Myeloid Acyl-CoA:Cholesterol Acyltransferase 1 Deficiency Reduces Lesion Macrophage Content and Suppresses Atherosclerosis Progression*

Acyl-CoA:cholesterol acyltransferase 1 (Acat1) converts cellular cholesterol to cholesteryl esters and is considered a drug target for treating atherosclerosis. However, in mouse models for atherosclerosis, global Acat1 knockout (Acat1−/−) did not prevent lesion development. Acat1−/− increased apoptosis within lesions and led to several additional undesirable phenotypes, including hair loss, dry eye, leukocytosis, xanthomatosis, and a reduced life span. To determine the roles of Acat1 in monocytes/macrophages in atherosclerosis, we produced a myeloid-specific Acat1 knockout (Acat1−M/−M) mouse and showed that, in the Apoe knockout (Apoe−/−) mouse model for atherosclerosis, Acat1−M/−M decreased the plaque area and reduced lesion size without causing leukocytosis, dry eye, hair loss, or a reduced life span. Acat1−M/−M enhanced xanthomatosis in apoe−/− mice, a skin disease that is not associated with diet-induced atherosclerosis in humans. Analyses of atherosclerotic lesions showed that Acat1−M/−M reduced macrophage numbers and diminished the cholesterol and cholesteryl ester load without causing detectable apoptotic cell death. Leukocyte migration analysis in vivo showed that Acat1−M/−M caused much fewer leukocytes to appear at the activated endothelium. Studies in inflammatory (Ly6Chi-positive) monocytes and in cultured macrophages showed that inhibiting ACAT1 by gene knockout or by pharmacological inhibition caused a significant decrease in integrin β 1 (CD29) expression in activated monocytes/macrophages. The sparse presence of lesion macrophages without Acat1 can therefore, in part, be attributed to decreased interaction between inflammatory monocytes/macrophages lacking Acat1 and the activated endothelium. We conclude that targeting ACAT1 in a myeloid cell lineage suppresses atherosclerosis progression while avoiding many of the undesirable side effects caused by global Acat1 inhibition.

In the early stages of atherogenesis, chronic inflammation increases the adherence of monocytes to activated endothelial cells of the artery, causing more monocytes to infiltrate the subendothelial space and transform into macrophages. In sustained hypercholesterolemia, resident macrophages continue to devour denatured lipoproteins, converting excess cholesterol to cholesteryl esters, and cause the cholesterol-loaded cells to become foamy. The cholesterol esterification process is carried out by acyl-coenzyme A:cholesterol acyltransferases (ACATs), 5 also known as sterol O-acyltransferases. There are two ACAT genes, Acat1 (1) and Acat2 (2)(3)(4). Most tissues, including macrophages, express ACAT1 as the major isoenzyme and express ACAT2 as the minor isoenzyme (5). For many years, targeting ACAT1 to reduce foam cell formation has been considered as a strategy to treat atherosclerosis. Pharmacological studies in mouse models have shown that partial inhibition of both ACAT1 and ACAT2 (6,7) or of ACAT1 only (8) reduces foam cell content in plaques and is beneficial in reducing atherosclerosis without much toxicity. In human trials, the ACAT inhibitors avasimibe and pactimibe (a second generation of avasimibe) were tested as supplements to statin (the potent HMG-CoA reductase inhibitor). However, neither drug showed efficacy in reducing plaque volume. These trials were inconclusive, likely because of the fact that avasimibe activates the expression of hepatic CYP3A4, which metabolizes various drugs, including statin. Co-administering avasimibe with statin might have neutralized the beneficial effects of statin. In addition, these trials did not examine whether ACAT inhibitors stabilized plaques. Mouse genetic studies were employed as an independent approach to evaluate the effect of ACAT1 inhibition in atherosclerosis. The results showed that, in hyperlipidemic mouse models, global Acat1 KO (Acat1 Ϫ/Ϫ ) or transplant study using bone marrow cells from Acat1 Ϫ/Ϫ mice did not prevent lesion development. In addition, global Acat1 Ϫ/Ϫ caused several additional undesirable phenotypes, including hair loss, dry eye, cutaneous xanthomatosis, and a shortened life span (9,10). It was not clear whether the effects observed using the Acat1 Ϫ/Ϫ mice were due to lack of ACAT1 in macrophages or in other cell types. We have shown recently that global Acat1 KO also increases proliferation of hematopoietic stem cells within the bone marrow (BM) and causes leukocytosis in mice (11). Increasing the production of myeloid and lymphoid lineages may affect atherosclerosis at various stages. A conditional Cre-recombinase transgenic mouse line under the control of the lysozyme-M promoter (LyzMCre) is available (12) and has been used successfully by many laboratories to delete specific genes in the myeloid lineage (i.e. monocytes, macrophages, dendritic cells, neutrophil granulocytes, and eosinophils) at 80 -95% efficiency with no significant deletion in lymphoid lineages (T and B cells) as well as other cell lineages. To study the roles of ACAT1 in the myeloid lineage, here we report the generation and detailed characterization of a myeloid-specific Acat1 KO mouse model (Acat1 ϪM/ϪM ). We also report the atherosclerotic effects of Acat1 ϪM/ϪM in hypercholesterolemic Apoe knockout (Apoe Ϫ/Ϫ ) mice in vivo and on macrophage integrin ␤ 1 expression in vitro. Additionally, we compare the effects of Acat1 Ϫ/Ϫ versus Acat1 ϪM/ϪM on leukocytosis, xanthomatosis, and life span.

