A goat anti-EMMPRIN polyclonal antibody (sc-9757), a rabbit polyclonal antibody (sc-7951) against cyclooxygenase-2 (COX-2), dithiothreitol, aprotinin, leupeptin, phenylmethylsulfonyl fluoride (PMSF), Tween 20 and Nonidet P-40 were purchased from Santa Cruz Biotechnology, Inc., (Dallas, TX, USA). The rabbit anti-glyceraldehyde 3-phosphate dehydrogenase (GAPDH) monoclonal antibody (NB100-56875) was purchased from Novus Biologicals, LLC (Littleton, CO, USA). The goat anti-rabbit (IRDye680LT) and donkey anti-goat (IRDye800CW) secondary antibodies were purchased from LI-COR Biosciences (Lincoln, NE, USA). The RNAiso Plus, SYBR Premix Ex Taq II and the PrimeScript RT Reagent kit with gDNA Eraser were purchased from Takara Bio, Inc., (Otsu, Japan). The goat anti-rabbit (ZB-5301) and rabbit anti-goat (ZB-2306) horseradish peroxidase (HRP)-conjugated secondary antibodies, a diaminobenzidine (DAB) color reagent kit (ZLI-9017) and an ELISA kit (EIA-1112) used for the determination of prostaglandin E2 (PGE2) levels were purchased from OriGene Technologies, Inc., (Beijing, China). Atorvastatin and ezetimibe, which is a selective inhibitor of cholesterol absorption, were gifts from Pfizer, Inc. (New York, NY, USA) and Merck & Co. (Kenilworth, NJ, USA), respectively. Polymerase chain reaction (PCR) amplification was performed using a thermal cycler PTC-200 (Bio-Rad Laboratories, Inc., Hercules, CA, USA), a nucleic acid protein analyzer DU800 (Beckman Coulter, Inc., Brea, CA, USA), a microplate reader FO39300 (Bio-Rad Laboratories, Inc.), a ChemiImager 5500 gel imaging analysis system (ProteinSimple, San Jose, CA, USA) and a Leica DMIRB inverted fluorescence microscope (Leica Microsystems, Inc., Buffalo Grove, IL, USA). RPMI 1640 medium and fetal bovine serum (FBS) were purchased from Gibco, Thermo Fisher Scientific, Inc. (Waltham, MA, USA). Phorbol 12-myristate 13-acetate (PMA), oxidized low-density lipoprotein (ox-LDL), NS-398 (an inhibitor of COX-2), PGE2 and sodium pentobarbitone (1507002) were obtained from Sigma-Aldrich (Merck Millipore, Darmstadt, Germany).

Apolipoprotein E knockout (ApoE^−/−) male mice (n=72; age, 8 weeks; weight, 24±1 g) were purchased from the Animal Centre of the Medical Department of Beijing University (Beijing, China). Mice were maintained under standard conditions of humidity (50–60%) and temperature (18–22°C), and were subjected to a 12 h hour light-dark cycle with ad libitum access to feed and water. Mice were inspected at least once every 24 h. Two types of feed were provided, which included a normal rodent diet and a high-fat content diet that contained 15% fat from lard and was supplemented with 1.25% (w/w) cholesterol.

All 72 mice were fed a normal rodent diet for one week. Subsequently, mice were distributed randomly and evenly into an early-start (ES) and a late-start (LS) treatment group (n=36 each).

