Reversal of hyperlipidemia with a genetic switch favorably affects the content and inflammatory state of macrophages in atherosclerotic plaques

Abstract

Background

We previously showed that the progression of atherosclerosis in the Reversa mouse (Ldlr^−/−Apob^100/100Mttp^fl/flMx1Cre^+/+) was arrested when the hyperlipidemia was normalized by inactivating the gene for microsomal triglyceride transfer protein. Here we tested whether atherosclerosis would regress if the lipid levels were reduced after advanced plaques formed.

Methods and Results

Reversa mice were fed an atherogenic diet for 16 weeks. Plasma lipid levels were then reduced. Within 2 weeks, this reduction led to decreased monocyte-derived (CD68+) cells in atherosclerotic plaques and was associated with emigration of these cells out of plaques. Also, the fall in lipid levels was accompanied by lower plaque lipid content and by reduced expression in plaque CD68+ cells of inflammatory genes and higher expression of genes for markers of anti-inflammatory M2 macrophages. Plaque composition was affected more than plaque size, with the decreased content of lipid and CD68+ cells balanced by a higher content of collagen. When a reduced lipid levels was combined with the administration of pioglitazone to simulate the clinical aggressive lipid management and PPARγ agonist treatment, the rate of depletion of plaque CD68+ cells was unaffected, but there was a further increase in their expression of anti-inflammatory macrophage markers.

Conclusion

The Reversa mouse model is a new model of atherosclerosis regression. After lipid lowering favorable changes in plaque composition were independent of changes in size. In addition, plaque CD68+ cells became less inflammatory, an effect enhanced by treatment with pioglitazone.

Introduction

Blocking progression of atherosclerotic lesions and inducing their regression are highly desirable goals. While mouse models have provided valuable information about atherosclerosis progression, comparatively little is known from them about the molecular mechanisms underlying regression. We previously allowed plaques to develop in Apoe-deficient mice and then transplanted segments of the diseased aorta into wild-type mice ^{1, 2}. Plaque regression (defined as a reduction in monocyte-derived cell content) rapidly ensued. Indeed, there was a remarkable 50% loss from the plaques of the monocyte-derived (CD68+) cells within three days ^{3, 4}. The loss was due to the migration of these cells to regional and systemic lymph nodes ³. We hypothesized that since dendritic cells (DCs) are migratory though a pathway dependent on CCR7 ⁵, the migratory CD68+ cells during regression acquired some features of DCs. Indeed, we found that during regression expression of CCR7 was induced in plaque CD68+ cells, and their migration was dependent on the CCR7-mediated pathway ⁴.

Recently another model, the Reversa mouse, has given us the opportunity to extend our studies to determine whether the findings in the transplant model were generally applicable to the regression process or model speicifc. Reversa mice (Ldlr^−/−Apob^100/100Mttp^fl/flMx1Cre^+/+) have severe hypercholesterolemia as a result of homozygosity for the LDL receptor knockout allele and an “apolipoprotein B100 (apoB100)–only” allele, but the hypercholesterolemia can be conditionally eliminated ⁶ by inducing the expression of the Mx1-Cre transgene, which inactivates Mttp (the gene for microsomal triglyceride transfer protein). In mammalian liver, microsomal triglyceride transfer protein is required for the secretion of atherogenic apoB-containing lipoproteins ⁷, so that after inactivation of its gene, plasma levels of VLDL and LDL fall ⁶.

We previously showed that by inactivating the Mttp gene early in life, atherosclerosis progression was retarded ⁶. We reasoned that the Reversa mouse could be used to study regression if the hypercholesterolemia were reduced after plaques developed. In the current study, we establish the Reversa mouse as a model of atherosclerosis regression, and then tested whether the reduction of hypercholesterolemia not only led to plaque depletion of CD68+ cells, as in the transplant model, but also to an alteration of their phenotype to “alternatively activated” M2 macrophages, which are considered to be anti-inflammatory (e.g., see ⁸). We also investigated whether the changes in atherosclerotic plaques occurring after the reduction in plasma cholesterol would be augmented by treatment of mice with the PPARγ agonist pioglitazone, given its ability to reduce cardiovascular risk ⁹ and its widespread clinical use in combination with lipid-lowering agents.

