miRNA could have a greater chance of correcting complex molecular derangements compared with targeting a specific gene. Hence, manipulation of miRNAs may open new avenues for molecular therapeutics of complex diseases, like diabetes mellitus and vasculopathies. Recently, some miRNAs have been shown to regulate EC functions implicated in angiogenesis, including EC proliferation, migration, and assembling in branched networks.14,15

The present study is the first to provide evidence for a role of miRNAs in diabetes mellitus–induced endothelial defects contributing to impaired postischemic angiogenesis. In fact, here we show that in vitro culture conditions mimicking diabetes mellitus and ischemia upregulate miR-503 in ECs and that, in vivo, diabetes mellitus increases miR-503 expression in ECs from ischemic limb muscles. We also show that increased miR-503 is responsible for repressed cdc25A and CCNE1 expression in ECs cultured under conditions mimicking diabetes mellitus and ischemia. Moreover, miR-503–forced expression inhibited EC proliferation, migration, and network formation on Matrigel and it additionally reduced vascular smooth muscle cell (VSMC) proliferation and migration, 2 processes that are instrumental for arteriogenesis.16 Furthermore, adenovirus-mediated local gene transfer with a decoy for miR-503 improved blood flow recovery and angiogenesis in diabetic mice with limb ischemia. Finally, miR-503 was highly expressed in both ischemic muscle specimens and the plasma of diabetic patients with critical limb ischemia. Our results suggest that miR-503 could be a novel therapeutic target in diabetic patients with peripheral ischemic complications.


An expanded version of Methods in the online-only Data Supplement includes detailed methods for the following: cells and cell culture; RNA extraction and Taq-Man Quantitative Real Time, lentivirus preparation, luciferase assays, preparation of an adenovirus for decoymiR-503 (Ad.decoymiR-503), Western blot analyses, cell cycle, cell biology assays, mouse limb ischemia and in vivo gene transfer, enrichment of ECs from limb muscles, and flow cytometry.

Cells and Cell Culture

Human umbilical vein ECs (HUVECs) and human microvascular ECs (HMVECs) were grown in EGM-2 (EBM-2 added with growth factors and other supplements) with 2% FBS. Human VSMCs cells were cultured in 10% FBS Dulbecco modified Eagles medium. To mimic hyperglycemia, ECs were incubated in 25 mmol/L D-glucose (high glucose [HG]). D-Mannitol was used as osmotic control (normal glucose condition [NG]). To mimic ischemia-induced tissue starvation, ECs were incubated in EBM-2 with 2% FBS, only (low growth factors [LGF]).

RNA Extraction and Taq-Man Quantitative Real-Time Analysis

Total RNA was extracted using TRIzol. Real-time quantification to measure miRNAs was performed with the TaqMan miRNA reverse transcription kit and miRNA assay (Applied Biosystems, Carlsbad, CA) with the DNA Engine Opticon 2 system (Bio-Rad Laboratories, Hercules, CA). miR-503 expression was normalized to the U6 small nucleolar RNA (snRU6). Primer identification numbers are: 4373228 for hsa-miR-503 and 4427975 for snRU6 (Applied Biosystems). For mRNA analysis, single-strand complementary DNA was synthesized from 1 μg of total RNA. Complementary DNA was amplified by real-time polymerase chain reaction (PCR) and normalized to 18S ribosomal RNA (endogenous control). Each reaction was performed in triplicate. Quantification was performed by the 2ΔΔCt method.17 Quantitative reverse transcription PCR (RT-PCR) was used to measure cdc25A, CCNE1, CCND1, CCND2, eGFP and 18S ribosomal RNA. RT-PCR primers are reported in the online-only Data Supplement.

Lentivirus Preparation and In Vitro Infection

Lentiviral vector-expressing premiR-503 (lenti.miR-503) and lenti.GFP (control) were generated. HUVECs, HMVECs and VSMCs were infected at 25 multiplicity of infection.

antimiRNA Transfection

A miR-503 inhibitor oligonucleotide (Applied Biosystems, 50 nmol/L) or a scrambled oligonucleotide was transfected into HUVECs or COS-7 cells using Lipofectamine 2000 (Invitrogen, Carlsbad, CA).

Identification of Target mRNAs of miRNA-503

To determine the gene targets of miR-503, 5 leading miRNA target prediction algorithms (TargetScan 4.1, miRanda, miRbase, Diana microT 3.0, and EIMMo2) were used (online-only Data Supplement Table IA).

Luciferase Assays

To investigate whether miR-503 directly regulates cdc25A and CCNE1 expression, portions of the 3′-UTR of these potential target genes were inserted downstream of a luciferase open reading frame. CCNE1 3′-UTR (S204537) and cdc25A 3′-UTR (S213182) vectors were purchased from SwitchGear Genomics (Menlo Park, CA). For controls, we prepared similar vectors in which 5 nucleotide mutations were inserted in the 3′-UTR sequences (cdc25a: 675 to 681; CCNE1: 248 to 254 and 486 to 492) complementary to the miR-503 “seed” sequence. For CCNE1, plasmids with single o double mutations in 3′UTR were prepared. Primers and mutation sequences are reported in online-only Data Supplement Table IIA and IIB. The different luciferase constructs were transfected into COS-7 cells together with either premiR-503 or a scrambled oligonucleotide sequence (control). COS-7 cells were chosen for their high efficiency of transfection and for the absence of endogenous miR-503 expression (data not shown). Cells were cultured for 48 hours and assayed with the Dual-Luciferase Reporter Assay System (Promega, Fitchburg Center, WI). Luciferase assay for CCND1 was omitted because previously performed by others.18 Luciferase assay for CCND2 was omitted because we found normal CCND2 protein level in HUVECs engineered to overexpress miRNA-503 (see below).

