Myeloid HMG-CoA (3-Hydroxy-3-Methylglutaryl-Coenzyme A) Reductase Determines Atherosclerosis by Modulating Migration of Macrophages

Introduction

Statins, which are inhibitors of HMGCR (3-hydroxy-3-methylglutaryl-coenzyme A reductase), the rate-limiting enzyme in cholesterol biosynthesis, are widely used as potent cholesterol-lowering agents to prevent coronary heart disease and other atherosclerotic diseases.¹ Most of the atheroprotective properties of statins are because of their ability to lower plasma levels of LDL (low-density lipoprotein)-cholesterol.² Statins have also been reported to exert cholesterol-independent (also called pleiotropic) effects that involve improving endothelial function and decreasing oxidative stress and vascular inflammation.³ The benefits of statins extend beyond prevention of coronary heart diseases, including prevention of ischemic stroke, dementia, osteoporosis, tumor growth, and inflammation associated with infection and rheumatic diseases.³ Most of these pleiotropic effects are conceivably mediated by their ability to block the synthesis of nonsterol isoprenoid intermediates,⁴ which are required for the attachment of small GTPases to membranes.⁵

Atherosclerosis is a chronic inflammatory disease that is characterized by arterial infiltration of monocytes/macrophages.⁶ LDL are deposited in the arterial wall where they undergo oxidative modification.⁷ Lipids generated during LDL modification initiate the recruitment of circulating monocytes to subendothelial spaces, where they subsequently differentiate into macrophages. Macrophages are converted to cholesterol ester–laden foam cells via unrestricted uptake of modified LDL mostly by scavenger receptors.⁸ Cholesterol ester in LDL is hydrolyzed in lysosomes to generate cholesterol, which is subsequently esterified by acyl-CoA cholesterol acyltransferase in the endoplasmic reticulum. Necrosis of the foam cells liberates large amounts of cellular cholesterol into the extracellular milieu, thereby provoking secretion of inflammatory cytokines. Therefore, each step of this pathway is a potential target for the treatment of atherosclerosis. It has been reasonably hypothesized that statins inhibit the development of atherosclerosis by inhibiting some of these steps involving monocytes/macrophages. Indeed, statins have been shown to modulate several functions of monocytes/macrophages in vitro.^9–13 However, it is unclear whether these functional changes occur in vivo and contribute to the atheroprotective effects of statins.

To test whether reduction of HMGCR in myeloid cells, including monocytes/macrophages, is atheroprotective in vivo, we generated mice lacking Hmgcr in cells expressing lysozyme (Hmgcr^m−/m− mice) using the Cre-loxP gene targeting technique and evaluated the effects of selective reduction of HMGCR on the development of atherosclerotic lesions in the absence of Ldlr (LDL receptor). Our results show that reduction of Hmgcr suppressed the ability of monocytes/macrophages to migrate, proliferate, and survive in vitro. Furthermore, reduction of Hmgcr in granulocytes and macrophages significantly reduced the size of atherosclerotic lesions in the absence of Ldlr without affecting either plasma lipoprotein profiles or the numbers of proliferating or apoptotic cells in vivo.

Materials and Methods

The data that support the findings of this study and study materials are available from the corresponding authors on reasonable request.

Mice

Mice heterozygous for the floxed Hmgcr allele were generated as described previously.¹⁴ Mice expressing Cre recombinase under the control of the lysozyme gene promoter (Cre; Jackson Laboratory)¹⁵ were backcrossed with C57BL/6J mice 6× before interbreeding. Hmgcr^+/f carrying 1 copy of the transgene were interbred with Hmgcr^+/f littermates lacking Cre to generate myeloid cell-specific Hmgcr^m−/m− mice and littermate controls (Hmgcr^fl/fl, Cre and wild-type [WT] mice). Age- and sex-matched littermates were used as the controls. Sex is known to influence atherosclerosis in mice.¹⁶ We used male mice in most of the experiments unless otherwise stated to compare the present findings with the results in our previous studies, where male mice were primarily used. Disruption of the floxed Hmgcr allele was confirmed by Southern blot. Genotyping was performed by PCR (polymerase chain reaction) using genomic DNA isolated from the tail tip. The primer sequences for the Lys-Cre transgenes were as follows: primer A: 5’GTGGTTAATGATCTACAG3’; primer B: 5’CCTGAACATGTCCATCAG3’. For floxed Hmgcr genotyping, we used primer A: 5’ GTCGACGTTGAA TCCTCTTGTCAGAC3’ and primer B: 5’CAAAGCAGACATGAGACTATTC3’. Ldlr^−/− mice¹⁷ were interbred with Hmgcr^fl/fl and Hmgcr^m−/m− mice to obtain Hmgcr^fl/fl;Ldlr^−/− and, Hmgcr^m−/m−;Ldlr^−/− mice.