Methods

Genetically Modified Mice and Experimental Design

All procedures were approved by the Institutional Animal Care and Use Committee. Four- to five-week-old “Reversa” mice (n = 86) were weaned onto a Western diet containing 0.15% cholesterol and 21% fat (WD; Research Diets) for 16 weeks. Mice were divided into “treated” and “control” groups. The treated group was switched to a chow diet and the expression of the Mx1-Cre transgene was induced with injections of polyinosinic-polycytidylic acid (pIpC) (Sigma, 500 μg every other day for a total of 4 injections). The pIpC injections led to inactivation of Mttp in the liver, resulting in the elimination of apoB100-containing lipoproteins from the plasma and normalization of plasma cholesterol levels nine days after the first injection. That time point is considered the “baseline” or day 0 time point for all experiments. Separate groups of mice were euthanized either at the baseline time point or 3, 6, 14, 28, or 56 days after the baseline time point. The control group, which was given injections of normal saline and continued on the western diet, was euthanized at the same time points. In some experiments, mice treated with pIpC were maintained on a chow diet or a chow diet containing pioglitazone (200 ppm), a dose that induces PPARγ target genes in CD68+ cells in plaques ⁴.

Aortic transplant experiments were done as previously described ⁴ with the only difference being that the donor mice were hyperlipidemic Reversa mice fed the western diet for 16 weeks, and the recipients were normolipidemic Reversa mice. The transplants were performed two weeks after the last pIpC injections in the recipients. Five days after the atherosclerotic aortic segment was transplanted, it was harvested from the recipient mouse, frozen in OCT, sectioned, and immunohistochemical and morphometric analyses conducted as described below.

Lipid and Lipoprotein Analyses

Blood samples were obtained from the retroorbital plexus. Plasma cholesterol levels were determined by colorimetric enzymatic assays that were adapted to 96-well formats (Infinity Total Cholesterol Reagent or Infinity Triglyceride Reagent, Sigma).

Labeling of Blood Monocytes

Ly6C^hi (CCR2-high expressing) and Ly6C^lo (CCR2-low expressing, CXCR1-high expressing) monocytes in sections of atherosclerotic plaques were labeled with fluorescent 0.5 μm microspheres (beads) as previously described ^{10, 11}. For selective labeling of the Ly6C^hi monocytes, 250 μl of liposomes containing clodronate were intravenously injected into mice to transiently deplete monocytes, followed by an injection of 250 μl of fluorescent beads 24 h later. To selectively label Ly6C^lo monocytes, 250 μl of the fluorescent beads were injected intravenously without the clodronate pretreatment. Briefly, circulating monocytes will take up the beads and then enter tissues, including atherosclerotic plaques. Fluorescent beads in the plaque are visualized by fluorescence microscopy and counted in a blinded fashion. Over time, the bead content of the plaque will decrease if the beads left in the same cells that brought them in or if they were transferred to another monocyte-derived cell that then left. In other words, a decrease in bead number indicates that there was emigration of monocyte-derived cells from the plaque.

Tissue Processing, Histology, Immunohistochemistry, and Morphometry

Mice were euthanized, and the hearts with attached aortic roots were collected and frozen in Optimum Cutting Temperature (OCT) compound as described ^{4, 12}. Serial frozen aortic sections (6-μm thick) were prepared and mounted on positively charged slides (Color Frost Plus; Fisher Scientific). Every fifth slide was stained with a CD68-specific antibody ¹³ and used for morphometric analysis and as a guide slide for laser-capture microdissection. Additional slides were stained with Oil-red-O ¹², MOMA-2, CCR7, MCP-1, or Arg I. For some studies, immunofluorescence microscopy was employed. Collagen content of lesions was assessed with Sirius Red–stained slides under polarizing light¹⁴. Intimal lesions andCD 68-immunostained areas were quantified by computer-aided morphometric analysis of digitized images (ImagePro Plus 3.0 software; Media Cybernetics, Silver Spring, MD).

Laser Capture Microdissection and RNA Extraction

For laser capture microdisssection (LCM), all procedures were performed under RNase-free conditions. Sections were fixed in 70% ethanol for 1 min, washed in H₂O, stained with Mayer’s hematoxylin (VWR Scientific) for 1 min, washed in H₂O, incubated in PBS for 15 sec (to develop the blue color), washed in H₂O, partially dehydrated in 70% ethanol followed by 95% ethanol, stained in eosin Y (VWR Scientific) for 5 sec, washed in 95% ethanol, and completely dehydrated in 100% ethanol (30 sec) and xylene (7 min). After air-drying for 10 min, foam cells could be identified by light microscopy and were verified by CD68 staining of guiding slides. Samples from 3 mice from the same treatment group and time point were pooled, and RNA was extracted with the QIAGEN micro RNA kit with on-column DNase I treatment. The concentration of RNA was determined by the Ribogreen RNA Quantitation kit (Molecular Probes), and RNA quality verified with the Agilent 2100 Bioanalyzer.

