AMP-activated protein kinase (cid:2) 1 promotes atherogenesis by increasing monocyte-to-macrophage differentiation

Monocyte-to-macrophage differentiation, which can be initi-ated by physiological or atherogenic factors, is a pivotal process inatherogenesis,adisorderinwhichmonocytesadheretoendo-thelial cells and subsequently migrate into the subendothelial spaces, where they differentiate into macrophages and macro-phage-derived foam cells and cause atherosclerotic lesions. However, the monocyte-differentiation signaling pathways that are activated by atherogenic factors are poorly defined. Here we report that the AMP-activated protein kinase (cid:2) 1 (AMPK (cid:2) 1) in monocytes promotes atherosclerosis by increasing monocyte differentiation and survival. Exposure of monocytes to oxidized low-density lipoprotein, 7-ketocholesterol, phorbol 12-myris-tate 13-acetate, or macrophage colony-stimulated factor (M-CSF) significantly activated AMPK and promoted monocyte-to-macrophage differentiation. M-CSF-activated AMPK is via M-CSFreceptor-dependentreactiveoxygenspeciesproduction. Consistently, genetic deletion of AMPK (cid:2) 1 or pharmacological inhibition of AMPK blunted monocyte-to-macrophage differentiation and promoted monocyte/macrophage apoptosis. Compared with E knock-out ( ApoE which show impaired clearing of plasma lipoproteins and spon-taneously develop atherosclerosis, ApoE (cid:3) AMPK showed reduced sizes of atherosclerotic lesions and lesser numbers of macrophages in the Furthermore, aortic lesions were decreased in ApoE (cid:3) / mice transplanted with ApoE (cid:2) 1 (cid:3) / (cid:3) bone marrow and in myeloid-specific AMPK (cid:2) 1-deficient ApoE (cid:3) / mice. rapamycin treatment, which abolished delayed monocyte differentiation in ApoE AMPK (cid:2) 1 (cid:3) / (cid:3) mice, lost its atherosclerosis-lowering effects in these mice. Mechanistically, we found that AMPK (cid:2) 1 expression of both LC3 and ULK1, which are two important autophagy-related markers. Rapamycin treatment increased FoxO3 activity as well as LC3 and ULK1 expressions in macrophages from AMPK (cid:2) 1 (cid:3) / (cid:3) mice. Our results reveal that and

reveal that AMPK␣1 deficiency impairs autophagy-mediated monocyte differentiation and decreases monocyte/macrophage survival, which attenuates atherosclerosis in ApoE ؊/؊ mice in vivo.
Peripheral monocytes circulate in the blood for 24 -48 h and undergo spontaneous apoptosis without appropriate stimulation (1)(2)(3). Atherogenic factors promote the adherence of monocytes to endothelial cells and their subsequent migration into the subendothelial spaces, where they are differentiated and transformed into macrophages and macrophage-derived foam cells (4,5). Thus, monocyte-to-macrophage differentiation is a pivotal process in atherogenesis (6 -8). Several factors, including oxidized low-density lipoprotein (ox-LDL) 3 particles, macrophage colony-stimulating factor (M-CSF), or granulocyte-macrophage colony-stimulating factor, play essential roles in this process (9 -11). In mice lacking both M-CSF and apolipoprotein E (apoE), atherosclerotic lesions are smaller than those in apoE-deficient mice (ApoE Ϫ/Ϫ ), probably due to the nearly complete lack of monocytes in the peripheral blood and impairments in monocyte-to-macrophage differentiation (7,8). M-CSF or ox-LDL has been known to activate phosphoinositide-3 kinase and mitogen-activated protein kinase pathways during differentiation (12,13). Recent studies also show that the induction of autophagy prevents monocytes from apoptosis and is critical for M-CSF-induced monocyte-to-macrophage differentiation (14,15). However, the precise molecular mechanisms by which they induce monocyte survival and differentiation are not fully understood.
AMP-activated protein kinase (AMPK) is an evolutionarily conserved serine/threonine kinase that consists of one catalytic subunit (␣) and two regulatory subunits (␤ and ␥) (16,17). There are two ␣ isoforms (AMPK␣1 and -␣2), which are differentially expressed in different tissues. The primary function of AMPK is to act as the energy and redox sensor that is involved in metabolic regulation and insulin sensitivity (16,17). Recent evidence also indicates novel roles for AMPK in the pathogen-esis of cardiovascular diseases. AMPK␣2 deletion increases atherosclerosis in ApoE Ϫ/Ϫ mice, probably via enhanced oxidative stress and endoplasmic reticulum stress (18). A recent study shows that AMPK activators metformin and AICAR significantly attenuated PMA-induced monocyte-to-macrophage differentiation in THP-1 cell and proinflammatory cytokine production (19). However, whether the effects of metformin or AICAR are AMPK-dependent is not clear because compound C, an inhibitor of AMPK, was ineffective in promoting monocyte-to-macrophage differentiation in the absence of PMA (19). The aim of the present study was to define the roles of AMPK␣1, the predominant isoform expressed by monocytes/ macrophages, in monocyte-to-macrophage differentiation and atherosclerosis. Our findings have revealed that the genetic deletion of AMPK␣1 suppressed autophagy and autophagymediated monocyte differentiation, resulting in increased monocytic cell death, all of which contributed to the reduction of atherosclerotic lesions in vivo.

