ETC-1002 regulates immune response, leukocyte homing, and adipose tissue inflammation via LKB1-dependent activation of macrophage AMPK.

ETC-1002 is an investigational drug currently in Phase 2 development for treatment of dyslipidemia and other cardiometabolic risk factors. In dyslipidemic subjects, ETC-1002 not only reduces plasma LDL cholesterol but also significantly attenuates levels of hsCRP, a clinical biomarker of inflammation. Anti-inflammatory properties of ETC-1002 were further investigated in primary human monocyte-derived macrophages and in in vivo models of inflammation. In cells treated with ETC-1002, increased levels of AMP-activated protein kinase (AMPK) phosphorylation coincided with reduced activity of MAP kinases and decreased production of proinflammatory cytokines and chemokines. AMPK phosphorylation and inhibitory effects of ETC-1002 on soluble mediators of inflammation were significantly abrogated by siRNA-mediated silencing of macrophage liver kinase B1 (LKB1), indicating that ETC-1002 activates AMPK and exerts its anti-inflammatory effects via an LKB1-dependent mechanism. In vivo, ETC-1002 suppressed thioglycollate-induced homing of leukocytes into mouse peritoneal cavity. Similarly, in a mouse model of diet-induced obesity, ETC-1002 restored adipose AMPK activity, reduced JNK phosphorylation, and diminished expression of macrophage-specific marker 4F/80. These data were consistent with decreased epididymal fat-pad mass and interleukin (IL)-6 release by inflamed adipose tissue. Thus, ETC-1002 may provide further clinical benefits for patients with cardiometabolic risk factors by reducing systemic inflammation linked to insulin resistance and vascular complications of metabolic syndrome.


ETC-1002 formulation / ETC-1002-CoA synthesis
For in vitro assays, ETC-1002 was formulated using aseptic technique at 30 and 100 mM in sterile dimethylsulfoxide (DMSO) and stored in sterile microcentrifuge tubes at 4°C for up to four weeks (stability was assessed). Working solutions of ETC-1002 were prepared in serum-free RPMI 1640 containing 12 mM HEPES, 10,000 U/ml penicillin, and 100 g/ml streptomycin. ETC-1002-CoA was synthesized using rat liver microsomes as described ( 32 ). For in vivo experiments, ETC-1002 dosing solutions were formulated by preparing a disodium salt aqueous solution using 2:1 molar ratio of NaOH to ETC-1002 in water. Carboxymethyl cellulose (CMC) and Tween-20 were added to make a fi nal solution containing 0.5% CMC and 0.025% Tween with a fi nal pH 7-8 . Compound concentrations in dosing solutions were administered at a volume of 10 ml/kg body.

Hepatocyte isolation
All animal procedures were approved by an Institutional Animal Care and Use Committee at the Michigan Life Science and Innovation Center. Nutritionally staged male Sprague-Dawley [Crl:CD (SD)] rats were anesthetized with isofl urane, and livers were perfused for hepatocyte isolation as previously described ( 12 ). Hepatocytes were plated in high-glucose DMEM containing 20% FBS, 14 mM HEPES, 0.2% bovine albumin, 2 mM L-glutamine, 1× MEM nonessential amino acids, 100 nM insulin, 100 g/ml dexamethasone, and 20 g/ml gentamicin at a density of 1.5 × 10 5 cells/cm 2 on collagen-coated 6-well plastic dishes. After an attachment period of 3-4 h, cells were washed once and cultured overnight in DMEM containing 10% FBS.

Macrophage preparation
Human blood was obtained from healthy volunteers under RCRC IRB (Independent Review Board) guidelines (Esperion IRB Protocol Number ESP001). For preparation of autologous serum, blood was allowed to clot and serum was separated with centrifugation at 2,500 rpm for 30 min. Peripheral blood mononuclear cells (PBMC) were isolated with density gradient using HISTOPAQUE®-1077 as previously described ( 33 ) and plated at 3 × 10 6 cells per 12-well plate. Monocytes were adherence purifi ed and differentiated for fi ve days in RPMI 1640 supplemented with 40% autologous serum, 14 mM HEPES, 100 U/ml penicillin, 50 U/ml streptomycin, and 2 mM L-glutamine.

