Oil red O (ORO), 1,10-dioctadecyl-3,3,30,30-tetramethylindocarbocyanine perchlorate (DiI) and human AB serum were obtained from Sigma (Sigma, United States). The intracellular cholesterol assay kit was from Cayman (Cayman Chemical Co, United States). Antibodies for CD36, Vimentin, tubulin, EEA1 and beta-actin were obtained from Abcam (Abcam, United States). Antibodies for SRA1 conjugated with APC and control antibody were manufactured by Bioyrt (Bioryt, United States). Antibodies for CD36 conjugated with APC and suitable control antibody were from Bio-Rad (Bio-Rad, United States). Sucrose, dextrose, SDS were obtained from Ducefa (Ducefa, Germany). Antibodies for rabbit IgG, IgM were purchased from bethyl chemisty (bethyl, United States). Antibody for p-PKAα/β/γ cat (Thr 198) was from Santa Cruz Biotechnology (sc-32968) (Santa Cruz, United States).

Pathogen-free, male C57BL/6 mice, 6–8 week old were purchased from Orient Bio (Orient Bio, South Korea). Vimentin null (Vim ^–/–) mice on a 129 background were provided by Dr. John E. Eriksson (Åbo Akademi University, Finland) and seven times backcrossed to C57Bl/6. LDL receptor null (LDLr ^–/–) mice were provided by Dr. Goo Taeg Oh (Ewha Womans University, Seoul, South Korea). In all experiments, age-matched (7–10 weeks old) male mice were used. The Institutional Animal Care and Use Committee (IACUC) of Ewha Womans University College of Medicine approved the experimental protocol (IACUC approval No. ESM-12-0198).

LDLr^−/–mice were reconstituted with bone marrow from Vim^+/+ mice (n = 11) and Vim^−/– mice (n = 10). To induce bone marrow aplasia, LDLr^−/–mice (male, age 6 weeks) were exposed to a single dose of 13 Gy (0.28 Gy/min, 200 kV, 4 mA) x-ray (total body irradiation) with a 4 mm aluminum filter, 1 day before the transplantation. Bone marrow cell suspensions were isolated by flushing the femurs and tibias from either male Vim^+/+ or Vim^–/– mice with phosphate-buffered saline. Irradiated recipients received 1.5 × 10⁷ bone marrow cells by intravenous injection into the tail vein. After 4 weeks, reconstituted Ldlr^−/– mice were fed a western diet for 15 weeks.

We generated Apoe^–/– Vim ^–/– (double null) mice by crossing the Apoe^–/– and Vim ^–/– mouse strains. Apoe^–/– Vim ^–/– mice and Apoe^–/– Vim^+/+ mice were fed a western diet for 15 weeks.

Western diet-fed mice were euthanized and perfused with cold 0.01 mol/L PBS through left ventricle. Aortae were dissected from the proximal ascending aorta to the bifurcation of iliac artery and fixed in 4% buffered paraformaldehyde for 24 h. After fixation, the aortae were split longitudinally and pinned open for surface lesion measurements with 0.5% Oil-Red O staining. For aortic sinus analysis, heart was embedded in optimal cutting temperature compound (OCT), and snap-frozen in liquid nitrogen. Serial 7-μm cryosections of the aortic sinus were cut using a Leica CM1950 cryostat. Four or more frozen section slides for aortic sinus were made per mouse. Cryosections were fixed in 4% buffered paraformaldehyde and stained in 0.5% Oil Red O for 25 min. Quantitative analysis of atherosclerotic lesions was performed using the Axiovision release 4.4 software (Carl Zeiss, Germany) program.

LDL was obtained from human plasma via density gradient ultracentrifugation (30). Oxidized LDL (oxLDL) was generated by dialysis of LDL with 5μM CuSO₄ in PBS for 6 h at 37^°C. To terminate oxidation of LDL, LDL was dialyzed with 100 μM EDTA in PBS.