Ad.decoymiR-503 Preparation and In Vitro Infection

The decoy for miR-503 was designed using multiple copies of complementary targets. This was intended to optimize the transgene repression in the presence of the miRNA (online-only Data Supplement Methods and Figure IA). ECs were infected with Ad.decoymiR-503 or Ad.Null (control) at 250 multiplicity of infection. To study if this titer of adenoviral vector affects EC behavior, HUVECs infected with 250 multiplicity of infection Ad.Null were compared to noninfected (PBS) HUVECs in proliferation and on Matrigel angiogenesis assays.

Western Blot Analyses

Western blots for CCND1, CCND2, CCNE1, cdc25A, and β-actin were performed.

Cell-Cycle Analysis

HUVECs were infected with lenti.miR-503 or lenti.GFP. After fixation in 70% ethanol and RNase A treatment, cells were stained with propidium iodide. DNA content was analyzed by flow cytometry.

BrdU Incorporation Assay

BrdU proliferation assays were performed on HUVECs, HMVECs, and VSMCs infected with lenti.miR-503 or lenti.GFP and then kept in their full media.

Adhesion Assay

At 48 hours from lentiviral infection, HUVECs were seeded on fibronectin-coated 96-well plates. One hour later, nonadherent cells were washed out and adherent cells were fixed and stained with DAPI.

Caspase-Activity Assay

Caspase-3 activity of HUVECs infected with lenti.miR-503 or lenti.GFP was measured.

Migration Scratch Assay

Migration capacities of HUVECs or VSMCs infected with lenti.miR-503 or lenti.GFP and growth arrested were measured using a scratch assay.

Matrigel Assay

HUVECs or HMVECs were infected with lentivirus and seeded in 48-well plates coated with growth factors–enriched Matrigel. Endothelial network formation was quantified by morphometry.

Mouse Limb Ischemia and In Vivo Gene Transfer

The experiments involving mice were performed in accordance with the Guide for the Care and Use of Laboratory Animals prepared by the Institute of Laboratory Animal Resources and with the prior approval of the UK Home Office and the University of Bristol ethics committee.

Type-1 diabetes mellitus was induced by streptozotocin injection in CD1 male mice. Mice at 3 months of diabetes mellitus (without insulin supplementation) and nondiabetic age- and gender-matched controls were anesthetized to induce left limb ischemia. Immediately after, Ad.decoymiR-503 or Ad.Null (109 plaque forming units) was delivered to the ischemic adductor muscle of diabetic mice. Nondiabetic mice received Ad.Null only. The superficial blood flow (BF) of the ischemic and contralateral foot was sequentially analyzed (at 30 minutes and at 7, 14, and 21 days) by color laser Doppler (n=at least 12 mice per group), and the ratio between BF in the ischemic foot and BF in contralateral foot was calculated and used as an index of percentage BF recovery. At 3 and 21 days postischemia, n=6 mice per group were euthanized for molecular biology analyses. At the latter time point, n=6 additional mice per group were perfusion-fixed, and adductors were paraffin-embedded for histological analyses.

Capillary and arteriole densities were determined using fluorescent microscopy on sections stained with Alexa 568-conjugated isolectin-B4 (an endothelial marker) and FITC-conjugated α-vascular smooth actin (VSMC marker).

EC Isolation From Mouse Limb Muscles

At 3 days postischemia induction, adductor muscles of diabetic and nondiabetic mice were harvested and enzymatically digested. ECs were immunomagnetically sorted using a CD146 antibody.19 CD146 is a transmembrane glycoprotein mainly expressed at the intercellular junction of ECs.20 To verify, EC enrichment by our CD146-positive separation, single cell suspensions were incubated with FITC-conjugated CD146 and APC-conjugated CD31 antibodies or the respective isotype for negative control. Unstained and single-stained controls were performed to define positivity. Flow cytometric analysis revealed that 82.5±2.8% of cells isolated from nondiabetic ischemic muscles coexpressed both CD146 and the classic EC marker CD31 and that diabetes mellitus does not significantly change this figure (83.4±2.9%, P=not significant; online-only Data Supplement Figure II).

Human Tissue Studies

Expressional studies were performed on plasma and on muscular specimens taken from either amputated legs of diabetic patients (n=11) or calf biopsies obtained during saphenous vein stripping surgery in nondiabetic patients (controls, n=11). Patient characteristics are reported in online-only Data Supplement Table III. Studies complied with the ethical principles stated in the Declaration of Helsinki and were covered by ethical approval number 11/2009 from IRCCS-Multimedica. Patients gave written informed consent to sample collection.