All mice were maintained in a temperature-controlled (25°C) facility with a 12 hours light/dark cycle and given free access to food (a chow diet) and water according to the regulations of the Animal Care Committees of Jichi Medical University.

Media and Buffer

We used media A to G to maintain cells and buffer A to collect and wash cells. The media were prepared as follows: medium A: DMEM (high glucose) containing 10% (v/v) FBS; medium B: medium A with 1 mmol/L sodium pyruvate, 2 mmol/L L-glutamine, and 10 ng/mL macrophage colony stimulating factor; medium C: DMEM (low glucose) with 2% (w/v) BSA; medium D: DMEM (high glucose) with 5 mg/mL of lipoprotein-deficient serum; medium E: DMEM (high glucose) with 5 mg/mL of BSA; medium F: medium A with 100 ng/mL lipopolysaccharide; and medium G: medium A with 5 mmol/L ATP. All media included 1% (w/v) penicillin G/streptomycin. Buffer A consisted of PBS with 5% (v/v) FBS.

Cells

TGEMs (thioglycolate-elicited macrophages) were obtained by peritoneal lavage with PBS from mice 3 days after a 2 mL intraperitoneal injection of 5% (w/v) thioglycollate broth. Nonelicited peritoneal cells were obtained from nontreated mice. TGEMs and nonelicited peritoneal cells were plated and cultured in medium A.

Bone marrow cell suspensions were isolated by flushing femurs and tibias of 12-week-old mice with buffer A. Cells were washed twice with buffer A and used for isolation of BMDMs (bone marrow-derived macrophages) and analysis of leukocyte populations.

BMDMs were prepared as follows. Briefly, bone marrow cells (5×10⁶ cells) were seeded and cultured with medium B at 37°C and 5% CO₂. After incubation for 3 days, the medium was replaced with new medium B and incubated for another 4 days. After incubation for 7 days total, cells were analyzed for flow cytometry.

Splenocytes, blood cells, and alveolar cells were prepared as follows. The spleen was homogenized into a single-cell suspension in buffer A by filtration through nylon mesh (70 μm). Blood was collected from an abdominal aorta after euthanasia. Alveolar cells were obtained by tracheal injection of PBS. After depletion of erythrocytes by osmotic lysis, splenocytes, blood cells, and alveolar cells were washed twice with buffer A and used for flow cytometry analysis.

Flow Cytometry

After depletion of erythrocytes by osmotic lysis, the cells were washed with PBS and resuspended in buffer A. After 5-minute incubation with purified anti-mouse CD16/CD32 antibody in buffer A to block the Fc receptor, antibody mixes were added and incubated for 30 minutes. The full list of antibodies used can be found in Table I in the online-only Data Supplement.

Flow cytometry was performed using a FACS Aria II Cell Sorter (BD Biosystems) or an LSRFortessa Cell Analyzer (BD Biosystems). All data were analyzed using FlowJo (Tree Star).

Monocytes were identified as CD11b⁺Gr-1⁻F4/80⁻; macrophages were identified as CD11b⁺F4/80⁺7AAD⁻ (7-aminoactinomycin D) or CD11b⁺Gr-1⁻F4/80⁺; BMDMs were identified as CD11b⁺F4/80⁺; granulocytes were identified as CD11b⁺Gr-1⁺7AAD⁻ or CD11b⁺Gr-1⁺; B cells were identified as B220⁺CD3e⁻; T cells were identified as B220⁻CD3e⁺. Furthermore, apoptotic, proliferating, or inflammatory macrophages were identified as CD11b⁺F4/80⁺annexin V⁺, CD11b⁺F4/80⁺Ki67 ⁺ or CD11b⁺F4/80⁺CD11c ⁺, respectively.

Immunoblot Analyses

Membrane and cytosolic fractions of TGEMs were used for immunoblot analyses of small GTPase proteins as previously described.¹⁴ Proteins (20 μg) were subjected to SDS-PAGE. Primary antibodies were as follows: RhoA (RAS homolog family member A; ARH 04; Cytoskeleton), RhoB (sc-180; Santa Cruz), Rac1 (RAS-related C3 botulinus toxin substrate 1; No. 05-389; Upstate), and Cdc42 (cell division cycle 42; 610928; BD Biosciences).