Quantitative Real-Time (qRT)-PCR

RNA abundance was determined by qRT-PCR with 100 pg of total RNA ⁴. The primer and probe sequences have been described previously ^{4, 15-17}. All data were normalized to cyclophilin A and expressed as fold change, compared with controls. For tissue samples generated by LCM, results were obtained from 2-3 independent samples, each pooled from plaques of 3 mice.

Statistics

All data are expressed as average ± SEM. For studies involving atherosclerotic lesions, the number of mice is indicated in the figure legends. Unless otherwise indicated, for RNA analyses two pooled samples were used for each experiment. PRISM software (GraphPad, San Diego) was used to analyze statistical differences typically by one-way ANOVA with the Bonferroni post-test for differences between selected pairs of means. In some cases, as noted in Results or figure legends, an unpaired t-test or ANOVA with Dunnett’s post-test was used to evaluate statistical differences. P values less than 0.05 were considered significant.

Results

Reversal of the hyperlipidemia decreases the content of CD68+ cells and lipids in plaques

Reversa mice (Ldlr^−/−Apob^100/100Mttp^fl/flMx1Cre^+/+) were placed on Western diet for 16 weeks, allowing them to develop severe hypercholesterolemia (1115 ± 215 mg/dl) and advanced atherosclerotic plaques. Nine days after the administration of pIpC, plasma cholesterol levels fell to 86 ± 5 mg/dl. That time point was designated the day 0 or the baseline time point, and mice were euthanized 3, 6, 14, 28, or 56 days later. Plasma cholesterol levels at these time points were measured, and the reduction in hyperlipidemia was sustained for the duration of the experiment (Supplemental Table 1).

The plaque area of CD68+ cells (predominately macrophages and macrophage-foam cells) at each time point is graphically summarized in Figure 1A. Based on ANOVA followed by the Bonferroni test, this area in pIpC-treated mice was reduced significantly at days 6, 14, and 28 (p<0.0001) compared to mice at the corresponding time points in the control group. Within each treatment group, plaque CD68+ content significantly decreased by day 3 after hyperlipidemia was reduced, whereas it significantly increased by day 6 in the control group (Figure 1A), as determined by Dunnett’s testing to compare results at day 0 to those at the other time points.

Figure 1. Correcting hyperlipidemia in Reversa mice reduces the content of CD68+ cells in plaques.

(A) CD68+ cell content in plaques of Reversa mice after plasma lipid normalization. Reversa mice (n=13/group) were fed a western diet for 16 weeks, at which point their high plasma lipid levels were either maintained (saline group) or normalized (pIpC group), as described in Methods. CD68+ cell content, quantified by morphometric techniques, was significantly lower 14 days after plasma lipids stabilized at the reduced levels. Statistical analysis within each treatment group was by ANOVA followed by Dunnett’s test to compare results at each time point to those on day 0. ⁺p < 0.03 or *p <0.0001 compared to day 0. (B) Immunohistochemical staining of CD68. By 56 days, CD68+ cells (red) in pIpC-treated mice were virtually undetectable.

As shown in Figure 1B, plaque CD68+ cell content in the pIpC group was negligible by day 56 (total cholesterol then averaged 80±11 mg/dL). In addition, using laser captured CD68+ cells from baseline and regressing plaques, we found that there was no significant difference in CD68 mRNA levels (Supplemental Figure 1A). In bone marrow-derived M1 and M2 macrophages, there was also no significant difference in CD68 mRNA abundance (data not shown). These results suggested that the decrease in CD68+ content in the pIpC-treated group was due to reduced cell number and not reduced cell expression. This would also be consistent with the immunostaining of plaque sections of another known macrophage marker, MOMA-2. Sections from only a subset of the mice were available for this analysis, but as shown in Supplemental Figure 1B, we found a high degree of correlation (r=0.96) between the areas stained by CD68 and MOMA-2 in these samples.