Mice and Generation of a Conditional Myeloid-specific Acat1
Knockout Mouse-WT, Apoe Ϫ/Ϫ , global Acat1 Ϫ/Ϫ (generated by Dr. Robert V. Farese, Jr. at the University of California, San Francisco and received in the C57BL/6J background from Dr. Sergio Fazio at Vanderbilt), Acat1 flox/flox , and Acat1 ϪM/ϪM (myeloid cell-specific Acat1 knockout) mice were all in the C57BL/6J background. The conditional Acat1 mouse (Acat1 flox/flox ) was created in collaboration with InGeneious Laboratories (Stony Brook, NY). The Acat1 LoxP construct was made by inserting two LoxP sites covering Acat1 exon 14, which includes amino acid His-450, known to be essential for ACAT1 enzymatic activity (13). The construct was injected into ES cells, and the correctly targeted clones, determined by Southern blotting and by diagnostic PCR, were injected in C57BL/6 blastocysts. The Neo marker was removed by backcrossing the resultant mice with C57BL/6 Frt mice. Mice with a heterozygous Acat1 flox allele were crossed to produce WT mice with an Acat1 allele (Acat1 ϩ/ϩ ), mice with a heterozygous Acat1 LoxP allele (Acat1 flox/ϩ ), and mice with a homozygous Acat1 LoxP allele (Acat1 flox/flox ). The alleles were confirmed by diagnostic PCR. Acat1 ϪM/ϪM mice were generated by crossing Acat1 flox/flox mice with the Cre-recombinase-transgenic mice under the control of the lysozyme-M promoter (LyzMCre) (The Jackson Laboratory).
The Apoe Ϫ/Ϫ mice were as described previously (11). The Apoe Ϫ/Ϫ /Acat1 Ϫ/Ϫ and Apoe Ϫ/Ϫ /Acat1 ϪM/ϪM mice were generated by crossing Acat1 Ϫ/Ϫ mice or Acat1 ϪM/ϪM mice with Apoe Ϫ/Ϫ mice. Mice were housed in a specific pathogen-free barrier facility under a regular light-dark cycle and fed standard chow (Harlan, catalog no. 2918, 0.02% cholesterol and 13.5 kCal% from fat) until they reached 8 weeks of age. Mice were then fed continuously with regular chow or a high-cholesterol, high-fat diet (designated here as Western diet) (Research Diet, catalog no. 99020201, 12.5%cholesterol and 35 kCal% from fat) (14) for various periods of time as indicated. Animal care and procedures were approved by the Dartmouth College Institutional Animal Care and Use Committee.
ACAT1 Protein and mRNA Expression and ACAT Activity Assay in Intact Cells-For ACAT1 expression in mouse adrenal glands, animals were killed by CO 2 asphyxiation. The adrenal glands were isolated rapidly and frozen on dry ice until ACAT1 immunoblotting analysis. Freshly thawed tissue samples were homogenized on ice in 50 mM Tris, 1 mM EDTA (pH 7.8) with protease inhibitors and then solubilized in detergent using 2.5% CHAPS and 1 M KCl. The homogenates were centrifuged at 100,000 ϫ g for 45 min at 4°C. The supernatants were used for immunoblotting analysis. For ACAT1 expression in peritoneal macrophages, monolayers of cells were solubilized with 10% SDS. The freshly solubilized lysate was used for immunoblotting analysis. For ACAT1 mRNA expression in peritoneal macrophages, total RNA was isolated with TRIzol reagent (Invitrogen) and stored at Ϫ80°C until use. Real-time PCR was performed using the DyNamo HS SYBR Green quantitative PCR kit (New England Biolabs) according to the protocol supplied by the manufacturer. The primer sequences used for detection of the Acat1 gene in the Acat1 ϪM/ϪM mouse were as follows: forward, 5Ј-TACATCATACTCCAACTACTACAG-3Ј; reverse, 5Ј-GACTGTCGTTAACAATGAAG-3Ј; amplicon size of 251 base pairs. For detection of the GAPDH gene, the primer sequences were as follows: forward, 5Ј-ATGGTGAAG-GTCGGTGTG-3Ј; reverse, 5Ј-CATTCTCGGCCTTGACTG-3Ј; amplicon size of 186 base pairs. The PCR temperature conditions included an initial denaturation at 94°C for 5 min. Subsequently, 40 cycles of amplification were performed that included denaturation at 94°C for 10 s and annealing and elongation at 60°C for 30 s. For measurement of ACAT activity in intact macrophage cells, peritoneal macrophages were isolated, plated, and fed with 100 g/ml aggregated LDL, as described previously (15), with or without 5 M F12511 (a small molecule isotype-nonspecific ACAT inhibitor) (16), for 12 h in a 6-well plate at 37°C. 20 l/well of 10 mM [ 3 H]oleate in 10% bovine serum albumin was added to the 2 ml of medium/well at 37°C for 60 min. The formation of [ 3 H]cholesteryl oleate was quantitated by scintillation counting after lipid extraction and separation by thin-layer chromatography as described previously (17).