In addition to CuSO₄-oxLDL, we generated oxLDL modified by myeloperoxidase (MPO) and used in several experiments as Figure 1D. LDL prepared from human plasma by density gradient ultracentrifugation was oxidatively modified by incubation in a buffer containing 50 mM sodium phosphate (pH 7.0) and 100 μM diethylene triamine pentaacetate (DTPA) with 30 nM MPO, 100 μg of glucose, glucose oxidase at 20 ng/ml (grade II; Boehringer Mannheim Biochemicals, Penzberg, Germany), and 0.5 mM NaNO₂ at 37^°C for 8 h (oxLDL) (31). The oxidation reaction was terminated by the addition of 40 μM butylated hydroxyl-toluene and 300 nM catalase to the reaction mixture.

Peritoneal macrophages were collected by peritoneal lavage of mice 4 days after intraperitoneal injection of 4% thioglycolate (1 ml). Mice were euthanized with CO₂ before harvesting macrophages. Cells were cultured in RPMI containing 10% fetal bovine serum and 1% penicillin-streptomycin. Media was changed to serum-free RPMI for treatment with oxLDL or other chemicals.

Primary bone marrow-derived macrophages (BMDMs) were collected from Vim ^–/– mice and wild type mice as described previously (32). Briefly, BMDMs were differentiated from murine bone marrow myeloid stem cells and cultured for 7 days with DMEM supplemented with 10% L929 supernatant containing 10% fetal bovine serum (FBS), 1 mM sodium pyruvate, 1% penicillin-streptomycin and 5 × 10–5M 2-mercaptoethanol at 37^°C and 5% CO₂.

SW13 cells were purchased from ATCC (ATCC, United States) and incubated following their protocol.

ORO staining was performed using murine peritoneal macrophages plated in 12-well plate as described previously (33). Cells were incubated with or without 50μg/ml of oxLDL for 18 h. Then macrophages were fixed in 4% paraformaldehyde in PBS for 20 min at room temperature. To stain the intracellular lipids, Cells were washed with PBS twice and treated with 60% isopropanol for 5 min and ORO solution (diluted 2:3 of H2O:0.3% ORO in isopropanol) for 15 min. After rinsing with distilled water, cells were examined microscopically.

We added 500 μl of isopropanol onto the ORO-stained macrophages to extract the ORO incorporated into the intracellular lipids. The amount of the ORO flowing out from the macrophages was measured by using spectrophotometry (absorbance at 510 nm).

Cells were lysed in buffer containing 1% triton X-100, 20 mM Tris-HCl (pH 7.5), 150 mM NaCl, 1 mM EDTA, 1 mM EGTA and protease inhibitor cocktail (Roche, Germany) and phosphatase inhibitors (10 mM phenylmethylsulfonyl fluoride/PMSF, 1% sodium pyrophosphate, 10 mM sodium fluoride, and 2 mM sodium vanadate). The lysates were separated by SDS-PAGE, transferred to PVDF membrane (Millipore), and analyzed by immunoblotting. Membranes were probed with antibodies for vimentin, CD36, phospho-vimentin (Ser72), and phospho-vimentin (Ser38). Antibodies for β-actin and GAPDH were used for normalization. Band intensities were quantified using Image J program (U. S. National Institutes of Health, United States).

Cells were incubated in 6 well plates with appropriate media containing 10% FBS. After rinsing twice with PBS, cells were harvested gently with scrapers and moved to round bottom glass tubes. After centrifuge at 1,000 × g in glass tubes, cells were incubated with 5% BSA for 30 min at 4^°C and then incubated with APC-conjugated anti-CD36 antibody (Bio legend, United States), FITC-conjugated anti-MSR1 antibody or isotype control (Bio legend, United States) diluted in PBS containing 5% BSA for 1 h at 4^°C. After three washes with PBS, cells were analyzed by flow cytometry.

Cells were lysed with RIPA buffer (1% triton x-100, 0.1% sodium dodecyl sulfate, 0.5% deoxycholate, 50 mM Tris-HCl (pH 7.5), 150 mM NaCl) and protease inhibitor cocktail. The protein concentrations of the lysates were quantified by BCA method and appreciate amounts of lysates were pre-incubated with Protein A or Protein A/G plus bead (Santa cruz, United States) for 2 h. The supernatant containing 350 mg protein was incubated with 2 μg of anti-CD36 antibody or anti-vimentin antibody overnight at 4^°C. Protein A or Protein A/G plus beads were added to the lysates for 4 h. Beads were extensively washed, boiled in laemmli buffer (Bio-rad, United States) and the bead-bound material was analyzed by immunoblotting for CD36, vimentin or PKA.

