The Endoplasmic Reticulum Stress-C/EBP Homologous Protein Pathway-Mediated Apoptosis in Macrophages Contributes to the Instability of Atherosclerotic Plaques

Mouse Models of Atherosclerotic Plaque Rupture

To generate atherosclerotic mice, male Chop^+/+/Apoe^−/− mice or Chop^−/−/Apoe^−/− mice, at the age of 13 weeks, were fed a high-cholesterol diet for 5 weeks. To generate the mouse models of atherosclerotic plaque rupture,²² 9-week-old male Chop^+/+/Apoe^−/− mice or Chop^−/−/Apoe^−/− mice fed normal chow were anesthetized and then their common carotid artery was ligated just proximal to the bifurcation. After the ligation of the common carotid artery or a sham operation, the mice were fed a high-cholesterol diet throughout the experiment. Four weeks after ligation, a polyethylene cuff (length, 2 mm) was placed around the common carotid artery, just proximal to the ligated site. Four days later, mice were euthanized and the lesions were analyzed. ER stress-activated indicator (ERAI) transgenic mice, which express green fluorescent protein (GFP) under ER stress conditions, were generated as previously described.²³

Hematoxylin-Eosin and Oil Red O Staining, and Immunohistochemistry

Anesthetized mice were perfused with PBS through the left cardiac ventricle and then tissues were fixed by perfusion with 4% paraformaldehyde containing PBS. After fixation, the common carotid artery was excised, embedded in optimal cutting temperature compound, and frozen in liquid nitrogen after removal of the cuff. Frozen (6-μm-thick) sections were cut and air dried. The precise methods for staining are described in the supplemental data.

Detection of Apoptosis and RT-PCR Analysis

Apoptosis was detected using a mitochondrial membrane potential–indicating dye (DePsipher; Trevigen Inc, Gaithersburg, Md) or an in situ apoptosis detection kit (Takara, Otsu, Japan).

Total RNA isolation and preparation, cDNA synthesis, and RT-PCR were performed as previously described.⁷

Measurement of the ER Stress Response in Living Cells

The ER stress response in living macrophages derived from ERAI mice was detected by measuring GFP expression, as previously described.²³

Quantitative results were expressed as mean±SEM and analyzed using the Mann-Whitney test or the t test.

CHOP Is Involved in the Rupture of Atherosclerotic Plaques

Male Apoe knockout (Chop^+/+/Apoe^−/−) mice and Chop and Apoe double-knockout (Chop^−/−/Apoe^−/−) mice were fed a high-cholesterol diet (Figure 1). There were no significant differences in body weight, plasma glucose level, or serum lipid concentration between the 2 groups of mice (supplemental Figure I). Overall, atherosclerosis in the aorta was estimated by oil red O staining (Figure 1A). The number of atherosclerotic lesions tended to be higher in the Chop^+/+/Apoe^−/− mice compared with the Chop^−/−/Apoe^−/− mice (Figure 1B), as previously reported,²⁰ but the difference was not significant. On the other hand, as shown in Figure 1C and D, the formation of atherosclerotic lesions in the brachiocephalic artery was significantly suppressed in the Chop^−/−/Apoe^−/− mice compared with the Chop^+/+/Apoe^−/− mice. As shown in Figure 1A, atherosclerotic lesions were advanced in the brachiocephalic artery in this model; advanced atherosclerotic lesions are also often observed in this area in human clinical cases. These results suggest that the ER stress-CHOP pathway plays a role in atherosclerosis but that its contribution to the early stages is limited. However, in advanced lesions, CHOP appears to play a major role (Figure 1C and D). The results in Figure 1 are in accordance with the previously reported results in the human clinical cases⁵ and in a mouse model.^{20,21,25–27}

• Download figure
• Open in new tab
• Download powerpoint

Figure 1. Comparison of the atherosclerotic lesion area in the aortae of Chop^+/+/Apoe^−/− and Chop^−/−/Apoe^−/− mice. A, Atherosclerosis mouse models were generated using male Chop^+/+/Apoe^−/− mice and Chop^−/−/Apoe^−/− mice, which were fed a high-cholesterol diet for 5 weeks. Aortae were excised from wild-type mice fed normal chow, and the aortae from wild-type and experimental mice were stained with oil red O. The bar represents 200 μm. B, The percentages of the positive area of oil red O staining in A were calculated and are presented as mean±SEM (n=4). NS indicates no significant difference between the 2 groups. C, Hematoxylin-eosin staining and oil red O staining of the brachiocephalic artery from representative mice from both models are shown. The bar represents 100 μm. D, The atherosclerotic area of the lesion in C was calculated, and the areas are shown as the mean±SEM (for each lesion, 3 sections from 10 mice were examined for calculation.). **P<0.01 between the 2 groups.