Isolation of Peritoneal Macrophages-Macrophages were recruited to the peritoneal cavity by injecting 1 ml of thioglycolate (3% w/v, Sigma) to induce acute aseptic peritonitis. 3 days after injection, the peritoneal cavities were infused with 6 ml of cold PBS. After agitation, 5 ml of fluid from the cavity was withdrawn carefully. Cells were collected as pellets after spinning down at 1000 ϫ g for 5 min at 4°C. The cell pellets were resuspended with DMEM containing 10% FBS and incubated in 6-well culture dishes with 2 ml/well in a 10% CO 2 incubator at 37°C for 2 h. After washing with PBS twice, the firmly attached macrophages were used to measure ACAT1 mRNA, ACAT1 protein, and ACAT enzyme activity in intact cells. These cells and the RAW macrophages were also used to monitor integrin ␤ 1 expression as described in the legend for Flow Cytometry and Immunoblot Analysis to Monitor Integrin ␤ 1 Expression-The methods used for flow cytometry analyses of blood cells and bone marrow cells were the same as those used in Ref. 11. The Ly6C hi monocyte subsets in the blood were isolated to measure the cell surface marker CD11b and integrin ␤ 1 (CD29). The antibodies used (CD29-PE (HM␤1-1) and CD11b-PE (M1/70)) were from Biolegend. Cells were stained with antibodies as indicated at room temperature for 15 min and washed once with PBS containing 3% FBS. Flow cytometry analysis was performed by using a BD FACSCanto TM flow cytometer (BD Biosciences). Data were analyzed with FlowJo Software (TreeStar). For immunoblotting analysis, monolayers of RAW macrophages or peritoneal macrophages were rinsed in Hanks' buffer and then solubilized in radioimmune precipitation assay buffer with protease inhibitor mixture. The anti-CD29 antibodies used for immunoblot analysis were from Santa Cruz Biotechnology (catalog no. sc-374429).
Measurement of CD11b ϩ Myeloid Cell Migration into Aortas in Vivo-This procedure was on the basis of the method described in Ref. (18), with some modifications. WT and Acat1 ϪM/ϪM myeloid cells (CD11b ϩ ) were isolated from blood leukocytes with a cell sorter and labeled with two different fluorescent dyes. The labeling involved the use of cell surface cross-linking reagents that carried two different fluorescent colors (PKH67GL and PKH26GL, Sigma). After labeling, equal numbers of WT and Acat1 ϪM/ϪM CD11b ϩ myeloid cells were mixed and served as the donor. The cell mixtures were transferred at 4 ϫ 10 6 cells/mouse into recipient Apoe Ϫ/Ϫ mice fed a Western diet for 8 weeks. After 16 h, mouse aortas were isolated. Aortas were removed and immersed in DMEM, penicillin-streptomycin (1:100 dilution of 10,000 g/ml stock), and amphotericin B (1:100 dilution of 250 g/ml stock). Adherent connective and fatty tissues were removed by dissection. Isolated mouse aortas were incubated in a sterile collagenase-elastase solution containing collagenase type II (315 units/ml, Sigma) and elastase (1.25 units/ml, Sigma) in Hanks' balanced salt solution at 37°C for 10 min to remove the adventitia. The remaining aortas were chopped into small pieces and incubated for 1 h in the collagenase-elastase solution at 37°C. After pipetting several times to generate single-cell suspensions, the residual undigested tissues were filtered with 40-m filters. The resultant single-cell suspensions were centrifuged (1000 ϫ g, 5 min) to collect the pellets. The cell pellets were resuspended in PBS containing 3% FBS for flow analyses.
Preparation of Mouse Aortas for en Face Analysis-The en face analysis was performed according to the procedure described in Ref. 14. Briefly, the animals were fasted for 5 h. Blood was drawn. Mice were euthanized and perfused with PBS. The innominate artery, carotid arteries, aortic arch, and descending aorta to the iliac bifurcation were removed surgically and fixed in 4% formalin. After staining with Sudan IV (a neutral lipid stain), the lesion areas were quantified separately as anatomical areas, including the aortic arch (from the aortic valve to the left subclavian artery), the thoracic aorta (from the left subclavian artery to the final intercostal arteries), and the abdominal aorta (from the intercostal arteries to the bifurcation of the iliac arteries). The lesion areas in these regions were then summed to quantity the total lesion area. Quantification of Sudan staining was done with Image J.
Plaque Cholesterol and Cholesterol Ester Content Measurement-Mice were euthanized, and descending aortas were removed and weighed. Lipids in these samples were extracted by using chloroform:methanol at (2:1 v/v). After evaporation of organic solvents by nitrogen blowing, lipid samples were redissolved in ethanol for cholesterol measurements with the Wako cholesterol kit. Cholesterol ester content was calculated by subtraction of the total cholesterol (free cholesterol and cholesterol ester) value from the free cholesterol value.
Tissue Procurement for Histology-Histological tissues (aortas and skin) were collected from mice following an 8-or 12-week Western diet treatment. The specimens were submerged immediately in 10% formalin solution and sent to the Department of Pathology at Dartmouth Hitchcock Medical Center for paraffin embedding, sectioning, and H&E staining. Tissues needed for Oil Red O and immunofluorescence histochemical staining were embedded in OCT (a cryo-embedding medium) and flash-frozen on dry ice. 10-m sections were cut with a cryostat and stored in a Ϫ80°C freezer until processed.
Aortic Lesion Morphology and Cellular Content Analyses-The lesions were analyzed using immunofluorescence methods adapted from Ref. 19. In short, 10-m frozen aortic sections were blocked with 3% BSA in Tris-buffered saline and Tween 20 (TBST) and then stained with primary antibodies (anti-CD68 or anti-␣ smooth muscle actin) overnight at 4°C. The sections were then rinsed with TBST and incubated with the appropriate Alexa Fluor-conjugated secondary antibodies for 1 h at room temperature. After five TBST washes, an antifade reagent and coverslip were added to the stained sections. Images were taken with a Zeiss LS510 confocal microscope. The cross-sectional area of aortic plaque was quantified with ImageJ software. To stain and visualize aortic lipids, a protocol described in Ref. 20 was utilized. Briefly, a 0.3% Oil Red O staining solution was made freshly in H 2 0 from a 0.5% isopropanol stock. Frozen OCT-embedded aortic sections were thawed and stained with Oil Red O for 15 min at room temperature. The sections were washed 20 times with H 2 0 and mounted with an antifade reagent and coverslip. Images were visualized with an Olympus IX73 inverted fluorescence microscope. The lipid area was measured and quantified with ImageJ software.