Protein arrays
At the end of differentiation period, macrophages were washed with PBS and switched to RPMI 1640 containing 5% autologous serum and supplemented with 14 mM HEPES, 100 U/ml penicillin, 50 U/ml streptomycin, and 2 mM L-glutamine. ETC-1002 at various concentrations (50 M and 100 M) was added to the media 1 h prior to stimulation with 100 ng/ml of lipopolysaccharide (LPS) from Escherichia coli 0111:B4 (Sigma-Aldrich, St. Louis, MO). Media conditioned by MDMs was collected 12 h following LPS stimulation and assayed with Proteome Profi ler Human Cytokine Array Kit, Panel A, and Human Matrix Metalloproteinase Array, according to the manufacturer's instructions. Data for cytokine and matrix metalloproteinase (MMP) arrays were captured and analyzed with a Kodak 4000MM Image Station. Net signal intensity for each analyte was expressed as percentage of the internal reference standard for each individual array membrane. Data are presented as mean ± SEM. Comparisons between groups dyslipidemia and other cardiometabolic risk factors. Benefi cial effects of ETC-1002 on lipid and glucose metabolism have been demonstrated in multiple animal models and have been linked to ETC-1002-mediated activation of AMP-activated protein kinase (AMPK) (12)(13)(14)(15). AMPK represents an attractive therapeutic target, as it plays a fundamental role in maintaining cellular energy homeostasis via coordinating glucose and lipid metabolism as well as downregulating proinfl ammatory signaling pathways (16)(17)(18).
Here we demonstrate that in primary human monocytederived macrophages (MDM), ETC-1002 mediates inhibition of a wide range of proinfl ammatory molecules that coincides with the modulation of AMPK and MAP kinase activities. AMPK phosphorylation and inhibitory effects of ETC-1002 on soluble mediators of infl ammation were signifi cantly abrogated by small inhibitory RNA (siRNA)mediated silencing of macrophage liver kinase B1 (LKB1), indicating that ETC-1002 activates AMPK and exerts its anti-infl ammatory effects via an LKB1-dependent mechanism. Furthermore, ETC-1002 treatment diminished leukocyte homing into infl ammatory sites in vivo as well as reduced adipose tissue infl ammation in a mouse model of diet-induced obesity (DIO). Thus, the anti-infl ammatory properties of ETC-1002 may offer additional clinical benefi ts by controlling immune responses in patients with metabolic abnormalities.

Thioglycollate-induced peritonitis
All animal procedures were approved by an Institutional Animal Care and Use Committee at the Michigan Life Science and Innovation Center. C57BL/6 male mice (Jackson Laboratory, Bar Harbor, ME) were injected intraperitoneally with 1 ml of 4% sterile thioglycollate. Mice (n = 6 per each experimental group) received either vehicle alone or vehicle containing ETC-1002 at 30 mg/kg/day by oral gavage 24 h prior to thioglycollate challenge and for the duration of the study. Peritoneal cavity leukocytes were obtained from three lavage washes collected 24 h and 72 h after thioglycollate injection with 3 ml of cold PBS. Leukocytes were pelleted from fl uid by centrifugation and resuspended in 5 ml of PBS. The number of cells migrated into peritoneal cavity 24 h (neutrophils) and 72 h (macrophages) after thioglycollate injection was determined with a hemocytometer. Data are expressed as mean ± SEM and comparisons between vehicle and ETC-1002-treated groups were performed using an unpaired twotailed Student t -test. Signifi cance was accepted at P р 0.05.

Epididymal fat-pad tissue explants and ex-vivo infl ammation studies
All animal procedures were approved by an Institutional Animal Care and Use Committee at the Michigan Life Science and Innovation Center. Male C57BL/6 mice (Jackson Laboratory, Bar Harbor, ME) were singly housed in environmental isolators on ALPHA-dri paper bedding. Animals were provided free access to 60% high-fat diet (HFD) upon arrival beginning at 11 weeks of age. At 12 weeks of age, mice were randomly assigned to treatment groups (n = 10/group) and administered with vehicle alone or vehicle containing ETC-1002 at 30 mg/kg/day via oral gavage for nine weeks. At the time of necropsy, epididymal fat pads from the chow-fed animals, HFD-fed animals, and animals on HFD treated with ETC-1002 were removed, and weight of the fat pads from individual mice was determined. Adipose tissues were immersed into ice-cold, serum-free RPMI 1640 media and dissected into approximately 5 × 5 mm fragments under sterile conditions. Adipose tissue protein extracts were prepared, and expression of AMPK, JNK1/2, p38, and F4/80 in adipose tissue was determined by Western blot. Remaining fragments of epididymal fat pads were then placed into a 12-well tissue culture plate (four fragments per well) and incubated in serum-free RPMI 1640 at 37°C in humidifi ed 0.5% CO 2 for 24 h. At the end of the incubation time, fat-pad fragments were removed from the plate and airdried, and then the collective weight of tissues in each well was determined. IL-6 levels in media conditioned by tissue explants were measured with mouse IL-6 Quantikine ELISA Kit (R and D Systems), and the total amount of IL-6 released per well was normalized to the corresponding dry tissue weight. The fat-pad mass for individual animals and IL-6 levels released by tissue explants are expressed as mean ± SEM. Comparisons between groups were performed by one-way ANOVA. A Bonferroni's post hoc multiple comparison test was used to assess signifi cant differences revealed by the ANOVA. Signifi cance was accepted at P р 0.05.

