miR-21 normalizes vascular smooth muscle proliferation and improves coronary collateral growth in metabolic syndrome

Abstract

Inadequate cell proliferation is considered a major causative factor for impaired coronary collateral growth (CCG). Proangiogenic growth factors (GFs) stimulate cell proliferation, but their administration does not promote CCG in patients. These GFs are increased in patients with metabolic syndrome and in animal models, where CCG is impaired. Here, we investigated whether excessive cell proliferation underlies impaired CCG in metabolic syndrome. Normal [Sprague-Dawley (SD)] and metabolic syndrome [James C. Russell (JCR)] rats underwent repetitive ischemia (RI; transient, repetitive coronary artery occlusion and myocardial ischemia). We have shown that CCG was maximal at d 9 of RI in SD rats but did not occur in JCR rats. The increase in cell proliferation (PCNA, Ki-67, cyclin A, phospho- cdc2, p21Waf, p27Kip) was transient (∼4-fold, d 3 RI) in SD rats but greater and sustained in JCR rats (∼8- to 6-fold, d 3–9 RI). In JCR rats, this was associated with increased and sustained miR-21 expression and accumulation of proliferating synthetic vascular smooth muscle cells in the lumen of small arterioles, which failed to undergo outward expansion. Administration of anti-miR-21 blocked RI-induced cell proliferation and significantly improved CCG in JCR rats (∼60%). miR-21-dependent excessive cell proliferation in the later stages of collateral remodeling correlates with impaired CCG in metabolic syndrome.—Hutcheson, R., Chaplin, J., Hutcheson, B., Borthwick, F., Proctor, S., Gebb, S., Jadhav, R., Smith, E., Russell, J. C., Rocic, P. miR-21 normalizes vascular smooth muscle proliferation and improves coronary collateral growth in metabolic syndrome.

Increase in cell proliferation is a widely accepted component of collateral growth induced by repetitive ischemia (RI; transient, repetitive coronary artery occlusion and myocardial ischemia) in normal, healthy animals. However, neither proliferation of specific cell types nor the time course of cell proliferation during collateral growth have actually been directly investigated. Furthermore, cell proliferation has not been studied in animal models in which collateral growth is impaired, including metabolic syndrome (1, 2). Endothelial cell (EC) proliferation in collateral growth has not been directly evaluated; however, impaired EC proliferation is a hallmark of impaired angiogenesis. Inhibitors of EC proliferation, endostatin and angiostatin, are up-regulated in type II diabetes (3), and inhibition of their production improves coronary collateral growth (CCG) in James C. Russell (JCR) rats (4). Thus, EC proliferation is likely decreased in metabolic syndrome. Endothelial progenitor cells (EPCs) are also decreased in metabolic syndrome (5). Thus, while it is fairly clear that EC and EPC proliferation are decreased when collateral growth is impaired, proliferation of other cell types essential to collateral growth in these scenarios has never been investigated. The idea that proliferation of all cell types is impaired in CCG in metabolic syndrome is based on incomplete observations and inferences.

The coronary vasculature of patients with metabolic syndrome and of JCR rats even very early in progression of vascular disease is characterized by neointimal lesions (6). In JCR rats, as in patients with metabolic syndrome, these lesions consist primarily of highly proliferative, synthetic vascular smooth muscle cells (VSMCs) and macrophages (6). Because growth factors (GFs) and cytokines that promote VSMC proliferation and monocyte and neutrophil infiltration are produced during collateral growth and are elevated in metabolic syndrome (7), we hypothesized that RI further increased VSMC proliferation and monocyte and neutrophil infiltration. In this study, we sought to determine whether excessive VSMC proliferation was associated with impaired CCG in metabolic syndrome.

MicroRNA-21 (miR-21) is now well established as a proproliferative, antiapoptotic miR that promotes proliferation and cell survival via down-regulation of PTEN and consequent up-regulation of PI3-kinase and Akt signaling, as well as by up-regulating mitochondrial prosurvival signals (Bcl-2; ref. 8). miR-21 is up-regulated in the neointima following vascular injury, and its down-regulation decreases neointima formation (8). miR-21 depletion decreased proliferation of cultured VSMCs (9). However, whether miR-21 is involved in regulation of collateral growth or myocardial angiogenesis has never been investigated, and its role in cancer angiogenesis is controversial. miR-21 induced tumor angiogenesis via Akt and ERK1/2 activation and HIF-1α expression (10). In contrast, miR-21 inhibited angiogenesis by decreasing RhoB expression and actin stress fiber formation (11). We hypothesize that these discrepancies are due to the fact that miR-21 is necessary for vascular growth, including angiogenesis and collateral growth, but that the amount and the timing of its expression must be strictly regulated. Thus, we also determined whether RI-induced VSMC proliferation in metabolic syndrome was miR-21 dependent.