In order to exclude as possible explanations for the results in Figure 1 either non-specific effects of pIpC on the immune system or the loss of MTP activity in cells of the arterial wall, we performed transplant studies using hyperlipidemic Reversa mice as donors, and pIpC-treated (normolipidemic) Reversa mice as recipients. The aortic grafts were harvested 5 days after transplantation. Note that the donors were not treated with pIpC, so it could not exert any effects on the immune system or effect the inactivation of the MTP gene in any tissue. Also, the recipients were used 2 weeks after the last injection of pIpC, well past the 48 h duration of its action in vivo ¹⁸.

The results are summarized in Figure 2. First, they show that as in apoE-/- mice ^{3, 4}, transfer of atherosclerotic aortae from another hyperlipidemic model to normolipidemic recipients causes decreased CD68+ cell content of plaques. Second, the % of reduction (~50%) found 5 days after exposure to the normolipidemic environment is quite comparable to that shown in Figure 1A after approximately the same duration of normolipidemia. Thus, the results demonstrate that the reduction in plaque CD68+ cells after hyperlipidemia is reduced is independent of the atherosclerotic mouse model, of MTP expression in the arterial wall, and of the effects of pIpC on the immune system. Though transplantation of an aortic arch to the abdominal aorta exposes the plaques to hemodynamic changes, as well as potential traumatic damage and inflammatory stimulation, given the similar results for the decreased content of CD68+ cells in the transplantation (Figure 2A) and the non-surgical experiments (Figure 1A), it is unlikely that these factors were significant confounders in the present study.

Figure 2. Transplantation of aortic segments from hyperlipidemic Reversa mice into normolipidemic Reversa mice leads to CD68+ cell depletion.

(A) Male hyperlipidemic Reversa mice were fed western diet for 16 weeks and aortic segments transplanted into normolipidemic Reversa (i.e., treated with pIpC and used 2 weeks thereafter). The transplanted aortic segments were then harvested 5 days later and morphometric analysis of plaque CD68 content was performed. *p=0.0045 compared to baseline by unpaired two-tailed t-test; n = 6/group. (B) Representative sections demonstrating a decrease in CD68+ cells as assessed by immunohistochemistry.

As shown in Figure 3A, the plaque content of neutral lipids in both baseline and pIpC-treated Reversa mice was mostly found in CD68+ areas. The lipid content of plaques in the pIpC-treated mice was significantly reduced by approximately 50% at day 14 after baseline (Figures 3B and 3C), similar to the reduction in CD68+ content at that time point (Figure 1A). In contrast, in the saline treated control group, at the same time point, there was an increase (~3X compared to Baseline) in the lipid content of the plaque as assessed by oil-red-O staining (data not shown). There was a concomitant increase in CD68+ area. Taken together, the changes in neutral lipid content (presumably cholesteryl ester) were directly related to the plaque CD68+ cell content, reflecting that this population includes the cholesterol-laden macrophage-foam cells.

Figure 3. Correcting hyperlipidemia in Reversa mice lowers neutral lipid content in aortic atherosclerotic plaques.

(A) Immunohistochemistry and co-localization studies were performed in the Baseline and pIpC-treated Reversa mice after 16 weeks of western diet feeding and 14 days after reduction of plasma lipids, respectively. The pictures are representative images showing Oil-red-O staining (neutral lipid, presumably cholesteryl ester). Note that although the staining is less for lipid and CD68+ cells in the pIpC-treated mouse, in both cases, there is co-localization of the 2 plaque components as assessed by confocal immunofluorescence microscopy. (B) Oil-red-O staining of the aortae of Reversa mice at baseline and 14 days after reduction of the plasma lipids (by pIpC injection). Reversa mice treated with pIpC had visibly reduced amounts of Oil-red-O staining. (C) Morphometric analysis of Oil-red-O staining. *p < 0.05 compared to baseline. n = 8/group. “L” indicates the lumen of the artery.

Expression of the chemokine receptor CCR7 is upregulated in CD68+ cells in plaques after reversing hyperlipidemia

We previously showed the depletion of CD68+ cells from apoe^−/− plaques 3 days after transplanting aortas into wild-type mice, and that this reduction was associated with the CCR7-dependent migration of monocyte-derived cells from the aorta to either regional lymph nodes or the systemic circulation. ^{3, 4}. To evaluate the importance of CCR7 in the regression of atherosclerosis in Reversa mice, we measured its expression in CD68+ cells laser captured from plaques 14 days after normalizing the plasma lipid levels. As shown in Figure 4A, CCR7 expression was increased significantly in plaque cells captured from pIpC-treated mice (relative to those from saline-injected control mice). These findings were confirmed by immunohistochemistry (Figure 4B).