AMPK␣1 inhibition impairs M-CSF-or ox-LDL-induced monocyte-to-macrophage differentiation in vitro
First, we detected the effects of AMPK inhibition on M-CSFinduced differentiation in human monocytes in vitro. As shown in Fig. 1A, M-CSF significantly increased CD71 expression, an indicator of monocyte-to-macrophage differentiation, whereas it decreased monocyte marker CD14 expression after 1 and 3 days of treatment. However, AMPK inhibition by compound C significantly suppressed M-CSF-induced CD71 expression and maintained CD14 expression at a higher level.
To examine whether AMPK is involved in ox-LDL-induced monocyte differentiation, THP-1 cells were treated with phorbol 12-myristate 13-acetate (PMA), ox-LDL, or 7-KC with or without AMPK inhibitor compound C. Ox-LDL or 7-KC significantly enhanced PMA-induced expressions of CD36 and CD11b, which are macrophage-differentiation markers. Meanwhile, compound C dramatically blocked the effects of ox-LDL and 7-KC on CD36 and CD11b expressions (Fig. 1, B and C).

AMPK␣1 deletion slows down monocyte-to-macrophage differentiation in vivo
Thioglycollate broth injection into the peritoneal cavity causes the recruitment of circulating monocytes to the peritoneum, where they transform into tissue macrophages within 4 days (20). This approach has been used to study monocyte differentiation in vivo (15). To detect whether or not AMPK␣1 deletion affects monocyte-to-macrophage differentiation in vivo, peritoneal Ly6C hi F4/80 lo monocytes and Ly6C lo F4/80 hi macrophages in CD11b hi CD90 lo B220 lo CD49b lo NK1.1 lo Ly-6G lo mononuclear cells (21) were sorted 1-4 days after thioglycollate administration. As shown in Fig. 1D, 1 day after thioglycollate injection, 60 -70% of cells in the peritoneal cavity were Ly6C hi F4/80 lo monocytes, whereas 20 -30% were Ly6C lo F4/ 80 hi mature macrophages, which are resident macrophages in the peritoneal cavity. The percentages of Ly6C hi F4/80 lo monocytes and Ly6C lo F4/80 hi macrophages at day 1 were comparable in WT and AMPK␣1 Ϫ/Ϫ mice. However, the percentages of Ly6C hi F4/80 lo monocytes at days 2 and day 3 were significantly higher, whereas the percentage of Ly6C lo F4/80 hi macrophages from day 2 to day 4 in the peritoneum of AMPK␣1 Ϫ/Ϫ mice were significantly lower than their WT counterparts (Fig. 1D).

AMPK is activated during monocyte differentiation
To test whether AMPK is activated during monocyte differentiation, phosphorylation of AMPK at threonine 172 was detected in bone marrow cells treated with M-CSF. As shown in Fig. 1, F and G, M-CSF significantly increased AMPK phosphorylation in both peripheral blood monocytes and isolated peritoneal monocytes/macrophages from ApoE Ϫ/Ϫ mice. M-CSF also largely enhanced AMPK phosphorylation in human monocytes (Fig. 1H) and AMPK activity in THP-1 cells (Fig. 1I). In addition, PMA, ox-LDL, and 7-KC all significantly increased AMPK phosphorylation in THP-1 cells, and 7-KC further increased PMA-induced AMPK phosphorylation (Fig. 1J).