MATERIALS AND METHODS

Rat model of CCG/RI

Male, 10- to 12-wk-old Sprague-Dawley (SD) rats (300–350 g; Charles River, Wilmington, MA, USA) or JCR:LA-cp rats (650–700 g; J. C. Russell and S. Proctor, University of Alberta, Edmonton, AB, Canada) were used for chronic (0–9 d) implantation of a pneumatic occluder over the left anterior descending coronary artery (LAD) as described previously (4, 12, 13). A suture was passed under the proximal portion of the LAD, and the occluder was sown onto the surface of the heart. The occluder catheter was externalized between the scapulae. When the occluder is inflated, the suture is pulled toward the surface of the heart, and the LAD is occluded. The LAD perfusion territory is termed the collateral-dependent zone (CZ) because perfusion in this area, while the LAD is occluded, depends on the development of coronary collaterals. The animals underwent the RI protocol, consisting of eight 40 s occlusions, once every 20 min (2 h, 20 min total) followed by a rest period of 5 h, 40 min. This 8 h cycle was repeated 3×/d for 0–9 d. Surgical procedures were performed in accordance with the Animal Welfare Act and are approved by the institutional animal care and use committees of the University of South Alabama, New York Medical College, and the University of Alberta.

The JCR rat is a cross between the lean LA/N Zucker rat and the spontaneously hypertensive obese (SHROB) rat developed in the laboratory of Dr. Carl Hansen (U.S. National Institutes of Health, Bethesda, MD, USA) and sent to Dr. James C. Russell. By 8 wk of age, the JCR rats develop obesity with fatty liver, insulin resistance with glucose intolerance, complex dyslipidemia (low HDL, high LDL and vLDL), and vasculopathy characterized by decreased endothelium-dependent and -independent vasorelaxation and intimal lesions morphologically identical to early atherosclerotic lesions in humans. By 12 wk, the rats exhibit widespread atherosclerosis, left ventricular hypertrophy, and myocardial and cerebral microinfarctions. At 16+ wk, the rats are prone to stroke and myocardial infarction, and at 18+ wk, they develop heart failure. Like the development of the metabolic syndrome and cardiovascular disease in humans, the apparent complexity of the cardiometabolic phenotype exhibited by the JCR rats is suspected to be multifactorial and polygenetic in etiology (6, 14).

VSMC culture

Coronary VSMCs were isolated from SD and JCR rats by Langendorff perfusion using 2 mg/ml collagenase and cultured in 10% FBS and 5% glucose DMEM (Sigma, St. Louis, MO, USA) overnight. VSMCs were then transferred to 0.1% FBS DMEM and treated with anti-miR-21 for 24 h where indicated, then transferred back to 10% FBS DMEM and exposed to intermittent hypoxia (3% O₂, 8 h, to mimic coronary occlusion), hyperoxia (80% O₂, 8 h, to mimic reperfusion), and normoxia (18% O₂, 8 h), to mimic RI, for a total of 48 h.

Anti-miR-21 delivery

Cultured VSMCs were treated with 60 nM locked nucleic acid (LNA)-modified anti-miR-21 (Exiqon, Woburn, MA, USA) for 24 h prior to exposure to hypoxia-hyperoxia-normoxia. JCR rats were treated with the LNA-modified anti-miR-21 at 2 mg/kg in 100 μl of sterile saline via intracardiac injection directly into the left ventricle (LV) cavity as described previously for anti-miR-145 (12) according to modification of previously used protocols for tail vein injection (15) on d 5 of RI. Scrambled LNA-anti-miR sequence was used as control.