OxLDL was complexed with DiI as described previously. OxLDL (500 μg/ml) was mixed with 10 μg/ml of DiI in DMSO at 37^°C for 16 h. DiI-OxLDL solution was dialyzed in PBS. Macrophages were cultured in 6 well plates and then incubated with oxLDL (50 μg/ml) at 37^°C for 30 min. Cells were treated with DiI-oxLDL 50 μg/ml for 5 min, fixed in 4% paraformaldehyde in PBS and washed with PBS. Cells were collected by scrapping and spinned at 1,000 g for cell debris removal. DiI-OxLDL taken up by macrophages was measured using flow cytometry.

Intracellular cholesterol was measured using a cholesterol fluorometric assay kit (Cayman). Vim^+/+ macrophages and Vim^–/– macrophages were cultured in 6 well plates with or without oxLDL (50 μg/ml) and cells were washed twice with PBS after 18 h. Assay buffer including 0.5% triton X-100 was added onto the cells. The cell lysates were centrifuged at 13,000 × g for 30 min at 4^°C. The supernatants were used for the cholesterol measurement. The value of intracellular cholesterol was normalized by comparison to the protein concentration of the sample.

Total RNA was extracted from macrophages using Trizol reagent (Invitrogen, United States) according to the manufacturer’s instruction. We used iScript cDNA synthesis kit (Bio-rad, United States) to synthesize cDNA. Quantitative real-time reverse transcriptase PCR (qRT-PCR) was performed with Power SYBR Green PCR Master Mix (Applied Biosystems, United States) and an ABI Real-Time PCR thermocycler. RNAs were analyzed by qRT-PCR with following primers (5′-CCC AGA GCA AAA AGC GAC TC-3′ and 5′- GGT CAT CAT CAC TTT GGT CCT TG-3′ for ABCA1, 5′-CAA GAC CCT TTT GAA AGG GAT CTC-3′ and 5′-GCC AGA ATA TTC ATG AGT GTG GAC-3′ for ABCG1, 5′-GGC TGC TGT TTG CTG CG-3′ and 5′-GCT GCT TGA TGA GGG AGG G-3′ for SR-B1, 5′-GAT CGG AAC TGT GGG CTC AT-3’ and 5′-GGT TCC TTC TTC AAG GAC AAC TTC-3’ for CD36, 5′-AAA GAA GAA CAA GCG CAC GTG G-3’ and 5′-GAG CAC CAG GTG GAC CAG TTT G-3’ for SR-A1, and 5′-TCC ATG ACA ACT TTG GCA TTG-3’ and 5′- TCA CGC CAC AGC TTT CCA-3’ for GAPDH.) GAPDH was used for internal control.

All plasmids were produced using standard cloning techniques. Vimentin sequence was cloned into pAcGFP-Hyro-N1 vector for imaging. pLVX-puro vector were used for mutagenesis. These vectors were provided by Dr. Youngho Ahn (Ewha Womans University, Seoul, South Korea). Vimentin construct was amplified by PCR with following primers (VIM -HindIII-F: GTCA AA GCT TCG ATG TCC ACC AGG TCC, GTG TC, VIM -HindIII-R: GTCA AA GC TTA TTC AAG GTC ATC GTG ATG CTG, VIM -EcoRI-Flag-F: GTCA GAATTC GCCACC ATG GAT TAC AAG GAT GAC GAC GAT AAG ATG TCC ACC AGG TCC GTG TC, VIM -EcoRI-R: GTCA GAATTC TTA TTC AAG GTC ATC GTG ATG CTG) in reactions as follows; denaturation at 94^°C for 5 min, followed by 30 cycles of denaturation at 94^°C for 2 min, annealing at 59^°C for 1 min, and extension at 72^°C for 2 min. A final elongation step was carried out at 72^°C for 10 min.