Measurements of Lipids

Blood was drawn from the jugular vein for separation of plasma. Cellular lipids were extracted by hexane/isopropyl alcohol. Concentrations of total cholesterol, free cholesterol, and triglycerides were measured using enzymatic methods.

Cholesterol Biosynthesis

Cholesterol biosynthetic rate was determined as described previously.¹⁸ Briefly, cholesterol of TGEMs (2×10⁶) was labeled with [2-¹⁴C]-sodium acetate (1 mCi/mL) for 3 hours, after incubation with medium C for 16 hours at 37°C. Cellular lipids were extracted and resolved by TLC (thin layer chromatography). [¹⁴C] cholesterol was measured by liquid scintillation counter.

Cholesterol Ester Formation

Cholesterol ester formation was determined as described previously.¹⁹ Briefly, after incubation in medium D for 24 hours, TGEMs were incubated with medium E containing 0.1 mmol/L [1-¹⁴C] oleate-albumin complex and either 50 or 100 µg/mL of acLDL (acetylated LDL) at 37°C for 24 hours.

Quantitative Reverse Transcription PCR

Total RNA was prepared from cells or tissues by using TRIzol. Relative amounts of mRNA were calculated using a standard curve or the comparative C_T method with the StepOnePlus quantitative reverse transcription PCR (qRT-PCR) cycler (Applied Biosystems) according to the manufacturer’s protocol. Mouse Gapdh mRNA was used as the invariant control. The primer-probe sets used for qRT-PCR were described previously.¹⁴

Adhesion Assay

Adhesion was determined as described previously.²⁰ Briefly, TGEMs (1×10⁵) were plated and incubated in medium A for 2 hours. Nonadherent TGEMs were removed, gently washed with PBS, and new medium A was added. At 0, 24, and 48 hours after incubation, the number of adherent TGEMs was counted by microscopy (IX70, Olympus) and Image J (National Institutes of Health).

Migration Assays

Cell migration assay was performed with the modified Boyden chamber trans-well system (3421, Costar).²¹ Infiltrating cells were stained with DAPI (4’6-diamidino-2phyenylindole; D9542, Sigma-Aldrich) and counted by Image J.

Phagocytosis Assay

Phagocytosis assay kit (IgG PE [phycoerythrin]) was purchased from Cayman Chemical (Catalog No. 600540). Assay was performed following manufacturer's protocol. Briefly, TGEMs (5×10⁵) were incubated in medium E for 16 hours. Two hours after the addition of latex-beads-PE, PE positive cells (phagocytic cells) were counted with an LSRFortessa Cell Analyzer.

Cytokine Production and IL-1β Secretion

IL (interleukin)-1β production and secretion from TGEMs was determined as described previously.²² Briefly, TGEMs (1×10⁶) were incubated in medium F for 2 hours. After the cells were washed with PBS, total RNA was prepared using TRIzol and used for qRT-PCR to measure the mRNA expression of various cytokines.

After incubation in medium F for 16 hours, TGEMs (2.5×10⁵) were incubated in medium G for 2 hours. IL-1β concentrations in the medium were measured by Mouse IL-1β/IL-1F2 DuoSet ELISA (DY 401, R&D Systems).

Histological Evaluation of Atherosclerotic Lesions

Hmgcr^m−/m−;Ldlr^−/− mice and littermate Hmgcr^fl/fl;Ldlr^−/− mice at the age of 12 weeks were fed an atherogenic diet containing 1.25% (w/v) cholesterol, 15% (w/v) fat, and 0.5% (w/v) cholic acid (Oriental Yeast Company; Tokyo, Japan). After feeding for 12 weeks, they were euthanized, and their hearts and aortas were isolated. The degree of atherosclerosis was assessed by determining lesion sizes on both pinned-open aortas and serial cross-sections through the aortic root as previously described.^23,24 To evaluate the total area of atherosclerotic lesions in the whole, pinned-open aorta (from the junction with the heart to the iliac bifurcation), each pixel of atherosclerotic lesion (as stained by Sudan IV) was measured by Adobe Photoshop. Atherosclerotic lesion area was calculated by the number of Sudan IV-stained pixels/(the number of pixels in the whole aorta)×100. To evaluate the area of atherosclerotic lesions in aortic root cross-sections, we modified by the method of Paigen et al.²⁵ Serial 6-μm aortic root cross-sections, from the beginning of the aortic sinus until the disappearance of valve cusps, were cut and mounted with 2 or 3 consecutive sections on each slide. Four slides, each separated by 60 μm, were used to evaluate the lesions: 2 at the end of the aortic sinus and 2 at the junctional site of the sinus and ascending aorta. Atherosclerotic lesion area was calculated by the number of Oil Red O-stained pixels/the number of pixels in the cross-section and converted to μm². Regions of interests were manually drawn to avoid incorrect measurement by automatic color-thresholding.