Mice in each group were randomly divided into four equal subgroups (n=9): Groups A (ES) and E (LS) were a normal diet control (NDC) group; groups B (ES) and F (LS) were a high-fat diet control (HDC) group; groups C (ES) and G (LS) were a low-dosage treatment (LDT) group receiving a high-fat diet and a low dosage of atorvastatin; and groups D (ES) and H (LS) were a high-dosage treatment (HDT) group receiving a high-fat diet and a high dosage of atorvastatin. Group C were orally administered 0.3 ml/day atorvastatin suspension, at a dosage of 5-mg/kg beginning in week 7, whereas atorvastatin was administered to group G beginning in week 11. Group D were orally administered 0.3-ml/day atorvastatin suspension, at a dosage of 10 mg/kg, beginning in week 7, whereas atorvastatin was administered to group H beginning in week 11. Groups A, B, E, and F were administered an isodose of normal saline. Following 8 weeks of statin intervention, mice in the ES and LS groups were humanely sacrificed in weeks 15 and 19, respectively.

Mice were surgically anesthetized via intraperitoneal injection of 1% sodium pentobarbitone (50 mg/kg). The ascending aorta was removed from each mouse and immersed in 4% paraformaldehyde at room temperature, for ≥24 h. Aortas were subsequently embedded in paraffin and cut transversally into serial sections 5-µm thick initiating at the aortic root. Remaining sections of the aortas (from the descending to the iliac aorta) were stored at −70°C.

Serum harvested from the mice was analyzed to determine total triglyceride (TG) cholesterol (TC) and LDL-C levels, using an Olympus AU2700 High-Volume Chemistry-Immuno Analyzer (Olympus Corp., Tokyo, Japan).

Total RNA was extracted from the frozen atherosclerotic carotid samples or THP-1 macrophages (American Type Culture Collection, Manassas, VA, USA) using RNAiso Plus reagent according to the manufacturer's protocol. Then, the total RNA was reverse-transcribed to cDNA using the PrimeScript^® RT reagent kit with gDNA Eraser (RR047A) according to the manufacturer's protocol. Primer Express^® software v3.0.1 (Applied Biosystems; Thermo Fisher Scientific, Inc.) and sequence information from the National Center for Biotechnology Information database (http://www.ncbi.nlm.nih.gov) were used to design the PCR primers for EMMPRIN, COX-2 and the housekeeping gene GAPDH. Primer sequences were as follows: EMMPRIN, forward 5′-GCAGAGGACACAGGCACTTAC-3′ and reverse 5′-ACAGGCTCAGGAAGGAAGATG-3′; GAPDH, forward 5′-AGGTCGGTGTGAACGGATTTG-3′ and reverse 5′-GGGGTCGTTGATGGCAACA-3′; COX-2, forward 5′-GATTGCCCGACTCCCTTGG-3′ and reverse 5′-AACTGATGCGTGAAGTGCTG-3′. A total of 1 µg cDNA was used as a template for Real-time PCR in 20-µl reaction volumes with SYBR Premix Ex Taq II according to the manufacturer's instructions on a Stratagene Mx3000P qPCR system (Agilent Technologies, Inc., Santa Clara, CA, USA). The amplification protocol used was as follows: An initial denaturation step for 5 min at 95°C followed by 40 cycles of denaturation for 10 sec at 95°C, annealing for 30 sec at 57°C and elongation for 30 sec at 72°C, and a final extension step at 72°C for 30 sec. Quantitative measurements were determined using the comparative −2ΔΔCq method (18). All samples were normalized against the endogenous level of GAPDH. All results were repeated in triplicate and were analyzed using MxPro-Mx3000P software (Agilent Technologies, Inc.). The results for the relative expression levels were expressed as the mean ± the standard deviation (SD).