Statistical Analysis

Group differences of continuous variables were compared by 1-way ANOVA or Student t test as appropriate. Relationships between variables were determined by the Pearson correlation coefficient. Continuous data are expressed as mean±SEM. A P value <0.05 was considered statistically significant. Analyses were performed with GraphPad Prism 5.0, SigmaStat 3.1 software (San Jose, CA). The overall false discovery rate (FDR) was estimated using the R package LBE.


HG/LGF Upregulates miR-503 Expression in Cultured ECs

miR-503 was increased in both HUVECs and HMVECs cultured in HG/LGF (Figure 1A and 1B). Culture in HG/LGF is intended to mimic advanced diabetes mellitus when hyperglycemia is accompanied with tissue starvation.

Figure 1.

Figure 1. miR-503 expression in ECs is increased by culture in HG and LGF. A and B, Relative expression of miR-503 in HUVECs and HMVECs cultured under NG, HG, LGF, or HG/LGF. miR-503 expression was normalized to snRU6 expression using the comparative Ct method.17 Experiments were performed in triplicate and repeated 3 times. miR-503 expression is reported to expression in the NG group. Values are means±SEM. *P<0.05 versus NG; #P<0.05 versus HG; §P<0.05 versus LGF. FDR q values: 0.0021 (A) and 0.0158 (B).

miR-503 Impairs EC Functions

To evaluate the functional consequences of high endothelial miR-503 levels, HUVECs were infected with lenti.miR-503. Successful miR-503 transduction was demonstrated by quantitative RT-PCR analysis (online-only Data Supplement Figure III). As shown inFigure 2A, miR-503–forced expression impaired HUVEC cycle, thus increasing the percentage of cells in G0/G1 and reducing the percentage of cells in S and G2/M phases. The antiproliferative effect of miR-503 on HUVECs was confirmed in a BrdU incorporation assay (Figure 2B). Notably, miR-503–forced expression for 48 hours did not affect EC apoptosis (online-only Data Supplement Figure IVA). The antiangiogenic potential of miR-503 was next tested in a Matrigel assay. HUVECs infected with lenti.miR-503 were impaired in their capacity to form cellular networks (Figure 2C). Furthermore, miR-503–overexpressing HUVECs showed reduced migratory capacity in an in vitro scratch assay (Figure 2D) and decreased adhesiveness to fibronectin (online-only Data Supplement Figure IVB). Forced miR503 expression also inhibited HMVEC proliferation and tube-like structure formation in vitro (online-only Data Supplement Figure VA and VB), as well as VSMC proliferation and migration (online-only Data Supplement Figure VIA and VIB).

Figure 2.

Figure 2. miR-503–forced expression impairs HUVEC functional capacities. HUVECs were infected with either lenti.miR-503 (miR-503) or lenti.GFP (GFP; control). A, Cell-cycle progression was assessed by flow cytometry. The percentage of cells in G1, S, and G2/M phases was calculated. B, HUVEC proliferation was measured by BrdU incorporation. C, Endothelial network formation on Matrigel was quantified at 24 hours from cell seeding and expressed as the total length of tube-like structures (graph on the left) and the number of branching points (graph on the right). Panels above show representative microphotographs (original magnification 100×; scale bar: 100 μm). D, HUVEC migration was assessed in a scratch assay. HUVEC monolayers were scratched. Images (original magnification 50×; scale bars: 250 μm) were acquired immediately after scratching and at 24 hours thereafter. Dashed lines indicate the front of migration. Gap closure was calculated. Experiments were performed in triplicate and repeated 3 times. All values are means±SEM. *P<0.05 versus GFP.

miR-503 Directly Targets cdc25A and CCNE1

In line with data from in vitro assays, bioinformatics interrogation of genes modulated by miR-503 confirmed a functional link with expressional control of cell cycle, adhesion, migration, and angiogenesis (online-only Data Supplement Table IB). We next analyzed putative miR-503 targets predicted by defined criteria.21 Four targets were consistently identified by 4 out of 5 prediction softwares: cdc25A, CCND1, CCND2, and CCNE1 (online-only Data Supplement Table IA). CCND1 has been already validated as direct target of miR-503.18Lenti.miR-503 reduced cdc25A and CCNE1 at mRNA and protein level in HUVECs in comparison with GFP-HUVECs. By contrast, CCND1 was downregulated at protein level only. We decided not to validate CCND2 as miR-503 target because our data indicated that its expression is not modulated by miR-503 in ECs (Figure 3A and 3B). As shown inFigure 3C, using luciferase assay, we validated cdc25A and CCNE1 as direct target genes of miR-503. Significant inhibition of the luciferase activity was observed in cells transfected with the constructs bearing an intact miR-503 binding site when compared with pLUC. Mutation of cdc25A target sequence prevented downregulation of luciferase activity by miR-503. For CCNE1, mutation of each conserved target sequence had only a partial effect, and mutation of both sequences fully rescued CCNE1 repression (Figure 3C).

Figure 3.