For immunohistochemistry, the sections were incubated with a primary anti-MOMA2 (monocyte+macrophage) antibody (0.8 µg/mL; Bio-Rad, Cat No MCA519G;) overnight at 4°C in a buffer. After washing, the sections were incubated with biotinylated secondary anti-rat antibody (7.5µg/mL; Vector Labs, Catalog No BA-9400) for 2 hours at 37°C. Finally, the sections were developed with 3,3’-diaminobenzidine tetrahydrochloride (Sigma, Catalog No. D-4293) and counterstained with hematoxylin.

Proliferating macrophages were determined by Ki-67 staining as previously described.¹⁴ Apoptotic cells were determined by TUNEL (TdT [terminal deoxynucleotidyl transferase]-mediated dUTP [deoxyuridine triphosphate nucleotides]-biotin nick-end labeling) using In situ Apoptosis Detection Kit (TAKARA, Catalog No. MK-500).

Adoptive Transfer of Hmgcrfl/fl and Hmgcrm⁻/m⁻ TGEM to Ldlr^−/− Mice

Mouse TGEMs were used as a source for the adoptive transfer study.²⁶ Red blood cells were lysed using BD Pham Lyse (BD Biosciences, No. 555899). Male Hmgcr^fl/fl and Hmgcr^m−/m− TGEMs were labeled with lipophilic fluorescent dye DiR (1,1’-dioctadecyl-3,3,3’,3’-tetramethylindotricarbocyanine iodide; excitation 745 nm/emission 800 nm; Caliper LifeScience, No. 125964) as described previously.²⁷ An equal number of labeled TGEMs (5×10⁶) was injected into the retro-orbital venous plexus of recipient male Ldlr^−/− mice fed the atherogenic diet for 12 weeks. Five days after injection, the aortas from recipient Ldlr^−/− mice were isolated and analyzed using an IVIS Spectrum computed tomography imaging system and Living Image Software (PerkinElmer). The signal intensity was expressed as average intensity per unit area after subtracting the tissue background signal from the control aorta and was normalized to the atherosclerotic plaque area. The lesion area was calculated in calibrated units which were converted pixels to squared millimeters using Image J.

Statistics

Results are presented as the mean±SD. For comparison between 2 groups, 2-tailed Student t test or Mann-Whitney U test was used for normally or nonnormally distributed data, respectively. Two-way repeated measures ANOVA followed by Sidak multiple was used for the analysis of the trends. For comparison between 4 groups, 1-way ANOVA followed by Tukey multiple comparisons test or Kruskal-Wallis test followed by Dunn multiple comparisons test was used for normally or nonnormally distributed data, respectively. All calculations were performed with GraphPad Prism version 6.0 for Macintosh (MDF).

Results

Reduction of Cholesterol Biosynthesis in Macrophages With Reduced Hmgcr

To determine whether expression of the Hmgcr gene was abrogated in macrophages, we performed a Southern blot analysis of DNA isolated from nonelicited cells obtained by peritoneal lavage (Figure 1A). Although only the 6.3-kb WT allele was observed in WT and only the 7.7-kb floxed allele was observed in Hmgcr^+/+ Lys-Cre (Cre) or Hmgcr^fl/fl mice, the 12.0-kb disrupted allele was observed in Hmgcr^m−/m− mice in addition to the floxed allele. Based on the intensities of the bands, 50% of the floxed allele was disrupted. Macrophages were isolated from TGEMs by flow cytometry as cells that were doubly positive for CD11b and F4/80 (Figure IA in the online-only Data Supplement). These cells (sorted macrophages) were used for measurements of mRNA by qRT-PCR. The mRNA expression of Hmgcr was decreased to 34% in the sorted macrophages from Hmgcr^m−/m− mice compared with those from Hmgcr^fl/fl mice (Figure 1B). Granulocytes were isolated from bone marrow cells as cells doubly positive for CD11b and Gr-1 (Ly-6G; sorted granulocytes; Figure IC in the online-only Data Supplement). The mRNA expression of Hmgcr was decreased to 29% in sorted granulocytes from Hmgcr^m−/m− mice compared with those from Cre mice (Figure 1C). Hmgcr expression was specifically decreased in macrophages and granulocytes from Hmgcr^m−/m− mice but did not differ in the liver between Hmgcr^fl/fl and Hmgcr^m−/m− mice (Figure 1D). The numbers of sorted macrophages and sorted granulocytes were not different between WT, Cre, Hmgcr^fl/fl, and Hmgcr^m−/m− mice (Figure IB and ID in the online-only Data Supplement). Cholesterol synthesis from acetate was decreased to 56% in the TGEMs from Hmgcr^m−/m− mice compared with those from Hmgcr^fl/fl mice (Figure 1E).