TUNEL Assay-A Roche in situ cell death detection kit was used to quantify apoptosis within aortic lesions. The assay was conducted according to the protocol of the manufacturer. In short, 10-m aortic root sections were rinsed with PBS, fixed with 4% paraformaldehyde for 10 min, and permeabilized with 2% Triton X-100 and 1% sodium citrate for 2 min. For a positive control, sections were incubated with 100 units of DNase for 15 min at room temperature and washed twice with PBS. After the permeabilization step, sections were incubated with 50 l of TUNEL reaction mixture (5 l of terminal deoxynucleotidyltransferase enzyme ϩ 45 l of fluorescein label solution) for 1 h at 37°C. To counterstain for plaque macrophages, a rat anti-CD68 primary antibody was added to the sections and incubated overnight at 4°C. After five TBST washes, sections were incubated with Alexa Fluor 568 goat anti-rat secondary antibody. For nuclear staining, sections were incubated with a Hoechst DNA stain (1:10,000) for 5 min, followed by three TBST washes. After the final wash step, antifade and coverslips were added to each slide. Images were taken with a Zeiss LS510 confocal microscope, and cells positive for all three stains (CD68, Hoechst, and TUNEL) were counted using ImageJ software.
Measurement of Body Weight and Food Intake-Body weight was monitored once a week. The measurements were done at 4 p.m. To measure food intake, a known amount of food was given to each animal. Every 2 weeks, the food was reweighed. The amount consumed was calculated by the difference between the original amount and the amount left.
Acat1 ϪM/ϪM Does Not Cause Leukocytosis in Normal Mice, nor Does It Cause Augmented Leukocytosis in Apoe Ϫ/Ϫ Mice-Bone marrow contains hematopoietic stem cells that give rise to cells in both the myeloid and the lymphoid lineages. These cell types are known to affect atherosclerosis at various stages. We have reported previously (11) that global Acat1 Ϫ/Ϫ increased the proliferation rates of hematopoietic stem cells and altered their composition within the bone marrow, resulting in significantly higher numbers of myeloid cells in their blood. We also found that Acat1 Ϫ/Ϫ increased the expression of interleukin 3 and interleukin 7 receptors in bone marrow cells, enabling these cells to proliferate more rapidly in response to interleukin 3/interleukin 7-mediated signaling (11). These results show that, to monitor the effects of Acat1 Ϫ/Ϫ in monocytes/macrophages during atherosclerosis development, a better Acat1 KO model is needed. Here we asked whether Acat1 ϪM/ϪM also leads to leukocytosis. The result showed that, in contrast to Acat1 Ϫ/Ϫ , Acat1 ϪM/ϪM does not cause leukocytosis (Figs. 1G and 2). Additional analyses show that, unlike Acat1 Ϫ/Ϫ mice, Acat1 ϪM/ϪM mice do not exhibit abnormalities in the following parameters: composition and proliferation rates of hematopoietic stem cells (Figs.  3 and 4), blood leukocyte composition and spleen weight (Fig. 2), and bone marrow cell expression of interleukin 3 and 7 receptors (Fig. 4).

Acat1 ϪM/ϪM Decreases the Plaque Area and Reduces Lesion Size in Apoe Ϫ/Ϫ Mice Fed a High-fat/High-cholesterol Diet-
The results described above show that the Acat1 ϪM/ϪM mouse is a more suitable model than the global Acat1 Ϫ/Ϫ mouse to investigate the function of Acat1 specifically in macrophages in atherosclerosis. We fed Apoe Ϫ/Ϫ /Acat1 ϪM/ϪM and Apoe Ϫ/Ϫ / Acat1 ϩM/ϩM mice (age and sex-matched) a Western diet for 12 weeks. Afterward, the mice were euthanized. The aortas were isolated, stained with Sudan IV, and cut longitudinally for en face analysis. The results (Fig. 6) show that, for male mice, the percent total lesion area is decreased significantly in Apoe Ϫ/Ϫ / Acat1 ϪM/ϪM mice compared with the value in Apoe Ϫ/Ϫ mice (5% versus 18%). Detailed analysis shows that Acat1 ϪM/ϪM consistently caused significant decreases in percent lesion area in various regions of the aorta (ascending, thoracic, and abdominal) (Fig. 6A, right panels). In female mice, Acat1 ϪM/ϪM also caused a significant decrease in percent total lesion area as well as in percent lesion area in the ascending region. However, for the thoracic and abdominal regions, the lesion areas in Apoe Ϫ/Ϫ mice were small, so the effect of Acat1 ϪM/ϪM could not be quantified reliably (Fig. 6B, right panels). We also monitored lesion size from cross-sectional samples isolated from the aortic root. The result (Fig. 6C) shows that Acat1 ϪM/ϪM caused a significant reduction in lesion size. Together, these results demonstrate that, in the Apoe Ϫ/Ϫ mouse, Acat1 ϪM/ϪM reduces the plaque area and lesion size during atherosclerosis progression.