M-CSF activates AMPK via M-CSF receptor-mediated reactive oxygen species (ROS) production
To explore whether M-CSF promoted AMPK activation through its receptor CD115, THP-1 cells were transfected with CD115 siRNA for 48 h. As shown in Fig. 1K, knockdown of CD115 significantly inhibited M-CSF-induced AMPK activation, implying that AMPK activation by M-CSF is CD115-dependent. It has been known that M-CSF stimulation results in the production of ROS (22,23). We confirmed that ROS levels were increased in M-CSF-treated peritoneal monocytes/ macrophages (Fig. 1L) and THP1 cells (Fig. 1M). Because ROS signaling is important in AMPK activation under various conditions (24), we further tested whether M-CSF activates AMPK via ROS production. M-CSF-treated peritoneal monocytes were incubated with or without ROS scavenger, 4-hydroxy-2,2,6,6-tetramethylpiperidine-N-oxyl (TEMPOL). As shown in Fig. 1N, TEMPOL significantly blocked the activation of AMPK induced by M-CSF.
To confirm that AMPK inhibition increases monocyte/ macrophage apoptosis during differentiation in humans, monocytes were isolated from human peripheral blood and incubated with compound C (5-10 M) in the presence of 40 ng/ml M-CSF for 1 or 3 days. As shown in Fig. 2F, inhibition of AMPK by compound C dramatically increased cleaved caspase 3 expression. In contrast, overexpression of AMPK␣1 in macrophages significantly inhibited 7-KC-, 25-OH cholesterol-, or staurosporine-induced apoptosis in peritoneal macrophages (Fig. 2G).
To investigate whether AMPK␣1 deletion impaired monocyte differentiation and whether increased apoptosis has impacts on macrophage numbers, peritoneal cells were isolated and counted 1-4 days after thioglycollate injection. As shown in Fig. 2H, peritoneal cell numbers in ApoE Ϫ/Ϫ /AMPK␣1 Ϫ/Ϫ mice are dramatically decreased at days 3 and 4 compared with those in ApoE Ϫ/Ϫ mice. Similar results were observed in AMPK␣1 Ϫ/Ϫ and WT mice injected with thioglycollate. (Fig.  2I).

Deletion of AMPK␣1 does not affect the functions of mature macrophages
Next, we determined whether AMPK␣1 deletion alters the functions of macrophages. The phagocytic ability of bone marrow-derived macrophages was detected by measuring the uptake of pHrodo red Escherichia coli bioparticles. As shown in Fig. 3A, the numbers of internalized bioparticles by bone marrow-derived macrophages were similar between ApoE Ϫ/Ϫ / AMPK␣1 Ϫ/Ϫ mice and ApoE Ϫ/Ϫ mice.

rus into bone marrow cells after 48 h of M-CSF treatment.
The amount of punctate GFP-LC3 structures in ApoE Ϫ/Ϫ / AMPK␣1 Ϫ/Ϫ cells was significantly decreased compared with those in ApoE Ϫ/Ϫ cells (Fig. 4D). Moreover, inhibition of AMPK with compound C significantly decreased LC3-I/II expression during human monocyte-to-macrophage differentiation (Fig. 4E). Because autophagy is a highly dynamic process, the decreased LC3 expression or the number of GFP-LC3 puncta could be due to either inhibiting autophagy induction or accelerating fusion of autophagosomes with lysosomes. To further explore the role of AMPK␣1 in autophagy flux, bone marrow-derived macrophages from WT and AMPK␣1 Ϫ/Ϫ mice were treated with or without bafilomycin A1 (5 nM), which blocks the fusion of autophagosomes with lysosomes. As shown in Fig. 4F, bafilomycin A1 increased LC3 accumulation both in WT and AMPK␣1 Ϫ/Ϫ macrophages. However, the total LC3 expressions were still lower in AMPK␣1 Ϫ/Ϫ macrophages after bafilomycin A1 treatment, suggesting that AMPK␣1 may regulate LC3 protein expression from the transcriptional level. We confirmed this result in THP1 cells treated with AMPK␣1 siRNA and bafilomycin A1 (Fig. 4G). To determine whether AMPK regulates LC3 at the transcriptional level, mRNA expression of LC3 was examined using real-time RT-PCR in THP1 cells treated with or without AMPK␣1 siRNA. As shown in Fig. 4H, the mRNA level of LC3 was significantly decreased in cells transfected with AMPK␣1 siRNA. Interestingly, another important autophagy-related kinase, UNC-51-like kinase-1 (ULK1) was also significantly reduced during M-CSF-induced bone marrow differentiation in AMPK␣1 Ϫ/Ϫ mice ( Fig. 4I). Silence of AMPK␣1 by siRNA in THP1 cells decreased both protein level and mRNA level of ULK1 expression (Fig. 4, J and K).

Forkhead box protein O3 (FoxO3) regulates AMPK␣1-dependent autophagy induction during monocyte-to-macrophage differentiation
To identify the signaling pathways responsible for regulation of autophagy-related genes by AMPK␣1, we focused on the role of FoxO3, which is known to be directly regulated by AMPK (27) and controls the transcription of autophagy-related genes, including LC3 and ULK1 (28,29). First, we detected the protein expression of FoxO3a in bone marrow-derived macrophages after 5 days of M-CSF treatment or peritoneal macrophages in WT or AMPK␣1 Ϫ/Ϫ mice. As shown in Fig. 5, A and B, FoxO3a expression dramatically decreased in AMPK␣1 Ϫ/Ϫ macrophages. We further confirmed this finding in THP1 cells transfected with AMPK␣1 siRNA (Fig. 5C). To determine whether AMPK␣1 regulates FoxO3 at the transcriptional level, mRNA expression of FoxO3 was detected in THP1 cells transfected with AMPK␣1 siRNA. Silence of AMPK␣1 significantly decreased FoxO3 mRNA expression (Fig. 5D). On the other hand, overexpression of AMPK␣1 in HEK293T cells significantly increased luciferase activity of the FoxO3 promoter (Fig. 5E).
To demonstrate that FoxO3 drives LC3 gene expression, mRNA and protein levels of LC3 were detected in THP1 cells transfected with control siRNA or FoxO3a siRNA. Silence of FoxO3a significantly suppressed LC3 mRNA and protein expression (Fig. 5, F and G). In addition, inhibition of FoxO3a by siRNA also suppressed ULK1 mRNA and protein expressions (Fig. 5, H and I).