For western blot analysis, mouse aortas were homogenized, ground in liquid nitrogen and subsequently lysed with radio-immunoprecipitation assay (RIPA) lysate and 100 µl phenylmethylsulfonyl fluoride (PMSF) in ice for 30 min. Following centrifugation at 20,000 × g for 30 min at 4°C, the protein was obtained from the deposition. Cells were washed twice with 1 ml ice-cold phosphate-buffered saline (PBS) and lysed by adding 100 µl ice-cold RIPA and 1 mmol/l PMSF. Protein concentration was determined using the Coomassie Brilliant Blue method with bovine serum albumin (BSA; 5 mg/ml, Beyotime Institute of Biotechnology, Haimen, China) as a standard. Equal amounts (30 µg/well) of protein were separated by 10% SDS-PAGE and the protein was subsequently transferred, using electroblotting for 30 min at 50 V, onto a polyvinylidene difluoride membrane. The membrane was blocked in a 6% non-fat milk solution in Tris-buffered saline with 0.5% Tween 20 (Santa Cruz Biotechnology, Inc.). The membrane was probed separately with a goat anti-EMMPRIN polyclonal antibody (1:1,000) at 4°C overnight, rabbit anti-goat HRP-conjugated secondary antibody (1:2,500) at 37°C for 1 h, a rabbit polyclonal antibody against COX-2 (1:1,000) at 4°C overnight, goat anti-rabbit HRP-conjugated secondary antibody (1:2,000) at 37°C for 1 h, rabbit anti-GAPDH polyclonal antibody (1:2,000) at 4°C overnight, then goat anti-rabbit HRP-conjugated secondary antibody (1:2,000) at 37°C for 1 h. EMMPRIN, COX-2 and GADPH were detected using Pierce™ ECL Western Blotting Substrate (Thermo Fisher Scientific Inc., 32109) according to the manufacturer's protocol. Densitometric signals were quantified using ImageQuant TL 1.1 software (General Healthcare Bio-Sciences, Pittsburgh, PA, USA) and GAPDH was used as a loading control.

Following the rehydration and dewaxing of the serial 5-µm thick paraffin sections, they were incubated with 3% hydrogen peroxide to inhibit endogenous peroxidase activity. The sections were subsequently blocked with 5% BSA and were incubated for 30 min at 4°C, prior to incubation overnight with the goat anti-EMMPRIN polyclonal antibody (1:500) in 1% (w/v) BSA in PBS. Sections were incubated at room temperature for 20 min with peroxidase-conjugated AffiniPure donkey anti-goat secondary antibodies (LI-COR Biosciences, IRDye800CW, 1:500) in 1% BSA in PBS. Hematoxylin staining was performed on the sections to reveal cell nuclei, using DAB as a substrate. Brown particles detected via light microscopy were considered positive. Image-Pro Plus 5.1 color microscopic image analysis software (Media Cybernetics, Inc., Rockville, MD, USA) was used to measure the positive staining integrated optical density (IOD) and provide semi-quantitative analysis. Five DAB-stained paraffin sections were randomly selected from each group and an average of five IOD measurements per group were calculated. Paraffin sections were immunostained for an rabbit anti-macrophage antibody (Abcam, Cambridge, MA, USA, ab56297, 1:100) at 4°C overnight and an anti-α-actin antibody (Santa Cruz Biotechnology, sc-21078, 1:100) at 4°C overnight to confirm the presence of macrophages and smooth muscle cells (SMCs) in the vascular wall and plaques. Paraffin sections were deparaffinized, rehydrated and subsequently treated with 0.3% hydrogen peroxide in methanol for 30 min followed by 2% rabbit serum (sc-2338, Santa Cruz Biotechnology, Inc.) to abolish endogenous peroxidase activity and to block non-specific antibody binding, respectively. Sections were subsequently incubated overnight at 4°C with the rabbit anti-macrophage (sc-21078, 1:100) and anti-α-actin antibody (ab56297, 1:100). Following incubation, the sections were washed several times with PBS and incubated for 30 min at 37°C with the goat anti-rabbit (IRDye680LT, 1:500) and donkey anti-goat (IRDye800CW, 1:500) secondary antibodies visualized with DAB Developer (OriGene Technologies, Inc.) and counterstained with 10% Mayer's hematoxylin. Subsequently, sections were mounted in Permamount and examined using light microscopy. Aortic paraffin sections were stained with hematoxylin and eosin (H&E) and Sirius red.