As shown in Figure 2A, CHOP is induced in macrophages that have invaded the atherosclerotic lesion. Pronounced macrophage infiltration was also detected in Chop^−/−/Apoe^−/− mice. Therefore, the invasion of macrophages into the atherosclerotic lesion is CHOP independent. The activation of caspase-3 (indicated by cleavage of the protein) can be used as a marker of apoptosis. Thus, we examined the expression of the caspase-3 active form to detect apoptotic cells in atherosclerotic lesions. Figure 2B shows that a large part of macrophages in the plaques of Chop^+/+/Apoe^−/− mice is undergoing apoptosis. In contrast, apoptotic cells were minimal in the Chop^−/−/Apoe^−/− mice, although there were many macrophage-derived cells. These results suggest that ER stress-CHOP–mediated apoptosis is induced in these infiltrating macrophages during advanced atherosclerosis.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2. ER stress-induced transcription factor CHOP and CHOP-dependent activation of caspase-3 are induced in advanced atherosclerotic lesions. A, Male Chop^+/+/Apoe^−/− mice (upper panels) and Chop^−/−/Apoe^−/− mice (lower panels) at the age of 8 weeks were fed a high-cholesterol diet for 5 weeks. The atherosclerotic lesions of the brachiocephalic artery were double immunostained with macrophage-specific monocyte/macrophage-2 (MOMA2) and anti-CHOP antibodies. Immunostaining for MOMA2 and CHOP was performed in the same field as the atherosclerosis plaque, and their merged images are shown. The bar indicates 30 μm. B, The atherosclerotic lesions of the brachiocephalic artery from both types of mice were double immunostained with macrophage-specific MOMA2 and anti–caspase-3 active form antibodies. Immunostaining for MOMA2 and the active form of caspase-3 in the same field as the atherosclerotic plaque, and their merged images, are shown. The bar represents 30 μm.

Next, we directly examined whether activation of the ER stress-CHOP pathway contributes to the instability of atherosclerotic plaques using an atherosclerotic plaque rupture model generated by the ligation and cuff placement method (Figure 3A, supplemental Figure II, and supplemental Figure III).²² The circumferential atherosclerotic lesions were formed in both the Chop^+/+/Apoe^−/− and Chop^−/−/Apoe^−/− mice. However, as shown in Figure 3B, the ratio of ruptured lesions was dramatically higher in the Chop^+/+/Apoe^−/− mice compared with the Chop^−/−/Apoe^−/− mice. Ruptures were observed in more than 60% of Chop^+/+/Apoe^−/− mice. In contrast, ruptures were detected in only approximately 15% of Chop^−/−/Apoe^−/− mice. In addition, the formation of a fibrous cap in the brachiocephalic artery was enhanced in Chop^−/−/Apoe^−/− mice compared with Chop^+/+/Apoe^−/− mice (supplemental Figure III). An immunohistochemical analysis showed that most of the infiltrated macrophage-derived cells present in the plaque express CHOP in the Chop^+/+/Apoe^−/− mice (Figure 3D). Because CHOP mediates ER stress-induced apoptosis, we examined whether apoptosis in the infiltrating macrophages is suppressed in Chop^−/−/Apoe^−/− mice compared with Chop^+/+/Apoe^−/− mice using TUNEL staining (Figure 3C) and staining for the active form of caspase-3 (Figure 3E). Figure 3E shows that most of the macrophage-derived cells were apoptotic in the Chop^+/+/Apoe^−/− mice. In contrast, the number of apoptotic cells was obviously decreased in the Chop^−/−/Apoe^−/− mice. The number of TUNEL-positive cells was also significantly lower in the Chop^−/−/Apoe^−/− mice compared with the Chop^+/+/Apoe^−/− mice. These results suggest that ER stress-CHOP–mediated apoptosis in macrophages plays a major role in the instability of plaques. To directly examine the role of CHOP in macrophages, we generated mice lacking CHOP only in bone marrow–derived cells, and mice expressing CHOP only in bone marrow–derived cells, by bone marrow transplantation as described elsewhere²⁴ (Figure 3C and Figure 4A). As shown in Figure 4B, in the case of Apoe^−/− mice lacking CHOP in bone marrow–derived cells, the rupture of atherosclerotic plaques was significantly suppressed compared with Chop^+/+/ Apoe^−/− mice and Apoe^−/− mice lacking CHOP in non–bone marrow–derived cells. The number of TUNEL-positive cells was also significantly suppressed in the case of Apoe^−/− mice lacking CHOP in their bone marrow–derived cells (Figure 3C). In the case of Apoe^−/− mice lacking CHOP, except in their bone marrow–derived cells, neither rupture formation nor apoptosis was suppressed. These results show that the ER stress-CHOP pathway in macrophages plays a major role in the instability of atherosclerotic plaques and that this instability is due to the increased apoptosis of the infiltrating macrophages.