Deletion of AMPK␣1 suppresses atherosclerotic lesion formation in vivo
To further investigate the effects of AMPK␣1 deletion on atherogenesis, age-and gender-matched ApoE Ϫ/Ϫ /AMPK␣1 Ϫ/Ϫ and ApoE Ϫ/Ϫ mice were fed a Western diet for 10 weeks. As depicted in Fig. 7, A and B, the areas of aortic lesions in both aortic roots and aortic arches were significantly lower in ApoE Ϫ/Ϫ / AMPK␣1 Ϫ/Ϫ mice than those in ApoE Ϫ/Ϫ mice, indicating that AMPK␣1 deletion reduces aortic lesions. Consistently, the necrosis core in the aortic root lesion was significantly smaller in ApoE Ϫ/Ϫ /AMPK␣1 Ϫ/Ϫ than that in ApoE Ϫ/Ϫ mice (Fig. 7C).
Next, we assayed the macrophage contents in aortic root lesions by staining of F4/80, a macrophage-specific marker. As shown in Fig. 7E, the positive area of F4/80 staining is significantly decreased in ApoE Ϫ/Ϫ /AMPK␣1 Ϫ/Ϫ mice when compared with their ApoE Ϫ/Ϫ counterparts, suggesting that the reduction of macrophages may contribute to smaller aortic lesions in ApoE Ϫ/Ϫ /AMPK␣1 Ϫ/Ϫ mice.

Decreased macrophage numbers and increased macrophage apoptosis without change in macrophage proliferation in the aortic lesion of ApoE ؊/؊ /AMPK␣1 ؊/؊ mice
We further determined the number of monocytes and macrophages in the aortas of ApoE Ϫ/Ϫ /AMPK␣1 Ϫ/Ϫ and ApoE Ϫ/Ϫ mice that were fed a Western diet for 10 weeks. There was no change in total Ly6C hi F4/80 lo monocyte numbers between the two groups, but Ly6C lo F4/80 hi macrophage numbers in the aortas of ApoE Ϫ/Ϫ /AMPK␣1 Ϫ/Ϫ mice were significantly decreased (Fig. 7I), suggesting that the reduced atherosclerotic lesion size in ApoE Ϫ/Ϫ /AMPK␣1 Ϫ/Ϫ mice might be due to the lesser number of differentiated macrophages in intima. In addition, apoptotic macrophages were significantly increased in the lesions of ApoE Ϫ/Ϫ /AMPK␣1 Ϫ/Ϫ mice (Fig.  7J), as detected by TUNEL and CD68 co-staining. Further, deletion of AMPK␣1 showed no effect on macrophage proliferation in lesions by analyzing positive immunostaining for proliferation marker Ki67 and macrophage marker Moma2 (Fig. 7K).

