E-selectin-targeting delivery of microRNAs by microparticles ameliorates endothelial inflammation and atherosclerosis

E-selectin is a surface marker of endothelial cell (EC) inflammation, one of the hallmarks of atherogenesis. Thus, we tested the hypothesis that delivery of microRNA (miR)-146a and miR-181b with an E-selectin-targeting multistage vector (ESTA-MSV) to inflamed endothelium covering atherosclerotic plaques inhibits atherosclerosis. Cy5-conjugated miR-146a and miR-181b were packaged in polyethylene glycol-polyethyleneimine (PEG/PEI) nanoparticles and loaded into ESTA-MSV microparticles. Both miRs were downregulated in tumor necrosis factor (TNF)-α-treated ECs. Transfection of TNF-α-treated mouse aortas and cultured ECs with miRs was more efficient with ESTA-MSV than with the PEG/PEI. Likewise, miR-146a/-181b packaged in ESTA-MSV efficiently suppressed the chemokines, CCL2, CCL5, CCL8, and CXCL9, and monocyte adhesion to ECs. Complementary in vivo tests were conducted in male apolipoprotein E-deficient mice fed a Western diet and injected intravenously with the particles prepared as above biweekly for 12 weeks. Treatment with miRs packaged in ESTA-MSV but not in PEG/PEI reduced atherosclerotic plaque size. Concurrently, vascular inflammation markers, including macrophages in aortic root lesions and chemokine expression in aortic tissues were reduced while the vascular smooth muscle cells and collagen increased in plaques from ESTA-MSV/miRs-treated vs. vehicle-treated mice. Our data supported our hypothesis that ESTA-MSV microparticle-mediated delivery of miR-146a/-181b ameliorates endothelial inflammation and atherosclerosis.

Cell adhesion assay. Transfected HMVECs were plated into 96-well plate overnight and then treated with TNF-α for 6 h. The human monocytic cell line, THP-1, was obtained from American Type Culture Collection (ATCC) and cultured in ATCC-formulated RPMI 1640 medium (GIBCO, Gaithersburg, MD) containing 10% fetal bovine serum. THP-1 cells were labeled with CellTracker Green CMFDA Dye (Life Technologies, Gaithersburg, MD) and washed twice with RPMI 1640 medium, and then 1 × 10 5 labeled cells were added to the EC monolayer, and incubated in a CO 2 incubator for 1 h on a rocker. Nonadherent cells were removed from the plate by gentle washing with PBS, and the number of THP-1 cells per view was also quantified from randomly acquired images.
Ex vivo aorta transfection. C57BL/6J (Jackson Laboratory) mice were sacrificed by CO 2 anesthesia. The aortas were freed from the surrounding connective tissue using a dissecting stereomicroscope. The vessels were cut into small segments (~2 mm length) and opened longitudinally with fine scissors and placed in 96-well plate. The en face preparations of aortas were cultured in Dulbecco's Modified Eagle Medium (DMEM, GIBCO) and transfected with Cy5-labeled miRs packaged in particles as described above for 24 h. The mean fluorescence intensity of Cy5 (Ex/Em = 650 nm/670 nm) was measured using an Infinite M1000 microplate reader (Tecan, Mechelen, Belgium).
In vivo delivery of miRs in mice. All protocols concerning animal use were approved by the Institutional Animal Care and Use Committee (IACUC) of Houston Methodist Research Institute and the methods were carried out in accordance with the approved guidelines. The 8-week-old male ApoE −/− mice, purchased from Jackson Laboratory, were fed a Western diet (TD.88137; Harlan Laboratories Inc., Indianapolis, IN) for 12 weeks. Meanwhile, mice were injected with PEG/PEI-vehicle, PEG/PEI-miR-146a, PEG/PEI-miR-181b nanoparticles, or ESTA-MSV-vehicle, ESTA-MSV-miR-146a, ESTA-MSV-181b microparticles biweekly for 12 weeks via the tail vein. Each particle injection contained 15 μg miRs in 100 μL, per mouse. Intraperitoneal glucose tolerance test (IPGTT). After 12-week feeding and particle injection, the IPGTT was performed as described 17 . Briefly, mice were fasted overnight for 12 h and then injected intraperitoneally with glucose (1 g/kg body weight). The tail vein blood glucose level was measured with an automated glucometer (Bayer, Elkhart, IN) at baseline and 15, 30, 60, and 120 min after the injection.