miR quantitation

Myocardial, lung, aorta, kidney, liver, and cultured VSMCs

Total RNA was isolated from VSMCs or the heart [normal zone (NZ) and CZ] with QIAzol followed by small RNA isolation with miRNeasy Mini Kits (Qiagen, Valencia, CA, USA). Total and small RNA concentration and quality were determined by absorbance at 260/280 nm. The ratio of 18S and 28S ribosomal RNA and the degree of DNA contamination were assessed by agarose gel electrophoresis with Sybr Green II staining. cDNA synthesis and quantitative RT-PCR were performed with TaqMan miR assays using 250 ng RNA. Absolute quantities of miR-21 in CZ and NZ cardiac tissue were obtained by quantitative RT-PCR using standards constructed from a dilution series of the miR-21 standard (Integrated DNA Technologies, Coralville, IA, USA). miR-21 was normalized to rat U6 RNA, extracted, and processed in the manner identical to miR-21, in lung, aorta, kidney, liver and cultured rat VSMCs and to Caenorhabditis elegans U6 RNA, which was added in identical quantities to all samples immediately after tissue homogenization, then processed in the manner identical to miR-21 in whole heart samples. Experiments were performed on n = 6 animals/time point (day of RI) and were analyzed by 2-way ANOVA followed by Bonferroni correction. P < 0.05 determined statistical significance.

Enterocyes

Rats were denied access to food overnight (16 h) before euthanasia under isofluorane anesthesia, and jejunal enterocytes were collected from the intestine as per an adapted protocol of the Weiser method (16). In brief, enterocytes were detached by incubation of jejunal tissue segments (∼30 cm) with dithiothreitol and citrate with shaking. Isolated cells were pelleted by centrifugation (1000 rpm, 2 min). Total RNA (TRIzol; Invitrogen, Carlsbad, CA, USA) was isolated from enterocytes, and the purity and integrity were checked using an RNA Agilent (Agilent RNA 6000 Nano; Agilent Technologies Inc., Santa Clara, CA, USA). Total RNA with an RNA integrity number ≥ 8 was deemed to be sufficient quality for subsequent microRNA (miRNA) analysis (Exiqon). Enterocyte miRNA profile was analyzed using the miRCURY LNA miRNA array (6th-generation rodent; Exiqon). Experiments were performed on n = 5 animals and were analyzed by Student's t test. P < 0.05 determined statistical significance. miR-21 was normalized to rat U6 RNA, extracted, and processed in a manner identical to myocardial miR-21.

Lymphatic cells

Rats were subjected to mesenteric lymph cannulation, as described previously (17). Briefly, following overnight food withdrawal, rats were fully anesthetized, and a cannula was inserted into the mesenteric lymph duct. A second cannula was inserted into the upper duodenum. Following surgery, rats were restrained in Bollman cages with free access to water and were closely monitored throughout the experiment. Following recovery from anesthesia, rats were administered a constant gastric infusion of saline for 6 h (unfed state). The subsequent total lymph secreted was collected over time. Lymph HDL fractions were separated by density ultracentrifugation. Briefly, a discontinuous NaCl/KBr gradient was formed by adjusting the density of lymph to 1.3 g/ml with KBr and underlying below-density solutions 1.020 and 1.063 g/ml (HDL), and separated by centrifugation (59,900 g; 4°C; 2 h). Single representative lymph samples were generated by pooling 5 biological replicates for each group. Total RNA was extracted from each representative sample using the Qiagen miRNeasy Mini Kit and reverse transcribed into cDNA (miRCURY LNA Universal RT miRNA PCR, polyadenylation, and cDNA synthesis kit; Exiqon). miRNA profile was assessed using the miRCURY LNA Universal RT miRNA PCR, rodent panel I and II (Exiqon) in a LightCycler 480 Real-Time PCR system (Roche Diagnostics, Indianapolis, IN, USA). Experiments were performed on n = 5 animals and were analyzed by Student's t test. P < 0.05 determined statistical significance. miR-21 was normalized to rat U6 RNA, extracted, and processed in a manner identical to myocardial miR-21.