Figure 3. cdc25A and CCNE1 are direct target genes of miR-503. A and B, Relative mRNA and protein expression of cdc25A, CCNE1, CCND1, and CCND2 in HUVECs infected with lenti.miR-503 (mi-R503) or lenti.GFP (GFP). Values are means±SEM. *P<0.05 and **P<0.01 versus GFP. C, Insert: sequences of miR-503 binding sites in cdc25A and CCNE1. Complementary nucleotides are indicated by vertical bars. Luciferase activity at 48 hours postcotransfection of COS-7 cells with either miR-503 or miR-scrambled oligonucleotides plus one of the following: pLuc, 3′-UTR-cdc25A, 3′-UTR-cdc25A with a mutation (mut) of the putative miRNA target site, 3′-UTR-CCNE1, 3′-UTR-CCNE1 with either single (either mut1 or mut2) or double (mut1+mut2) mutation of the putative miRNA target sites. Experiments were performed in triplicate and repeated 3 times. Values are means±SEM. *P<0.05; **P<0.01 versus related miR-scramble; ††P<0.01 versus nonmutated 3′-UTR-cdc25a; #P<0.05 and ##P<0.01 versus nonmutated 3′-UTR-CCNE1. FDR q value: 0.0157.

Effect of miR-503 Inhibition in ECs Cultured in HG/LGF

Incubation in HG/LGF for 24 hours decreased HUVEC proliferation (Figure 4A and 4B) and networking capacity on Matrigel (Figure 4C and 4D). In order to understand whether miRNA-503 contributes to these altered functions, we inhibited miR-503 using 2 different approaches: Ad.decoymiR-503 and antimiR503 (antisense oligonucleotide). DecoymiR-503 transcript targeting was preliminarily validated by flow cytometry, which showed a specific reduction of eGFP in COS-7 cells cotransfected with premiR-503. This was consistent with miR-503 binding to the eGFP 3′-UTR (online-only Data Supplement Figure IB and IC). Moreover, reduction of Ad.decoymiR503-associated eGFP in HUVECs cultured in HG/LGF (versus HUVECs cultured in NG) was also revealed by fluorometry (online-only Data Supplement Figure ID). Adenoviral infection of HUVECs did not alter their proliferative and tube-like structure–forming capacities (comparisons: Ad.Null versus PBS; online-only Data Supplement Figure VIIA and VIIB). Most important, Ad.decoymiR-503 rescued HUVEC proliferation (Figure 4A) and networking capacities (Figure 4C) under HG/LGF. Similar results were obtained with antimiR-503 (Figure 4B and 4D). Furthermore, Ad.decoymiR-503 restored cdc25A and CCNE1 protein expression in HG/LGF-cultured HUVECs (Figure 4E). Because CCND1 expression was not inhibited in HG/LGF-cultured HUVECs (data not shown), this cyclin was not further considered.

Figure 4.

Figure 4. miR-503 inhibition restores HUVEC proliferation and tube-like structure formations in HG/LGF. HUVECs were either infected with Ad.decoymiR-503 (decoymiR-503) or Ad.Null (Null) control (A, C, and E) or transfected with antimiR503 or a scrambled sequence control (B and D) and then cultured in HG/LGF or NG for 24 hours. A and B, HUVEC proliferation was assessed by BrdU incorporation. C and D, HUVEC network formation on Matrigel was quantified by measuring the total length of tube-like structures. E, Representative Western blot bands and relative quantification of cdc25A and CCNE1 protein expression. Relative values are calculated after normalization with α-actin levels (loading control). Experiments in A, B, C, and D were performed in triplicate and repeated 3 times. Western blot: n=6 per group. All values are means±SEM. *P<0.05 and **P<0.01 versus NG+Null; +P<0.05 and ++P<0.01 versus HG/LGF+Null; #P<0.05 versus NG+scramble; §P<0.05 versus HG/LGF+scramble. FDR q values: 0.0184 (A), 0.0175 (B), 0.0165 (C), 0.0135 (D), and 0.0073 (E).

Inhibition of miR-503 Normalizes Postischemic BF Recovery and Muscular Neovascularization in Diabetic Mice

miR-503 was remarkably higher in ischemic muscles (3 days postischemia) of diabetic mice in comparison to nondiabetic and/or nonischemic controls (Figure 5A). Importantly, miR-503 was expressed by CD146-positive ECs freshly isolated from murine ischemic limb muscles, with mi-R503 expression being enhanced by diabetes mellitus (Figure 5B).

Figure 5.