En face staining of aorta. Mice were sacrificed by excess inhalation of carbon dioxide. Hearts and aortas were perfused in situ with phosphate-buffered saline (PBS) and then removed and placed on ice cold PBS. The perivascular fat and connective tissue around the aorta were carefully cleaned, and the aortic arch and thoracic aorta were cut longitudinally, fixed with isopropanol, and stained with oil red O (ORO) for 15 min. The images were captured by a digital camera.
ESTA-MSV enhances transfection efficiency. We transfected the TNF-α -treated C57BL/6J mouse aortas with PEG/PEI nanoparticles and ESTA-MSV microparticles loaded with Cy5-conjugated miRs, and according to the mean fluorescence intensity, transfection efficiency was significantly higher in microparticle vs. nanoparticle groups (Fig. 1D,E). Transfection of the TNF-α -treated HMVECs with same particles revealed that both markedly increased the miR levels. Importantly, the miR levels were higher in microparticle-vs. nanoparticle-transfected ECs (Fig. 1F,G).
ESTA-MSV/miRs improve endothelial function of ApoE −/− mice. ApoE −/− mice were fed a Western diet and intravenously injected with particles biweekly for 12 weeks. Considering the hypothesis that endothelial dysfunction is the initial atherogenic event, we began treatment during early atherogenesis, and assessed endothelial function by testing vascular reactivity in aortas and carotid arteries ex vivo. We found that ESTA-MSV/miRs, but not PEG/PEI/miRs, significantly improved the ACh-induced endothelium-dependent relaxation of both aortas and carotid arteries of ApoE −/− mice (Fig. 3A-D). SNP-induced endothelium-independent relaxations in these arteries were comparable in all groups ( Supplementary Fig. S4A-D). We also observed that the ESTA-MSV/ miRs inhibited the ACh-induced contraction, a marker of endothelial dysfunction, in the presence of nitric oxide (NO) synthase inhibitor, while the PEG/PEI/miRs had little effect (Fig. 3E,F). ESTA-MSV/miRs decrease plaque size and stabilize plaques of ApoE −/− mice. In PEG/PEI/ miRs-treated groups, plaque size was moderately lower compared to vehicle-treated group while according to the en face ORO-staining, both miRs packaged in ESTA-MSV microparticles significantly decreased plaque size in aortic arches and branches and in the thoracic aortas (Fig. 4A,B). However, the miR-181b packaged in MSV non-targeted system did not have effects, and so did scrambled RNA (Supplementary Fig. S5). ESTA-MSV-miR146a appeared to be the most potent effector of reduced lesion size. Similar results were obtained from the ORO-stained frozen sections of aortic roots (Fig. 4C,D). In addition, ESTA-MSV/miRs markedly increased the collagen deposition in plaques shown by Masson trichrome-stained sections of aortic roots (Fig. 4E,F).
ESTA-MSV/miRs decrease macrophages in plaques of ApoE −/− mice. According to double immunostaining plaques contained CD68-positive macrophages and α -SM-actin-positive vascular SMCs. What is more, the number of macrophages decreased while that of the vascular SMCs increased in plaques of ESTA-MSV/ miRs-treated vs. vehicle-treated mice (Fig. 5A,B). However, PEG/PEI/miRs had no effect (Fig. 5A,B).

ESTA-MSV/miRs downregulate adhesion molecules in aortas of ApoE −/− mice.
Aortic tissue-miR-146a/-181b expression was higher in ESTA-MSV-vs. PEG/PEI-treated mice, whereas the opposite occurred in the liver and spleen, which are the major sites of PEG/PEI-miRNAs deposition (Fig. 6A-F). In addition, expression of Ccl2, Ccl5, Ccl8, and Cxcl9 in the abdominal aortas was lower in both ESTA-MSV/ miRs-treated vs. vehicle-treated control mice (Fig. 6G-J), while PEG/PEI/miRs only minimally affected expression of these chemokines (Fig. 6G-J). Neither microparticle treatment affected the serum lipids nor the glucose tolerance in mice ( Supplementary Fig. S6A-D) nor the serum ALT and AST activities ( Supplementary Fig.  S7A,B) and serum creatinine level (Supplementary Fig. S7C).

