Macrophage AMPK β1 activation by PF-06409577 reduces the inflammatory response, cholesterol synthesis, and atherosclerosis in mice

Summary Atherosclerotic cardiovascular disease is characterized by both chronic low-grade inflammation and dyslipidemia. The AMP-activated protein kinase (AMPK) inhibits cholesterol synthesis and dampens inflammation but whether pharmacological activation reduces atherosclerosis is equivocal. In the current study, we found that the orally bioavailable and highly selective activator of AMPKβ1 complexes, PF-06409577, reduced atherosclerosis in two mouse models in a myeloid-derived AMPKβ1 dependent manner, suggesting a critical role for macrophages. In bone marrow-derived macrophages (BMDMs), PF-06409577 dose dependently activated AMPK as indicated by increased phosphorylation of downstream substrates ULK1 and acetyl-CoA carboxylase (ACC), which are important for autophagy and fatty acid oxidation/de novo lipogenesis, respectively. Treatment of BMDMs with PF-06409577 suppressed fatty acid and cholesterol synthesis and transcripts related to the inflammatory response while increasing transcripts important for autophagy through AMPKβ1. These data indicate that pharmacologically targeting macrophage AMPKβ1 may be a promising strategy for reducing atherosclerosis.


INTRODUCTION
Atherosclerotic cardiovascular disease is one of the leading causes of death worldwide.Risk factors for atherosclerotic disease include high levels of low-density lipoprotein (LDL) cholesterol, and triglyceride rich lipoproteins (TGRL). 1,2In addition to hyperlipidemia, inflammation, including both the innate and adaptive immune system, are fundamental drivers of atherosclerosis. 3,4Atherosclerotic plaques are rich in both cholesterol and immune cells, specifically macrophages, resulting in the accumulation of lipid laden macrophages (also referred to as foam cells) within the intima of large blood vessels.These macrophage rich atherosclerotic plaques can be further characterized by lipid accumulation, inflammatory cytokine release, recruitment of additional immune cells, cellular death (apoptosis or necrosis) and fibrosis within the arterial walls.
The AMP-activated protein kinase (AMPK) is a ubiquitously expressed heterotrimeric protein that consists of an a, b, and g subunit.2][13][14][15] AMPK is activated by a wide range of small molecules and xenobiotics that disrupt mitochondrial function leading to increases in AMP/ADP (i.e., berberine, metformin, and canagliflozin) or mimic the effects of AMP/ADP (i.e., AICAR) (for more details see reviews by Day et al. 16 and Steinberg et al. 17 ).][23][24] Over the last decade, several small molecules which directly bind to the regulatory AMPKb1 isoform and allosterically activate the kinase without disrupting mitochondrial function have been developed for treating metabolic diseases including type 2 diabetes, metabolic ll OPEN ACCESS associated fatty liver disease (MAFLD, formerly known as NAFLD) and autosomal dominant polycystic kidney disease (ADPKD) (for more details see reviews by Day et al. 16 and Steinberg et al. 17 ).The b1 subunit of AMPK is highly expressed in the liver, adipose tissue, and immune cells.AMPK b1 is critical for suppressing lipid synthesis in the hepatocytes and macrophage inflammation and increasing fatty acid oxidation in both hepatocytes and macrophages, 5,21,[25][26][27] suggesting that targeting this subunit may be beneficial for cardiometabolic diseases such as SLD and atherosclerosis.
PF-06409577 is an orally bioavailable, indole carboxylic acid, that selectively binds to the allosteric drug and metabolite (AdaM) site of the AMPK b1 isoform, leading to allosteric activation with an EC50 of approximately 3.3 nM with limited off target effects. 28,29In contrast to previous b1 selective AMPK activators such as A769662, PF-06409577 is more potent and has more favorable pharmacokinetic properties for oral administration, in rats, dogs, and monkeys. 28Specifically, dose response studies in mice indicated that maximal inhibition of liver de novo lipogenesis occurred with oral delivery in methylcellulose at 100 mg/kg, a dose which corresponded with serum concentrations of $8 nM, and is consistent with the EC50 for maximally activating AMPKb1 complexes in cell free assays. 29And while previous studies have shown that treatment of mice and non-human primates with PF-06409577 lowers liver steatosis and serum cholesterol 29 the effects of targeting the AMPK b1 subunit with PF-06409577 in macrophages and mouse models of atherosclerosis is currently unknown.
In the current study we find that PF-06409577 reduces atherosclerosis in ApoE À/À and PCSK9 over-expression mouse models through a pathway requiring the AMPK b1 isoform.Surprisingly, these reductions in atherosclerosis were not associated with changes in hepatic or serum lipid levels.Instead, we found that PF-06409577 induced reductions in atherosclerosis required AMPK b1 in myeloid cells and this dose dependently activated AMPK in bone marrow-derived macrophages (BMDMs) reducing cholesterol and fatty acid synthesis, lowering the expression of inflammatory response and I-kappaB kinase/NF-kB regulated transcripts and increasing the expression of transcripts important for autophagy.These data suggest that activating AMPK b1 complexes in macrophages may be important for reducing atherosclerosis.