Figure 5. miR-503 inhibition normalizes postischemic neovascularization and blood flow recovery in diabetic mice. Unilateral limb ischemia was induced in diabetic and nondiabetic mice. A, Relative expression (to nonischemic and nondiabetic mice) of miR-503 in ischemic (3 days postischemia) and contralateral limb muscles of diabetic and nondiabetic mice (n=6 per group). B, Relative expression of miR-503 in CD146-positive ECs isolated from ischemic limb muscles of diabetic and nondiabetic mice (n=6 per group). In A and B, miR-503 expression was normalized to snRU6 expression using the comparative Ct method.17 Then, miR-503 expression was reported to expression in group of the nonischemic nondiabetic muscles. C and D, Ad.decoymiR-503 (decoymiR-503) or Ad.Null (Null) was delivered to the ischemic adductors of diabetic mice. Nondiabetic mice received intramuscular Ad.Null (Null). C, Line graph shows the time course of postischemic foot BF recovery (calculated as the ratio between ischemic foot BF and contralateral foot BF) in nondiabetic mice given Ad.Null (black dotted line) and in diabetic mice given either Ad.Null (red line) or Ad.decoymiR-503 (yellow line) (n=12 mice per group). Representative color laser Doppler images taken at 21 days postischemia induction are shown. D, Column graphs show densities of capillaries and small arterioles (diameter <50 μm) in ischemic adductors at 21 days postischemia (n=6 per group). Representative pictures show capillaries stained by isolectin-B4 (red fluorescence) and arterioles stained by isolectin-B4 and α-smooth actin (green fluorescence) (original magnification 500×; scale bar: 500 μm). E, Representative Western blot analyses for cdc25A and CCNE1 in adductors at 3 days and 21 days postischemia (n=6 per group). Bar graphs show relative protein quantification of cdc25A and CCNE1. Relative values are normalized by α-actin levels. All values are means±SEM. ++P<0.01 versus nonischemic muscle of nondiabetic mice; †P<0.05 and ††P<0.01 versus ischemic muscle of nondiabetic mice; #P<0.05 versus nonischemic muscle of diabetic mice; §P<0.05 and §§P<0.01 versus nondiabetic ischemic mice given Ad.Null; *P<0.05 and **P<0.01 versus diabetic ischemic mice given Ad.Null. FDR q values: 0.0375 (A); 0.0121 (C; Doppler analysis at 14 days); 0.0061 (Doppler analysis at 21 days), 0.0120 (capillary density), and 0.0252 (arteriole density) (D); 0.01921 (cdc25A; 3 days), 0.0106 (cdc25A; 21 days), and 0.0223 (CCNE1; 3 days) (E). Non Diab Isch indicates nondiabetic ischemia; Diab Isch, diabetic ischemia.

Next, we tested if local miR-503 inhibition improves postischemic reparative neovascularization and blood flow recovery in diabetic mice. The capacity of Ad.decoymiR-503 to transduce limb muscles was preliminarily demonstrated by RT-PCR of the eGFP tag in Ad.decoymiR-503–injected nonischemic limb muscles of nondiabetic mice (online-only Data Supplement Figure VIII).

As expected, postischemic foot BF recovery was impaired in diabetic mice receiving Ad.Null versus nondiabetic mice given Ad.Null (Figure 5C). Importantly, Ad.decoymiR-503 normalized BF recovery in diabetic mice (Figure 5C). Likewise, Ad.decoymiR-503 increased capillary and arteriolar densities in ischemic muscles of diabetic mice (Figure 5D). Moreover, as shown inFigure 5E, at both 3 and 21 days postischemia, cdc25A protein was downregulated by diabetes mellitus but normalized by Ad.decoymiR-503. Diabetes mellitus downregulated CCNE1 at 3 days postischemia, with this response being prevented by Ad.decoymiR-503. CCND1 expression was not modulated by diabetes mellitus (data not shown), and consequently we did not check if Ad.decoymiR-503 could alter CCND1 expression.

Expression of miR-503 and Target Genes in Human Samples

miR-503 expression was remarkably higher in limb muscles of diabetic patients undergoing major amputation for critical ischemia in comparison with calf biopsies of nondiabetic and nonischemic control subjects (Figure 6A). Moreover, miR-503 and cdc25A expression were inversely correlated in the diabetic patients (Pearson correlation r=−0.784; P=0.0025) (Figure 6C through 6E). By contrast, no correlation was found between miR-503 and either CCND1 or CCNE1 (data not shown). Finally, plasma levels of miR-503 were ≈15 times higher in diabetic patients with critical ischemia compared with control subjects (Figure 6B).

Figure 6.

Figure 6. miR-503 and cdc25A expression in human samples. A and B, Relative expression of miR-503 in plasma and limb muscles of diabetic patients with critical limb ischemia undergoing lower limb amputation and of nondiabetic and nonischemic control patients undergoing saphenous vein stripping (n=11 patients per group). miRNA expression was analyzed using the TaqMan Real Time PCR assay. Relative expression was calculated using the comparative Ct method.17 Values were normalized to snRU6 expression. Values are means±SEM. *P<0.05 versus nondiabetics. C, Plasma miR-503 expression in each diabetic patient. D, Muscular miR-503 expression in each diabetic patient. E, Muscular cdc25A protein expression (assessed by Western blot) in each diabetic patient. Relative cdc25A values were obtained by normalizing for α-actin; F, Line graph showing the inverse correlation between muscular expression of miR-503 and cdc25A protein in diabetic patients (Pearson correlation r=−0.784; P=0.0045).


miRNAs comprise a class of endogenous small noncoding RNAs that control gene expression by acting on target mRNAs for degradation and/or translational repression. Although increasing evidence indicates that miRNAs are key players in cardiovascular disease, only minor studies have been carried out on the relationship between miRNAs and diabetes mellitus in the vascular system. One study showed that both miR-503 and miR-320 are upregulated in myocardial ECs of Goto–Kakizaki diabetic rats and that miR-320 impairs angiogenesis in vitro.22 The other investigation reported that HG upregulates miR-221 expression in HUVECs and that miR-221 inhibits HUVEC migration in vitro.23 In our study, we demonstrated that miR-503 is involved in diabetic endothelial dysfunction.