• Download figure
• Download PowerPoint

Effects of Reduction of Hmgcr in Myeloid Cells on the Number of Tissue Macrophages

To determine whether reduction of Hmgcr in myeloid cells affects whole body metabolism, we compared body weight and plasma levels of lipids among the male and female WT, Cre, Hmgcr^fl/fl, and Hmgcr^m−/m− mice (Table 1). There were no differences in these parameters in either male or female mice. Next, we compared the numbers of leukocytes in bone marrow cells, blood cells, alveolar cells, and splenocytes between Hmgcr^fl/fl and Hmgcr^m−/m− mice. Granulocytes, monocytes, macrophages, B cells, and T cells were isolated by flow cytometry by gating for Gr-1⁺CD11b⁺, Gr-1⁻CD11b⁺F4/80⁻, Gr-1⁻CD11b⁺F4/80⁺, B220⁺CD3e⁻, and B220⁻CD3e⁺, respectively (Figure II in the online-only Data Supplement). There were no differences in these cell populations in bone marrow, blood, or spleen (Figure 1F). Interestingly, however, the numbers of macrophages and monocytes in alveolar cells from Hmgcr^m−/m− mice were decreased to 65% and 56%, respectively, compared with those from Hmgcr^fl/fl mice. Hmgcr^m−/m− mice and Hmgcr^fl/fl mice (control) were used for subsequent experiments.

Effects of Reduction of Hmgcr on the Ability of Monocytes/Macrophages to Differentiate, Undergo Apoptosis, Proliferate, or Polarize

To evaluate the potency of monocytes to differentiate into macrophages in response to macrophage colony stimulating factor,²⁸ we assessed the number of BMDMs with flow cytometry. Seven days after incubation with macrophage colony stimulating factor, the number of CD11b and F4/80 doubly positive BMDMs from Hmgcr^m−/m− mice was reduced to 52% compared with those from Hmgcr^fl/fl mice (Figure 2A).

• Download figure
• Download PowerPoint

Recently, it has been reported that macrophages can proliferate in atherosclerotic lesions.²⁹ To evaluate the proliferating activity of macrophages, we measured the number of CD11b⁺F4/80⁺Ki-67⁺ cells (proliferating macrophages) in TGEMs with flow cytometry. The percentage of proliferating macrophages in TGEMs was decreased to 66% in Hmgcr^m−/m− mice compared with Hmgcr^fl/fl mice (Figure 2B).

To determine whether reduction of Hmgcr affects the ability of macrophages to undergo apoptosis, we measured the number of CD11b⁺F4/80⁺annexin V⁺ cells (apoptotic macrophages) in TGEMs by flow cytometry. There was no difference in the number of apoptotic TGEMs from Hmgcr^fl/fl and Hmgcr^m−/m− mice (Figure 2C). Thapsigargin, a sarcoendoplasmic reticulum calcium ATPase inhibitor, induces apoptosis of macrophages.³⁰ After incubation of TGEMs with thapsigargin, the number of apoptotic TGEMs from Hmgcr^m−/m− mice increased to 150% compared with those from Hmgcr^fl/fl mice.

Inflammatory (M1) macrophages in an atherosclerotic lesion secrete certain cytokines, such as TNF (tumor necrosis factor)-α, and accelerate the development of atherosclerosis.³¹ To evaluate the population of M1 TGEMs, we measured the number of CD11b⁺F4/80⁺CD11c⁺ cells using flow cytometry. There was no difference in the number of M1 TGEMs between Hmgcr^fl/fl and Hmgcr^m−/m− mice (Figure 2D).