For site-directed mutagenesis, we used polymerase chain reaction (PCR) with oligonucleotide mutation primers and the template vimentin cDNA. The primers for site-directed mutagenesis are VIM -S72A-F: GCC GTG CGC CTG CGG GCC AGC GTG CCC GGG GTG, VIM -S72A-R: CAC CCC GGG CAC GCT GGC CCG CAG GCG CAC GGC, VIM -S72D-F: GCC GTG CGC CTG CGG GAC AGC GTG CCC GGG GTG, VIM -S72D-R: CAC CCC GGG CAC GCT GTC CCG CAG GCG CAC GGC. The mutation sites were confirmed by sequencing by Cosmo Genetech.

Genetically modified lentiviruses were produced by transient transfection of 293T cells with Plvx -VIM, Plvx -V72A, Plvx -V72D, Plvx-puro, PAX2 packaging plasmid (containing gag and pol gene of HIV) and MDG2 envelop plasmid (containing vesicular stomatitis viral glycoprotein expressing vector) using Lipofectamine 2000 reagents (Thermo Fisher Scientific, United States).

A293T cells were seeded onto 6 well plates at 50∼70% confluence 1 day before the transfection. Transfection mixture containing 6 μg target vector, 5 μg PAX2, 3 μg MDG2, and 10 μg/ml Lipofectamine was added to the A293T cells. The media was collected after 24 h, filtered through a 45 μm pore filter and added to the recipient cells. To enhance the transfection efficiency, virus-containing media was mixed with polybrene 5 μg/ml (Millipore). We performed second transfection repeating the same procedure 24 h after the first transfection to promote viral incorporation.

Cells grown on glass coverslips were fixed with 4% formaldehyde in ice-cold PBS for 20 min and then 0.1% triton X-100 in ice-cold PBS for 5 min. Cells were incubated with 5% BSA in PBS for 1 h to reduce non-specific signals and then incubated with primary antibodies diluted in PBS containing 5% BSA for 16 h at 4^°C. After 3 washes with PBS, cells were incubated with appropriate fluorescent dye-conjugated secondary antibodies for 1 h and washed with PBS. Then nuclei were stained with DAPI-containing mounting solution (Vector laboratories, United States). Fluorescently stained cells were examined under a Zeiss confocal microscope and analyzed by Zeiss imaging processor.

Aortic sinus tissues were embedded in optimal cutting temperature compound (SAKURA Tissue-Tek, United States) for frozen section and snap frozen at –80^°C. These frozen tissues were sectioned on a cryostat, transferred to slides and then dried to preserve morphology. Sections were dried for 5 min at room temperature and were fixed by pre-cold acetone for 15 min. The following process was equal to immunocytochemistry.

Mouse peritoneal macrophages were incubated with oxLDL (50 μg/ml) for 10 min. After the designated time point, cells were washed twice with RPMI followed by the application of 0.5 ml of homogenization buffer [250 mM sucrose, 1 mM EDTA, 1 mM phenylmethylsulfonyl fluoride (PMSF)], in which cells were gently detached using a cell scraper, lysed, and further processed for examination by a sucrose flotation assay. Specifically, after centrifugation (1,000 × g), the post-nuclear supernatant was collected and adjusted to a concentration of 25% sucrose and 1 mM EDTA in 1 ml total volume. In 1 ml increments, 2.4 ml of 45% sucrose was transferred to the bottom of a SW41Ti tube and successively overlaid with 5.2 ml of 35% sucrose, 3.9 ml of 25% sucrose, and 1 ml of post-nuclear supernatant in 25% sucrose. These fractions were further analyzed using endosomal markers, CD36 and vimentin.

Lipid peroxidation of LDL was measured by Thiobarbituric Acid Reactive Substance (TBARS) formation. Briefly, LDL and oxLDL were incubated with ice cold 10% Trichloroacetic acid to precipitate protein for 15 min on ice. And centrifuge samples 2,200 × g for 15 min at 4^°C. Place supernatant into new labeled tube and add equal volume of 0.67% (w/v) Thiobarbituric Acid (TBA). Incubate in a boiling water bath for 10 min and record absorbance at 532 nm.

Data are expressed as mean ± standard error of the mean (SEM). Student’s t-test was used for comparisons between two sample means. ANOVA test was used to compare three or more groups. We also used non-parametric Kruskal–Wallis test for post-comparison hoc test. A p-value less than 0.05 were considered statistically significant. All experiments were repeated at least 3 times independently and all the measurements were done three times for a set. Analyses were performed using GraphPad Prism Software (GraphPad Software).