Figure 4. Effects of the correction of hyperlipidemia on CCR7 expression levels in plaque CD68+ cells and on the emigration from plaques of macrophages derived from Ly6C^hi and Ly6C^lo circulating monocytes.

(A) Mice were treated as in Figure 1. CCR7 mRNA levels in plaque CD68+ cells laser captured 14 days after pIpC or saline treatments were assessed by qRT-PCR. (B) CCR7 protein expression in aortic plaques under the conditions in panel A, as judged by immunohistochemistry with a CCR7-specific antibody. (C, D) Quantification of the relative levels in plaques of macrophages derived from circulating Ly6C^hi (“CCR2+”) and Ly6C^lo (“CX3CR1+”) monocytes. One week before the first pIpC or saline injection, the two major subsets types of circulating monocytes were labeled in vivo with fluorescent beads (Methods). Fourteen days after the baseline time point, the number of fluorescent beads in plaques was counted. Fewer beads were present in sections from pIpC-treated mice, indicating emigration of the macrophages containing them. *p < 0.05; n >10/group in panel A; n = 5/group for panels C and D. “L” indicates the lumen of the artery.

In order to investigate the possibility that pIpC, independent of reducing plasma lipid levels may be inducing CCR7, we performed additional control experiments. Ldlr^−/− mice maintained on a WD diet for 16 weeks were injected with pIpC and the aortas were examined by immunohistochemistry. We did not observe significant CCR7 staining of plaques in those mice, suggesting that the changes that we observed in the Reversa mice were regression-related and not an artifact of pIpC treatment (data not shown).

Both Ly6C^hi and Ly6C^lo-derived macrophages emigrate from atherosclerotic plaques

The changes in CCR7 expression suggested that plaque CD68+ cell depletion could be mediated by cell emigration. There are two major circulating mouse monocyte subpopulations, CCR2⁺CX3CR1⁺Ly6C^hi and CCR2⁻CX3CR1⁺⁺Ly6C^lo, and both are normally recruited to atherosclerotic plaques ^{10, 11, 19-22}. To determine whether the plaque macrophages derived from either subset emigrate during regression, we took advantage of the ability to differentially label the subsets of circulating monocytes ^{10, 11, 22} and then follow the plaque content of labeled cells (which are CD68+; Supplemental Figure 2) after injections of pIpC or normal saline. The labeling of the Ly6C^hi population requires the use of clodronate to first deplete the circulating monocytes ¹¹. To address whether clodronate treatment itself may be affecting monocyte/macrophage trafficking, we performed morphometric analysis on histological sections from mice that were treated with pIpC and clodronate, and compared the CD68 areas to the corresponding values obtained when only pIpC was used. As shown in Supplemental Figure 3, it is clear that clodronate had no effect on plaque CD68+ cell content independent of the reduction of plasma lipids.

As shown in Figures 4C and 4D, macrophages derived from both the Ly6C^hi (CCR2+) and Ly6C^lo (CX3CR1+) subsets emigrated from plaques, as assessed by the numbers of beads remaining in the plaque 14 days after reversal of hyperlipidemia.

Collagen content is increased in atherosclerotic plaques after reversal of hyperlipidemia

As shown in Figure 5A, after the reduction in hyperlipidemia the change in total plaque area did not follow the monotonic decline in CD68+ content (Figure 1A). We therefore predicted that other plaque components, particularly collagen, increased over time, given that human atherosclerosis plaque stabilization is pictured to involve both macrophage depletion and increased fibrosity (e.g., see ²³ for a recent review). Indeed, we found a significant increase in collagen content in plaques after reversal of the hyperlipidemia (Figure 5B). As shown in Figure 5C, there is mainly a diffuse mesh of collagen with a fibrous cap observed under the regression conditions.

Figure 5. Correction of the hyperlipidemia in Reversa mice has little effect on the size of plaques, but increases their collagen content.