The Effect of Myeloid ACAT1 Deletion on Atherosclerotic Lesion Composition and Apoptosis-To further characterize the effect of depleting myeloid ACAT1 on atherosclerosis progression, we examined the aortic plaque lipid content in Apoe Ϫ/Ϫ and Apoe Ϫ/Ϫ /Acat1 ϪM/ϪM mice fed a Western diet for 12 weeks. Oil Red O-stained aortic sections revealed that Apoe Ϫ/Ϫ /Acat1 ϪM/ϪM mice accumulate less atheroma lipid compared with the Apoe Ϫ/Ϫ control (Fig. 7, A and B). Similarly, the levels of free cholesterol and cholesteryl esters extracted from whole aorta tissue of Apoe Ϫ/Ϫ /Acat1 ϪM/ϪM mice were significantly lower than those of Apoe Ϫ/Ϫ mice (Fig. 7C). Taken together, these data suggest that knocking out Acat1 in myeloid cells results in reduced aortic lipid content. Interestingly, immunofluorescence microscopy revealed that Apoe Ϫ/Ϫ mice harboring ACAT1-null myeloid cells exhibited significantly less plaque macrophages, as demonstrated by a reduction in macrophage marker CD68-positive staining in the aortic intimal layer compared with that of Apoe Ϫ/Ϫ mice (Fig. 7, D and E). Lesion smooth muscle content, detected with an anti-␣ smooth muscle-specific actin antibody, was also analyzed by immunofluorescence. Quantification of the ␣ smooth muscle actin-positive area in aortic cross sections showed that there is no significant difference in smooth muscle content between the two genotypes (Fig. 7, D and F). It has been speculated that inhibiting ACAT1 by genetic or pharmacological methods may benefit atherosclerosis because it may result in macrophage cell death, which leads to less plaque formation. To investigate whether a reduced macrophage presence within the aortic lesions of Apoe Ϫ/Ϫ /Acat1 ϪM/ϪM is a result of increased macrophage apoptosis, we measured lesion apoptosis by doublestaining with an anti-CD68 antibody and TUNEL. Cells positive for CD68 (red), TUNEL (green) and exhibiting a yellow color in the overlay microscopy image were depicted as TUNEL-positive macrophages. The result of the colocalization analysis shows that, compared with a positive control sample (DNasetreated lesion section, Fig. 7G, bottom panels), Apoe Ϫ/Ϫ and Apoe Ϫ/Ϫ /Acat1 ϪM/ϪM mice on 12 weeks of a Western diet did not show detectable evidence for lesion cell death (Fig. 7G, first  and second columns). This result is consistent with the work of Rong et al. (7), who demonstrated that partial pharmacological inhibition of ACAT1 and ACAT2 was beneficial for atherosclerosis and that ACAT inhibition was not associated with increased lesion cell death.

. Acat1 ؊M/ϪM does not affect IL3R␤ expression in LSK cells and IL7R␣ expression in B cells and B progenitor cells.
A and B, cell surface IL3R␤ (A) and IL7R␣ (B) expression was monitored by flow cytometry using the same procedure as described previously (11). A, expression in LSK cells and in total BM cells, and in B and B progenitor cells as indicated. Results are from nine WT mice and nine Acat1 ϪM/ϪM mice. Mice fed a chow diet were analyzed when they were 2 months old. Results are reported as mean Ϯ S.E.

Acat1 ϪM/ϪM Causes Decreased Appearance of Leukocytes at the Activated Endothelium in Vivo and Decreased Expression of Integrin ␤ 1 (CD29) in Activated Inflammatory Monocytes in
Vitro-The result described above (Fig. 7D) showed that, in atherosclerotic Apoe Ϫ/Ϫ mice, Acat1 ϪM/ϪM caused fewer foamy macrophages within the lesion. Previously, others have reported that global Acat1 KO caused fewer foamy macrophages within the lesion; however, the cause for their observations remained unexplained (9,10). Because we could not demonstrate significant macrophage cell death within lesions from Apoe Ϫ/Ϫ mice or the Apoe Ϫ/Ϫ /Acat1 ϪM/ϪM mice (Fig. 7G), we hypothesized that the lack of Acat1 in monocytes/macrophages might retard the ability of inflammatory monocytes to move into the lesion area. To test this possibility, we first performed an in vivo leukocyte migration assay (18) to monitor the number of exogenously labeled leukocytes appearing at the aortas of atherosclerotic mouse. Chow-fed, 2-month-old male WT and Acat1 ϪM/ϪM mice were used as blood leukocyte donors. Male Apoe Ϫ/Ϫ mice fed a Western diet for 8 weeks were used as leukocyte recipients. The results showed that, 16 h after injecting equal amounts of WT and Acat1 ϪM/ϪM leukocytes into Apoe Ϫ/Ϫ mice, significantly fewer Acat1 ϪM/ϪM leukocytes appeared at the aorta than WT leukocytes (Fig. 8A). The leukocytes used in the experiment described in