Constitutively low miR-503 expression in human ECs is remarkably upregulated by culture in HG/LGF to mimic advanced diabetes mellitus when hyperglycemia is accompanied by tissue starvation. Moreover, we found that mouse ECs freshly isolated from ischemic limb muscles express miR-503 and that diabetes mellitus upregulates miR-503 in these ECs. In addition, miR-503 expression was remarkably increased in ischemic muscles of both type-1 diabetic mice and diabetic patients undergoing foot amputation for critical ischemia. Plasma miR-503 was also elevated in the diabetic and ischemic patients.

We have shown that miR-503–forced expression impairs EC proliferation, migration, adhesion, and network formation capacities. miR-503 also reduced VSMC proliferation and migration, which are instrumental for arteriogenesis.16 Conversely, miR-503 inhibition restored normal EC proliferation and in vitro angiogenesis under HG/LGF, without affecting EC function under normal culture conditions. Thus, miR-503 might be considered a suppressor of postischemic neovascularization in diabetes mellitus and a potential therapeutic target for improving healing of diabetic ischemic tissues. To gain further insights into this issue, we employed a gene therapy approach based on the adenovirus-mediated local delivery of a 3′-UTR decoy for miR-503 to downregulate miR-503 activity in ischemic muscles. Decoys, at least when delivered by viral vectors, can stably antagonize a miRNA without requiring multiple administrations.24 Moreover, local injection of a decoy vector should minimize off-target effects that could follow the systemic administration of other miRNAs inhibitors.24 Decoys have not been previously employed in angiogenesis studies. By contrast, miRNA antisense oligonucleotides, which include cholesterol-conjugated antagomirs, have been already used to demonstrate the role of miR-92 and miR-126 in the postischemic neovascularization process in nondiabetic mice.25,26 Our data show that adenoviral-mediated local delivery of the miR-503 decoy normalizes postischemic muscular neovascularization and blood flow recovery in diabetic mice. To the best of our knowledge, this is the first demonstration of a therapeutic approach targeting a specific miRNA and resulting in relevant benefit for healing of limb ischemia in diabetes mellitus.

Each miRNA is believed to directly bind to and regulate the translation or stability of many target mRNAs. Recognizing these gene targets is essential before interfering with miRNA expression for therapeutic purposes. Using a Web-based computational tool developed to identify molecular pathways potentially altered by miRNAs,27 we identified cell cycle as the main function affected by miR-503. To find the genes directly regulated by miR-503, we performed an in silico analysis using 5 target prediction software programs and considering targets that were identified by at least 4 algorithms. Experimental evidence identified cdc25A, CCND1, and CCNE1 as direct targets of miR-503. However, CCND1 expression was not modulated by either HG/LGF in HUVECs or by diabetes mellitus and ischemia in vivo. Notably, miR-503 inhibition prevented cdc25A and CCNE1 downregulation in ECs cultured in HG/LGF. Similarly, Ad.decoymiR-503 normalized cdc25A and CCNE1 protein levels in the ischemic limb muscles of diabetic animals.

One of the most striking findings in this study was the dramatic increase in miR-503 expression in limb muscle of diabetic patients. Interestingly, in these human samples, miR-503 and cdc25 expression were inversely correlated. By contrast, miR-503 and CCNE1 expression did not correlate, possibly because CCNE1 is under the control of divergent transcriptional commands.28 The high level expression of miR-503 is compatible with the advanced vasculopathy of these patients. Our data suggest that miR-503 and cdc25A may represent therapeutic targets for diabetic patients who are developing muscular ischemia.

miRNAs are protected from RNAse and stable in the blood.29 Therefore, their stability makes miRNAs concentrations well suited for being tested in patient samples as potential biomarkers of different pathological conditions.30 In line with this insight, the plasma levels of miR-503 were dramatically increased in diabetic patients with critical limb ischemia in comparison with controls. Further studies are required to investigate miR-503 as a potential circulating biomarker of ongoing ischemia in diabetic subjects.

In conclusion, we have demonstrated that miR-503 is antiangiogenic in the context of diabetes mellitus. miR-503 is upregulated by diabetes mellitus and ischemia, and it plays a pathogenic role in the diabetes-induced impairment of reparative angiogenesis. The findings of this study highlight important clinical implications of miR-503 in diabetes-associated vascular complications.


Dr Ben Carter and Dr Alessandro Cardinali (Bristol Heart Institute Clinical Trial Office) performed some statistic analyses. We thank Orazio Fortunato, BSc, and Daniela Cordella, BSc (IRCCS-Multimedica), for technical assistance.

Sources of Funding

This work was funded by Medical Research Foundation grant G0901764, British Heart Foundation grants BS/05/001, PG/06/146/21946, and RG/09/005/27915, and European Community integrated project RESOLVE! FP7 HEALTH-F4–2008. Dr Emanueli is a British Heart Foundation Basic Science Senior Research Fellow.