Discussion
In the present study, we have for the first time demonstrated that AMPK␣1 activation promotes monocyte-to-macrophage differentiation and survival by promoting autophagy. We fur-ther demonstrate that AMPK␣1 controls autophagy by regulating the FoxO3a transcription factor activity, which is involved in the regulation of autophagy-related genes. Consistently, deletion of AMPK␣1 reduced plaque growth and intimal macrophage accumulation in the early lesions of ApoE Ϫ/Ϫ mice. Results from bone marrow transplantation experiments and myeloid specific knock-out of AMPK␣1 further suggest that myeloid AMPK␣1 plays an important role in atherosclerotic formation. Further, aortic lesions in both aortic roots and aortic arches are significantly lower in ApoE Ϫ/Ϫ /AMPK␣1 Ϫ/Ϫ mice than those in ApoE Ϫ/Ϫ mice. Importantly, rapamycin treatment, which normalized monocytes differentiation and macrophage survival in ApoE Ϫ/Ϫ /AMPK␣1 Ϫ/Ϫ mice, ablated its atherosclerosis-suppressing effects in ApoE Ϫ/Ϫ mice. Overall, our results suggest that AMPK␣1-dependent and autophagy-mediated monocyte differentiation and survival promote the initiation and progression of atherosclerosis in The induction of autophagy is known to be critical for M-CSF-induced monocyte-to-macrophage differentiation (14,15). There are numerous studies demonstrating the role for AMPK in autophagy induction in response to various cellular stresses (26,36,37). M-CSF and its receptor, CD115, are critical for monocyte-to-macrophage differentiation and cell survival (38 -40). However, the molecular mechanism underlying how AMPK regulates autophagy in monocyte differentiation has not yet been fully understood. In this study, we show that ox-LDL, 7-KC, and M-CSF increased the conversion of LC3-II, which indicates the induction of autophagy. Consistently, we found that M-CSF, PMA, and ox-LDL, all of which promote the differentiation of monocytes into macrophages, effectively activated AMPK in macrophages. Furthermore, we demonstrate that M-CSF-activated AMPK is via M-CSF receptor-dependent ROS production. Importantly, deletion of AMPK␣1 significantly suppressed LC3 and ULK1 mRNA and protein expressions via FoxO3 transcription factor. Autophagy activation with rapamycin improved monocyte differentiation in AMPK␣1 Ϫ/Ϫ mice, confirming that defected autophagy is responsible for defected differentiation and survival observed in AMPK␣1 Ϫ/Ϫ mice. Our findings provide convincing evidence to support a  Table 1 Body weight, lipid profiles, blood glucose, and blood pressures in Western diet-fed ApoE ؊/؊ and ApoE ؊/؊ /AMPK␣1 ؊/؊ mice Body weight, n ϭ 9; glucose, cholesterol, triglyceride, n ϭ 8 -12; systolic and diastolic blood pressure, n ϭ 6. BP, blood pressure.

Parameter
ApoE ؊/؊ ApoE ؊/؊ /AMPK␣1 ؊/؊ promoting role of AMPK␣1 in controlling autophagy induction in monocyte-to-macrophage differentiation and survival. A recent study reported that metformin or AICAR inhibits monocyte differentiation into macrophage and therefore attenuates Ang-II-induced atheromatous plaque formation (19). Although the authors showed that metformin or AICAR activated AMPK, it could not exclude the AMPK-independent effects of metformin or AICAR on monocyte differentiation. In fact, we found that metformin protected against hyperglycemia-induced atherosclerosis through AMPK␣2 activation, which has no effect on monocyte differentiation (41). In contrast with a reduced atherosclerosis observed in this study, Cao et al. (42) reported that myeloid AMPK␣1-deleted LDL receptor knock-out mice (Ldlr Ϫ/Ϫ ) increased formation of atherosclerotic plaque with enhanced macrophage inflammation and higher plasma triglyceride and cholesterol content. Although ApoE Ϫ/Ϫ and Ldlr Ϫ/Ϫ mice have been used extensively as two mouse models of atherogenesis, Ldlr Ϫ/Ϫ mice have much lower plasma cholesterol levels and develop less severe atherosclerotic lesions than ApoE Ϫ/Ϫ mice on a normal chow diet (43,44). Therefore, the elevation of lipid level by deletion of myeloid AMPK␣1 in Ldlr Ϫ/Ϫ mice may be a major contributor to the manifestation of atherosclerosis in this mouse model. Indeed, pharmacological AMPK activator S17834 has been reported to attenuate hyperlipidemia and suppress aortic atherosclerosis in insulin-resistant Ldlr Ϫ/Ϫ mice in part through suppression of SREBP-1c-and SREBP-2-dependent lipogenesis (45). However, in our models, AMPK␣1 deletion or myeloid AMPK␣1 deletion in ApoE Ϫ/Ϫ background did not alter lipid levels. Therefore, the impaired monocyte differentiation and increased monocyte/macrophage apoptosis due to AMPK␣1 deletion may play the key roles in alleviation of atherosclerosis in an ApoE knock-out mouse model. Overall, our results indicate that AMPK␣1 activation during monocyte differentiation via its promotion of monocyte differentiation and survival increases the numbers of macrophage, which promotes the initiation and progression of atherosclerosis in ApoE Ϫ/Ϫ mice. Why AMPK␣1 deletion has opposite impacts in plasma lipid levels in ApoE Ϫ/Ϫ mice versus LDLr Ϫ/Ϫ mice is unknown and warrants further investigation.
A growing body of evidence indicates that the mTOR inhibitor rapamycin has pleiotropic anti-atherosclerotic effects. Administration of rapamycin systemically has been demonstrated to prevent the development of atherosclerosis in different animal models (31)(32)(33)(34)(35). Although the exact mechanisms accounting for the beneficial effects of mTOR inhibition on atherosclerosis are not yet clear, the potential beneficial effects of rapamycin and its derivatives include inhibition of smooth muscle cell proliferation (46), reduced MCP-1 expression within the injured arterial wall (34), and inhibition of lipid accumulation in macrophages and smooth muscle cells (47,48). Our current results also show that although rapamycin slightly promoted monocyte differentiation, the net effect of rapamycin treatment was anti-atherosclerotic in ApoE Ϫ/Ϫ mice, suggesting that multiple mechanisms may be involved in the regulation of atherosclerosis development. However, in ApoE Ϫ/Ϫ / AMPK␣1 Ϫ/Ϫ mice, the percentage increase in monocyte differentiation into macrophage was significantly higher than that of ApoE Ϫ/Ϫ mice when compared with vehicle treatment, and rapamycin treatment showed no effect on plaque size in ApoE Ϫ/Ϫ / AMPK␣1 Ϫ/Ϫ mice, implying that improved differentiation by rapamycin in ApoE Ϫ/Ϫ /AMPK␣1 Ϫ/Ϫ mice may counteract the other anti-atherosclerosis effects of rapamycin.
In summary, we demonstrate that AMPK␣1 plays a key role in modulating monocyte-to-macrophage differentiation and survival through FoxO3-dependent autophagy induction by regulating expression of genes involved in autophagy. AMPK␣1 activation accelerates monocyte differentiation and survival, which participates in the initiation and progression of atherosclerosis.