Effects of Reduction of Hmgcr on the Ability of Macrophages to Produce Cytokines, Adhere, Migrate, Phagocytose, or Store Lipids

To evaluate the inflammatory activation of macrophages in response to lipopolysaccharide, we incubated TGEMs with lipopolysaccharide and measured the mRNA expression of inflammatory cytokines by qRT-PCR (Figure 3A) and the amount of IL-1β secretion (Figure 3B). Although there were no differences in the mRNA expression and secretion of IL-1β in TGEMs between Hmgcr^fl/fl and Hmgcr^m−/m− mice, the expression of other cytokines, such as TNF-α, IL-6, MCP-1 (monocyte chemoattractant protein 1) and GM-CSF (granulocyte-macrophage colony-stimulating factor), was significantly increased (Figure 3A).

• Download figure
• Download PowerPoint

There were no differences in the ability of TGEMs to adhere to plastic dishes between Hmgcr^m−/m− and Hmgcr^fl/fl mice (Figure 3C). However, the migratory activity of TGEMs from Hmgcr^m−/m− mice was reduced to 51% compared with those from Hmgcr^fl/fl mice (Figure 3D). The addition of mevalonate reversed the reduction of the migratory activity, but the addition of squalene did not, indicating that reduced supply of the substrates for the nonsterol pathway underlies the reduced migratory activity of TGEMs from Hmgcr^m−/m− mice. There was no difference in phagocytic activity between Hmgcr^fl/fl and Hmgcr^m−/m− mice (Figure 3E).

Intracellular contents of free cholesterol, cholesterol ester, and triglycerides were not changed (Figure 3F). There were no differences in ability to esterify cholesterol upon incubation with acLDL between Hmgcr^fl/fl and Hmgcr^m−/m− mice (Figure 3G). Consistent with this result, there were no statistically significant differences in the mRNA expression levels of genes controlling cholesterol or fatty acid metabolites except for Hmgcr (Figure 3H).

Effects of Reduction of Hmgcr on Subcellular Translocation of Small GTPase Proteins

To test whether the phenotypic changes of the monocytes/macrophages are mediated by alteration of small GTPase proteins, we compared the amounts of RhoA, RhoB, Rac1, and Cdc42 in plasma membrane and cytosol fractions of TGEMs between Hmgcr^fl/fl;Ldlr^−/−, and Hmgcr^m−/m−;Ldlr^−/− mice (Figure III in the online-only Data Supplement). The ratios of RhoA, RhoB, Rac1, and Cdc42 proteins in the plasma membrane to those in cytosol were 2.5, 1.3, 1.3, and 1.1, respectively. In particular, the ratios of RhoA and Rac1 were significantly higher in TGEMs from Hmgcr^m−/m−;Ldlr^−/− than in TGEMs from Hmgcr^fl−/fl−;Ldlr^−/, indicating that reduction of HMGCR rather promotes the membrane translocation of small GTPase proteins.

Reduction of Hmgcr Attenuates Atherosclerosis

To evaluate the effects of reduction of Hmgcr in myeloid cells on the development of atherosclerosis, we generated Ldlr-deficient mice which lacked Hmgcr in a myeloid cell-specific manner (Hmgcr^m−/m−;Ldlr^−/−) and used Hmgcr^fl/fl;Ldlr^−/− mice as a control. After 12 weeks of feeding an atherogenic diet, the aorta and heart were collected to evaluate atherosclerotic lesion areas of male and female mice (Figure 4; Figure IV in the online-only Data Supplement). The en face lesion areas in the aortas from male Hmgcr^m−/m−;Ldlr^−/− mice were decreased to 66% compared with those from Hmgcr^fl/fl;Ldlr^−/− mice (Figure 4A and 4B). Similarly, the cross-sectional lesion areas of the aortic sinuses from male Hmgcr^m−/m−;Ldlr^/− mice were decreased to 60% compared with those from Hmgcr^fl/fl;Ldlr^−/− mice (Figure 4C and 4D). The results were essentially the same in female mice (Figure IV in the online-only Data Supplement). Further, in lesions of the aortic sinuses, the areas that positively stained for MOMA-2, a macrophage marker, were decreased by 52% in male Hmgcr^m−/m−;Ldlr^/− mice (Figure 4E and 4F). MOMA-2-positive areas relative to total plaque areas were comparable between male Hmgcr^fl/fl;Ldlr^−/− and Hmgcr^m−/m−;Ldlr^/− mice as shown in Figure 4G, suggesting that the smaller lesions of Hmgcr^m−/m− mice resulted from a decrease in the influx or migration of monocytes/macrophages into the lesions. However, there were no differences in body weight and plasma lipid parameters between Hmgcr^fl/fl;Ldlr^−/−, and Hmgcr^m−/m−;Ldlr^/− mice (Figure V in the online-only Data Supplement; Table 2).