H&E staining was performed on the aortic paraffin sections to observe the atherosclerotic plaque morphology. Computer-assisted morphometry (Image-Pro Plus 5.1; Media Cybernetics, Inc.) was used to analyze the cross-sectional area of each plaque, the relative plaque area (the ratio of the plaque area to the lumen area) and the fibrous cap thickness (FCT). The FCT was measured at 10 equidistant points around the cap of each slice. For each mouse, three sections (with an interval of 10 sections) from the thickest part of the plaque were selected and the mean value obtained was used for subsequent statistical analysis. Three consecutive paraffin sections from each mouse were used to detect macrophages and SMCs using an immunohistochemistry (IHC) assay, and to detect collagen using Sirius red staining. Lipoid vesicles were averaged directly from the H&E sections. The vulnerability index (VI) of the plaque was calculated using the following formula (19): VI = (macrophages + lipoid vesicles) / (SMCs + collagen).

Human monocytic THP-1 macrophages were cultured in suspension in RPMI 1640 containing 100 U/ml penicillin, 10% FBS and 100 µg/ml streptomycin at 37°C in an atmosphere containing 5% CO₂. A total of 1×10⁶ THP-1 cells per well were seeded in 6-well plates and stimulated with 0.1 µm PMA for 48 h until they adhered to the wells and exhibited macrophage-like morphology. Plates were washed twice with 1 ml PBS following culture. Cells were cultured in serum-free medium for an additional 24 h and incubated with the corresponding stimuli: 100-µm/ml ox-LDL was used to induce COX-2 and EMMPRIN expression and no ox-LDL was used as a control. For inhibition experiments, the cells were pre-incubated with 5 µm atorvastatin, 5 µm ezetimibe or 10 µm NS-398 for 1 h.

A PGE2 ELISA kit (EIA-1112) was used to measure PGE2 levels in the culture medium, according to the manufacturer's protocol. Absorbance was measured at 405 nm with a microplate reader (Thermo Fisher Scientific Inc. Waltham, MA, USA).

Image-Pro Plus 5.1 software was used to count gray-scale and intensity values, and areas of interest. SPSS 13.0 software (SPSS, Inc., Chicago, IL, USA) was used for statistical analyses. The data were presented as the mean ± the SD of three independent measurements. Groups of data were compared by performing one-way analysis of variance, followed by Tukey's multiple comparison tests. Pearson's product-moment correlation coefficient was used to analyze linear correlations. P<0.05 was considered to indicate a statistically significant difference.

Under the high-fat feeding conditions, there were no statistically significant differences observed in TC, LDL-C and TG levels among the three groups of ApoE^−/- mice.

Plaque in the HDC groups (groups B and F) was characterized by thin fibrous caps, extensive acellular collagenous masses and large lipoid vesicles that extended to and narrowed the lumen under the microscope. By contrast, the plaque in the NDC groups (groups A and E) contained thick fibrous caps and small lipoid vesicles (Fig. 1A). The relative plaque area in the HDC groups was significantly larger than that of the NDC groups (P<0.01; Fig. 1B). VP was defined as an FCT <65 mm (20,21). The mean FCT in the HDC groups was <65 mm and was significantly lower than that of the NDC groups (P<0.05; Fig. 1C). The macrophages and the SMCs in the plaque were stained brown during the IHC assay. Sirius Red stained type I collagen yellow and red, and type III collagen green (Fig. 1A). The HDC groups exhibited significantly increased percentages of macrophages (P<0.05) and lipoid vesicles (P<0.01), and decreased percentages of SMCs (P<0.05) and collagen (P<0.01) in plaques compared with the NDC groups (Fig. 1D). Therefore, the HDC groups were classified as having VP. Mice in the HDC groups exhibited decreased FCT (P<0.05; Fig. 1C) and increased VI (P<0.01; Fig. 1E). A positive histological correlation was detected between EMMPRIN expression and the VI (P<0.05; Fig. 1F).