• Download figure
• Open in new tab
• Download powerpoint

Figure 3. CHOP and CHOP-dependent activation of caspase-3 are induced in ruptured atherosclerotic lesions. A, Schedule for generating the mouse plaque rupture model by the ligation and cuff placement method. B, Quantitative comparisons of rupture events of the common carotid artery between Chop^+/+/Apoe^−/− mice (8 sections each from 13 mice were examined.) and Chop^−/−/Apoe^−/− mice (3 sections each from 12 mice were examined) were performed. Data are shown as the mean±SEM. **P<0.01 between the 2 groups. C, Chop^+/+/Apoe^−/− and Chop^−/−/Apoe^−/− mice were treated to generate the mouse plaque rupture model, as in Figures 3 and 4. The generated plaque lesions of the common carotid artery from each mouse were stained using the TUNEL method to detect apoptotic cells and were stained with methyl green at the same time. TUNEL-positive apoptotic cells were calculated, and the percentage of positive cells per 1000 cells is shown as the mean±SEM. **P<0.01 between the 2 groups. D, The generated plaque lesions of the common carotid artery from both mouse models were double immunostained with macrophage-specific MOMA2 and anti-CHOP antibodies in the same field, and the merged images are shown. The bar represents 30 μm. E, The generated plaque lesions of the common carotid artery from both mouse models were double immunostained with macrophage-specific MOMA2 and anti–active caspase-3 antibodies in the same field, and the merged images are shown. The bar represents 30 μm. HCD indicates high-cholesterol diet.

• Download figure
• Open in new tab
• Download powerpoint

Figure 4. The expression of CHOP in bone marrow–derived cells plays a crucial role in the rupture of atherosclerotic plaques. A, Schedule of bone marrow transplantation, followed by the generation of the mouse plaque rupture model by ligation and cuff placement. B, Quantitative analysis of rupture events of the common carotid artery after the treatment shown in A (9 sections each from 3 mice were examined.). Genotypes of recipient mice and donor mice (in parenthesis) are shown on the bottom. Data are given as the mean±SEM. **P<0.01 between the 2 groups. HCD indicates high-cholesterol diet.

