Background
Acute coronary syndrome (ACS) is the leading cause of cardiovascular morbidity and mortality. Most major adverse events occur independent of the plaque size and the degree of luminal stenosis and most ACS arises from vulnerable plaques which are more prone to rapid plaque progression [
1]. The vulnerable lesion has several structural and functional hallmarks different from stable plaque including large necrotic core, inflammation, thin fibrous cap, depletion of smooth muscle cells and collagen, and increased formation of neo microvessels [
2]. However, the precise mechanisms underlying plaque destabilization are still unclear.
Accumulating evidences have shown a strong association between obesity and cardiovascular disease [
3]. Obesity is associated with the inflammation in perivascular adipose tissue (PVAT) which not only serves a structural support for most arteries but also secretes series of molecules to actively modulate vascular function [
4]. In 1991, it was the first time to report that PVAT had an anticontractile function [
5]. In 2002, PVAT was found to release a vasoactive factor [
6]. Since then, the list of factors released from PVAT has expanded and now includes adipokines (e.g., leptin and adiponectin), angiogenesis factors and inflammatory factors (e.g., MCP-1, IL-6). From twentieth century on, scientists revealed that chemokines production from PVAT played a role in the pathogenesis of atherosclerosis [
7]. These molecules might contribute to the alterations of the function and structure of vascular wall, including chronic inflammation, alterations of vascular tone, smooth muscle cell dysfunction, neo-angiogenesis and hence to the development of atherosclerosis. Epidemiological studies show that PVAT correlates with plaque burden in human [
8‐
10]. Experimental animal studies also demonstrate that PVAT accelerated atherosclerosis in mice [
11,
12]. Additionally, clinical observations suggest that coronary perivascular adipose tissue is related to the presence of lipid core, macrophage infiltration and severity of atherosclerotic plaque which are the characteristics of high risk plaque [
13,
14]. These findings urged us to hypothesize that PVAT contributes to plaque vulnerability through paracrine effects on the vasculature from ‘outside to inside’.
There are compelling evidences that endoplasmic reticulum stress (ER stress) plays fundamental roles in atherogenesis and atherosclerotic progression [
15,
16]. Endoplasmic reticulum (ER) is the cellular organelle in which protein folding, calcium homeostasis, and lipid biosynthesis occurs [
15]. Perturbations in ER homeostasis cause ER stress. High fat feeding and obesity [
17,
18] and several other high-risk factors of atherosclerosis such as hypertension [
19,
20], cigarette smoking [
21], high glucose [
22] and hyperhomocysteinemia [
23], can lead to increased ER stress in adipose tissue. ER stress induces inflammatory phenotypic alteration of adipose tissue which plays an important role in atherogenesis [
17]. Thus, we speculated that unresolved ER stress might induce the dysfunction of PVAT by leading to aberrant chemokines secretion and contributed to atherosclerotic plaque progression.
In this study, we found that transplanted PVAT promoted plaque vulnerability in the setting of high-fat diet (HFD) which could be ameliorated by 4-PBA at least in part dependent on decreased GM-CSF released locally by transplanted PVAT. These findings demonstrate a direct relationship between ER stress in PVAT and plaque destabilization and implicate GM-CSF secretion by PVAT as a mediator of this pathological process.
Methods
Animal model
Eight weeks old male apoE
−/− mice (Guangdong medical laboratory animal center) were used. To investigate the impact of PVAT on atherosclerotic stability, we executed carotid collar placement as described [
24,
25] and these mice were maintained on a HFD (Guangdong medical laboratory animal center, 0.15% cholesterol and 21% fat). The carotid collar was removed after 6 weeks, followed by adipose tissue transplantation surgery. 2 weeks after transplantation, ER stress inhibitor 4-PBA (5 μg per mouse) was locally administrated using pluronic gel to the transplanted PVAT and then animals were euthanized 4 weeks later. The left common carotid arteries and the transplanted adipose tissue were collected for histological and molecular biological analysis.
All procedures were approved by the Institutional Animal Care and Use committee of Sun Yat-sen University at Guangzhou.