The results of the RT-qPCR (Fig. 2), the western blot analysis (Fig. 3) and the IHC assay (Fig. 4) revealed significantly increased EMMPRIN expression in the HDC groups compared with the NDC (P<0.01), LDT (ES, P<0.05; LS, P<0.01) and HDT (P<0.01) groups, in ES and LS rats. The significant difference between the HDC and NDC groups indicates that EMMPRIN expression was higher in VPs than in stable plaques.

Significantly decreased expression of EMMPRIN was observed in the LDT (ES, P<0.05; LS, P<0.01) and HDT (P<0.01) groups compared with the HDC group, according to the mRNA (Fig. 2D-F) and protein levels (Fig. 3B, E and F). This was confirmed by the histological findings shown in Fig. 4 (Fig. 4A, C and D). Histological EMMPRIN expression was positively correlated with the VI (P<0.01; Fig. 4F).

High dose atorvastatin induced greater downregulation of EMMPRIN expression than the low dose (P<0.05). EMMPRIN expression was evaluated using RT-qPCR (Fig. 2F), western blotting (Fig. 3B and D) and an IHC assay (Fig. 4A and E). In the HDT and LDT groups, EMMPRIN expression was significantly lower in the ES than in the LS group, according to mRNA (LDT, P<0.05; HDT, P<0.01; Fig. 2F) and protein levels (P<0.05; Fig. 3B and D). The histological findings also indicated significantly lower EMMPRIN expression in the ES HDT group than in the corresponding LS group (P<0.05; Fig. 4E).

Treatment of THP-1 macrophages with 100 µg/ml ox-LDL for 6 h induced a rapid increase in PGE2 secretion, which peaked at 12 h (Fig. 5A). RT-qPCR and western blotting were used to examine whether COX-2 and EMMPRIN mRNA and protein expression were upregulated in THP-1 macrophages stimulated with ox-LDL in vitro. The mRNA and protein levels of COX-2 and EMMPRIN were increased significantly at 6 h compared with the levels in the control group (P<0.05; Fig. 5B and C). A positive correlation was observed between changes in mRNA and protein expression of EMMPRIN and COX-2 (data not shown).

Following pretreatment with atorvastatin or ezetimibe (both 5 µM) for 1 h, THP-1 macrophages were stimulated with 100 µg/ml ox-LDL for an additional 12 h. COX-2 and EMMPRIN protein expression levels were subsequently measured using western blotting. Atorvastatin markedly inhibited the ox-LDL-induced expression of COX-2 and EMMPRIN, whereas ezetimibe did not (Fig. 5D). These results suggest that the upregulation of EMMPRIN expression may be associated with increased COX-2 levels in THP-1 macrophages.

Additionally, the results demonstrated that ox-LDL was able to markedly upregulate COX-2 and EMMPRIN expression; however, further research on the role of the COX-2/PGE2 pathway is required. Therefore, these findings elucidate the possible effect of COX-2/PGE2 on ox-LDL-induced EMMPRIN expression.

Following pretreatment with 5 µM atorvastatin or 10 µM NS-398 for 1 h, THP-1 macrophages were stimulated with 1 µM ox-LDL or 0.1 µM ox-LDL and PGE2, for a further 12 h. Notably, when atorvastatin or NS-398 was used as a co-pretreatment with PGE2, the inhibitory effects of atorvastatin or NS-398 on ox-LDL-induced EMMPRIN mRNA and protein expression were significantly reversed (P<0.05; Fig. 5E and F). Furthermore, PGE2 levels were significantly increased by ox-LDL (P<0.05) and decreased by atorvastatin and NS-398 (Fig. 5G). These results indicate that EMMPRIN expression in THP-1 macrophages was elevated due to an increase in PGE2 levels, whereas atorvastatin and NS-398 each suppressed PGE2 expression. This inhibition was reversed following the addition of PGE2.