Primary human MDMs do not form ETC-1002-CoA thioester in vitro
Recent studies have identifi ed AMPK and ATP citrate lyase (ACL) as distinct molecular targets for ETC-1002, mediating its benefi cial effects on glucose and lipid metabolism ( 12 ). While ETC-1002-free acid activates liver AMPK, hepatocytes also readily convert the parent compound into were performed by one-way ANOVA. A Bonferroni's post hoc multiple comparison test was used to assess signifi cant differences revealed by the ANOVA. Signifi cance was accepted at P р 0.05.

siRNA-mediated silencing of macrophage LKB1
Macrophages were differentiated in 40% autologous serum for six days as described. At the end of the differentiation period, cells were washed with HBSS and switched to RPMI containing 5% autologous serum for 1 h prior siRNA treatment. LKB1 (STK11) (#4392420) and Negative control (Mock) (#4390843) siRNAs (Silencer Select PreDesigned siRNA, Invitrogen/ Ambion, Logon, UT) were incubated for 15 min with polymerbased GenMute siRNA transfection reagent (SignaGen Laboratories, Ijamsville, MD) to form a siRNA transfection complex. LKB1 siRNA and mock siRNA transfection complexes were then added to MDMs at 50 nM siRNA fi nal concentration for 5 h. At the end of 5 h incubation period, media containing siRNA transfection complexes was removed, cells were washed three times with HBSS, and then supplemented with RPMI containing 5% autologous serum. MDMs were allowed to recover for 24 h before stimulation with LPS and/or ETC-1002 treatment. Cell lysates were prepared as described in Western blot sections, and LKB1 expression and AMPK phosphorylation were assessed by Western blot. Conditioned media was collected, and levels of IL-6 and chemokine (C-C motif) ligand 2 (CCL2) / monocyte chemoattractant protein-1 (MCP-1) were determined by ELISA (Human MCP-1 and IL-6 DuoSet, R and D Systems, Minneapolis, MN).

HPLC-UV for ETC-1002-CoA determination
A 15 l aliquot from the extraction procedure described above was injected into the HPLC system utilizing an Alltima® C8 5 , 250 × 4.6 mm ID HPLC column (Alltech Associates, Inc., Deerfi eld, IL) running 15-40% acetonitrile in 25 mM K 2 HPO 4 (pH 7.0) gradient before UV detection at 254 nm on a G1314A photodiode array detector (PDA) (Agilent Technologies, Inc., Santa Clara, CA). ETC-1002-CoA concentrations were determined by comparing the sample peak area to the peak area of an ETC-1002-CoA calibration standard. infl ammation ( 22,23,35 ). These effects of AMPK activation are linked, at least in part, to inhibition of Toll-like receptor (TLR)-mediated signaling and downregulation of proinfl ammatory molecules released by immune cells ( 23,30,31 ). To determine whether ETC-1002-mediated phosphorylation and activation of AMPK in primary human MDMs yields similar anti-infl ammatory effects, macrophages were stimulated with 100 ng/ml of LPS in either the absence or presence of varying concentrations of ETC-1002. Secretion of proinfl ammatory cytokines and chemokines ETC-1002-CoA thioester, which has been shown to directly inhibit ACL ( 12 ). To determine whether macrophages metabolize ETC-1002 to the CoA thioester at levels comparable to primary hepatocytes, intracellular ETC-1002-CoA was measured in both cell types following incubation with ETC-1002 at concentrations of 30 M or 100 M. The model of primary human MDMs differentiated in autologous serum was selected since, unlike immortalized monocyte or macrophage cell lines, these cells more accurately refl ect macrophage behavior at sites of chronic infl ammation in human disease states ( 33 ).
Consistent with previously published data ( 12 ), ETC-1002-CoA was readily detectable in cell lysates from cultured hepatocytes ( Fig. 1A ). In contrast, analysis of MDM extracts failed to detect any measurable amounts of ETC-1002-CoA thioester at concentrations up to 100 M, demonstrating that, unlike primary rat hepatocytes, macrophages do not to convert ETC-1002-free acid into ETC-1002-CoA thioester at detectable levels. ( Fig. 1A ). Thus, in primary human MDMs, ETC-1002-free acid is likely to mediate pharmacological effects of ETC-1002.