Animals
The animal protocol and procedures were reviewed and approved by the University of Oklahoma Institute Animal Care and Use Committee.

Generation of ApoE and AMPK␣1 double-knock-out mice
AMPK␣1 Ϫ/Ϫ mice that had been generated and backcrossed onto a C57BL/6J background, as described previously (49), were crossed with ApoE Ϫ/Ϫ mice (Jackson Laboratory) to generate ApoE Ϫ/Ϫ /AMPK␣1 Ϫ/Ϫ mice. Mice were housed in temperature-controlled cages under a 12-h light-dark cycle and given free access to water and chow. Accelerated atherosclerosis was induced by feeding the mice with a Western diet containing 0.21% cholesterol and 21% fat (Research Diets Inc., D12079B). This diet was administered at 8 weeks of male mice and continued for 10 weeks. In some experiments, mice were fed with either a microencapsulated rapamycin-containing Western diet, which contains 14 mg/kg food of rapamycin providing a dose of ϳ2.24 mg of rapamycin/kg of body weight/day (50,51), or a control Western diet (with empty microcapsules) for 10 weeks. Microencapsulated rapamycin and empty microcapsules were purchased from Rapamycin Holdings Inc. (San Antonio, TX) and incorporated into the Western diet by Research Diets Inc.

Monocyte isolation and monocyte-to-macrophage differentiation in vitro
Human peripheral blood mononuclear cells were separated by Ficoll density gradient centrifugation from venous blood of healthy consenting volunteers, after approval from the ethical committee of the University of Oklahoma Health Sciences Center. Lymphocyte/monocyte-rich layer was washed with PBS and then suspended in DMEM supplemented with 2 mM L-glutamine, 100 units/ml penicillin/streptomycin, and 10% FBS for 2 h at 37°C and. Non-adherent cells were then removed, and the adherent cells were cultured in DMEM supplemented with 2 mM L-glutamine, 100 units/ml penicillin/streptomycin, 10% FBS serum, and 40 ng/ml human M-CSF in the presence or absence of AMPK inhibitor compound C for the indicated time.
Blood from 10 -12 weeks old ApoE Ϫ/Ϫ /AMPK␣1 Ϫ/Ϫ and ApoE Ϫ/Ϫ mice was collected with heparin, and red blood cells were removed by ammonium chloride lysis. The monocytes were then enriched using the immunomagnetic EasySep mouse monocyte enrichment kit (STEMCELL Technologies Inc.) according to the manufacturer's instructions. The purity of CD11b ϩ Ly6C ϩ cells was assessed by flow cytometry and ranged from 75 to 85%.
THP-1 cells were cultured in RPMI 1640 with 10% FBS. The cells were differentiated with PMA, ox-LDL, or PMA and ox-LDL combinations with or without compound C for 3 days.

Bone marrow transplantation
Eight-week-old male ApoE Ϫ/Ϫ and ApoE Ϫ/Ϫ /AMPK␣1 Ϫ/Ϫ mice were subjected to 11-gray lethal total-body irradiation (two doses of 5.5 grays within an interval of 4 h) to eliminate endogenous bone marrow stem cells and bone marrow-derived cells. Mice were retro-orbitally injected with ApoE Ϫ/Ϫ or ApoE Ϫ/Ϫ /AMPK␣1 Ϫ/Ϫ bone marrow cells (5 ϫ 10 6 cells). Four weeks after transplantation, the mice were placed on the Western diet for 8 weeks. The hematologic chimerism of transplanted ApoE Ϫ/Ϫ mice was confirmed in genomic DNA from bone marrow by PCR analysis.