• Download figure
• Download PowerPoint

To determine whether reduction of Hmgcr changes proliferative or apoptotic activity of macrophages in lesions, we performed immunohistochemical staining for Ki-67 and TUNEL. We could hardly detect any Ki-67-positive macrophages in either genotype (Figure VIA in the online-only Data Supplement). The number of TUNEL-positive macrophages in the lesions were not significantly different between Hmgcr^fl/fl;Ldlr^−/−, and Hmgcr^m−/m−;Ldlr^/− mice (Figure VIB and VIC in the online-only Data Supplement). In advanced lesions, the development of necrosis reflects the degree of apoptosis. The extent of necrotic area per atherosclerotic lesion was comparable between genotypes (Figure VID and VIE in the online-only Data Supplement).

Reduction of Hmgcr ImpairsIn Vivo Migration of TGEMs to the Lesions

To determine the role of HMGCR in the accumulation of macrophage in atherosclerotic lesions, we compared the in vivo migration of adoptively transferred male Hmgcr^fl/fl and Hmgcr^m−/m− TGEMs to atherosclerotic lesions of male Ldlr^−/− mice (Figure 5A). The in vivo migration of Hmgcr^m−/m− TGEMs was reduced by 72% compared with Hmgcr^fl/fl TGEMs (Figure 5B).

• Download figure
• Download PowerPoint

Discussion

In the present study, we showed that macrophages with genetically reduced Hmgcr had impaired ability to migrate, proliferate, and survive in vitro. However, reduction of Hmgcr did not cause any difference in the macrophages’ ability to polarize to M1, adhere, phagocytose or store lipids. In addition, plasma membrane–associated small GTPase proteins, such as RhoA were increased in Hmgcr^m−/m− macrophages. In the setting of Ldlr deficiency, Hmgcr^m−/m− mice developed significantly smaller atherosclerotic lesions than Hmgcr^fl/fl mice, although there were no differences either in the plasma lipoprotein profiles or in the numbers of proliferating or apoptotic cells in the lesions between the 2 types of mice in vivo. The in vivo migration of Hmgcr^m−/m− macrophages to the lesions was reduced by 72% compared with Hmgcr^fl/fl macrophages. Thus, genetic reduction of HMGCR in myeloid cells may exert atheroprotective effects primarily by decreasing the migratory activity of monocytes/macrophages to atherosclerotic lesions.

Why was the disruption of Hmgcr incomplete? A similar incomplete disruption of Hmgcr was observed in 3-week-old liver-specific Hmgcr knockout mice that were obtained by crossing the same floxed mice with Alb (albumin)-Cre mice.¹⁴ However, disruption of lipoprotein lipase or squalene synthase was almost complete when the mice with the floxed gene were crossed with Lys-Cre mice²⁴ or Alb-Cre mice,³² respectively. Therefore, the incomplete disruption is likely not attributable to the Cre mice but to the targeted gene, Hmgcr. Indeed, Hmgcr is essential for survival, as demonstrated by cytocidal effects of high concentrations of HMGCR inhibitors.³³ It is reasonable to speculate that cells whose Hmgcr were completely eliminated died out and that only cells whose Hmgcr remained were able to survive. Partial deletion can be possible if only a single allele is deleted. Interestingly, similar incomplete disruption was reported for several genes, such as c-FLIP (FLICE-like inhibitory protein), an anti-apoptotic gene,³⁴ and IL-4Rα.³⁵

Hmgcr^m−/m− mice had fewer alveolar monocytes and macrophages (Figure 1F). Why were only alveoli affected? We speculate that this question is related to the fact that tissue-resident macrophages originate from 2 distinct progenitors: yolk sac–derived erythro-myeloid progenitors and hematopoietic stem cells.³⁶ Before birth, hematopoietic stem cells almost entirely replace monocytes/macrophages, which develop from erythro-myeloid progenitors during embryogenesis. In alveoli, this replacement takes place more gradually after birth. If macrophages derived from erythro-myeloid progenitors express higher levels of lysozyme than those from hematopoietic stem cells, macrophages derived from erythro-myeloid progenitors may express higher activity of Cre recombinase, thereby abrogating HMGCR more completely. Further studies are needed to test this hypothesis.

