This study was designed to test the hypothesis that the therapeutic effects of the AMPK agonist AICAR against insulin resistance involve its anti-inflammatory function, which requires macrophage SIRT1. The plausibility of this hypothesis was driven by several prior findings on AMPK’s anti-inflammatory functions and AICAR’s beneficial effects on insulin resistance/type 2 diabetes. First, we and others have shown that AMPK plays an important role in the regulation of macrophage inflammation 
. Second, activation of AMPK by AICAR has been shown to improve insulin sensitivity and glucose homeostasis 
. It is not clear, however, whether the anti-inflammatory function of AMPK contributes to its insulin-sensitizing effects. We have previously shown that AMPK’s anti-inflammatory function depends on macrophage SIRT1 
. We therefore generated MSKO mice, whose insulin resistant phenotype should mainly originate from inflammation due to myeloid-specific deficiency of SIRT1. We then administrated AICAR to MSKO mice to test our hypothesis. We reasoned that if quenching inflammation is required for the full strength of AICAR to prevent insulin resistance, AICAR will not be effective in prevention of insulin resistance in MSKO mice, since AICAR likely fails to suppress the enhanced inflammation due to the absence of SIRT1. Indeed, we found that AICAR injection was able to suppress inflammation and reduce insulin resistance in control mice but not in MSKO mice. It has been known for years that AICAR has insulin-sensitizing effects 
. However, the pleiotropic effects of AICAR in multiple metabolic tissues have made it difficult to determine the contribution of AMPK’s anti-inflammatory capacity to its insulin-sensitizing effects. Instead, these beneficial effects are largely attributed to AMPK actions on glucose and lipid metabolism in skeletal muscle and liver 
. Our data clearly support the hypothesis that the full capacity of AICAR to reduce insulin resistance requires its inflammation-suppressing ability as an essential component, in addition to other beneficial effects including lipid and glucose metabolism.
Although in our study we used a low dose of AICAR that did not affect the mouse adiposity, a confounding factor often seen to affect insulin sensitivity, AICAR administration still likely exerted its insulin-sensitizing effects through direct regulation of energy metabolism in muscle and liver. In addition, AICAR may regulate inflammation independent of AMPK 
. Additional genetic studies, involving deletion of AMPK in specific tissues, are required to separate the systemic effects on energy metabolism from the direct effect of anti-inflammation, and to determine, to what extent, the contribution of AMPK’s anti-inflammation to its insulin-sensitizing functions. As such, the macrophage is a very good target tissue to address whether the anti-inflammatory effect of AMPK is required for its insulin sensitizing and glucose-reducing effects. Indeed, Steinberg’s group was the first to investigate the role of macrophage AMPK in regulation of obesity-induced inflammation and insulin resistance 
. They showed that genetic deletion of macrophage AMPK β1 subunit in hematopoietic cells including macrophages via bone marrow transplantation enhanced adipose tissue macrophage inflammation and insulin resistance 
. This study discovers the importance of macrophage AMPK in regulation of obesity-induced inflammation and insulin resistance.