To determine if vimentin plays a role in macrophage foam cell formation, we performed ORO staining with Vim^+/+ and Vim^−/– murine peritoneal macrophages incubated with oxLDL for 18 h (Figure 1A). We extracted ORO from the ORO-stained macrophages and measured the absorbance to quantify the amount of intracellular lipids. The ORO incorporated in the Vim^−/– macrophages was 10% less than that in the Vim^+/+ macrophages (Figure 1A). In accordance, the increment of intracellular cholesterol after the oxLDL treatment which was measured by cholesterol oxidase assay was 6% less in Vim^−/– macrophages than in Vim^+/+ macrophages (Figure 1B). These data indicate that vimentin deletion reduces macrophage foam cell formation.

Previous reports revealed that vimentin functions in uptake of virus and metal ion such as zinc (13, 14, 34). To evaluate roles of vimentin in lipid uptake, peritoneal macrophages from Vim^−/– and Vim^+/+ mice were treated with oxLDL complexed with diI (DiI-oxLDL) for 5 min and fixed with 4% paraformaldehyde. Our data from flow cytometry showed that the mean fluorescence intensity of macrophages representing DiI-oxLDL uptake was 50% less in Vim^−/– macrophages than in Vim^+/+ macrophages (Figure 1C).

To clarify in which stage of oxLDL-internalization vimentin functions, we treated macrophages with U18666a (10 nM), Niemann-Pick C1 (NPC-1) inhibitor, to block the later process of endocytosis, especially the transport of LDL-derived cholesterol from lysosomes to endoplasmic reticulum (35). Vim^+/+ and Vim^−/– mouse peritoneal macrophages were pre-incubated with U18666a (10 nM) for 90 min, incubated with or without oxLDL (50 μg/ml) for 18 h and stained with ORO (Figure 1D). The absorbance of ORO eluted from Vim^−/– macrophages was 1.3 fold increased after the oxLDL treatment while Vim^+/+ macrophages showed 2.0 fold-increase after the oxLDL treatment (Figure 1E). The result suggested that less intracellular lipids in Vim^−/– macrophages were not caused by increased trafficking of free cholesterol from late endosome/lysosomes to other cellular compartments, but is caused by diminished uptake of lipoproteins.

Macrophages express various receptors including CD36 and scavenger receptor-A (SR-A) to uptake oxLDL. ABCA1 and ABCG1 mediate cholesterol efflux to apolipoprotein A1 and high density lipoprotein (36–38). Dysregulation of uptake and efflux of lipids leads to foam cell formation. We performed western blot analyses and quantitative real time PCR (RT-PCR) for receptors known to regulate cholesterol influx and efflux. Quantitative RT-PCR results showed no significant difference in expressions of CD36, LOX-1, SR-A, ABCA1, ABCG1, and SR-B1 between Vim^+/+ and Vim^−/– mouse macrophages (Figure 2A). The western blot showed that CD36 protein levels were not different between Vim^+/+ and Vim^−/– macrophages (Figure 2B).

However, plasma membrane-localized CD36 was significantly different between Vim^+/+ and Vim^−/– macrophages. Subcellular fractionation and western blots using the cytoplasmic and membrane fractions of macrophage lysates revealed that plasma membrane-localized CD36 in Vim^−/– mouse peritoneal macrophages was 60% less than in Vim^+/+ mouse peritoneal macrophages (Figure 2C). In accordance, flow cytometry showed that Vim^−/– macrophages exhibited less cell surface-localized CD36 than Vim^+/+ mouse peritoneal macrophages (Figure 2D). Bone marrow-derived macrophages (BMDM) from Vim^−/– mice also showed 50% less cell surface-localized CD36 than BMDM from Vim^+/+ mice in our flow cytometry data (Figure 2E). However, cell surface-localized SR-A measured by flow cytometry did not show difference between Vim^+/+ and Vim^−/– mouse peritoneal macrophages (Figure 2F).

OxLDL via CD36 activates c-Jun N-terminal kinase (JNK) 1/2 through MKK4 in macrophages (3). OxLDL treatment induced phosphorylation of MKK4 and JNK 1/2 in Vim^+/+ macrophages, however, Vim^−/– macrophages showed diminished responses to oxLDL (Figure 2G).