Fig. 8A consisted mainly of monocytes/macrophages and neutrophil granulocytes. In the remainder of this work, we focused on monocytes/ macrophages. In inflammatory monocytes, integrin ␤ 1 (CD29), along with its heterodimeric partners (i.e. integrins ␣ 1, ␣ 3, ␣ 4, etc.), is known play key roles in facilitating the interaction between monocytes and the activated endothelium (23). The result described in Fig. 8A led us to speculate that Acat1 ϪM/ϪM might decrease the CD29 expression in monocytes. It is known that, during inflammation, monocytes, but not neutrophil granulocytes, express the specific antigen Ly6C hi at their cell surface (24). Therefore, to test this possibility, we isolated Ly6C hi -positive monocytes (the major subset of monocytes under inflammatory conditions) from the blood of Apoe Ϫ/Ϫ and Apoe Ϫ/Ϫ /Acat1 ϪM/ϪM mice on a Western diet for 8 weeks and then quantitated the expression levels of CD29 and the monocyte cell surface marker CD11b by flow cytometry. The results show that, compared with Apoe Ϫ/Ϫ monocytes, the surface expression of CD29 was decreased significantly in Apoe Ϫ/Ϫ /Acat1 ϪM/ϪM monocytes (Fig. 8C), whereas the expression of CD11b remained the same in these two cell types (Fig. 8D). The result of an additional control experiment showed that the total number of Ly6C hi monocytes in the blood is the same in Apoe Ϫ/Ϫ and Apoe Ϫ/Ϫ /Acat1 ϪM/ϪM mice (Fig.  8B). This result suggests that Acat1 KO decreases integrin ␤ 1 expression in inflammatory monocytes. To further investigate this finding, we asked whether the link between ACAT1 blockage and integrin ␤ 1 expression could also be demonstrated in cultured macrophages in vitro. Previous studies showed that, in the mouse macrophage cell line RAW 264.7, the CD29 expression level was up-regulated in response to various inflammatory cytokines. Interestingly, treating RAW macrophage cells acutely with manganese ions (for 30 min) mimics the stimulatory effects of the cytokines (25). Here we show that treating RAW macrophages (maintained in medium without inflammatory stimulants) with the ACAT1-specific inhibitor K604 at 1 M for 24 h, known to inhibit ACAT1 activity by more than 80% without inhibiting ACAT2 activity (8), did not cause any detectable change in CD29 expression (Fig. 9A). Adding manganese ions to cells not pretreated with the ACAT1 inhibitor significantly increased CD29 protein content. However, adding manganese ions to cells pretreated with the ACAT1 inhibitor failed to elicit the same stimulatory response (Fig. 9B). We performed similar experiments in peritoneal macrophages isolated from WT and Acat1 ϪM/ϪM mice (Fig. 9C), from WT and global Acat1 KO mice (Fig. 9D), or from Apoe KO mice treated with or without the ACAT1 inhibitor K604 at 1 M (Fig. 9E). The results consistently show that inhibiting ACAT1, either by gene KO or by small molecule inhibition, decreases CD29 expression FIGURE 5. Apoe ؊/؊ /Acat1 ؊/؊ mice, but not Apoe ؊/؊ /Acat1 ؊M/ϪM mice on a Western diet display leukocytosis and higher LSK cell proliferation in bone marrow. Black, white, and gray columns represent results from Apoe Ϫ/Ϫ , Apoe Ϫ/Ϫ /Acat1 ϪM/ϪM , or Apoe Ϫ/Ϫ /Acat1 Ϫ/Ϫ mice, respectively. A and B, peripheral blood leukocyte numbers (A) and CD11b ϩ Gr1 ϩ myeloid cell numbers (B) were measured by flow cytometry. Horizontal bars represent mean values. Male Apoe Ϫ/Ϫ , Apoe Ϫ/Ϫ /Acat1 ϪM/ϪM , and Apoe Ϫ/Ϫ /Acat1 Ϫ/Ϫ mice were fed a Western diet for 6 weeks after they reached 8 weeks of age. Each circle represents a result from one animal (n ϭ 9 -20 mice/group). C and D, LSK, common myeloid progenitor (CMP), granulocyte/macrophage progenitor (GMP), and megakaryocyte/erythrocyte-restricted progenitor (MEP) cells in BM were quantified by flow cytometry. Results are reported as total number (C) and percentage of total cell numbers within the BM (n ϭ 9 -20 mice/group) (D). E and F, proliferation and apoptosis of LSK cells were quantified by bromodeoxyuridine (BrdU, E) or Annexin V staining (F) (n ϭ 8 -10 mice/group). Results are reported as mean Ϯ S.E. *, p Ͻ 0.05; **, p Ͻ 0.01; ***, p Ͻ 0.001. in peritoneal macrophages activated by manganese ions (Fig. 9, C-E). Together, these results suggest that inhibiting ACAT1 enzyme activity decreases CD29 expression in inflammatory monocytes/macrophages when they are in an activated state. These results also show that the effect of ACAT1 inhibition on CD29 expression does not require the presence of the ApoE protein in macrophages.