Western blots

Western blots were performed as described previously (2, 4, 12, 18). Hearts were excised; the LV was dissected, the CZ was separated from the nonischemic NZ, and the heart was snap-frozen in liquid nitrogen before homogenization in lysis buffer containing 0.1% SDS and 1% Triton. Cultured VSMCs were lysed in lysis buffer containing 0.1% SDS and 1% Triton. Equal amounts of protein (30 μg) were separated by SDS-PAGE and transferred to Hybond-ECL nitrocellulose membranes (GE Healthcare Biosciences, Pittsburgh, PA, USA). Anti-PCNA, cyclin BI, cyclin A, p21Waf, p27Kip, phospho-specific anti-cell division control protein 2 (cdc2; Thr14/Tyr15) and Rb(Ser807/Ser811/Ser780) (Millipore, San Diego, CA), and anti-osteopontin (Abcam, Cambridge, MA, USA) were used for Western blotting. Bands were visualized by enhanced chemiluminescence (GE Healthcare Biosciences) and quantified using Un-Scan-It Image software (Silk Scientific Corp., Orem, UT, USA). Experiments were performed on n = 6 animals/time point (day of RI) and were analyzed by 2-way ANOVA followed by Bonferroni correction. P < 0.05 determined statistical significance.

Immunohistochemistry (IHC)

IHC analysis was performed as described previously (12). Formalin-fixed, paraffin-imbedded cardiac tissue was cut into 10 μm sections. Primary anti-smooth muscle myosin heavy chain (SM-MHC; 1:200; Abcam), anti-nonmuscle myosin heavy chain B (NMHC-B; 1:200, Abcam), anti-proliferating cell nuclear antigen (PCNA; 1:200, Millipore, San Diego, CA, USA), anti-Ki-67 (1:200; Millipore), anti-von Willebrand factor (vWF; 1:100; Abcam), and anti-osteopontin (1:200; Abcam) and secondary Alexa350-, Alexa488-, and Alexa568-conjugated (Invitrogen) antibodies were used. Blue (Alexa350), green (Alexa488), and red (Alexa568) fluorescence were visualized and representative images collected using a Nikon fluorescent microscope equipped with Nikon Elements software (Nikon, Tokyo, Japan). Images are representative of n = 6 animals/time point (day of RI) from 5 consecutive cardiac cross-sections per animal and 5 separate fields per slide.

Histology

Arteriolar and capillary density, lumen diameter, wall thickness, and wall thickness to lumen diameter ratio were measured in cardiac cross-sections on d 9 of RI as described previously (12). Formalin-fixed, paraffin-embedded cardiac tissue was cut into 5 μm sections. For measurements of arteriolar and capillary densities, a 1 mm² grid was superimposed over SM-MHC-stained (for arterioles/arteries) and hematoxylin- and eosin-stained (for capillaries and arterioles/arteries) cardiac cross sections and vessels <20 μM in diameter and >20 μM in diameter inside the grid were counted as capillaries and arterioles/arteries, respectively. Fully formed arterioles/arteries were additionally identified by the presence of SM-MHC-positive VSMCs. Larger arteries were distinguished from veins by assessment of vessel wall thickness, where wall thickness/lumen diameter ratio >0.25 signified an artery and <0.25 a vein (19). Average collateral diameter was determined by averaging all SM-MHC- and PCNA-positive vessels in SD rats and all NMHC-B-, SM-α-actin-, and PCNA-positive vessels in JCR rats on d 9 of RI. NMHC-B and SM-α-actin were used as VSMC markers due to lack of contractile, SM-MHC-positive VSMCs in the JCR rats. Data were collected from n = 6 animals/group from 5 consecutive cross-sections per animal and 5 separate 1 mm² grids per slide.

In situ hybridization (ISH) was performed using the Exiqon miRCURY LNA miRNA ISH Optimization Kit 7 FFPE (Exiqon) using double DIG-labeled miRCURY LNA miR-21 detection probes at a concentration of 40 nM. Scrambled miRNA probe was used for a negative control, and 5′-DIG-labeled U6 snRNA was used for a positive control. Hybridization was performed on formalin-fixed, paraffin-embedded tissue, cut into 5 μM cross-sections, mounted on slides, and handled according to the manufacturer's protocol. Images are representative of n = 6 animals/group from 5 consecutive cardiac cross-sections per animal and 5 separate fields per slide.

Myocardial and collateral-dependent blood flow measurements were performed as described previously (4, 12, 20). Color microspheres (5×10⁵, 15 μM diameter) labeled with samarium (d 0 RI; initial surgery) or gold (d 10 RI) were injected into the LV during LAD occlusion. Arterial reference blood samples (carotid) and heart tissue from the NZ and the CZ were collected, weighed, and sent to BioPal (Worcester, MA, USA) for analysis. Blood flows to the NZ and the CZ (ml/min/g) were calculated as [(radioactive count in myocardial tissue) × (blood reference withdrawal rate)/(radioactive count in blood reference)]/(weight of myocardial tissue). Blood flows were measured in the following groups of animals: SD RI, JCR RI, JCR RI + anti-miR-21, and JCR + scrambled anti-miR. Results are expressed as the CZ/NZ flow ratio on d 9 of RI. All experiments were performed on n = 8 animals/group. Results were analyzed by 2-way ANOVA followed by Bonferroni correction. P < 0.05 determined statistical significance.