Discussion
We described a novel and effective targeted delivery system for miRs-based therapy for atherosclerosis. Although the miRs have potential in the treatment of atherosclerotic diseases 19 , critical challenges about in vivo delivery must be overcome prior to its translation to clinical applications; e.g., once entering the blood stream, miRs can be rapidly cleared. Several viral vectors and non-viral delivery systems have been formulated to address the limitations of poor in vivo stability and inappropriate biodistribution 20 . Compared with viral vectors and liposomes, the PEG/PEI nanoparticles are a safer and more efficacious delivery system in vivo 11,21,22 , although non-specific distribution has been an unresolved problem. The ESTA-MSV targeting delivery system was previously used to treat bone marrow metastatic breast cancer 15 . However, this is the first attempt to inhibit endothelial inflammation and atherosclerosis via an E-selectin-targeting carrier to deliver miRs. In this study, the ESTA-MSV microparticle-packaged miRs showed excellent therapeutic effects on atherosclerotic lesions; while non-specific PEG/PEI nanoparticle-packaged miRs had little or no effect, a difference that may be attributed to the differential distribution due to ESTA-MSV affinity. Compared to PEG/PEI, the ESTA-MSV-mediated delivery gave higher, stable expression of exogenous miRs in aorta but lower levels in liver and spleen thereby indicating specificity even if delivered systemically. Our novel findings suggest that the ESTA-MSV delivery system not only has higher selectivity of distribution but also carries the agents for controlled release.
Sun et al. 23 reported that injection with 1 nmol of miR-181b packaged with lipofectamine once a week for 12 weeks through tail vein attenuated the atherosclerosis in ApoE −/− mice. In the present study, mice were injected with 2 nmol of miR-181b packaged in ESTA-MSV system every other week and showed decreased atherosclerosis. The average dosage of miR-181b was similar between these two studies. It suggests that ESTA-MSV may achieve pharmacological effects with a lower administration frequency. E-selectin, also known as endothelial-leukocyte adhesion molecule 1, is a cell adhesion molecule expressed on inflamed endothelial cells in response to inflammatory cytokines 24,25 . We confirmed, in this study, that the E-selectin is almost absent in healthy endothelium but is apparently induced in irregular endothelium covering atherosclerotic plaques in mice and humans. E-selectin is also expressed in newly-formed microvessels, which are usually within an inflammatory microenvironment. This feature makes it feasible in treating bone marrow metastatic breast cancer by targeting the tumor neovessels as in our previous study 15 . Neovascularization also exists in the established atherosclerotic plaques, and is thought to promote lesion progression, including destabilization, hemorrhage, erosion and ultimately rupture 26,27 . Therefore, inhibition or normalization of the plaque angiogenesis to stabilize the lesion could be the mechanism underlying the therapeutic effectiveness of ESTA-MSV 28 . However, miRs delivery to neovessels within the plaque via ESTA-MSV microparticles remains an open question worth addressing. This hypothesis requires further investigation of the ESTA-MSV delivery system in established atherosclerotic plaques.