Acat1 ϪM/ϪM Exacerbates Cutaneous Xanthomatosis but Does Not Cause a Shortened Life Span in Hyperlipidemic Apoe Ϫ/Ϫ Mice-In Apoe Ϫ/Ϫ mice on a high-fat/high-cholesterol diet, Acat1 Ϫ/Ϫ increase atherosclerotic lesions and exac-erbates cutaneous xanthomatosis. Atherosclerosis involves the accumulation of foamy macrophages in the intima of the arterial wall, whereas xanthomatosis is caused by large accumulations of lipids in macrophage-like cells at the skin area. To test the involvement of macrophage ACAT1 in xanthomatosis, we fed Apoe Ϫ/Ϫ , Apoe Ϫ/Ϫ /Acat1 Ϫ/Ϫ , and Apoe Ϫ/Ϫ /Acat1 ϪM/ϪM mice a Western diet for 8 weeks and examined xanthomas after H&E staining. The results show (Fig. 10A) that xanthomas (characterized by the presence of abundant cholesterol crystals) are amply present in both Apoe Ϫ/Ϫ /Acat1 Ϫ/Ϫ and Apoe Ϫ/Ϫ / Acat1 ϪM/ϪM mice but not readily detectable in Apoe Ϫ/Ϫ mice. Additional analyses showed that xanthomas were absent from Acat1 Ϫ/Ϫ mice or Acat1 ϪM/ϪM mice (data not shown). Therefore, similar to Acat1 Ϫ/Ϫ , Acat1 ϪM/ϪM also exacerbates cutaneous xanthomatosis. When placed on a high-cholesterol/ high-fat diet, Acat1 Ϫ/Ϫ reduces the life span of Apoe Ϫ/Ϫ mice. We tested the effects of Acat1 ϪM/ϪM versus Acat1 Ϫ/Ϫ on the life span of Apoe Ϫ/Ϫ mice placed on a Western diet for various time periods as indicated. The results (Fig. 10B; left panel,   ious time periods as indicated and found no significant difference between the two genotypes (Fig. 11). We also monitored weight gain in Apoe Ϫ/Ϫ , Apoe Ϫ/Ϫ /Acat1 Ϫ/Ϫ , and Apoe Ϫ/Ϫ / Acat1Ϫ M /Ϫ M mice on a Western diet for up to 8 weeks (Fig.  12). The result showed that, compared with Apoe Ϫ/Ϫ mice, the weight gain in Apoe Ϫ/Ϫ /Acat1 Ϫ/Ϫ mice occurred faster from the third to the seventh week, but the mice dropped their weights in the eighth week. In contrast, Apoe Ϫ/Ϫ /Acat1 ϪM/ϪM mice gained weight more slowly than Apoe Ϫ/Ϫ mice. The causes of the weight changes observed in Apoe Ϫ/Ϫ /Acat1 Ϫ/Ϫ and Apoe Ϫ/Ϫ /Acat1 ϪM/ϪM mice relative to Apoe Ϫ/Ϫ mice are currently unknown. As described earlier, both Apoe Ϫ/Ϫ / Acat1 Ϫ/Ϫ and Apoe Ϫ/Ϫ /Acat1 ϪM/ϪM mice, but not Apoe Ϫ/Ϫ mice, exhibited xanthomatosis (Fig. 10A). Xanthoma can lead to edema. The changes in weights observed (Fig. 12) may reflect the differences in the degree of swelling around the skin tissues of these two mouse genotypes. Further investigations are needed to clarify this issue.

Discussion
In the Apoe Ϫ/Ϫ mouse model for atherosclerosis, Acat1 ϪM/ϪM decreases the plaque area, reduces lesion size, and does not increase lesional cell death. Most interestingly, Acat1 ϪM/ϪM also leads to less foamy macrophages in the lesion. To provide a mechanism for the latter finding, we performed a leukocyte migration assay in vivo and showed that lack of ACAT1 causes less leukocyte adherence at the activated endothelium. We also performed cell culture studies in vitro and showed that Acat1 ϪM/ϪM causes decreased surface expression of integrin ␤ 1 (CD29) in inflammatory monocytes, but only in an activated state. Our results suggests that the sparse presence of Acat1 Ϫ/Ϫ macrophages in plaque is, at least in part, caused by the decreased surface expression of integrin CD29 in Acat1 Ϫ/Ϫ monocytes, which weakens the interactions between monocytes and endothelial cells. The mechanism by which blocking of ACAT1 reduces the expression of CD29 in activated monocytes/macrophages is currently unknown. Studies using model cell lines have shown that, without cytokine/ chemokine stimulation, the ␤1-bearing heterodimers located at the plasma membrane of monocytes stay at a low-affinity state. At a stimulated state, the integrin dimers become activated at the plasma membrane and undergo rapid endocytosis, lysosomal degradation, and recycling back to the plasma membrane (26,27). We are currently testing whether ACAT1 blockage affects one or more of these processes. In addition, in the future, it would be interesting to test whether Acat1 Ϫ/Ϫ alters the surface expression of other ligands that are also known to be involved in recruiting macrophages into plaques.
The Lys MCre technology used in this study is expected to delete Acat1 in all cells in the myeloid lineage (i.e. monocytes, macrophages, dendritic cells, neutrophil granulocytes, and eosinophils) at 80 -95% efficiency. This work focused on the effect of Acat1 deletion in monocytes and macrophages (Figs. 1; 7; 8, B and C; and 9), and the results strongly suggest that depleting Acat1 in monocytes/macrophages contributes to the atherosclerosis resistance phenotype observed in Apoe Ϫ/Ϫ / Acat1 ϪM/ϪM mice. In addition to monocytes, neutrophils are abundantly present in the blood. Because of their short life span, only a sparse amount of neutrophils can be found within an atherosclerotic lesion. However, emerging evidence implies that neutrophils play multiple roles in atherosclerosis, as reviewed in Ref. 28. Both dendritic and eosinophils are specialized immune cells present in the blood. Their roles in atherosclerosis remain largely unexplored. Further investigations are needed to investigate the effects of deleting Acat1 in other myeloid cells in atherosclerosis development.
Previous work has shown that, in hyperlipidemic mouse models, global Acat1 Ϫ/Ϫ causes an increase in lesion size (9). Why global Acat1 Ϫ/Ϫ failed to prevent atherosclerosis development is not clear at present; global Acat1 is expected to also diminish the presence of macrophages in atherosclerotic lesions. To offer one clue: unlike Acat1 ϪM/ϪM , which only depletes cells in the myeloid lineage, global Acat1 Ϫ/Ϫ is expected to delete Acat1 from all cells involved in the lesion progression process, including endothelial cells (which act as a barrier for the infiltration of various immune cells), activated T cells, and smooth muscle cells, etc., as reviewed in Ref. 22. In addition, global Acat1, but not Acat1 ϪM/ϪM , causes augmented leukocytosis in apoe Ϫ/Ϫ mice. It is possible that deletion of Acat1 in endothelial cells and/or lymphoid cells coupled with augmented leukocytosis may cause enlargement of the lesion.