RESULTS
PF-064095777 reduces atherosclerosis via AMPK b1 in ApoE À/À mice independently of reductions in circulating triglycerides or cholesterol Whole body ApoE À/À or ApoE À/À AMPKb1 À/À mice were fed a Western diet and treated daily with PF-06409577.PF-06409577 did not alter body mass, adiposity, lean mass, or insulin sensitivity in ApoE À/À or ApoE À/À AMPKb1 À/À mice over the 6-week treatment period (Figures S1A-S1E) while having a very minor effect on improving glucose tolerance (Figures S1F and S1G).Despite similar adiposity and insulin sensitivity, PF-06409577 reduced atherosclerotic plaques size by $ 45% in the aortic root of ApoE À/À but not ApoE À/À AMPKb1 À/À mice (Figures 1A and 1B).Similarly, although not significant, PF-06409577 treatment tended to reduce necrotic area within the plaques of ApoE À/À but not ApoE À/À AMPKb1 À/À (Figure 1C).In contrast to observations in hyperlipidemic rats or non-human primates where PF-06409577 reduced serum triglycerides and cholesterol, surprisingly there were no effects on these parameters in ApoE null mice (Figures 1D and 1E).Consistent with the lack of liver steatosis and inflammation in ApoE mice there was no change in liver lipids or the mRNA expression of markers of liver monocyte or macrophage infiltration (F4/80) or inflammatory cytokines (Tnfa or Il1b) following treatment with PF-06409577 in either genotype (Figures S2A-S2E).These data indicate that PF-06409577 reduces atherosclerosis in ApoE null mice through a mechanism requiring AMPKb1 but is independent of reductions in serum cholesterol in this model.