It has been reported that OxLDL/CD36 interaction induces release of cytokines including TNF-α (39). We measured TNF-α production in response to oxLDL in Vim^+/+ and Vim ^–/– macrophages. Vim^+/+ macrophage secreted 1.47 fold more TNF-α than Vim ^–/– macrophages (Figure 2H).

The results revealed that vimentin deletion decreases oxLDL/CD36-provoked signaling via less CD36 trafficking to plasma membrane.

Since we have found decreases in oxLDL uptake and CD36 trafficking that are attributed to vimentin deficiency, we tested if vimentin and CD36 are physically associated. The immunostaining showed that CD36 and vimentin are co-localized. The co-localization of vimentin and CD36, especially near the plasma membrane, was increased within 10 min after the oxLDL treatment compared with oxLDL-untreated macrophages (Figure 3A). Moreover, immunoprecipitaion assay results indicated that oxLDL increased association between vimentin and CD36 (Figure 3B).

We isolated endosomal fractions using sucrose gradient fractionation. Early endosomes were labeled by early endosome marker EEA1 and are in the 10th, 11th, and 12th fractions. There were CD36 and vimentin in early endosome fractions. In addition, the amounts of CD36 and vimentin in the early endosome fractions were increased when the cells were treated with oxLDL (50 μg/ml) (Figure 3C).

Overall, these results reveal that oxLDL promotes colocalization of CD36 and vimentin.

The major regulatory mechanism of vimentin is site-specific phosphorylation (18). We found that oxLDL induced phosphorylation of serine72 (ser72) in a dose-dependent manner (Figures 4A,B). OxLDL induced more vimentin phosphorylation (Ser72) than native LDL (Supplementary Figure 1). However, serine38 of vimentin was not affected by oxLDL (Figure 4A). Subcellular fractionation study showed that phosphorylated vimentin was increased in response to oxLDL both in the cytosolic and the membrane fractions of macrophages (Figure 4C). Vimentin phosphorylation (Ser72) by oxLDL was dependent on CD36. Western blot showed that oxLDL induced vimentin phosphorylation (Ser72) in macrophages from wild type mice while vimentin phosphorylation (Ser72) was not induced by oxLDL in macrophages from CD36 null mice (Figure 4D).

Vimentin phosphorylation at Ser72 is known to impair the assembly of vimentin intermediate filaments during mitosis in fibroblasts (18). To demonstrate the effect of vimentin phosphorylation (Ser72) in macrophages, we separated macrophage lysates into soluble/insoluble fractions based on solubility in triton X-100. In the soluble fraction which contains monomeric vimentin, vimentin was increased after the oxLDL treatment. In accordance, the level of vimentin in insoluble, filamentous fraction was decreased after the oxLDL treatment (Figure 4E).

To identify the specific kinase that phosphorylates vimentin in response to oxLDL, we blocked various kinases using inhibitors as staurosporine, IPA-3, LY294002 and H-89. We found that H-89, a specific inhibitor for Protein Kinase A (PKA) (40), reduced oxLDL-induced vimentin phosphorylation (Ser72). We also found that oxLDL induced phosphorylation of PKA (PKA α/β/γ-c, Thr198) (Figure 5A). In accordance, H-89-treated wild type macrophages internalized 50% less DiI-oxLDL than H-89-untreated macrophages (Figure 5B).

BMDMs transfected with PKA-specific siRNA showed less oxLDL-induced phosphorylation of vimentin (Ser72) than control siRNA-transfected cells (Figure 5C). However, BMDMs transfected with siRNA against protein kinase C (PKC) which is also known to phosphorylate vimentin (41) did not show changes in oxLDL-induced vimentin phosphorylation (Ser72) (Figure 5D). To see if PKA is physically associated with vimentin, we performed immunoprecipitaion using anti-vimentin antibody and immunoblotting with anti-PKA antibody. The result showed that oxLDL treatment for 5 min induced association between PKA and vimentin (Figure 5E).

These data demonstrated that PKA is activated by oxLDL and mediates vimentin phosphorylation (Ser72).