RESULTS

Cell proliferation in response to RI is excessive in metabolic syndrome

Total myocardial cell proliferation, measured by PCNA expression, was ∼2-fold higher in JCR vs. SD rats at baseline (d 0 RI), then increased transiently (∼4±1-fold d 3 RI; ∼2±1-fold d 6 RI) in the CZ of SD animals and returned to baseline by d 9 of RI. In contrast, in the CZ of JCR rats, total cell proliferation increased to a greater extent on d 3 of RI (∼8±1-fold) and remained elevated (∼6±1-fold) on d 6 and 9 of RI (Fig. 1A).

Figure 1.

SD or JCR rats were euthanized on d 0, 3, 6, or 9 of RI. Tissue samples were collected from the NZ or the CZ. A) Left panels: representative anti-PCNA and β-tubulin (loading control) Western blots. Right panels: cumulative data, n = 6. B) Representative images of cardiac cross-sections (CZ) stained with anti-Ki-67 and anti-vWF.

Ki-67-positive staining indicates that this proliferation was specific to coronary arteries (Fig. 1B). Furthermore, prominent Ki-67 staining persisted in the CZ of JCR rats for the duration of the RI protocol, whereas it was nearly completely absent after d 3 of RI in SD rats (Fig. 1B). Changes in regulators of cell cycle progression further support these findings. p27Kip down-regulation and Rb phosphorylation were transient (d 6 RI) in SD rats, but sustained in JCR animals. p21Waf increased in SD but decreased in JCR rats on d 9 of RI. Cyclin A expression and phosphorylation increased and cdc2 phosphorylation decreased transiently and returned to baseline by d 6 of RI in SD rats, but were sustained in JCR rats (Fig. 2). These results demonstrate an early and transient increase in actively proliferating cells in SD rats, but a sustained increase in these cells in JCR rats. Colocalization of NMHC-B (which is expressed by synthetic but not by contractile VSMCs) and PCNA, shown in Fig. 4, identifies a significant portion (∼60%) of neointimal cells on d 6 of RI and the majority (∼90%) of neointimal cells on d 9 of RI to be synthetic proliferating VSMCs. High NMHC-B and very low SM-MHC (which is expressed exclusively by contractile VSMCs) expression confirm the findings of our recent study in which VSMCs of JCR rats failed to convert to the contractile phenotype in the later days of the RI protocol (12). This correlated with formation of large, mature collaterals on d 9 of RI in CZs of SD rats, whereas CZs of JCR rats exhibited immature collaterals that failed to undergo luminal expansion (Fig. 3).

Figure 2.

Top panels: representative Western blots of anti-p21Waf, p27Kip, phospho-Rb, cyclin A, phospho-cdc2, and β-tubulin (loading control). Bottom panels: cumulative data, n = 6.

Figure 4.

JCR rats were euthanized on d 6 and 9 of RI. Representative images of cardiac cross-sections (CZ) stained with anti-SM-α-actin (red), DAPI (blue), PCNA (green), SM-MHC (red), and NMHC-B (blue) are shown.

Figure 3.

Representative images of cardiac cross-sections (CZ) stained with hematoxylin and eosin.

RI-induced miR-21 up-regulation correlates with excessive cell proliferation in metabolic syndrome, and miR-21 down-regulation blocks cell proliferation