Previous in vitro studies have shown that statins affect several characteristics of macrophages. Some of these effects were recapitulated in our study. For example, the reduction of Hmgcr impaired differentiation (Figure 2A) and migration (Figure 3D), as was reported for statin treatment.^37,38 However, other effects of statins such as alterations in adhesion,³⁹ phagocytosis,⁴⁰ and cytokine secretion⁴¹ were not observed in our study. Of course, these differences might result from the different methods of inhibiting HMGCR: drugs and genetic manipulation. It is also possible that statins modulate cell functions via their off-target effects independently of inhibition of HMGCR.

HMGCR is indispensable for cell survival because it supplies isoprenoids, which are used for prenylation of essential proteins.⁴ Surprisingly, reduction of Hmgcr increased the levels of plasma membrane–associated small GTPase proteins (Figure III in the online-only Data Supplement). These results were apparently paradoxical, but similar activation of small GTPase proteins and stimulated production of proinflammatory cytokines were reported in myeloid cell-specific GGTase-I (geranylgeranyltransferase type I) deficient mice.⁴² In these mice, GTP-bound active forms of small GTPase proteins were increased, although the small GTPase proteins were underprenylated and their subcellular localization was not affected. Because GGTase-I is downstream of HMGCR, myeloid cell-specific reduction of Hmgcr may mimic the phenotype of myeloid cell-specific deficiency of GGTase-I in this respect. Moreover, the impaired migratory activity of TGEMs from the Hmgcr^m−/m− mice might be caused by increased activity of RhoA, as random migration was reported to be increased in Rho-deficient neutrophils.⁴³ Therefore, most of the phenotypic changes of macrophages with genetically reduced Hmgcr, such as impaired ability to migrate, proliferate, and survive in vitro, can be explained by altered functions of these small GTPase proteins. In this respect, it is interesting to note that a farnesyltransferase inhibitor, which reduced the amount of active Ras, prevented the development of atherosclerosis.⁴⁴

In vitro studies showed both potentially proatherogenic and antiatherogenic phenotypes in the macrophages with reduced Hmgcr. Stimulated synthesis of proinflammatory cytokines, such as TNF-α, in response to lipopolysaccharide (Figure 3A) might be proatherogenic, whereas decreases in either differentiation (Figure 2A), proliferation (Figure 2B), or migration (Figure 3D) and an increase in apoptosis under ER (endoplasmic reticulum) stress (Figure 2C) might be antiatherogenic. The in vivo phenotype of reduced atherosclerosis indicates that the antiatherogenic phenotypes of macrophages collectively dominated over any proatherogenic phenotype. Based on the results of Ki-67 or TUNEL staining, as well as the adoptive transfer experiment of TGEMs, the decrease in migratory activity appears to play a more important role in the atheroprotection in Hmgcr^m−/m− mice than the changes in proliferation and apoptosis. Decreased cell proliferation was not detected in vivo because the cell proliferation was negligible under the experimental condition. Increased apoptosis was not detected in vivo, likely because ER stress in the lesional cells was not as severe as that induced by thapsigargin in vitro.

Our study has several limitations. First, we cannot completely define which cell lineage was responsible for the atheroprotection because HMGCR was deficient in both macrophages and granulocytes. There is increasing evidence that granulocytes are involved in the development of atherosclerosis,⁴⁵ although evidence for the involvement of monocytes/macrophages is overwhelming. Second, we did not address the molecular pathway by which reduction of HMGCR affects the functions of monocytes/macrophages, although it is highly probable that isoprenylation was the key intermediary linking HMGCR deficiency to altered functions of macrophages. Third, it is unknown whether statins used in the clinical setting inhibit HMGCR in monocytes/macrophages to the same degree accomplished in the current study.

In conclusion, macrophages with genetically reduced Hmgcr have an impaired ability to migrate, proliferate, and survive in vitro. Impaired migration in particular, may contribute to the atheroprotective effect of the genetic reduction of Hmgcr in myeloid cells in vivo. Targeting this specific pathway may be a promising option for developing novel pharmacotherapies for atherosclerosis.

Acknowledgments

We thank Yukiko Hoshino, Mika Hayashi, and Nozomi Takatsuto for excellent technical assistance and Dr Timothy F. Osborne at Sanford-Burnham-Presbys Medical Discovery Institute for critical reading of the article. We also thank the Core Center of Research Apparatus (Jichi Medical University) for the management of FACSAria II and LSRFortessa. LSRFortessa was subsidized by JKA Foundation through its promotion funds from KEIRIN RACE.

Footnotes