PF-06409577 induced reductions in atherosclerosis involves myeloid AMPKb1
Given serum cholesterol levels were unchanged in the ApoE À/À mice, we hypothesized that PF-06409577 may reduce atherosclerosis by activating AMPKb1 complexes within immune cells, including macrophages.To test this hypothesis, we crossed mice with Cre-recombinase expression driven by the LysM promoter to mice with the AMPKb1 allele floxed (hereafter referred to as AMPKb1 fl/fl or AMPKb1 LysM ) to generate myeloid specific AMPK b1 À/À mice.Isolation of peritoneal macrophages showed an $80% reduction in AMPKb1 with no change in AMPKb2 expression and a greater than 50% reduction in phosphorylation of acetyl-CoA carboxylase (ACC) (Figures 2A-2D), indicating significant reductions in AMPK activity in the macrophage compartment.
To induce atherosclerosis in AMPKb1 fl/fl or AMPKb1 LysM mice, we injected a gain-of-function PCSK9 AAV via the tail vein to over-express PCSK9 in the liver, which leads to reductions in the liver LDLr and increases in serum cholesterol when mice are fed a Western diet. 24,30,31As anticipated, injection with the PCSK9 AAV resulted in a 20-fold increase in serum PCSK9 in both AMPKb1 fl/fl or AMPKb1 LysM mice (Figure S3A).Daily oral gavage with PF-06409577 resulted in a subtle reduction in body weight in both AMPKb1 fl/fl and AMPKb1 LysM mice compared to vehicle controls (Figure S3B).Treatment with PF-06409577 did not alter serum or liver cholesterol and exerted only modest effects on serum and liver triglycerides (Figures S3C-S3F).Despite similar serum and liver lipid profiles, PF-06409577 reduced atherosclerotic plaque size in AMPKb1 fl/fl but not AMPKb1 LysM mice (Figures 2E and 2F).Consistent with previous studies, 5 AMPKb1 LysM mice had smaller atherosclerotic plaques than WT controls (Figures 2E and 2F).These data indicate that PF-06409577 reduces atherosclerosis in mice injected with a PCSK9 AAV and that AMPKb1 in myeloid cells may be important for mediating this effect.
Treatment of macrophages with PF-06409577 activates AMPK and markers of autophagy while reducing fatty acid and cholesterol synthesis and genes related to the inflammatory response and I-kappaB kinase/NF-kB To evaluate the molecular pathways being affected in macrophages we isolated BMDM from WT and AMPKb1 null mice and conducted a dose response with PF-06409577.PF-06409577 increased the phosphorylation of downstream targets of AMPK, ACC, and ULK1, following 1.5 h of treatment at 1 mM, 3 mM, and 10 mM doses in WT but not AMPK b1 À/À macrophages (Figures 3A-3C).These effects are consistent with previous dose responses with PF-06409577 in murine, rat, monkey, and human hepatocytes. 29Additionally, when compared to A-769662 at 100 mM, the first generation AMPKb1 selective activator, 24,32,33 there was a greater increase in phosphorylation of ACC and ULK1 with just 10 mM of PF-06409577, indicating much greater potency for activating AMPK.
RNA sequencing in WT and AMPKb1 À/À BMDMs indicated that PF-06409577 was highly specific for AMPKb1, with distinct clustering found by principal component analysis (PCA) analysis (Figure 4A).Further, examination of differentially regulated genes found that in WT mice PF-06409577 resulted in upregulation of 879 genes and downregulation of 674 genes with minimal changes to gene expression in AMPKb1 À/À macrophages (Figures 4B and 4C).When examining the top 50 most differentially expressed genes (DEGs) (Figure 4D), it was notable that several of these genes are involved in regulation of transcription including RNA Pol II, including Klf8, Zfp322a, and Klf7.To gain a more descriptive understanding of the pathways being altered we performed gene ontology analysis of the DEGs.Interestingly, GO terms that were most significantly downregulated included structural elements including regulation of cell shape and actin cytoskeleton organization, but also inflammatory response, immune system processes, phagocytosis and I-kappaB kinase/NF-kB signaling (Figure 4E).Specifically, downregulated transcripts regulated by I-kappaB kinase/NF-kB included Rela, also known as p65 of NFkB, and Nfkb2, Tlr9, Mapk14, and IRF8 (Table S1).
PF-06405977 also led to the AMPK-dependent upregulation of several biological processes including protein transport, and surprisingly an upregulation of catabolic processes including protein catabolic processes and sterol and cholesterol biosynthetic processes and autophagy (Figure 4F).Consistent with AMPK activation of ULK1, PF-06409577 upregulated the expression of genes critical for autophagy including Atg5, Atg12, and Lamp2.Surprisingly, PF-06409577 also upregulated genes in cholesterol synthesis including Hmgcs, Hmgcr, Fdft1, Cyp51, and Insig (Table S2).PF-06409577 inhibits cholesterol and fatty acid synthesis in hepatocytes suggesting upregulation of cholesterol synthesis genes may have been to compensate for reduced metabolic flux through this pathway.To directly test this hypothesis we assessed fatty acid and sterol synthesis in BMDMs and found that PF-06409577 dose dependently reduced both fatty acid and cholesterol synthesis in an AMPKb1-dependent manner (Figures 4G and 4H).These data show that PF-06409577 induced reductions in atherosclerosis were associated with the suppression of fatty acid and cholesterol synthesis, induction of markers of autophagy (increased ULK1 phosphorylation and Atg5, Atg12, and Lamp2) and the suppression of genes related to inflammatory responses, immune system processes, and IkappaB kinase/NF-kB signaling.