Atherosclerotic lesion analysis
After being fed the Western diet for 10 weeks, the mice were fasted for 14 h and then anesthetized and euthanized. The heart and aortic tissues were removed from the ascending aorta to the ileal bifurcation and placed in 4% paraformaldehyde for 24 -48 h. After fixation, the adventitia was thoroughly cleaned under a dissecting microscope. To analyze the lesion area in the aortic root, the heart was dissected from the aorta, embedded in OCT compound, and sectioned (8-m thickness). Four serial cryosections were collected from each mouse and stained with Oil Red O for neutral lipids and then counterstained with hematoxylin to visualize nuclei. Images of plaques were captured under the Olympus microscope, which was connected to a QImaging Retiga CCD camera, and quantitative analysis was performed with ImageJ software (National Institute of Health) by averaging the lesion areas in the four sections. To analyze the lesion area in the aortic arch, the intimal surface was exposed by a longitudinal cut from the ascending arch to an area that was 5 mm distal to the left subclavian artery. This allowed the lumen of the aortic arch to be laid flat. The aorta was rinsed for 5 min in 75% ethanol, stained with 0.5% Sudan IV in 35% ethanol and 50% acetone for 15 min, destained in 75% ethanol for 5 min, and then rinsed with PBS. Digital images of the aorta were captured under a stereomicroscope, and the lesion area was quantified from the aortic arch to the area that was 5 mm distal to the left subclavian.

Macrophage staining and macrophage proliferation in situ
Staining of macrophages in aortic root lesion was performed with rat anti-mouse F4/80 from Abcam. Apoptotic macrophages in atherosclerotic lesions were detected by using the DeadEnd TM fluorometric TUNEL system (Promega) and rat anti-mouse CD68 primary antibody (AbD Serotec) followed by goat anti-rat IgG-conjugated Alexa Fluor 594.
To detect in situ macrophage proliferation, sections of aortic root were labeled with a primary rat anti-mouse MOMA-2 antibody (AbD Serotec) and a rabbit anti-mouse antibody, Ki-67 (Abcam), detected by Alexa Fluor 488 goat anti-rat IgG antibody and Alexa Fluor 594 donkey anti-rabbit IgG antibody (Life Technologies, Inc.), respectively. Images were recorded using a Zeiss LSM 710 inverted confocal microscope.

Determination of serum cholesterol, triglyceride, and blood glucose levels
Blood glucose levels were determined by applying tail blood to a OneTouch Ultra blood glucose monitoring system (LifeScan). Serum cholesterol and triglyceride levels were measured enzymatically using Infinity reagents from Thermo Scientific, according to the manufacturer's instructions.

Mouse peritoneal macrophages and bone marrow-derived macrophages
ApoE Ϫ/Ϫ /AMPK␣1 Ϫ/Ϫ or ApoE Ϫ/Ϫ mice (8 -10 weeks old) were intraperitoneally injected with 1.5 ml of 4% BBL thioglycollate brewer (BD Biosciences). After 4 days, peritoneal macrophages were collected from peritoneal exudates. Cells were cultured in DMEM (Corning Cellgro), which was supplemented with 2 mM L-glutamine, 100 units/ml penicillin, 0.1 mg/ml streptomycin, and 10% FBS for 2 h. Non-adherent cells were removed by washing with PBS, and adherent cells were taken as peritoneal macrophages. To obtain bone marrow-derived macrophages, mouse femurs and tibias were isolated, and both ends of the bones were cut with scissors. The cells from the bone marrow were flushed with 10 ml of DMEM. Cells were seeded at a density of 2 ϫ 10 5 marrow cells/cm 2 in 6-well culture plates in DMEM containing 10% FBS, penicillin, streptomycin, 2 mM L-glutamine, and 40 ng/ml mouse M-CSF (eBioscience). After 3 days, the medium was replaced with fresh medium. After another 6 days in culture, differentiated macrophages were used for various experiments in DMEM with 10% FBS.

Phagocytosis assay
The phagocytic activity of macrophages was measured by the uptake of red fluorescent pHrodo E. coli bioparticles (Invitrogen). Briefly, 5 ϫ 10 6 macrophages were suspended in 100 l of Hanks' balanced salt solution containing 20 mM HEPES, pH 7.4, and mixed with 20 l of pHrodo E. coli bioparticles. The mixture was incubated for 30 min at either 37°C for uptake activity or 0°C for background activity.

Chemotaxis assay
Macrophage chemotaxis assays were performed using 8-m Corning transwells. Bone marrow-derived macrophages from ApoE Ϫ/Ϫ /AMPK␣1 Ϫ/Ϫ or ApoE Ϫ/Ϫ mice were placed in the upper wells of the chamber, and the lower wells contained 100 ng/ml MCP-1. Following incubation at 37°C for 16 h, cells remaining on the upper surface of the filter were removed mechanically. Migrated cells were stained with hematoxylin and then counted manually under the microscope.