Carotid collar placement
A silastic tube (0.3 mm inner diameter, 0.5 mm outer diameter, and 2.5 mm long) was placed around the left common carotid near its bifurcation. Briefly, mice were anesthetized and the left common carotid arteries were dissected from the surrounding connective tissue. Collars were placed carefully around the left carotid arteries and then tied with three circumferential silk ties at their axial edges. Then, the entry wound was closed.
Adipose transplantation to carotid artery
Thirty mg of PVAT was collected from the thoracic aorta of donor C57BL/6J mice (Guangdong medical laboratory animal center) fed a HFD for 4 weeks. The collected PVAT, or inguinal subcutaneous adipose tissue (SQAT) from the same group of donor mice as a control, was implanted adjacent to left common carotid artery of apoE−/− mice. Sham-operated apoE−/− mice underwent the same surgery without a fat transplant. Skin was sutured with 6-0 nylon filament.
To analyze the expression profiles of angiogenesis-related proteins, we used the mouse angiogenesis antibody array (RayBiotech, USA), according to the manufacturer’s instructions. They can detect 24 antibodies directed to proteins involved in angiogenesis. Briefly, 10 mg tissue of PVAT and SQAT were incubated in 1 ml serum-free DMEM medium. After 24 h the supernatants from tissue cultures were centrifuged at 14,000 rounds/min for 3 min. The centrifuged supernatants were stored at − 80 °C until further processing.
Supernatants from tissue cultures were mixed with 70 μl of biotin-conjugated detection antibodies for 2 h at room temperature with gentle shaking. Following a washing step, streptavidin-fluor was added to each sub-array. Make sure the glass slides were absolutely dry before scanning or storage. Data were captured by GenePix 4000B Microarray Scanner (Molecular Devices, USA) and analyzed by analysis tool software for RayBio antibody array (RayBiotech).
Cell culture
Mouse preadipocytes 3T3-L1 cells (Shanghai institute for biological sciences, Shanghai, China) were cultured in growth medium (DMEM supplemented with 10% FCS). Two days postconfluence, the cells were induced to differentiate with standard cocktail consisting of medium (DMEM supplemented with 10% FBS) with 1 μmol/l dexamethasone, 10 μg/ml insulin and 0.5 mmol/l isobutyl-methylxanthine (Sigma). After 4 days in differentiation medium, the cells were treated with medium containing 10 μg/ml insulin for 4 days and then maintained in DMEM supplemented with 10% FBS alone. Then, cells were considered mature adipocytes. ER stress was induced by tunicamycin (TM) and suppressed by 4-PBA.
Human umbilical vein endothelial cells (HUVECs) (Shanghai institute for biological sciences) were maintained in endothelial cell growth medium-2 (EGM-2; Lonza) supplemented with 10% FBS (Biological Industries, Israel), at 37 °C in humidified incubator (5% CO2). HUVECs were cultured to 2–6 passages for experiment.
Mouse aortic ring assay
Three-dimensional ex vivo mouse aortic ring angiogenesis assay allows analysis of cellular proliferation, migration, tube formation, microvessel branching, perivascular recruitment and remodeling, providing a more complete picture of angiogenic processes compared with traditional in vitro tube formation assays. We performed the aortic ring assay as previously described [
26], with minor modifications. Briefly, the rings from the aortas of 4 weeks old C57BL/6 mice were inserted between two layers of matrigel basement membrane matrix (BD Biosciences) and cultured in EGM-2 in the presence of DMEM or supernatants from PVAT cultures for 7 days. Endothelial sprouts were stained with antibody to CD31 (abcam) by immunofluorescence and quantitative analysis was performed with software Image-Pro Plus 6.0. Three separate aortic sections were quantified for each aorta, and the results were averaged for each animal.
96-well plates were coated with 50 μl Matrigel (Millipore) and incubated at 37 °C for 60 min to allow the Matrigel to solidify. HUVECs were plated at a density of 5 × 10
4 cells/well with supernatants from PVAT cultures or vehicle and incubated at 37 °C for 8 h. The cells were then photographed using a Nikon digital camera. Five randomly selected view fields were photographed in each well. The average of five fields was taken as the value for each sample. Tube formation was quantified by measuring the length of capillary structures using the software Image-Pro Plus 6.0 [
27].