Free Cholesterol Accumulates in the ER of Macrophages and Induces ER Stress

Next, we prepared mouse peritoneal macrophages as previously described,¹⁴ to examine how the ER stress pathway is induced in macrophages. In macrophage-derived foam cells, oxidized LDL-derived cholesterol is esterified by acyl–coenzyme A and cholesterol acyltransferase (ACAT) and accumulates as esterified cholesterol in cytosolic droplets.^6,28,29 Esterified cholesterol is intoxic to the cells, although the cells are filled with numerous lipid-containing droplets. However, as the atherosclerotic lesions progress, the amount of unesterified free cholesterol increases.³⁰ The suppression of ACAT-mediated cholesterol esterification and/or activation of neutral cholesteryl esterase is induced during the progression of atherosclerosis.^6,29,30 Therefore, we directly examined whether unesterified free cholesterol in living macrophages induced the ER stress pathway (Figure 5). Peritoneal macrophages were prepared from wild-type and ERAI mice and used to model infiltrating macrophages, as previously reported.²⁰ Supplemental Figure IV shows that macrophages used in these experiments are infiltrated type as the foam cells in atherosclerotic lesions. In Figure 5A, the distribution of intracellular cholesterol was detected by oil red O staining in peritoneal macrophages. When the macrophages were treated with acetyl-LDL, cholesterol was detected in the peripheral area of the cells. In contrast, when cells were exposed to acetyl-LDL plus the ACAT inhibitor CI976, cholesterol was detected in the perinuclear area. These results clearly show that free cholesterol accumulated in different compartment(s) compared with esterified cholesterol in living cells. Free cholesterol was distributed in the ER but esterified cholesterol was not (Figure 5B). Then, we examined whether free cholesterol accumulation induced the ER stress pathway in vivo. Induction of the active form of X-box binding protein 1 (XBP1) is a hallmark of the ER stress response.^31–33 The cells derived from ERAI transgenic mice express a variant of GFP under XBP1-activated conditions and can be used to directly monitor ER stress in vivo.^23,34,35 Peritoneal macrophages from ERAI transgenic mice were treated with acetyl-LDL plus CI976 or acetyl-LDL alone, then the ER stress response was monitored by evaluating the intensity of GFP emission (Figure 5C and D). As shown in Figure 5D, the levels of ER stress response after treatment with acetyl-LDL plus CI976 were about twice compared with those after treatment with acetyl-LDL alone. These results show that the intracellular accumulation of free cholesterol strongly activates the ER stress response pathway. To our knowledge, this is the first case directly demonstrating the activation of the ER stress pathway by free cholesterol in living macrophages. Other ER stress markers, such as CHOP, Derlin(s), and ER degradation enhancer, mannosidase alpha-like 1 (EDEM), were also induced by free cholesterol accumulation (supplemental Figure VA and B).^36–39 Also, we showed that accumulation of free cholesterol induces apoptosis in peritoneal macrophages (supplemental Figure VC and D). Treatment with caspase inhibitor significantly suppressed free cholesterol–induced cell death in macrophages (supplemental Figure VIA and B). This result supports our conclusion that apoptosis is mediated by the accumulation of free cholesterol in infiltrating macrophages.

• Download figure
• Open in new tab
• Download powerpoint

Figure 5. Free cholesterol induces the ER stress pathway in primary cultured peritoneal macrophages. A, Mouse peritoneal macrophages were untreated or treated with 100-μg/mL acetyl (Ac)-LDL or 10-μmol/L ACAT-inhibitor CI976 or 100-μg/mL acetyl-LDL plus 10-μmol/L CI976 for 24 hours. Next, the cells were stained with oil red O to detect cholesterol deposition. The bar represents 50 μm. The inset in each panel shows a representative photograph of the distribution of intracellular cholesterol in macrophages stained by oil red O. The bar represents 10 μm. B, Mouse peritoneal macrophages were treated with 25-μg/mL DiI (1,1′-dioctadecyl-3,3,3′,3′-tetramethyl-indocarboeyanine perchlorate)-labeled acetyl-LDL alone (upper panels) or 25-μg/mL DiI-labeled acetyl-LDL plus 10-μmol/L CI976 (lower panels) for 48 hours. Then the cells were stained with ER-Tracker dye and observed with a fluorescence microscope. DiI images (left panels), ER-Tracker dye images (middle panels), and their merged images (right panels) of the same field are shown. The bar represents 15 μm. C, Peritoneal macrophages were prepared from ERAI-transgenic mice, then were either untreated or treated with 100-μg/mL acetyl-LDL plus 10-μmol/L CI976 for 24 hours, or 100 μg/mL acetyl-LDL for 24 hours, as indicated in the figure. Next, the expression of GFP was observed using fluorescence microscopy. The bar represents 30 μm. D, Peritoneal macrophages from ERAI-transgenic mice were treated with acetyl-LDL, 100 μg/mL, with CI976 (black bars) or without CI976 (white bars), 10 μmol/L, for the indicated times. The expression of GFP was calculated using a spectrofluorometer and was standardized to the induction levels of β-galactosidase. The results of 3 independent experiments were calculated, and the results of the cells treated with acetyl-LDL plus CI976 for 48 hours were set at 100; all data are shown as the mean±SEM (n=3). **P<0.01 between the 2 groups.