During the course of characterizing the phenotypes of MSKO mice, Schug et al reported a similar mouse model of myeloid SIRT1 deletion 
. Our findings agree with the majority of the conclusions of that study, including the activated inflammation in macrophages with SIRT1 deficiency. However, we have explored the other potential mechanisms underlying the enhanced inflammation in SIRT1-deficient macrophages. Schug and colleagues attribute the enhanced inflammation in SIRT1-deficient macrophages mainly to the hyperacetylation of the NF-κB subunit p65 at lysine 310 (K310) 
. Although acetylation of p65 at lysine 310 is a logical target due to previous findings on the ability of SIRT1 to deacetylate this lysine site and suppress NF-κB transcriptional activity 
, we believe that the pro-inflammatory phenotype of SIRT1-deficient macrophages may not be solely explained by lysine 310 hyper-acetylation. In fact, the very first study targeting the role of lysine acetylation in the regulation of p65 functions revealed that acetylation of lysine 310 is required for full transcriptional activity of p65, but has no effects on DNA binding ability of p65 
. Both Schug’s and our ChIP assays indicate that macrophage SIRT1 deficiency increases p65 DNA binding to its consensus promoters 
, which may not be attributed to lysine 310 hyper-acetylation. Moreover, we found that SIRT1 deletion promotes iKKα/β phosphorylation, an upstream signal of p65 nuclear translocation, and also stimulates the phosphorylation of JNK, an inflammatory signal that parallels the iKK/NF-κB pathway. All these data suggest that SIRT1 deficiency alters the global inflammatory networks. We therefore explored the inflammatory pathways involving macrophage alternative activation, which has been known to regulate systemic inflammation and play important roles in the development of metabolic disorders 
. Macrophage SIRT1 may be involved in macrophage alternative activation. We found that SIRT1 expression is higher in anti-inflammatory M2 macrophages than pro-inflammatory M1 macrophages, and that SIRT1 deficiency coordinately stimulates M1 macrophage conversion and inhibits M2 macrophage alternative activation. As a result, myeloid deletion of SIRT1 increases infiltration of classically activated M1 macrophages and decreases alternatively activated M2 macrophage content in fat. On the other hand, ER stress has emerged as a key upstream signal that activates macrophage inflammatory networks, including both JNK and NF-κB 
. We found that SIRT1 deficiency elevated the total protein and phosphorylation of IRE1α (Fig. S9
), a key ER stress sensor, in response to saturated fatty acid stearate (C18) and thapsigargin (Tg) (Fig. S8A and S8B
middle), two potent inducers of macrophage ER stress 
. In sum, our data demonstrate that the altered macrophage polarization and probably ER stress pathways may contribute to the pro-inflammatory phenotype featuring activated systemic inflammatory networks in SIRT1-deficeint macrophages. Although the activated NF-κB pathway (through p65 hyperacetylation at K310) itself may partially explain the M1 macrophage tendency in SIRT1-deficient macrophages, further studies will be required to address how SIRT1 regulates macrophage polarization and ER stress pathways.
AMPK and SIRT1 share striking similarities in nutrient sensing and regulation of energy metabolism. Recent studies have disclosed a crosstalk between these two in regulation of metabolic pathways. For instance, AMPK can be an upstream signal to increase SIRT1 activity via inducing fatty acid oxidation and increasing the agonist NAD+
levels, leading to the deacetylation and activation of PGC-1α in muscle 
. On the other hand, SIRT1 can also be in driving position to activate AMPK via deacetylating and activating LKB1, the upstream kinase of AMPK 
. No matter which one is the upstream or downstream signal between the two, AMPK and SIRT1 are coordinately regulated and cooperate to regulate downstream pathways. It appears that AMPK and SIRT1 can activate each other and feed off ensuing signaling between them. Which one is the upstream or downstream signal may depend on different types of cells or biological pathways. In regulation of the macrophage inflammation, we previously found that AMPK antagonizes inflammation through SIRT1 by increasing the SIRT1 activator NAD+
. Interestingly, Galic et al demonstrated a key role of fatty acid oxidation in mediating AMPK inhibition of macrophage inflammation 
. Fatty acid oxidation appears to be a novel mechanism underlying nutrient-induced inflammation in macrophages and fits well into the scenario where AMPK interacts with SIRT1 to regulate macrophage inflammation, because increasing fatty acid oxidation may enhance cellular NAD+
content, which may further activate SIRT1. In the present study, we also found that myeloid SIRT1 may serve as the downstream signal that mediates the anti-inflammatory of the AMPK agonist AICAR in vivo
. It is likely that activation of AMPK may induce fatty acid oxidation and increase cellular NAD+
, which further lead to activation of SIRT1.
In summary, our data demonstrate that AICAR treatment decreases adipose macrophage inflammation, thereby contributing to the protection against HF diet-induced insulin resistance. In contrast, myeloid deletion of SIRT1 increases macrophage inflammation through promotion of pro-inflammatory M1 polarization coupled with inhibition of M2 activation. The beneficial effects of AICAR in antagonizing inflammation and insulin resistance require myeloid SIRT1. We conclude that myeloid SIRT1 is a therapeutic target of the anti-inflammatory and insulin-sensitizing effects of AICAR. The anti-inflammatory property of AMPK and SIRT1 may contribute to their beneficial effects in antagonizing obesity-induced insulin resistance.