Cell migration assay
The scratch wound migration assay was performed as described before [
28]. Briefly, confluent HUVEC sheets were starved for 6 h before starting the experiments. Confluent cell monolayer was then scraped with a 200 μl pipette tip to generate scratch wounds and rinsed twice with PBS. Cells were photographed immediately and 24 h after the scratch with a Nikon digital camera (Nikon ECLIPSE Ti). The wound area was then measured to determine cell migration.
Chromatin immunoprecipitation (ChIP)
ChIP analysis was performed using Chromatin IP kit (Pierce). Anti-p65(pNF-κB) antibody (CST 8242S) was used for immunoprecipitation and then pNF-κB association with GM-CSF promoters was estimated using SYBR-Green RT-PCR. For positive control polymerase II antibody was used (Cell Signaling Technology). Corresponding normal IgG was used as negative control. ChIP was performed overnight at 4 °C with rotation. Input DNA was purified along with ChIP probes. Association with pNF-κB and GM-CSF promoters was estimated using SYBR-Green RT-PCR. Sequence of primers used for ChIP is given in Additional file
1: Table S1.
Ready-To-Glow™ NF-κB Secreted Luciferase Reporter System
We used the Ready-To-Glow™ NF-κB Secreted Luciferase Reporter System (Clotech, Catalog No.631743), according to the manufacturer’s instructions. Mature 3T3-L1 cells were transiently transfected with pMetLuc2-control vector or pNFκB- MetLuc2-reporter vector. 24 h after transfection, the media was removed and replaced by media with vehicel or TM (1 μg/ml) or pretreatment of 4-PBA (5 mM) for half hour to activate the NF-κB signal transduction pathway. After 8, 16 and 24 h, 160 μl media samples were removed and analyzed for Metridia luciferase activity in a luminometer (Molecular devices, SpectraMax M5/M5e). The fold induction was calculated for different time points following substrate addition.
Statistical analysis
The statistical significance of differences between groups was determined by one-way ANOVA for multiple comparisons, or Student’s t test when comparisons were made between two groups. Values are expressed as mean ± SEM, p < 0.05 are considered significant.
Discussion
A consensus is emerging that PVAT is related to plaque vulnerability, but direct evidence is lacking. Here, we have established a model in which PVAT was transplanted to carotid artery with atherosclerosis to test the effects of PVAT transplantation on plaque vulnerability. In addition, we locally administrated 4-PBA to the transplanted PVAT to investigate the role of ER stress in paracrine effects of PVAT. Our data suggest that ER stress in PVAT could increase plaque vulnerability, partly by increasing GM-CSF paracrine via transcription factor NF-κB.
Adipose tissue not only stores energy but also secrets numerous growth factors, cytokines, and hormones. PVAT has recently been recognized as a novel factor in vascular biology implicated in the pathophysiology of cardiovascular disease. Emerging researches suggest that PVAT may contribute to the pathogenesis of vascular disease by producing large numbers of biologically active molecules [
6]. Our study may explain its biological activity from the perspective of the microstructures. Even though PVAT was a mixed population of white and brown adipocytes, ultrastructural detection by transmission electron microscopy showed that there were abundant big mitochondria no matter with small lipid droplets or one big lipid droplet in PVAT while SQAT was made up of a single lipid vacuole with rare mitochondria (Fig.
2). Therefore, the rich mitochondria in perivascular adipocytes might be the structure basis of its biological activity.
Vulnerable plaque is an important culprit mechanism for the development of acute coronary syndrome (ACS), which is influenced by several factors such as vasa vasorum and intraplaque neovascularization, large necrotic core, inflammation and intraplaque protease activity [
2]. However, the underlying mechanism of plaque vulnerability remains unclear and need to be further elucidated. Clinical observations suggest PVAT increases severity of atherosclerotic plaque and correlates with coronary plaque burden [
13]. We demonstrated PVAT promoted plaque vulnerability in PVAT-transplanted carotid arteries (Figs.
3,
4 and Additional file
1: Figure S2). To examine the impact of PVAT, endogenous PVAT in femoral artery was removed and replaced with transplanted fat in some studies [
30,
31], or exogenous PVAT was transplanted to the carotid artery without PVAT in other studies [
11,
12]. Fat transplantation was used in our study to reduce injuring artery. Carotid collar placement was used to study plaque vulnerability and produced atherosclerosis plaque in fixed location and facilitate administration to artery.