Since miR-21 induces VSMC proliferation in culture, we investigated whether it was induced by RI. RI increased miR-21 transiently in SD rats (∼2.5-fold d 3, ∼2-fold d 6 RI) but in a sustained manner in JCR rats (∼3-fold d 3, ∼4-fold d 6 RI) (Fig. 5A). Consistent with previous reports in the literature, miR-21 is expressed ubiquitously in all cell types; however, its expression during CCG is increased in proliferating neointimal cells (Fig. 5B), proportional to levels measured by RT-PCR in Fig. 5A. The neointimal cells are mostly synthetic VSMCs, which proliferate, and neutrophils, which are recruited to the CZ, but once there do not proliferate (Fig. 4). Thus, miR-21 expression in SD and JCR rats correlates with VSMC proliferation. Exposure to repetitive hypoxia-hyperoxia-normoxia induced an ∼3-fold increase in miR-21 in JCR but not in SD VSMCs in primary cultures (Fig. 5C). miR-21 down-regulation by anti-miR-21 blocked hypoxia-hyperoxia-normoxia-induced coronary VSMC proliferation in primary culture (Fig. 6B) and RI-induced total cell and VSMC proliferation in vivo (Fig. 6B–D), conclusively demonstrating miR-21 dependence of excessive RI-induced cell proliferation in metabolic syndrome. Scrambled, nontargeting anti-miR had no effect on cell proliferation (Fig. 6B, D). Anti-miR-21 delivery does not result in miR-21 degradation but in formation of stable, inert miR-21/anti-miR-21 duplexes, as shown in Fig. 6A.

Figure 5.

A) SD or JCR rats were euthanized on d 0, 3, 6, or 9 of RI. Tissue samples were collected from the NZ or the CZ. Mature miR-21 levels were determined by RT-PCR, n = 6. B) Representative ISH images of cardiac cross-sections using miR-21-specific probes (blue staining). C) Primary SD and JCR VSMCs were subjected to intermittent hypoxia-hyperoxia-normoxia for 48 h. miR-21 levels were determined by RT-PCR, n = 3. D) Mature miR-21 levels were determined by RT-PCR in enterocytes, total lymph, and the HDL lymph fraction of SD and JCR rats at baseline (no RI) as indicated, n = 5–6.

Figure 6.

A) Top left panel: primary VSMCs (JCR) were treated with anti-miR-21 where indicated, then subjected to intermittent hypoxia-hyperoxia-normoxia for 48 h (SD and JCR). Top right panel: rats (JCR) were treated with anti-miR-21 on d 5of RI where indicated and euthanized on d 9 of RI (SD and JCR). Tissue samples were collected from the NZ or the CZ. miR-21 levels were determined by RT-PCR, n = 3 for cultured VSMCs, n = 6 for rats. Representative Northern blot shows miR-21/anti-miR-21 duplex formation. Bottom panel: representative ISH images of cardiac cross-sections using miR-21-specific probes (blue staining). B) VSMCs and rats were treated as in A. PCNA expression was measured by Western blot. Cumulative data are n = 3 for cultured VSMCs, n = 6 for rats. C) Rats were treated with anti-miR-and euthanized as in A. Representative images of cardiac cross-sections (CZ) stained with anti-PCNA, SM-MHC, and NMHC-B are shown. D) Rats (JCR) were treated with anti-miR-21 on d 5 of RI where indicated and euthanized on d 9 of RI (SD and JCR). Tissue samples were collected from the NZ or the CZ. Cyclin A, phospho-cdc2, p21Waf, and p27Kip expression was measured by Western blot (bottom). Cumulative data are n = 6. See panel B for loading control (blots were sequentially stripped and reprobed).

Basal miR-21 level was also increased in the myocardium of JCR rats (Fig. 5A) and in JCR VSMCs (Fig. 5C) without RI, which correlated with increased basal cell proliferation in JCR rats (Figs. 1 and 2). Interestingly, miR-21 expression was likewise significantly increased (∼3.5-fold) in jejunal enterocytes from JCR rats vs. control (Fig. 5D). Lymphatic concentration of miR-21 in JCR rats was also increased in both total lymph (∼3-fold) and in the lymph-HDL (∼3.75-fold) fraction (Fig. 5D) collected from the mesenteric lymphatic duct. miR-21 levels in enterocytes were specifically examined because in the rat the intestine contributes most of the lipid content to lymphatics as either native chylomicrons (CMs) or HDL. We have tested for miRs to reside on CMs; however, our results indicated that 95% of miRs in lymphatics are associated with the HDL fraction and not the CM fraction (unpublished results). From the mesenteric lymphatic duct, the contents then drain directly into the portal vein and into the circulation.

miR-21 down-regulation correlates with significant improvement in CCG in metabolic syndrome