ETC-1002 activates AMPK in primary human MDMs
In human hepatocellular carcinoma (HepG2) cells treated with ETC-1002, profound and concentration-dependent activation of AMPK has been previously attributed to ETC-1002-free acid as, similar to primary human MDMs, these cells do not metabolize parent molecules to ETC-1002-CoA thioester ( 12 ). To determine whether ETC-1002 activates AMPK in human macrophages, MDMs differentiated in autologous serum were treated with various concentrations of the compound, and cell lysates were probed with anti-phosphorylated AMPK ␣ (T172) antibody. In vehicle-treated MDMs, basal levels of AMPK phosphorylation were readily detectible as demonstrated by the appearance of an immunoreactive band corresponding to the expected molecular mass of ‫ف‬ 62 kDa ( Fig. 1B ). Concentration-dependent increases in phospho-AMPKto-total-AMPK ratio in ETC-1002-treated cells indicate that ETC-1002 induces AMPK (T172) phosphorylation at levels comparable to those observed previously in HepG2 cells ( 12 ). ACC (serine 79) is a unique AMPK phosphorylation site and is commonly used as a marker of AMPK activity ( 34 ). As such, when cell lysates from MDMs treated with ETC-1002 were probed with anti-phosphorylated ACC (S79) antibodies, sustained and concentration-dependent increases in ACC (S79)-specifi c immunoreactivity were observed at the expected molecular mass of ‫ف‬ 280 kDa. Consistently, the phospho-ACC to total-ACC ratio increased by 25% and 60% in cells treated with 50 µM and 100 µM ETC-1002, respectively ( Fig. 1B ). Taken together, these data confi rm the AMPK-activating properties of ETC-1002 in primary human MDMs ( Fig. 1B ).

ETC-1002 inhibits release of proinfl ammatory cytokines and chemokines by stimulated human MDMs
Pharmacological activation of AMPK has been shown to reduce infl ammatory responses and minimize organ damage in multiple animal models of acute and chronic Primary human MDMs were differentiated in autologous serum for fi ve days, followed by stimulation with 100 ng/ml of LPS in either the absence or presence of ETC-1002 (n = 4 for each experimental condition). Proteins in cell lysates were separated on SDS-PAGE, transferred to PVDF membrane, and blotted with antibodies raised against phosphorylated AMPK ␣ (T172) and ACC (S79) as well as total AMPK and ACC. Density of AMPK-and ACC-specifi c immunoreactive bands was captured and quantitated with a Kodak 4000MM Image Station. Data are presented as ratio of phosphorylated to total proteins (mean ± SEM). Comparisons between groups were performed by one-way ANOVA and Bonferroni's post hoc multiple comparison tests, * P < 0.05. BDL, below detection limit. cytokines, ETC-1002 suppressed the LPS-induced expression of CCL1, CCL2, CCL5, and CXCL10 chemokines to an even greater extent, with concentration-dependent inhibition ranging from 17% to 96% ( P < 0.05) ( Fig. 3 ). Likewise, other markers of vascular and adipose tissue infl ammation, including ICAM1, CD40L, MIP1 ␤ as well as the activated complement component C5a, were signifi cantly attenuated in ETC-1002-treated cells ( Fig. 4A ).

Anti-infl ammatory effects of ETC-1002 coincide with modulation of AMPK and MAPK signaling pathways
To investigate mechanism(s) underlying the anti-infl ammatory properties of ETC-1002, activation of AMPK, ACC, and MAP kinases JNK and p38 were assessed in LPSstimulated MDMs. Consistent with the established ability of proinfl ammatory stimuli to inhibit AMPK activity in macrophages ( 30,31 ), LPS-challenged human MDM significantly reduced levels of phosphorylated AMPK ␣ and ACC by 70% and 80%, respectively ( Fig. 5 ). The LPSinduced reduction in AMPK phosphorylation coincided into MDM-conditioned media was then evaluated with cytokine arrays.
In resting MDMs, levels of proinfl ammatory cytokines and chemokines remained at the lower limits of detection. In contrast, stimulation with LPS triggered a strong and consistent infl ammatory response as determined by a dramatic increase in the secretion of a variety of proinfl ammatory molecules ( Figs. 2, 3, and 4 ). Acute-phase proteins, such as TNF-␣ , IL-1 ␤ , IL-6, and IL-8, are believed to be a hallmark of low-grade infl ammation associated with metabolic abnormalities ( 2,3 ). As shown in Fig. 2 , treatment with ETC-1002 strongly suppressed the release of TNF-␣ , IL-1 ␤ , IL-6, and IL-8 by stimulated MDMs in concentration-dependent manner with inhibitory effects ranging from 15% to 77% ( P < 0.05).
As leukocyte fl ux into infl ammatory sites is controlled by chemotactic gradients ( 36,37 ), we next investigated the effects of ETC-1002 on the expression of CC-and CXCchemokine family members that are known to govern monocyte, lymphocyte, and neutrophil traffi cking ( 36,37 ). Similar to the stimulatory effects exerted by endotoxin on the upregulation of acute-phase proteins, LPS treatment increased the production of a wide array of chemotactic molecules ( Fig. 3 ). Consistent with the inhibitory effects exerted by ETC-1002 on the expression of acute-phase  LKB1 is required for ETC-1002-mediated AMPK activation ( 12 ). To determine whether the activation of macrophage AMPK by ETC-1002 is also LKB1-dependent, siRNA interference was utilized to reduce endogenous LKB1 protein levels and assess the effects of LKB1 knockdown on ETC-1002-mediated AMPK phosphorylation and inhibition of immune response.
Primary human MDMs were transfected with negative control "mock" or LKB1 siRNA for 5 h followed by a 24 h recovery period. Unstimulated or LPS-induced cells were then treated with vehicle or 100 µM ETC-1002, and AMPK phosphorylation as well as production of proinfl ammatory cytokines were assessed by Western blot and ELISA ( Fig. 6 ). Macrophages transfected with LKB1 siRNA demonstrated a 63% reduction ( P = 0.0002) in LKB1 protein with increased phosphorylation of MAP kinases p38 and JNK as shown by increased phosphorylated-to-total p38 and JNK protein ratio ( Fig. 5 ). By contrast, when cells were stimulated with LPS in the presence of ETC-1002, AMPK and ACC phosphorylation was increased relative to LPS alone, whereas phosphorylated p38 and JNK levels were attenuated by 73% and 47%, respectively, when treated with 50 M of ETC-1002 and by 90% and 71%, respectively, at 100 µM of ETC-1002 ( Fig. 5 ).