(A) Plaque areas at baseline (day 0) and at various time points after treating the mice with pIpC or normal saline, as measured by morphometric analysis of hematoxylin and eosin–stained sections. Statistical analysis of the means in the pIpC group was based on N=13 for each time point; * indicates p<0.01 for the comparison between day 0 and any other time point; ns indicates no statistical difference among the other time points. (B) Quantification of collagen content in plaques at baseline (day 0) or 14 days after saline or pIpC treatment, as judged by morphometric analysis of Sirius Red–stained sections examined under polarizing light. Statistical analysis of the means was based on N=11 for each group; * indicates p<0.0001 for the comparison of the pIpC group to either the baseline (day 0) or saline treatment group. (C) Representative section showing a diffuse intimal mesh and a sub-endothelial collection of collagen under regression conditions from a pIpC-treated mouse used for the analysis shown in panel B. “L” and “M” indicate the lumen and the medial layer of the artery, respectively.

Lipid lowering promoted decreased inflammation in plaque CD68+ cells, an effect enhanced by PPARγ activation

Increased expression of inflammatory genes in macrophages has been shown to be associated with plaque progression (reviewed in ²⁴). We have previously determined that the expression of 2 inflammation-related genes, VCAM-1 and MCP-1, are decreased in plaque CD68+ cells in the transplant model of regression ⁴. We found consistent findings in the Reversa mouse model. As shown in Figure 6, transcripts for VCAM-1 and MCP-1, as well as ICAM-1 and TNFα, were reduced in pIpC-treated Reversa mice at day 14.

Figure 6. Correction of the hyperlipidemia in Reversa mice is associated with reduced expression of inflammatory markers in plaque CD68+ cells.

CD68+ cells in plaques were obtained by laser-capture microdissetion 14 days after baseline. Levels of gene expression were compared in CD68+ cells from pIpC- and saline–treated Reversa mice. Total RNA was isolated, and transcript levels of VCAM-1, ICAM-1, MCP-1, and TNFα were measured by qRT-PCR. Data were from two pools of RNA, each from three mice, and are expressed as fold change compared with samples obtained from mice at baseline. *p < 0.05.

We also examined gene-expression markers for M2 macrophages. Whereas M1 macrophages are activated cells that produce mainly pro-inflammatory cytokines, such as tumor necrosis factor α (TNFα), IL-6, and IL-12 ^{8, 25}, M2 macrophages (“alternatively activated”) dampen inflammatory responses by producing anti-inflammatory factors such as IL-10, transforming growth factor β (TGF-β), and IL-1 receptor antagonist (IL-1Ra) ⁸. Interestingly, markers of M2 macrophages (Arginase I, Mannose Receptor, CD163, C-lectin, Fizz-1) in CD68+ cells within plaques were significantly increased in Reversa mice after lipid lowering (Figure 7). These markers were further increased in mice treated during the regression phase with pioglitazone (Figure 7). Although the effects of pioglitazone on gene-expression were unequivocal, we were unable to detect any additional effect of the drug on the content of CD68+ cells within plaques (Figure 8A).

Figure 7. M2 macrophage markers are enriched in CD68+ cells in the plaques of Reversa mice under regression conditions, with further enhancement by PPARγ activation.

Mice were treated as in Figure 1 and CD68+ cells were harvested by laser-capture microdissection and their RNA isolated as in Figure 6. qRT-PCR was then used to measure the levels of the indicated M2 markers. Note that in pIpC-treated mice, the expression of these markers were higher than in baseline mice, with further enhancement found with pIpC and PPARγ activator pioglitazone co-treatment. *p < 0.05 compared with baseline; **p < 0.05 compared with pIpC. Data are from three pools of RNA, each from three mice, and are expressed as fold change compared with samples obtained from mice at baseline.

Figure 8. PPARγ activation does not increase the rate of CD68+ cell depletion, but further reduces PAI-1 expression, in Reversa mouse plaques after hyperlipidemia was corrected.

(A) Reversa mice were treated with pIpC to reverse the hyperlipidemia as in Figure 1; one-half of the mice received pioglitazone in their diet. Plaque CD68+ cell content was measured by morphometric analysis of sections stained by immunohistochemistry. n = 8/group. (B) Reversa mice were treated as in Figure 1 and CD68+ cells were obtained by laser-capture microdissection. The level of PAI-1 transcripts was measured by qRT-PCR and normalized to cyclophilin A. *p < 0.05 compared to baseline mice. **p < 0.05 compared with the mice treated with pIpC alone. Data are from three pools of RNA, each from three mice.