The online-only Data Supplement is available with this article at

Correspondence to Costanza Emanueli, PhD, FAHA,
Chair of Vascular Pathology and Regeneration, Regenerative Medicine Section, School of Clinical Medicine, University of Bristol, Bristol, BS2 8HW, England, UK
. E-mail


  • 1. King H, Aubert RE, Herman WH. Global burden of diabetes, 1995–2025: prevalence, numerical estimates, and projections. Diabetes Care. 1998; 21:1414–1431.CrossrefMedlineGoogle Scholar
  • 2. Spinetti G, Kraenkel N, Emanueli C, Madeddu P. Diabetes and vessel wall remodelling: from mechanistic insights to regenerative therapies. Cardiovasc Res. 2008; 78:265–273.CrossrefMedlineGoogle Scholar
  • 3. Stout RW. Glucose inhibits replication of cultured human endothelial cells. Diabetologia. 1982; 23:436–439.CrossrefMedlineGoogle Scholar
  • 4. Doyle JW, Smith RM, Roth TP. The effect of hyperglycemia and insulin on the replication of cultured human microvascular endothelial cells. Horm Metab Res. 1997; 29:43–45.CrossrefMedlineGoogle Scholar
  • 5. Lorenzi M, Nordberg JA, Toledo S. High glucose prolongs cell-cycle traversal of cultured human endothelial cells. Diabetes. 1987; 36:1261–1267.CrossrefMedlineGoogle Scholar
  • 6. Baumgartner-Parzer SM, Wagner L, Pettermann M, Grillari J, Gessl A, Waldhausl W. High-glucose–triggered apoptosis in cultured endothelial cells. Diabetes. 1995; 44:1323–1327.CrossrefMedlineGoogle Scholar
  • 7. Satyanarayana A, Kaldis P. Mammalian cell-cycle regulation: several Cdks, numerous cyclins and diverse compensatory mechanisms. Oncogene. 2009; 28:2925–2939.CrossrefMedlineGoogle Scholar
  • 8. Ray D, Kiyokawa H. CDC25A phosphatase: a rate-limiting oncogene that determines genomic stability. Cancer Res. 2008; 68:1251–1253.CrossrefMedlineGoogle Scholar
  • 9. Wilson DW, Lame MW, Dunston SK, Segall HJ. DNA damage cell checkpoint activities are altered in monocrotaline pyrrole-induced cell cycle arrest in human pulmonary artery endothelial cells. Toxicol Appl Pharmacol. 2000; 166:69–80.CrossrefMedlineGoogle Scholar
  • 10. Huang S, Chen CS, Ingber DE. Control of cyclin D1, p27(Kip1), and cell cycle progression in human capillary endothelial cells by cell shape and cytoskeletal tension. Mol Biol Cell. 1998; 9:3179–3193.CrossrefMedlineGoogle Scholar
  • 11. Bartel DP. MicroRNAs: genomics, biogenesis, mechanism, and function. Cell. 2004; 116:281–297.CrossrefMedlineGoogle Scholar
  • 12. Couzin J. MicroRNAs make big impression in disease after disease. Science. 2008; 319:1782–1784.CrossrefMedlineGoogle Scholar
  • 13. Small EM, Frost RJ, Olson EN. MicroRNAs add a new dimension to cardiovascular disease. Circulation. 2010; 121:1022–1032.LinkGoogle Scholar
  • 14. Wang S, Olson EN. AngiomiRs: key regulators of angiogenesis. Curr Opin Genet Dev. 2009; 19:205–211.CrossrefMedlineGoogle Scholar
  • 15. Wu F, Yang Z, Li G. Role of specific microRNAs for endothelial function and angiogenesis. Biochem Biophys Res Commun. 2009; 386:549–553.CrossrefMedlineGoogle Scholar
  • 16. Schaper W. Collateral circulation: past and present. Basic Res Cardiol. 2009; 104:5–21.CrossrefMedlineGoogle Scholar
  • 17. Schmittgen TD, Livak KJ. Analyzing real-time PCR data by the comparative C(T) method. Nature Protoc. 2008; 3:1101–1108.CrossrefMedlineGoogle Scholar
  • 18. Jiang Q, Feng MG, Mo YY. Systematic validation of predicted microRNAs for cyclin D1. BMC Cancer. 2009; 9:194. Abstract.CrossrefMedlineGoogle Scholar
  • 19. Schrage A, Loddenkemper C, Erben U, Lauer U, Hausdorf G, Jungblut PR, Johnson J, Knolle PA, Zeitz M, Hamann A, Klugewitz K. Murine CD146 is widely expressed on endothelial cells and is recognized by the monoclonal antibody ME-9F1. Histochem Cell Biol. 2008; 129:441–451.CrossrefMedlineGoogle Scholar
  • 20. Kebir A, Harhouri K, Guillet B, Liu JW, Foucault-Bertaud A, Lamy E, Kaspi E, Elganfoud N, Vely F, Sabatier F, Sampol J, Pisano P, Kruithof EK, Bardin N, Dignat-George F, Blot-Chabaud M. CD146 short isoform increases the proangiogenic potential of endothelial progenitor cells in vitro and in vivo. Circ Res. 2010; 107:66–75.LinkGoogle Scholar
  • 21. Sethupathy P, Megraw M, Hatzigeorgiou AG. A guide through present computational approaches for the identification of mammalian microRNA targets. Nat Methods. 2006; 3:881–886.CrossrefMedlineGoogle Scholar
  • 22. Wang XH, Qian RZ, Zhang W, Chen SF, Jin HM, Hu RM. MicroRNA-320 expression in myocardial microvascular endothelial cells and its relationship with insulin-like growth factor-1 in type 2 diabetic rats. Clin Exp Pharmacol Physiol. 2009; 36:181–188.CrossrefMedlineGoogle Scholar
  • 23. Li Y, Song YH, Li F, Yang T, Lu YW, Geng YJ. MicroRNA-221 regulates high glucose-induced endothelial dysfunction. Biochem Biophys Res Commun. 2009; 381:81–83.CrossrefMedlineGoogle Scholar
  • 24. Brown BD, Naldini L. Exploiting and antagonizing microRNA regulation for therapeutic and experimental applications. Nat Rev Genet. 2009; 10:578–585.CrossrefMedlineGoogle Scholar
  • 25. Bonauer A, Carmona G, Iwasaki M, Mione M, Koyanagi M, Fischer A, Burchfield J, Fox H, Doebele C, Ohtani K, Chavakis E, Potente M, Tjwa M, Urbich C, Zeiher AM, Dimmeler S. MicroRNA-92a controls angiogenesis and functional recovery of ischemic tissues in mice. Science. 2009; 324:1710–1713.CrossrefMedlineGoogle Scholar
  • 26. Wang S, Aurora AB, Johnson BA, Qi X, McAnally J, Hill JA, Richardson JA, Bassel-Duby R, Olson EN. The endothelial-specific microRNA miR-126 governs vascular integrity and angiogenesis. Dev Cell. 2008; 15:261–271.CrossrefMedlineGoogle Scholar
  • 27. Maragkakis M, Reczko M, Simossis VA, Alexiou P, Papadopoulos GL, Dalamagas T, Giannopoulos G, Goumas G, Koukis E, Kourtis K, Vergoulis T, Koziris N, Sellis T, Tsanakas P, Hatzigeorgiou AG. DIANA-microT web server: elucidating microRNA functions through target prediction. Nucleic Acids Res. 2009; 37:W273–W276.CrossrefMedlineGoogle Scholar
  • 28. Moroy T, Geisen C. Cyclin E. Int J Biochem Cell Biol. 2004; 36:1424–1439.CrossrefMedlineGoogle Scholar
  • 29. Mitchell PS, Parkin RK, Kroh EM, Fritz BR, Wyman SK, Pogosova-Agadjanyan EL, Peterson A, Noteboom J, O'Briant KC, Allen A, Lin DW, Urban N, Drescher CW, Knudsen BS, Stirewalt DL, Gentleman R, Vessella RL, Nelson PS, Martin DB, Tewari M. Circulating microRNAs as stable blood-based markers for cancer detection. Proc Natl Acad Sci U S A. 2008; 105:10513–10518.CrossrefMedlineGoogle Scholar
  • 30. Chen X, Ba Y, Ma L, Cai X, Yin Y, Wang K, Guo J, Zhang Y, Chen J, Guo X, Li Q, Li X, Wang W, Zhang Y, Wang J, Jiang X, Xiang Y, Xu C, Zheng P, Zhang J, Li R, Zhang H, Shang X, Gong T, Ning G, Wang J, Zen K, Zhang J, Zhang CY. Characterization of microRNAs in serum: a novel class of biomarkers for diagnosis of cancer and other diseases. Cell Res. 2008; 18:997–1006.CrossrefMedlineGoogle Scholar