ETC-1002 activates macrophage AMPK and inhibits immune response via an LKB1-dependent mechanism
Liver kinase B1 is an essential serine-threonine kinase that directly phosphorylates and activates AMPK ( 16,18,38,39 ). Previously, we have demonstrated that in HepG2 cells,

ETC-1002 attenuates homing of leukocytes into infl ammatory site in vivo
Migration of circulating monocytes into adipose tissue followed by differentiation into proinfl ammatory macrophages is a hallmark of chronic, low-grade infl ammation associated with excess of visceral adipose tissue ( 2,40 ). As ETC-1002 inhibits the generation of proinfl ammatory chemokines in vitro ( Fig. 3 ), the ability of ETC-1002 to affect chemotactic potential was assessed in response to thioglycollate-induced peritonitis. As shown in Fig. 7A , compared with cells transfected with "mock" siRNA ( Fig. 6A ). As expected, unstimulated and LPS-induced "mock"-transfected MDMs showed increased AMPK (T172) phosphorylation in response to ETC-1002 treatment ( Fig. 6B ). In contrast, AMPK (T172) phosphorylation was largely abolished in cells transfected with LKB1 siRNA ( Fig. 6B ). Furthermore, consistent with anti-infl ammatory effects of AMPK activation, ETC-1002 inhibited release of IL-6 and CCL2/MCP-1 by "mock"-transfected macrophages whereas cells transfected with LKB1 siRNA not only remained insensitive to ETC-1002 treatment but also revealed slightly stronger proinfl ammatory response ( Fig. 6C ). Thus, these data demonstrate that in primary human macrophages, Fig. 5. ETC-1002-mediated modulation of AMPK and MAPK signaling pathways. MDMs were differentiated in autologous serum for fi ve days, and ETC-1002 at various concentrations was added to the media 1 h prior to LPS stimulation (100 ng/ml). Cell lysates were collected 3 h after LPS challenge (n = 3 for each experimental condition). (A) Proteins in cell lysates were separated on SDS-PAGE, transferred to PVDF membrane blotted with antibodies specifi c for phosphorylated and total AMPK, ACC, p38, and JNK1/2. (B) Phosphorylation status of AMPK, ACC, p38, and JNK1/2 in MDM cell lysates was quantitated by densitometry of the respective immunoreactive bands. Data are presented as ratio of phosphorylated to total proteins (mean ± SEM). Comparisons between groups were performed by one-way ANOVA and Bonferroni's post hoc multiple comparison tests, * P < 0.05. both 24 h and 72 h, indicating that ETC-1002 is likely to inhibit transmigration of both neutrophils and macrophages during the inception and propagation of infl ammatory response ( Fig. 7A ).
During transmigration into infl ammatory disease sites, monocytes and macrophages express a wide array of MMPsproteolytic enzymes, which not only allow cells to negotiate extracellular matrix barriers but also may control chemotactic gradient by processing soluble mediators of infl ammation and directly regulating infl ammatory signaling pathways (42)(43)(44). To investigate whether inhibitory effect of ETC-1002 on leukocyte homing can, at least in part, be attributed to reduced expression of MMPs, media conditioned by human MDMs was assayed with MMP detection following thioglycollate challenge, leukocyte numbers in peritoneal lavage increased after 24 h and 72 h (26.4 × 10 6 / ml lavage fl uid and 4.4 × 10 6 /ml lavage fl uid, respectively). By contrast, when mice were treated with 30 mg/kg of ETC-1002 prior to thioglycollate administration, leukocyte infl ux into the peritoneal cavity was signifi cantly reduced by 48% at 24 h (13.7 × 10 6 /ml lavage fl uid; P < 0.008) and by 61% at 72 h (1.8 × 10 6 /ml lavage fl uid; P < 0.002) ( Fig. 7A ). Infl ammatory response in mouse peritoneal cavity has defi ned sequential phases of the disease in which infl ux of neutrophils predominates at the early stages of infl ammation (up to 24 h) followed by a macrophage-enriched infi ltrates (48-72 h) ( 41 ). In ETC-1002-treated mice, the number of leukocytes in peritoneal lavage was reduced at Fig. 6. ETC-1002 activates macrophage AMPK and inhibits immune response via LKB1-dependent mechanism. Primary human MDMs differentiated for fi ve days in autologous serum were transfected with negative control "mock" or LKB1 siRNA for 5 h followed by 24 h recovery period. Unstimulated macrophages or cells induced with 100 ng/ml of LPS cells were then treated with vehicle or 100 µM ETC-1002. Proteins in cell lysates were separated on SDS-PAGE, transferred to PVDF membrane, and blotted with (A) anti-LKB1 and anti-␤ -Actin antibodies or (B) antibodies specifi c for phosphorylated and total AMPK. (C) Media conditioned by macrophages was collected, and levels of IL-6 and CCL2/MCP-1were determined by ELISA. LKB1 expression and phosphorylation status of AMPK in MDM cell lysates were quantitated by densitometry of the respective immunoreactive bands. For LKB1, data are presented as ratio of total LKB1 to total ␤ -actin (mean ± SEM). For AMPK, data are presented as ratio of phosphorylated to total AMPK protein. Levels of IL-6 and CCL2/MCP-1 in macrophageconditioned media are expressed as picograms per milliliter (mean ± SEM). Comparisons between groups performed by two-tailed nonpaired Student t -test, * P < 0.05, or by one-way ANOVA and Bonferroni's post hoc multiple comparison tests, * P < 0.05. HFD-fed versus 0.89 ± 0.07 g, HFD-fed/ETC-1002; P < 0.05). To evaluate the status of adipose tissue-associated infl ammation, fragments of epididymal fat pads were maintained as ex vivo explants, and IL-6 levels were measured in media conditioned by tissue explants for 12 h. As shown in Fig. 8A , HFD-induced expansion of adipose tissue was accompanied by approximately 2-fold increase in IL-6 levels (16.6 ± 7.8 pg/mg chow-fed versus 35.7 ± 3.6 pg/mg HFDfed; P < 0.05). Importantly, and consistent with the in vitro anti-infl ammatory effects of ETC-1002, IL-6 release was reduced by 63% in mice receiving ETC-1002 (IL-6 release: 35.7 ± 3.6 pg/mg HFD-fed versus 13.3 ± 2.2 pg/mg HFDfed/ETC-1002; P < 0.05).
Signifi cantly, AMPK-activating properties of ETC-1002 were further demonstrated in vivo as AMPK phosphorylation attenuated by HFD was substantially restored in adipose tissues of ETC-1002-treated animals ( Fig. 8B ). Likewise, decreased phosphorylation of adipose JNK and p38 in tandem with reduced expression of macrophage-specifi c marker F4/80 strongly supports the role for AMPK/MAPK axis in mediating anti-infl ammatory effects of ETC-1002 in vivo ( Fig. 8B, Fig. 9 ).