The expression of plasminogen activator inhibitor-1 (PAI-1) is reduced in regressing plaques and this effect is enhanced by PPARγ activation

One marker of the atherothrombotic state is an increased plasma level of plasminogen activator inhibitor-1 (PAI-1) ²⁶, which is expressed by macrophages, as well as other cell types. During regression, PAI-1 gene expression in plaque CD68+ cells decreased, with a further reduction observed in the pioglitazone group (Figure 8B). The latter result is consistent with a report that PPARγ agonists reduce PAI-1 gene expression ²⁷.

Discussion

The underlying cause of hyperlipidemia in the Reversa mouse is LDL receptor deficiency, resulting in elevated plasma LDL levels, the statistically strongest risk factor for atherosclerotic disease in people. It is also this lipoprotein fraction that is most affected by the conditional ablation of the Mttp gene. That the Reversa mouse is a clinically relevant model is suggested, then, by its characteristics mimicking both the common finding of elevated plasma LDL in hyperlipidemic patients and the treatment strategy to lower these levels in order to reduce the risk of atherosclerotic disease. We previously showed that correction of the hyperlipidemia in Reversa mice early in life retarded the development of atherosclerosis ⁶. In the current study, we show that reversal of hyperlipidemia after atherosclerosis develops to an advanced stage leads to lower levels of CD68+ cells (which are mainly monocyte-derived macrophages and foam cells) in plaques. There were also phenotypic changes in plaque CD68+ cells, including 1) evidence for migratory behavior in vivo; 2) decreased expression of genes encoding inflammatory and prothrombotic factors; and 3) enrichment in markers of the M2 macrophage state. The activation of PPARγ by pioglitazone treatment did not further decrease the plaque CD68+ cell content, but led to other presumably beneficial effects, such as an enhancement of the M2 state and a further reduction in PAI-1 expression levels. Furthermore, the regression process was characterized by reduced plaque lipid and increased content of collagen.

In earlier work, by transplanting aortas from apoe^−/− mice into wild-type mice, the reversal of dyslipidemia also reduced plaque CD68+ cell content ^{1, 3, 4, 28}. As in the present study, this reduction was accompanied by increased CCR7 mRNA in lesional CD68+ cells and by evidence of their migration out of plaques. We have now extended these results to show that this emigration was independent of the regression model and the subset of circulating monocytes from which the plaque CD68+ cells were derived. There also may have been decreased ongoing recruitment of monocytes into plaques after lipid reduction, but this was not examined in the present study. The slower kinetics in the depletion of the CD68+ cells in the Reversa mice may relate to the more gradual course of lipid reduction or the reduced HDL levels after inactivating Mttp in the liver ⁶.

A probable link between lipid lowering and CD68+ cell emigration is suggested by the structure of the mouse and human CCR7 promoters. Both contain potential sterol response elements (SREs), which are functional in vitro and promote chemotaxis of macrophages in response to the CCR7 ligands CCL19 or 21 (J.E. Feig, L. Shang, M.J. Garabedian, E.A. Fisher, manuscript submitted). The finding of reduced Oil-red-O staining raises the possibility that the “regression environment” within the arterial wall might be associated with CCR7 gene expression through its SRE. This would be consistent with our finding increased nuclear localization of SREBP2 in regressing plaques (J.E. Feig, L. Shang, M.J. Garabedian, E.A. Fisher, manuscript submitted). This regulation is also consistent with a recent report that loading of THP-1 human monocytes with oxidized LDL suppresses CCR7 expression ²⁹.

The finding that correcting hyperlipidemia also led to an enrichment of markers of M2 macrophages in plaque CD68+ cells was striking. Macrophage heterogeneity in human and mouse atherosclerotic plaques has been documented ¹⁷, but to our knowledge this is the first report of dynamic changes during regression. Interestingly, as shown in Supplemental Figure 4, we found that there is some co-localization to the same plaque area of markers of M1 and M2 macrophages, but also spatial distinction between the markers. In a recent study ³⁰ examining macrophage plasticity in progressing atherosclerotic lesions, the authors concluded that M2 cells can convert to the M1 state. The present results suggest that in addition to a possible conversion of the M1 to M2 state, another source of M2 cells in regressing plaques may be the continued recruitment of circulating monocytes ³ that become M2 cells, a possibility we are currently testing.