Clinical Perspective

MicroRNAs (miRNAs) are post-transcriptional inhibitory regulators of gene expression that bind to complementary messenger RNA transcripts. After initial studies in developmental biology and cancer, miRNAs have recently come into focus of cardiovascular diagnostics and therapeutics. Because each miRNA can repress many target mRNAs, it is possible that dysregulation of a single miRNA might account, at least in part, for complex pathological situations. Here, we report for the first time the importance of miR-503 in diabetes mellitus–associated ischemic disease, which currently represents a major cause of morbidity and mortality in diabetic patients. In vitro, the combination of high glucose and starvation remarkably enhances the expression of miR-503 in human endothelial cells, and so does diabetes mellitus in endothelial cells extracted from murine ischemic limb muscles. In vitro experiments showed that forced expression of miR-503 inhibits endothelial cell proliferation and endothelial network formation. Because miR-503 represses cell cycle–associated genes, we investigated whether miR-503 activation may impinge on postischemic reparative angiogenesis. In a diabetic mouse model of limb ischemia, local inhibition of miR-503 activity accelerated vascular healing and blood flow recovery. Importantly, miR-503 was found up-regulated in muscular biopsies and peripheral blood–derived plasma of diabetic patients with critical limb ischemia. From a therapeutic perspective, manipulation of miR-503 may represent a novel molecular means to foster reparative angiogenesis in diabetic patients. In the diagnostic context, more studies are necessary to determine if miR-503 could be exploited as a biomarker of progressive vascular disease.


eLetters should relate to an article recently published in the journal and are not a forum for providing unpublished data. Comments are reviewed for appropriate use of tone and language. Comments are not peer-reviewed. Acceptable comments are posted to the journal website only. Comments are not published in an issue and are not indexed in PubMed. Comments should be no longer than 500 words and will only be posted online. References are limited to 10. Authors of the article cited in the comment will be invited to reply, as appropriate.

Comments and feedback on AHA/ASA Scientific Statements and Guidelines should be directed to the AHA/ASA Manuscript Oversight Committee via its Correspondence page.