Anti-miR-21 treatment correlated with significant collateral expansion on d 9 of RI (41±7 μM average collateral lumen diameter in JCR + anti-miR-21 vs. 22±4 μM in untreated JCR vs. 60±15 μM in SD; Fig. 6C). This was confirmed by measurements of coronary and collateral-dependent blood flow. Myocardial blood flow in the CZ and the NZ was measured in the same rats in which anti-miR-21 was shown to decrease cell proliferation (Fig. 6). At d 9 of RI, mean CZ flow was 0.52 ± 0.05 ml/min/g in anti-miR-21-treated JCR rats vs. 0.12 ± 0.04 ml/min/g in untreated JCR rats and 1.84 ± 0.06 ml/min/g in untreated SD rats. CZ/NZ flow ratio was 0.56 ± 0.05 (JCR RI+anti-miR-21), 0.16 ± 0.03 (JCR RI) and 0.85 ± 0.05 (SD RI) (Fig. 7B). This increase was due to increased collateral artery formation and not angiogenesis, since arteriolar density increased from ∼1 to ∼2 > 20 μM lumen diameter arteriol/artery per square millimeter, but capillary density did not increase (Fig. 7A). Because anti-miR-21 very significantly decreased cell proliferation and we used PCNA-positive staining to identify collateral arteries, it is possible that collateral number in anti-miR-21-treated JCR rats is in fact underestimated. Scrambled, nontargeting anti-miR had no effect on coronary blood flows or arterial or capillary densities (Fig. 7).

Figure 7.

JCR rats were treated with anti-miR-21 or scrambled (nontargeting) anti-miR on d 5 of RI where indicated, and SD or JCR rats underwent 9 d of RI, n = 6. A) Arteriolar (SM-MHC- and PCNA-positive vessels >20 μM) and capillary (vessels <20 μM) densities were determined in cardiac (CZ) cross-sections on d 9 of RI. B) Coronary blood flow was measured in the CZ and the NZ using microspheres during LAD occlusion and is expressed as the ratio between CZ and NZ flows on d 9 of RI.

Anti-miR-21 itself is neither cell nor tissue specific. However, using our delivery protocol (dose and timing), anti-miR-21 was preferentially taken up by neointimal cells, as demonstrated by decreased or absent miR-21-positive staining of the neointimal cells but not of the cells of the vascular media, adventitia, or surrounding myocytes in anti-miR-21-treated JCR animals (Fig. 6A). Furthermore, because it was not delivered systemically (via tail vein injection), but directly into the LV cavity, anti-miR-21 down-regulated miR-21 most strongly in the heart, specifically in the CZ (∼70%) vs. NZ (∼50%) vs. aorta (thoracic ∼25%, abdominal ∼10%), liver (∼25%), kidney (∼5%), or lung (∼0%) (Supplemental Fig. S1). Therefore, the effects of anti-miR-21 delivery on CCG in this study are largely related to its effects on neointimal cells, which are synthetic, proliferating VSMCs, as demonstrated in this study, and neutrophils (unpublished results).

DISCUSSION

Increased proliferation is a necessary part of early collateral remodeling (21). This is consistent with our findings demonstrating transient cell proliferation in early CCG in normal animals in this study. The later phases of CCG include luminal expansion, and VSMCs return to the contractile phenotype with negligible cell proliferation (21). This is the first study to demonstrate that, contrary to belief, in metabolic syndrome RI in fact induces a persistent increase in total cell proliferation, which is incompatible with luminal expansion and increased collateral-dependent blood flow.

The most significant portion of this excessive and temporally inappropriate proliferative response appears to be taken up by synthetic VSMCs. We did not seek to identify the other predominant cell types that accumulate in the lumen of immature collaterals of JCR rats during the late days of RI (d 6–9) in this study. However, in a closely related study, we have identified these cells to be primarily (>90%) neutrophils, which contribute to impaired CCG primarily by their abnormal, miR-21-mediated extended survival (unpublished results). The remaining PCNA-positive unidentified cells visible on d 6 RI along the internal lamina in the JCR collaterals are likely the small number of monocytes or macrophages also observed in our other study (unpublished results).