DISCUSSION
In the present studies, we demonstrate that in stimulated primary human MDMs, ETC-1002 exerts broad antiinfl ammatory effects consistent with LKB1-dependent array. As shown on Fig. 7B , release of MMP-1 and MMP-10 into MDM-conditioned media was unaffected by ETC-1002 treatment, while levels MMP-8 and MMP-13 were upregulated. Taken together, these data indicate that ETC-1002mediated inhibition of leukocyte chemotaxis is unlikely to result from reduced macrophage proteolytic potential but instead is governed by altered chemotactic gradient.

ETC-1002 reduces epididymal fat-pad mass and adipose tissue infl ammation in a mouse model of diet-induced obesity
Recent studies have demonstrated that in a mouse model of DIO, a two-week treatment with ETC-1002 was suffi cient to reduce body weight as well as to lower fasting plasma glucose and insulin levels ( 12 ). Since adipose tissue macrophages are believed to play a critical role in governing immune responses and insulin resistance in DIO ( 2,40,45 ), we next evaluated whether the benefi cial effect of ETC-1002 on glycemic control can be linked to reduced levels of adipose tissue-associated infl ammation. As such, male C57BL/6 mice were placed on HFD and orally dosed with either vehicle alone or ETC-1002 at 30 mg/kg/day for nine weeks. At the termination of the study, visceral adipose tissue from mice placed on HFD was macroscopically larger ( Fig. 8A ), with the average epididymal fat pad mass signifi cantly increased relative to chow-fed animals (0.31 ± 0.04 g, chow-fed versus 1.3 ± 0.04 g, HFD-fed; P < 0.05) ( Fig. 7A ). By contrast, ETC-1002-treated mice displayed a 32% reduction in fat-pad mass (fat-pad mass: 1.3 ± 0.04 g, Cell migration into peritoneal cavity was estimated by number of leukocytes in the peritoneal lavage fl uid and expressed as mean ± SEM. Group comparison performed with nonpaired two-tailed Student t -test, * P < 0.05. For MMP array, net signal intensity for each analyte was expressed as percentage of the internal reference standard for each individual array membrane (mean + SEM). Comparisons between groups were performed by one-way ANOVA and Bonferroni's post hoc multiple comparison tests, * P < 0.05. polyphenols, has been shown to yield anti-infl ammatory effects (22)(23)(24)(25), resulting in diminished leukocyte accumulation in target organs, limited interstitial edema, and reduced levels of TNF ␣ , IL-1 ␤ , IL-6, and IL-8 ( 22,23 ). Similar to the anti-infl ammatory effects demonstrated here with ETC-1002-treated human MDMs, pharmacological activation of AMPK in macrophage cell lines or bone activation of macrophage AMPK and the downregulation of MAPK signaling. Furthermore, in vivo, ETC-1002 attenuates leukocyte homing into infl ammatory sites while inhibiting adipose tissue infl ammation.
In multiple models of inflammation, AMPK activation by 5-aminoimidazole-4-carboxamide-1-␤ -D-ribofuranoside (AICAR) or synthetic, as well as plant-derived For AMPK, p38, and JNK1/2, data are presented as ratio of phosphorylated to total proteins (mean ± SEM). For F4/80 expression, data are normalized to ␤ -actin (mean ± SEM). Comparisons between groups were performed by one-way ANOVA and Bonferroni's post hoc multiple comparison test, * P < 0.05. Data are expressed as mean ± SEM. slightly higher infl ammatory response as shown by increased expression of IL-6 and CCL2/MCP-1. This may refl ect inability of cell to adequately compensate for stress associated with LPS and/or compound treatment when the LKB1/AMPK axis is disrupted.