Several researches have also revealed the pathogenic role of PVAT in atherosclerosis [
32‐
34]. David Manka et al. [
11] transplanted PVAT to the mouse carotid artery and found that PVAT could enhance the neointimal response to vascular injury. On the contrary, some studies suggest that PVAT is vasculoprotective. For instance, removal of endogenous PVAT enhanced the neointimal response to wire injury in the femoral artery [
30]. Additionally, research in the SMPG knockout mouse suggest an atheroprotective effect of PVAT at a cool temperature [
35]. However, the results in current study are not strictly comparable with the previous researches. Some of these researches transplanted visceral adipose tissue rather than real PVAT to the carotid artery. Some studies removed endogenous PVAT, which might make the artery injured. Therefore, these conclusions might be dependent on the experimental model. However, the model of fat transplantation used in this research might be more proportionate to study the impact of PVAT.
Several high-risk factors of atherosclerosis can cause PVAT dysfunction and increase ER stress in adipose tissue [
7]. In current research, we locally treated the transplanted PVAT with ER stress inhibitor 4-PBA and found that reduction ER stress in PVAT would attenuate the effect of PVAT on plaque destabilization and angiogenesis (Fig.
3,
4 and Additional file
1: Figure S2). Thus, we speculated that ER stress might contribute to the dysfunction of adipose tissue and lead to aberrant adipokine secretion.
Intimal angiogenesis, as a source of intraplaque hemorrhage, is associated closely with plaque vulnerability. From the in vitro, ex vivo and in vivo evidences, we concluded that PVAT could increase angiogenesis which would be attenuated by ER stress inhibitor. So then, we screened out the angiogenesis factors produced by transplanted adipose tissue with mouse angiogenesis antibody array kit. The results showed that several factors might play a role in the paracrine effect of PVAT, mainly including MCP-1, IL-6 and GM-CSF (Fig.
6). We focused on GM-CSF, because there were several reports on the role of MCP-1 and IL-6 secretion in PVAT.
GM-CSF is a hematopoietic growth factor, stimulating hematopoietic stem cells to produce granulocytes and monocytes. It is also a proinflammatory and pro-angiogenic factor. Additionally, GM-CSF could recruit and activate M1 macrophages and thus contributes to adipose tissue inflammation in response to a HFD [
36]. The current data indicate that 4-PBA impaired production of GM-CSF in PVAT accompanied by improved plaque stabilization. Interestingly, another study also found that GM-CSF promoted plaque vulnerability due to promotion of macrophage apoptosis and plaque necrosis [
37]. In addition, several studies suggest GM-CSF is correlated with plaque burden [
38‐
40]. Therefore, GM-CSF might play an important role in the impact of PVAT on plaque vulnerability. However, this conclusion was based on the previous study and direct evidence is needed for better understanding the link between GM-CSF and plaque vulnerability.
Next, we used the transcriptional inhibitor Actinomycin D to demonstrate that ER stress upregulated GM-CSF expression in adipocytes by a transcriptional mechanism. Previous researches found that ER stress activated transcriptional factor nuclear factor-κb (NF-κB) [
41,
42] and other studies suggested NF-κB was a transcription factor of GM-CSF [
43,
44]. NF-κB is a transcription factor that controls the expression of genes involved in immune responses, inflammation, and cell cycle, which can be activated by a variety of stimuli, including cytokines, T and B cell mitogens, viral proteins, and stress inducers [
45]. Therefore, we hypothesized that ER stress might increase NF-κB binding to the promoter of GM-CSF gene in adipocytes. Our present results showed that NF-κB could bind to the promoter of GM-CSF in the absence of stimulation, and the combination would be enhanced by ER stress inducer. Thus, NF-κB was a transcription factor of GM-CSF in adipocytes. However, 3T3-L1 cell line was used in this part of research, which might make the significance of the results limited.
Authors’ contributions
RY, SWL and JYC performed the experiments, analyzed the data, and drafted the manuscript. YY carried out q-PCR. MXW carried out Western blotting. HFZ and ZJG participated in animal experiments. JFW and YXC conceived the study, analyzed the results and wrote the manuscript. All authors read and approved the final manuscript.