GFs and proliferation-promoting cytokines augment collateral growth in normal animals (7). VEGF is required for CCG (22). However, GFs and cytokines failed to stimulate collateral growth in clinical trials in which retrospective analysis discovered that the majority of patients satisfied diagnostic criteria for metabolic syndrome (23, 24). Likewise, VEGF did not restore CCG in a rat model of metabolic syndrome (25). Unlike in normal animals, where proangiogenic GFs increase transiently early in CCG, then decline as collaterals develop and ischemia lessens (26), VEGF and PDGF are consistently elevated in metabolic syndrome (7). Furthermore, increased endothelial permeability in metabolic syndrome provides continual access for GFs to VSMCs. Our study identifies VSMCs as major contributors to RI-induced excessive cell proliferation in metabolic syndrome. These data correlate with our recently published study demonstrating failure of JCR VSMCs to return to the quiescent, contractile phenotype (12). The sustained elevation in GFs could drive this sustained VSMC proliferation. Our study definitively identifies miR-21 as the mediator of excessive RI-induced VSMC and total cell proliferation in metabolic syndrome. miR-21 down-regulation has recently been shown to decrease EGFR expression (27) suggesting an interplay between this miR and GF signaling potentially relevant to CCG regulation.

We propose that a partial explanation for failure of GFs to stimulate collateral growth in metabolic syndrome is their propensity to restore EC and EPC proliferation, which is impaired, but simultaneously cause excessive proliferation and luminal accumulation of other cell types, in large part synthetic VSMCs. Continual proliferation and extracellular matrix deposition by these cells may interfere with collateral expansion in later stages of collateral growth in part by creating a physical obstacle to blood flow and thus shear stress-induced NO release and NO-dependent signaling activation, which is essential for initiation of collateral expansion.

In addition to normalizing VSMC proliferation, down-regulation of miR-21 significantly improved CCG in our study. We believe that the mechanism by which miR-21 regulates CCG is more complex than via its effect on cell proliferation alone. In the study parallel to this one, the effect of anti-miR-21 on CCG was nearly identical to that of blocking neutrophil infiltration (unpublished results). However, a major factor that increases neutrophil survival in the periphery from mere hours to days (28) is an extracellular matrix protein, osteopontin, which is synthesized and secreted chiefly by proliferative, synthetic VSMCs. Thus, we believe that these two processes, excessive and sustained VSMC proliferation and prolonged neutrophil survival, which are regulated by the same upstream mediator, miR-21, are not independent of each other but are instead completely interconnected, so that one promotes the other, and both together lead to CCG impairment.

This study suggests that while the initial, inward remodeling phase of coronary collateral growth remains intact in metabolic syndrome, the later, outward remodeling is compromised. The mechanisms underlying outward remodeling in collateral growth remain to be identified. Matrix metalloproteinases (MMPs) could be involved. We have previously shown that MMP2, MMP9, MMP7, and MMP14 (MT1-MMP) were activated during coronary collateral growth in normal but not in JCR rats (4, 20) on d 3 RI, which precedes significant outward remodeling in the rat by 3 d, and thus could be involved in preparing the extracellular matrix for outward expansion of the vessel. We have also shown that inhibition of MMP12 activation significantly improved CCG in JCR rats (20). Restoration of CCG by MMP12 inhibition was fully accounted for by inhibition of MMP12-dependent angiostatin and endostatin production. Angiostatin and endostatin effects are restricted to EC survival, proliferation, and, less so, migration. The precise mechanisms by which EC survival and proliferation effect outward remodeling are unknown, but it is known that they will increase NO production in response to VEGF and shear stress. NO release in response to VEGF and increased shear stress and their downstream signaling, leading to decreased endothelial activation and the switch of VSMCs from synthetic to contractile phenotype, has been proposed as a mechanism of outward remodeling. JCR rats and patients with metabolic syndrome are characterized by endothelial dysfunction with elevated oxidative stress (2, 29, 30). We have also recently shown that JCR VSMCs are incapable of converting to the contractile phenotype in the later stages of CCG due to low expression of miR-145 (12).

Our study also showed elevated basal levels of miR-21 in the myocardium of JCR animals. This was not due to increased fibrosis in the JCR myocardium (unpublished results). However, elevated basal cardiac miR-21 correlated with elevated miR-21 levels in the total lymph, HDL-lymph fraction, and enterocytes from JCR animals. HDLs have been identified as a major transporter of miRNA (31, 32) in the circulation. More recently, the lymphatic system was revealed to be critical for the metabolic turnover of HDL (33) and the reverse cholesterol transport system. These results suggest that altered enterocyte function and/or HDL-dependent lymphatic transport underlie elevated cardiac miR-21 levels. Therefore, our study highlights the connection between altered metabolism, gastrointestinal system, and cardiovascular disease in the metabolic syndrome, which requires further study.