Free Cholesterol Accumulation Induces ER Stress-CHOP-Bax–Mediated Apoptosis in Peritoneal Macrophages

CHOP plays a central role in ER stress-mediated apoptosis.^{15–17,19,40,41} We showed that CHOP plays a crucial role in the induction of apoptosis by the accumulation of free cholesterol (supplemental Figure VI). Peritoneal macrophages from wild-type or Chop^−/− mice were treated with acetyl-LDL plus CI976 or acetyl-LDL alone. We evaluated apoptosis by using a dye indicating the mitochondrial membrane potential and lactate dehydrogenase release. In Chop^−/− cells, apoptosis induced by treatment with acetyl-LDL plus CI976 was significantly suppressed.

Because the activation and translocation of Bax, a proapoptotic member of the B-cell lymphoma (Bcl)-2 family, from the cytosol to the mitochondria is important in CHOP-mediated apoptosis,¹⁵ we next examined whether Bax was activated and translocated to the mitochondria in a CHOP-dependent manner using acetyl-LDL plus CI976-treated wild-type (Figure 6A) and Chop^−/− (Figure 6B) primary cultured peritoneal macrophages. The Bax antibody (N-20) predominantly recognizes the activated proapoptotic state of Bax, whereas the Bax antibody (B-9) reacts with Bax regardless of its conformation.¹⁵ Nuclear staining with Hoechst dye 33258 was also performed to detect nuclear apoptotic morphological changes. In untreated cells, the nucleus was finely stained with nuclear staining. When untreated macrophages from wild-type or Chop^−/− mice were immunostained with Bax antibodies, the cytoplasm was diffusely stained with the B-9 antibody, whereas immunofluorescence was weak with the N-20 antibody. In contrast, when wild-type macrophages were treated with acetyl-LDL plus CI976, apoptotic chromatin condensation was seen. (Cells are indicated by arrows in Figure 6A.) When immunostaining for Bax was performed in the acetyl-LDL plus CI976-treated wild-type cells using the B-9 or N-20 antibody, particulate structures were observed in the cytoplasm, indicating translocation of Bax to the mitochondria. Positive staining with the N-20 antibody after this treatment indicated that there was a conformation change in Bax. In contrast, when Chop^−/− macrophages were treated with acetyl-LDL plus CI976, apoptotic changes were barely detected on nuclear staining. Under these conditions, the cytoplasm of Chop^−/− cells was diffusely stained with the B-9 antibody and only weakly immunostained with the N-20 antibody. These results show that the translocation of Bax to the mitochondria and the induction of its conformation change by acetyl-LDL plus CI976 treatment are CHOP dependent. The activation and translocation of Bax were only minimally detected in CI976- or acetyl-LDL–treated cells at 24 hours. As shown in Figure 6 and supplemental Figure VIC, low levels of apoptosis induction were detected in the cells that had accumulated the esterified form of cholesterol for a long period. This apoptosis was CHOP independent. These results are consistent with those in Figure 3B. In Figure 3B, rupture formation was still detected in a few cases, despite the lack of CHOP expression. From these results, it can be deduced that at least a small part of macrophage apoptosis and plaque rupture is induced by the accumulation of esterified cholesterol in a CHOP-independent manner.

• Download figure
• Open in new tab
• Download powerpoint

Figure 6. Treatment with acetyl-LDL plus ACAT-inhibitor CI976 induces CHOP-dependent activation and translocation of Bax to the mitochondria of macrophages. A, Peritoneal macrophages from wild-type mice were untreated or were treated with acetyl-LDL, 100 μg/mL, plus CI976, 10 μmol/L, or CI976, 10 μmol/L, or acetyl (Ac)-LDL, 100 μg/mL, for the indicated times. The cells were then fixed, double immunostained with a monoclonal antibody against full-length Bax (Bax B-9), and a polyclonal antibody against the N-terminal portion of Bax (Bax N20), stained with Hoechst dye 33258 and observed by fluorescence microscopy. The arrows indicate nuclear condensation in the apoptotic cells. The bar represents 10 μm. B, Peritoneal macrophages from Chop^−/− mice were untreated or were treated with acetyl-LDL, 100 μg/mL, plus CI976, 10 μmol/L, or CI976, 10 μmol/L, or acetyl-LDL, 100 μg/mL, for the indicated times. The cells were then fixed and double immunostained with a monoclonal antibody against Bax B-9 and a polyclonal antibody against Bax N20, stained with Hoechst dye 33258, and observed by fluorescence microscopy. The bar represents 10 μm.