While numerous reports have demonstrated that AMPK can govern infl ammatory response in immune cells, the signaling events leading to these effects remain unclear. Many signaling pathways, including Akt, GSK ( 29 ), SIRTs ( 28 ), NF-kB ( 17 ), and MAPK ( 30 ), have been proposed to mediate the anti-infl ammatory effects that arise as a consequence of AMPK activation. However, the primary intracellular signaling cascades vary among published studies, depending on cell type, nature of proinfl ammatory stimuli, and experimental design. In our model using primary human MDMs, AMPK activation by ETC-1002 inhibits LPSinduced MAPK signaling as refl ected in the reduced phosphorylation of JNK and p38. Similar inhibition of JNK and p38 phosphorylation has been previously demonstrated for AMPK activators in macrophage cell lines as well as in human monocytes where pharmacological activation of AMPK offsets immune responses in a MAPK-dependent fashion ( 26,30 ).
Chronic low-grade infl ammation has been regarded as a major pathogenic factor in obesity, T2D, and CAD ( 2,3 ). Along with acute-response proteins, macrophage infi ltration in adipose and vascular tissues orchestrates the expression of a complex network of chemokines and adhesion molecules that coordinate extravasation of leukocyte subsets into infl ammatory sites ( 47,48 ). Previous studies have established that AMPK activation reduces expression of CCL2/MCP-1 in macrophage-like cell lines as well as in adipose tissues of obese mice ( 30,49 ). While ETC-1002 also attenuated macrophage CCL2/MCP-1 expression in our studies, it exerted a potent inhibitory effect on an array of chemotactic and adhesion molecules that play important roles in controlling vascular and adipose tissue infl ammation, including ICAM-1, CXCL10 (IP-10), CCL5 (RANTES), and CCL1 (I-309) ( 2,48 ). In addition to the inhibition of canonical mediators of infl ammation, ETC-1002 significantly attenuated expression of CD40L and C5a. Interestingly, in clinical studies, plasma levels of CD40L and C5a are associated with an increased risk for the onset of acute cardiovascular events in humans ( 50,51 ). Further, the ability of ETC-1002 to attenuate macrophage chemotactic potential in vitro was confi rmed in a mouse model of thioglycollate-induced peritonitis in which administration of ETC-1002 largely reduced neutrophil and macrophage infl ux. Importantly, this inhibition of leukocyte homing likely to refl ect reduced chemotactic gradient and be independent of proteolytic events, as MMP release by stimulated macrophages was either unaffected (MMP-1, MMP-10) or even upregulated (MMP-8, MMP-13) upon compound treatment. It is noteworthy that production of some macrophage gene products (IL-1ra, MIP1 ␣ , TIMP-1) remained unchanged or was upregulated (TIMP-2, MMP-8, MMP-13) in drug-treated cells. While highlighting the complexity of the macrophage immune response, this may also indicate that ETC-1002-mediated inhibition of cytokine and marrow-derived macrophages largely reduces expression of TNF ␣ , IL-1 ␤ , IL-6, and IL-8 ( 30,31 ). Further, mechanistic links between AMPK activation and the inhibition of immune responses have been confi rmed in recent genetic studies utilizing RNAi or dominant negative versus constitutively active forms of AMPK ␣ ( 28,29 ). Consistently, genetic AMPK ␤ 1 deletion supports the critical role for AMPK in regulating macrophage polarization/infl ammatory status in vitro and in vivo ( 31 ). In the present study, siRNA interference targeting endogenous levels of LKB1 was utilized to confi rm the mechanistic link between ETC-1002mediated AMPK activation and its anti-infl ammatory effects. LKB1 is a well-established, physiologically relevant kinase that phosphorylates and activates AMPK ( 16,18,38,39 ). AMPK activation by agonists, such as AICAR, metformin, or phenformin, or in response to stress was absent in LKB1-deficient cells, identifying LKB1 as an obligatory AMPK-activating kinase ( 38,46 ). Therefore, silencing of LKB1 in macrophages is likely to prevent activation of otherwise intact AMPK ␣ protein. Indeed, 60% reduction of macrophage LKB1 expression was suffi cient to diminish AMPK phosphorylation in ETC-1002-treated cells and to largely revoke ETC-1002-mediated inhibition of immune response, further confi rming that ETC-1002 activates AMPK and exerts its anti-infl ammatory effects via mechanisms dependent on the LKB1/AMPK axis. It is noteworthy that LKB1-defi cient macrophages revealed soluble ICAM-1, as well as lymphocyte and monocyte release of IL-1, IL-2, IL-6, IL-8, INF-␥ , TNF-␣ , and MCP-1 ( 58,59 ). These recent clinical fi ndings reveal a remarkable similarity to the anti-infl ammatory "signature" of ETC-1002 ( 12,13,15 ). Given the clinical safety and efficacy profi les of ETC-1002, including consistent reduction in both LDL-C and hsCRP levels ( 19 ), these data reinforce its potential promise as a novel therapy for cardiovascular and metabolic diseases.