To evaluate the role of phosphorylated vimentin (Ser72), we generated Ser72-to-Asp phospho-mimicking mutant (V72D) and Ser72-to-Ala (V72A) non-phospho-mimicking mutant using site-directed mutagenesis. Vim^−/– BMDM showed 40% less DiI-oxLDL uptake than Vim^+/+ BMDM and Vim^−/– BMDM that restored vimentin expression showed increased uptake of oxLDL. Vim^−/– BMDM transfected with V72D showed 1.3 and 2.0-fold higher rates of DiI-oxLDL uptake than vimentin or V72A-transfected macrophages, respectively (Figure 6A).

The site-directed mutagenesis experiments were repeated using SW13 cells, renal epithelial cell line not expressing vimentin. SW13 cells transfected with phospho-mimicking mutant (V72D) showed 1.4 fold increased uptake of DiI-oxLDL compared with SW13 cells with wild type vimentin (Figure 6B).

Since Vim^−/– mouse macrophages showed less CD36 in plasma membrane than Vim^+/+ mouse macrophages, we hypothesized that vimentin phosphorylation may influence CD36 trafficking from cytosol to membrane. To prove the hypothesis, we stained plasma membrane CD36 with fluorescently labeled anti-CD36 antibody and performed flow cytometry to measure the fluorescence intensities that represent the amount of CD36 localized in plasma membrane. CD36 localized in plasma membrane was 2.0 fold more in the macrophages with V72D compared with macrophages with V72A (Figure 6C). These data suggest that vimentin (Ser72) phosphorylation promotes translocation of CD36 to plasma membrane.

To evaluate if atherosclerotic arterial lesion has higher levels of phosphorylated vimentin (Ser72) in macrophages, we performed immunostaining of cross-sections of the aortic sinuses of Ldlr ^–/– mice fed a normal chow diet and Ldlr ^–/– mice fed a western diet for 14 weeks. The result showed that macrophages in the atherosclerotic plaque of the western diet-fed mice had higher levels of phosphorylated vimentin (Ser72) compared with macrophages in the aortic sinuses of the chow-diet fed mice (Figure 6D).

To investigate whether deletion of vimentin in macrophages affects the formation of atherosclerotic lesions, we transplanted bone marrow from Vim^+/+ and Vim^−/– mice to Ldlr^−/– mice and placed these mice on a western diet for 15 weeks. En face aortae with ORO staining showed that Ldlr^−/– mice with Vim^−/– bone marrow had less atherosclerotic lesion than Ldlr^−/– mice with Vim^+/+ bone marrow (13.13 4.42% vs. 7.464 1.95%, p-value = 0.0019, Figure 7A). The cross sections of the aortic sinuses stained with ORO also showed that atherosclerotic plaques are smaller in Ldlr^−/– mice with Vim^−/– bone marrow than in Ldlr^−/– mice with Vim^+/+ bone marrow (Figure 7B). Macrophage counts obtained from F4/80 staining of the cross sections showed that Ldlr^−/– mice with Vim^−/– bone marrow had 60% less macrophage number in the atherosclerotic plaque area compared to Ldlr^−/– mice with Vim^+/+ bone marrow (Figure 7C).

Blood VLDL/LDL and total cholesterol concentrations of these two groups of mice were not different, however, the HDL concentrations of Ldlr^−/– mice with Vim^−/– bone marrow were 20% less than Ldlr^−/– mice with Vim^+/+ bone marrow (Figure 7D). The body weight of Ldlr^−/– mice with Vim^−/– bone marrow and Ldlr^–/– mice with Vim^+/+ bone marrow were not significantly different (Figure 7E).

These data suggest that macrophage vimentin contributes to the development of atherosclerosis.

We generated Apoe^–/– Vim ^–/– mice and fed a high fat diet for 15 weeks. Atherosclerotic lesions of aortae were analyzed by ORO staining of the en face bloc. Apoe^–/– Vim ^–/– mice showed 57% less atherosclerotic lesions in the aortae than Apoe^–/– Vim^+/+mice. The cross sections of the aortic sinuses stained with ORO also showed that atherosclerotic plaques are smaller in Apoe^–/– Vim ^–/– mice than Apoe^–/– Vim^+/+mice (Supplementary Figure 2).

We concluded that vimentin promotes atherosclerosis in hypercholesterolemic mice.