In addition to increasing lesion size and inducing augmented leukocytosis, Acat1 Ϫ/Ϫ is associated with several additional FIGURE 8. A, Acat1 ϪM/ϪM retards leukocyte movement to the atherosclerotic endothelium. Chow-fed male WT and Acat1 ϪM/ϪM mice 2 months old were employed as blood leukocyte donors. Male Apoe Ϫ/Ϫ mice fed a Western diet for 8 weeks were employed as leukocyte recipients. Blood leukocytes were withdrawn, sorted by CD11b antibody, and labeled with fluorescent dye (FITC) from a WT donor and with fluorescent dye (PE) from Acat1 ϪM/ϪM donor as described under "Experimental Procedures." After injecting equal numbers of mixed CD11b ϩ leukocytes from both genotypes intravenously for 16 h, mice aortas were collected, and cells from the aortas were extracted and analyzed by flow cytometry (n ϭ 6 mice). The percentage of fluorescent FITCor PE-labeled cells in the total amount of extracted aortic cells is plotted in the y axis. B, Acat1 ϪM/ϪM does not alter the total number of Ly6C hi monocytes in the blood. C, Acat1 ϪM/ϪM decreases the cell surface level of the key integrin ␤ 1 in Ly6C hi monocytes. Ly6C hi monocytes were isolated from the blood of male Apoe Ϫ/Ϫ and Apoe Ϫ/Ϫ /Acat1 ϪM/ϪM mice fed a Western diet for 8 weeks. Cell surface expression levels of integrin ␤ 1 CD29 and the surface marker CD11b in these monocytes were quantified by flow cytometry (n ϭ 20 -25 mice/group). MFI, mean fluorescence intensity. Results are reported as mean Ϯ S.E. *, p Ͻ 0.05; ***, p Ͻ 0.001. undesirable phenotypes, including hair loss, dry eye, early lethality, and cutaneous xanthomatosis. Acat1 ϪM/ϪM in hyperlipidemic Apoe Ϫ/Ϫ mice does not cause any of the undesirable phenotypes described above, with the exception that Acat1 ϪM/ϪM also exacerbates xanthomatosis. The cellular mechanisms that lead to xanthomatosis are not well understood at present. To speculate why Acat1 ϪM/ϪM decreases the plaque area/size while augmenting xanthomatosis: macro-phages are derived from multiple cell lineages. Therefore, it seems plausible that, although most of the macrophages in the plaque originate from monocytes in the blood, the macrophages present in xanthoma may reside locally within the skin tissue. The interactions between monocytes and activated endothelium, which is a required step for atherogenesis, may not be part of the rate-limiting steps that produce xanthoma. This and other possibilities need to be investigated further in the FIGURE 9. Inactivating ACAT1 by gene KO or by pharmacological inhibition decreases the integrin ␤ 1 (CD29) level in manganese ion-activated macrophages. A and B, the effect of ACAT1 inhibition on CD29 expression level in RAW 264.7 mouse macrophage cells depends on manganese treatment. RAW 264.7 macrophages were plated in 6-well dishes in DMEM plus 10% fetal bovine serum and grown in the same medium. Cells in DMEM plus 10% fetal bovine serum were treated with 1 M of an ACAT1-specific inhibitor (K604) and 0.02% ethanol or with 0.02% ethanol only for 24 h until ϳ90 -100% confluency. A, cells were harvested without manganese ion treatment. B, cells were rinsed twice with Hanks' buffer without calcium or magnesium ions and then incubated with the same buffer and 5 mM manganese ions for 30 min at 37°C. Afterward, cells were harvested in radioimmune precipitation assay buffer with protease inhibitor mixture. The extracts were used for immunoblotting analysis using a mouse monoclonal antibody against CD29 (Santa Cruz Biotechnology, catalog no. sc-374429) to monitor CD29 expression. C-E, Acat1 KO or ACAT1 enzyme activity inhibition decreases CD29 expression in manganese-activated peritoneal macrophages. Peritoneal macrophages isolated from WT and Acat1 ϪM/ϪM mice (C), WT and global Acat1 Ϫ/Ϫ mice (D), or Apoe KO mice treated without or with 1 M ACAT1-specific inhibitor K604 for 24 h (E) were subjected to manganese activation and then harvested for immunoblot analysis in the same manner as described in A and B. Quantitation of immunoblots was performed by ImageJ analysis. *, p Ͻ 0.05; ***, p Ͻ 0.001. E, p ϭ 0.09. future. Would inhibiting ACAT1 in macrophages produce xanthomas as a mechanistic side effect in patients affected by atherosclerosis? We think this is unlikely because blocking ACAT1 in myeloid cells (Acat1 ϪM/ϪM ) alone without Apoe deficiency does not cause xanthomatosis. In humans, xanthomatosis can only be observed in patients who genetically lack Apoe or the low-density lipoprotein receptor. Xanthomatosis also occurs in certain rare genetic diseases, but it has not been reported in patients who suffer from diet-induced atherosclerosis.
Our results suggest that, in a mouse model, targeting ACAT1 specifically in a myeloid lineage may benefit atherosclerosis progression by reducing the infiltration of foamy macrophages. Atherosclerotic plaques can be classified as initial, intermediate, and advanced lesions. Unlike initial and intermediate lesions, advanced lesions contain a much larger necrotic core and lipid-laden smooth muscle cells in addition to foamy macrophages. In the future, it will be interesting to examine the effects of Acat1 ϪM/ϪM in mouse models for advanced lesions.