DISCUSSION
In the current study, we show that PF-06409577 reduces atherosclerosis in two mouse models.Further, these effects are dependent on myeloid AMPK b1, as the effects of PF-06409577 on atherosclerosis were absent in both whole body AMPKb1 knock-out mice and myeloid specific AMPK b1 knockout mice.These data are in contrast to studies with another direct AMPKb1 activator, A-769662, which failed to reduced atherosclerosis in mouse models, 24 likely reflects that PF-06409577 is a much more potent activator of AMPK than A-769662 (as shown in Figure 3).The absence of changes in hepatic and serum lipid levels in the current study is in contrast to effects seen with the same dose of PF-06409577 in C57Bl/6J mouse model fed a high-fat diet as well as studies in non-human primates. 29While unexpected, it is potentially related to the hyperlipidemic nature of these atherosclerotic models and alterations to classical LDL clearance, which is consistent with statins typically not yielding robust reductions in atherosclerosis or serum LDL in the ApoE À/À mouse model. 34Furthermore, these data demonstrate that extra-hepatic targets can alter atherosclerotic plaque development.ApoE À/À mice have drastically altered lipoprotein metabolism, serum lipid profiles and hepatic lipid levels, compared to other model systems; however, the atherosclerotic plaque morphology seen in this model is most similar to that of humans. 34Future studies examining the effects of PF-06409577 in other models of atherosclerosis, including LDL receptor null mice and mini pigs, will be important to confirm whether observations are translatable across distinct models of atherosclerosis.
We conducted a dose response in BMDMs and found that PF-06409577 increased the phosphorylation of two downstream substrates of AMPK, ULK1, and ACC in an AMPK b1-dependent manner.Consistent with the phosphorylation of ACC, PF-06409577 dose dependently reduced fatty acid synthesis and cholesterol synthesis in bone marrow macrophages from WT but not AMPK b1 KO mice.Similarly, consistent with the phosphorylation of ULK1, RNAseq analysis indicated one of the most upregulated GO terms was autophagy.2][13][14][15] Treatment of macrophages with PF-06409577 lowered GO terms related to the inflammatory response, immune system process, phagocytosis and IkappaB kinase/NF-kB signaling.These data indicate that PF-06409577 inhibits lipid synthesis and markers of inflammatory response while promoting markers of autophagy in macrophages suggesting this may be important for the effects of this compound to lower atherosclerosis in mouse models.
In conclusion, this study shows that oral delivery of the small molecule AMPK activator PF-06409577 reduces the development of atherosclerotic plaques in a myeloid AMPKb1 dependent manner.Furthermore, these reductions in atherosclerosis occurred independently of changes in serum cholesterol.Instead, PF-06409577 activation of AMPK in macrophages reduced transcripts related to the inflammatory response, phagocytosis, immune system processes and IkappaB kinase/NF-kB signaling while suppressing cholesterol and fatty acid synthesis and increasing markers of autophagy.In the setting of cardiometabolic disease where MASLD, insulin resistance and atherosclerosis are commonly linked, these data suggest that an AMPKb1 specific activator such as PF-06409577 may be effective not only at reducing liver steatosis and insulin resistance but also by targeting macrophages and atherosclerosis.Future studies examining whether the effects of other AMPKb1 selective compounds which have been shown to have positive effects on MASLD and insulin resistance in mouse models 35 and phase 2 clinical trials 36 also have effects on atherosclerosis will be informative to confirm these observations.

Limitations of study
A limitation of our findings is that we did not complete a dose response examining the effects of PF-06409577 on atherosclerosis.Previous studies have completed in vivo dose responses examining the effects of PF-06409577 on increasing AMPK activity in the kidney 28 and inhibiting liver de novo lipogenesis 28,29 and found that maximal effects are observed at a dose of 100 mg/kg.Therefore, while we selected this same dose and delivery method of PF-06409577, we did not complete a dose response study examining the effects of PF-06409577 on atherosclerosis or activation of AMPK in macrophages.Therefore, future dose response studies in mice will be important.

STAR+METHODS
Detailed methods are provided in the online version of this paper and include the following:

Figure 1 .
Figure1.PF-064095777 reduces Atherosclerosis via AMPK b1 in ApoE -/-mice independently of reductions in circulating triglycerides or cholesterol ApoE À/À and ApoE À/À AMPK b1 À/À mice were fed a Western diet and treated with vehicle or 100 mg/kg PF-06409577 by oral gavage.(A) Representative images of plaques in the aortic root, with plaques outlined in red, scale bar represents 100 mM, (B) plaque quantification, (C) necrotic area, (D) serum cholesterol, and (E) triglycerides.Data are presented as mean G SEM, * indicates p < 0.05 by two-way ANOVA with Fisher's LSD post-hoc testing.

Figure 2 .Figure 3 .
Figure 2. PF-06409577 reduces atherosclerosis via macrophage AMPK activation (A-D) Peritoneal macrophages were isolated from AMPK b1 fl/fl and AMPK b1 LysM mice.(A) AMPK levels and phosphorylation of ACC following treatment with PF-06409577 (10 mM, 90 min) were detected by western blot and (B-D) quantified using ImageJ data are presented as mean G SEM, * indicates p < 0.05 and *** indicates <0.005 by unpaired t test.AMPK b1 fl/fl and AMPK b1 LysM mice were injected with PCSK9 AAV, fed a Western diet and treated with Vehicle or 100 mg/kg PF-06409577 by oral gavage for 6 weeks.(E and F) Representative plaque images, scale bar represents 100 mM and quantification.Data are presented as mean G SEM, # indicates p < 0.05 in overall group effect my two-way ANOVA, * indicates p < 0.05 by two-way ANOVA with Fisher's LSD post-hoc testing.
d RESOURCE AVAILABILITY B Lead contact B Materials availability B Data and code availability d EXPERIMENTAL MODEL AND STUDY PARTICIPANT DETAILS B Animals B Cell culture d METHOD DETAILS B Atherosclerotic measurements B ELISAs B Rt-q-PCR B Plasma and liver triglycerides and cholesterol d QUANTIFICATION AND STATISTICAL ANALYSIS SUPPLEMENTAL INFORMATION