Binding of ox-LDL to macrophages
Peritoneal macrophages or bone marrow-derived macrophages were plated on 12-well plates at a concentration of 2.5 ϫ 10 5 cells/well. The cells were incubated with 10 g/ml DiI-oxLDL (Kalen Biomedical) on ice for 1 h and washed three times with ice-cold PBS. DiI-oxLDL binding was assayed by either fluorescence microscopy or flow cytometry using excitation and emission wavelengths of 514 and 565 nm, respectively.

Foam cell formation assay
Peritoneal macrophages from ApoE Ϫ/Ϫ /AMPK␣1 Ϫ/Ϫ or ApoE Ϫ/Ϫ mice were incubated with either 50 g/ml acetylated-LDL or 50 g/ml ox-LDL for 48 h. After incubation, the cells were washed, fixed with 4% paraformaldehyde, stained with Oil Red O to detect intracellular neutral lipids, counterstained with hematoxylin, and photographed under the microscope.

Flow cytometry
Alexa 488-conjugated anti-human CD14, phycoerythrinconjugated anti-human CD71, Alexa 488-conjugated antimouse CD36, and phycoerythrin-conjugated anti-mouse/ human CD11b antibodies were purchased from Biolegend. Fluorescence intensity was quantified by flow cytometry using a FACSCalibur cytometer (BD Biosciences), and data were analyzed by Summit software (Beckman Coulter). To quantify apoptosis in peritoneal macrophages in the in vivo differentiation study, 10 5 cells were rinsed with PBS, incubated with Annexin V-FITC/propidium iodide at room temperature for 30 min, and then counted by flow cytometry. To measure intracellular reactive ROS, cells were loaded with 20 M 5-(and-6)-chloromethyl-2Ј,7Ј-dichlorodihydrofluorescein diacetate, acetyl ester (H2DCFDA, Molecular Probes) in the dark for 30 min at 37°C in a 5% CO 2 atmosphere. The cells were then washed twice, and intracellular fluorescent products were measured immediately by flow cytometry with 450 -490-nm excitation and 505-520-nm emission.

Measurement of cytokines
Serum levels of TNF-␣, IL-1␤, IL-6, MCP-1, and M-CSF in ApoE Ϫ/Ϫ /AMPK␣1 Ϫ/Ϫ and ApoE Ϫ/Ϫ mice, which were fed a Western diet for 10 weeks, were measured using the Millipore multiplex kit by the Bio-Plex 200 system (Bio-Rad). MCP-1 levels in the peritoneal lavage and serum after 4 h of thioglycollate injection were measured by the enzyme-linked immunosorbant assay kit (BioLegend), following the manufacturer's protocol.

Western-blotting analysis
Cell lysates or tissue homogenates were subjected to Western-blotting analysis, as described previously (52).

AMPK activity assay
THP-1 cells were treated with 40 ng/ml M-CSF for the indicated times, and endogenous AMPK was immunoprecipitated using the AMPK antibody. AMPK activity was assayed using the SAMS peptide, as described previously (53).

Luciferase assay
HEK293T cells were co-transfected with pGL3-enhancer FoxO3a promoter vector (FHRE-Luc was a gift from Dr. Michael Greenberg (Addgene plasmid 1789) in combination with the pCMV-AMPK␣1 plasmid (Origene, catalog no. RC218572) or empty plasmid. Renilla luciferase vector was used as an internal control. After 24 h of transfection, the cells were harvested and analyzed for firefly luciferase and Renilla luciferase activity using the Dual-Luciferase reporter assay (Promega) as per the manufacturer's protocol. Data were expressed as the -fold increase over mock-transfected cells. HEK293T cells were transfected with pGL3-enhancer FoxO3a promoter vector and Renilla luciferase vector for 24 h. The cells were then treated with or without 500 nM rapamycin for 4 h. Foxp3 promoter activity was measured as described above.
Real-time quantification was performed by SYBR Green (Bio-Rad) with the C1000 thermal cycler, CFX96 detection system (Bio-Rad). Relative gene expression was normalized to 18S rRNA and compared using the ⌬⌬Ct method.

Statistical analysis
Data are reported as mean Ϯ S.E. Statistical comparisons were performed with the unpaired Student's t test to compare two groups or one-way analysis of variance to compare three or more groups. The Bonferroni multiple-comparison test was used for post hoc analysis. p Ͻ 0.05 was considered significant.
Author contributions-M. Z., H. Z., and Y. D. designed and performed the experiments and analyzed the data with the help of all authors, wrote the manuscript, and acquired partial funding. Z. L. and Z. C. helped with the planning and performance of some experiments. M.-H. Z. designed and supervised the project, revised the manuscript, and acquired funding.