Classically activated M1 macrophages have long been recognized to be inflammatory ^31-33. In contrast, the M2 state is thought to represent an anti-inflammatory state ⁸ associated with tissue remodeling and repair, which certainly occur during plaque regression. Although the mechanisms for reduced inflammatory markers and increased M2 markers remain to be explored, these changes may be related to recent findings by others ³⁴, ³⁵. They found that increased macrophage sterol content led to heightened macrophage activation in vitro and in vivo. Moreover, this effect was reversed by efflux of sterols from cells. The reduced Oil-red-O staining of plaques in our studies may indicate that there was a sufficient loss of lipid from the plaques in the regression environment to not only activate CCR7 through its SRE, but also to have an anti-inflammatory effect on CD68+ cells.

That pioglitazone treatment increased the expression of the M2 markers is consistent with recent studies showing that PPARγ activation promotes the polarization of macrophages to an anti-inflammatory state ^{17, 36}. Although this would generally be considered to be beneficial, Tabas and colleagues have shown that PPARγ activation in Ldlr^−/− mice with advanced atherosclerosis had a negative impact on plaque macrophages by increasing cell death ³⁷. If those effects had occurred in the current study, we would have expected an increased rate of CD68+ cell loss from plaques as well as histological evidence of necrosis, but neither was observed. The two studies can be reconciled by recognizing that the effects of the lipid burden of the cells, since once they become foam cells, PPARγ activation can no longer skew them towards the M2 phenotype ¹⁷. Thus, in the present study, it is likely that the reversal of the hyperlipidemia and the consequent decrease in plaque lipid content allowed for the M2-polarizing activity of PPARγ. Another potentially protective benefit of PPARγ activation was the reduction of PAI-1 expression in CD68+ cells, consistent with the ability of PPARγ to repress PAI-1 gene transcription ²⁷.

After the correction of the hyperlipidemia, the plaques of Reversa mice had an increased content of collagen. This finding is consistent with observations from the transplantation model, where regression was accompanied by increased connective tissue in plaques ². We suspect that these observations are likely due to decreased matrix degradation because we have observed reduced MMP-2 and MMP-9 mRNA levels in CD68+ cells after reversal of hyperlipidemia (J.E. Feig, E.A. Fisher, unpublished). Additionally, there may have been some stimulation of collagen synthesis related to the increase in arginase I expression ³⁸.

Before closing, we should also note that the reduction in the plasma level of LDL particles is clinically achieved usually by increasing their clearance with statins. New strategies based on decreasing hepatic apoB-lipoprotein production akin to the strategy taken in the Reversa mouse, however, have been shown to be effective hyoplipidemic therapies in clinical trials (e.g., ^{39, 40}). Though end point studies are not next available, based on the present results and the analyses of clinical trial data showing event reduction independent of the means of LDL lowering (e.g., ⁴¹), it is reasonable to expect benefits on plaques in patients treated with the novel agents.

In summary, we have introduced a new model of atherosclerosis regression in which identified were significant changes in the composition of plaques after the correction of hyperlipidemia, including reduced numbers of CD68+ cells and increased collagen content. Besides quantitative changes in numbers of CD68+ cells, there were also effects on their gene expression, with evidence for the induction of migratory machinery, enrichment in the M2 state, and reduced expression of inflammatory and atherothrombotic genes. Pioglitazone treatment enhanced these changes, implying additive clinical benefits of aggressive lipid management and PPARγ stimulation. Continued study of this model should lead to an improved understanding of plaque regression at the molecular level.

Perspective for the practising physician.

The ultimate cure for atherosclerosis would be the regression of arterial plaques. Discovery research towards this goal has been hampered by limited and sometimes cumbersome animal models. The Reversa mouse combines a standard model of human atherosclerosis, the hyperlipidemic LDL receptor-deficient mouse, with a genetic switch that electively shuts off LDL production. In the present study, arterial plaques were allowed to develop in Reversa mice to a stage mimicking advanced human coronary artery disease, then the elevated LDL was severely reduced, thereby simulating aggressive lipid management. The major findings after such lipid reduction were decreases in the content and inflammatory state of the central cell of plaques, macrophages, with the change in total plaque size more modest because of compensatory increases in collagen content. The improvement in macrophage inflammatory status was augmented by treatment with pioglitazone, consistent with the effects of PPAR gamma agonists on macrophages in vitro. The results may explain why plaque volume decreases have been modest in recent statin trials, despite significant reduction in events, as well as provide one basis for the cardioprotective effects of pioglitazone in clinical studies. Continued study of this convenient model should lead to an improved understanding of plaque regression at the molecular level.