Several kinds of nanoparticles and microparticles have been developed and tested in the experimental atherosclerosis. Based on their physical features, PLGA-b-PEG copolymer and sugar-based amphiphilic nanoparticles should target macrophages 29,30 . PLGA-b-PEG copolymer encapsulated liver X receptor agonist delivery in LDL receptor −/− mice shows localized effects on macrophage without targeting liver thus sparing adverse effect on lipid metabolism 29 . A sugar-based amphililic macromolecule effectively blocks oxidized LDL uptake through binding and regulating macrophage scavenger receptors 30 . A nanoparticle carrier system (MaxSuppressor In vivo RNA Lancer Kit) delivers miR-126-5p to vascular ECs and reduces atherosclerosis 31 . Moreover, many attempts have been made to modify the particle surface-ICAM1-targeting, VCAM1-targeting, collagen IV-targeting, and vascular targeting particles-so that they target the plaque [32][33][34][35][36] . VCAM-1-targeted particles carried and delivered miR inhibitors to ECs and prevented atheroma formation in a mouse model of atherosclerosis 37,38 . Collagen IV-targeted nanoparticles delivered a small peptide, which would otherwise have been quickly renally cleared, to advanced plaques 34 . Another peptide, PREY, which has high fibronectin and filamin-A affinity when conjugated to liposomes, targets the atheroprone vasculature by delivering BH 4 , which reduces vascular superoxide 35 . Additionally, silica-gold nanoparticle alone may achieve therapeutic effects without anti-atherosclerotic agents in patients diagnosed with coronary artery disease 39 . The present study identified the ESTA-MSV microparticle as promising therapeutic alternative to combat atherosclerosis.
MiRs are endogenous small non-coding RNAs (containing ~22 nucleotides), which function in post-transcriptional regulation of gene expression 40 . Several miRs have been identified as regulators of EC, vascular SMC, and macrophage functions, and thereby regulate the development and progression of atherosclerosis 41 . Moreover, the miRs have emerged as potent therapeutic agents for atherosclerosis. In endothelial cells, miR-146a and miR-181b produced anti-inflammatory effects by suppressing the nuclear factor-κ B pathway 23,42,43 . In the present study, we found that the atheroprotective effects of these two shear stress-inducible miRs are mediated through downregulated adhesion molecule expression, which reduces adhesion of monocytes to ECs. However, we could not exclude other possible mechanisms. As these two miRs may also be involved in the glucose and lipid metabolism 44 , we evaluated the serum lipid levels and glucose tolerance and detect no differences among the groups. Impairment ACh-induced relaxation and enhanced ACh-induced contraction are markers of endothelial dysfunction. Although the mechanisms underlying the miRs-elicited improvement of endothelial function in ApoE −/− mice are not thoroughly understood, one might assume that the inhibition of macrophage infiltration and inflammation of vascular wall might improve NO-dependent vasodilation. In addition, miR-146a also targets NADPH oxidases that produce reactive oxygen species (ROS) 45,46 , a major NO quencher that is important role in the regulation of endothelial function. Therefore, ROS inhibition might improve endothelial NO-mediated vasodilation and contribute to the atheroprotective effect of miR-146a; this needs further investigation. The expression of miR-146a in aorta (A) liver (C) and spleen (E) of ApoE −/− mice after intravenously injected with PEG/PEI/miR-146a or ESTA-MSV/miR-146a biweekly for 12 weeks. The expression of miR-181b in aorta (B) liver (D) and spleen (F) of ApoE −/− mice after injected with PEG/PEI/miR-181b or ESTA-MSV/miR-181b for 12 weeks. *P < 0.05, **P < 0.01 vs. PEG/PEI group. The expression of Ccl2 (G), Ccl5 (H), Ccl8 (I), and Cxcl9 (J) in aortic tissues of ApoE −/− mice after intravenously injected with vehicle, miR-146a, and miR-181b (15 μg) loaded in PEG/PEI nanoparticles or ESTA-MSV microparticles biweekly for 12 weeks. Data are shown as the means ± SEM (n = 5). *P < 0.05, **P < 0.01 vs. vehicle group of the same particle, # P < 0.05, ## P < 0.01 vs. PEG/ PEI groups of the miR.
Scientific RepoRts | 6:22910 | DOI: 10.1038/srep22910 Although the ESTA-MSV system was designed to specifically target E-selectin on inflamed endothelium, we cannot rule out the participation of other types of cells, which may also be affected, including SMCs and macrophages in the vascular wall.
In summary, given that E-selectin expression is enhanced on inflamed EC surface, E-selectin-targeting ESTA-MSV microparticles enrich endothelium with therapeutic miRs. Importantly, we found that miR-146a and miR-181b decreased plaque size and macrophage infiltration, while increased vascular SMCs and collagen deposition.