In summary, the present report describes anti-infl ammatory properties of ETC-1002, a novel investigational drug being developed for the treatment of dyslipidemia and other cardiometabolic risk factors. We demonstrate that in stimulated primary human MDMs, ETC-1002 exerts strong anti-infl ammatory effects consistent with LKB1-dependent activation of macrophage AMPK and downregulation of MAPK signaling. In vivo, ETC-1002 attenuates homing of leukocytes into the infl ammatory site and inhibits adipose tissue infl ammation in a mouse model of DIO. Thus, in addition to regulating imbalances of lipid and carbohydrate metabolism ( 12,19 ), ETC-1002 may provide further clinical benefi ts for patients with cardiometabolic risk factors by reducing systemic infl ammation associated with insulin resistance and vascular complications of metabolic syndrome (Fig. 9). chemokine release is unlikely to result from off-target effects and/or cytotoxicity of the compound.
In DIO, circulating monocytes enter adipose tissue and differentiate into macrophages that display a proinfl ammatory phenotype and often reside in a form of crown-like structures juxtaposed to blood vessels and metabolically active as well as damaged adipocytes ( 2,40,52 ). Adipose tissue infl ammation is believed to be a critical factor in promoting insulin resistance associated with visceral obesity ( 2,3 ). Indeed, ablation of adipose tissue macrophages (ATM) or genetic disruption of the TLR4/JNK axis protects mice from developing obesity-related activation of the immune system and insulin resistance ( 45,53 ). Likewise, selective therapeutic inactivation of JNK's downstream effector molecules, TNF ␣ or IL-1 ␤ , curbs hyperglycemia, lowers the proinsulin-to-insulin ratio, and improves insulin sensitivity in rodent models of obesity ( 54,55 ). Similarly, ETC-1002 diminishes adipose JNK phosphorylation in mice on HFD and limits ATM content as shown by reduced expression of macrophage-specifi c marker 4F/80. Furthermore, adipose tissue mass was reduced in obese mice treated with ETC-1002 in tandem with decreased IL-6 levels in explant cultures of epididymal fat pads. Improved adipose tissue infl ammation coincided with increased AMPK phosphorylation and was consistent with previously reported data showing that ETC-1002 reduces body weight along with plasma fasting glucose and insulin levels in a mouse model of DIO ( 12 ).
Since the pathophysiology of the immune response associated with metabolic abnormalities in humans is very complex, targeting more broad upstream signaling events, as opposed to single infl ammatory effector molecule, may provide greater clinical benefi t in patients with cardiometabolic disorders ( 9,10 ). In this context, the anti-infl ammatory benefi ts of AMPK activation have not been appreciated fully. Currently, the AMPK activator metformin is the most successful and widely prescribed drug to control hyperglycemia in subjects with T2D. In addition to well-documented effects of metformin on fasting plasma glucose and HbA1c ( 56 ), administration of metformin in T2D subjects is associated with signifi cant lowering of plasma hsCRP levels ( 57 ). Interestingly, recent clinical studies suggest that these benefi ts may be directly linked to effects of metformin on monocyte and lymphocyte immune function ( 58,59 ). Metformin administered to patients with impaired fasting glucose reduced plasma levels of hsCRP,