α1AMPK deletion in myelomonocytic cells induces a pro-inflammatory phenotype and enhances angiotensin II-induced vascular dysfunction

Abstract

Aims

Immune cell function involves energy-dependent processes including growth, proliferation, and cytokine production. Since the AMP-activated protein kinase (AMPK) is a crucial regulator of intracellular energy homeostasis, its expression and activity may also affect innate and adaptive immune cell responses. Therefore, we aimed to investigate the consequences of α1AMPK deletion in myelomonocytic cells on vascular function, inflammation, and hypertension during chronic angiotensin II (ATII) treatment.

Methods and results

We generated a mouse strain with α1AMPK deletion in lysozyme M⁺ myelomonocytic cells. Compared to controls, chronic ATII infusion (1 mg/kg/day for 7 days) lead to increased vascular oxidative stress and aggravated endothelial dysfunction in LysM-Cre⁺ x α1AMPK^fl/fl mice. This was accompanied by an increased aortic infiltration of CD11b⁺F4/80⁺ macrophages and enhanced pro-inflammatory cytokine release (tumour necrosis factor-alpha, interferon-gamma, and interleukin-6). Mechanistically, we found that increased expression of C-C chemokine receptor 2 (CCR2) in α1AMPK deficient myelomonocytic cells facilitated their recruitment to the vascular wall. In addition, expression of the ATII receptor type 1a and the oxidative burst was increased in these cells, indicating an increased susceptibility towards pro-oxidant stimuli.

Conclusions

In summary, α1AMPK deletion in myelomonocytic cells aggravates vascular oxidative stress and dysfunction by enhancing their recruitment to the vascular wall and increasing their susceptibility towards pro-oxidant stimuli. Our observations suggest that metabolic control in myelomonocytic cells has profound implications for their inflammatory phenotype and may trigger the development of vascular disease.

Introduction

Among all established cardiovascular risk factors, arterial hypertension has the strongest impact on morbidity and mortality worldwide.¹ Activation of the renin angiotensin aldosterone system plays an important role in this setting as its inhibition (e.g. by angiotensin-converting enzyme inhibitors) is highly effective in preventing atherosclerosis progression and the development of heart failure.² Mechanistically, angiotensin II (ATII) increases vascular reactive oxygen species (ROS) production and causes endothelial dysfunction,³ a condition that is associated with increased cardiovascular events in patients at risk.⁴ Important sources of oxidative stress in the vasculature include the NADPH oxidase, which is expressed in vascular cells but may also originate from myelomonocytic cells that accumulate in the vasculature.⁵ ATII may directly activate vascular expressed NADPH oxidases,⁶ but it is also known as a strong pro-inflammatory stimulus that enhances the recruitment of myelomonocytic cells to the vascular wall.⁷

Recent evidence suggests that toxin-mediated depletion of lysozyme M⁺ cells largely protected from ATII-induced hypertension and associated vascular dysfunction, indicating a fundamental role of these cells in the development of vascular disease.⁸ However, the oxidative burst of these cells plays an important role in host defence, and a complete suppression of their function (e.g. by Nox2 gene silencing) will result in the inability of the body to eliminate pathogens, a condition known as chronic granulomatous disease.^9,10 Therefore, a complete elimination of myelomonocytic cells is not feasible for the prevention of cardiovascular disease, but a modulation of their stress response may be an option in order to reach this goal.

In this respect, it was recently shown that the AMP-activated protein kinase (AMPK)—initially known as a metabolic stress sensor that coordinates energy-consuming and energy producing pathways within the cell—also modulates inflammatory responses.^11–15 For example, activation of AMPK was shown to promote anti-inflammatory and to suppress pro-inflammatory signalling events in macrophages.^16,17 Recent work highlighted the role of α1AMPK in macrophage polarization and associated cytokine release that was required to resolve inflammation in a model of muscle injury in vivo.¹⁸

Therefore, we aimed to investigate whether deletion of α1AMPK specifically in LysM⁺ myelomonocytic cells may modulate vascular inflammation and dysfunction during chronic ATII treatment.

Methods

Animals and blood pressure recordings

All animal studies were in accordance with the Guide for the Care and Use of Laboratory Animals as adopted and promulgated by the U.S. National Institutes of Health and were granted by the ethics committee of the University Hospital Mainz (Az 23 177-07/G 13-1-032). Mice with a deletion of α1AMPK in myelomonocytic cells were created by breeding LysM-Cre⁺ mice⁸ with α1AMPK^fl/fl mice on a C57Bl/6 background.¹⁶ The offspring were genotyped and backcrossed with α1AMPK^fl/fl mice to obtain LysM-Cre⁺ x α1AMPK^fl/fl mice. Experiments were conducted in 8–12 weeks old male LysM-Cre⁺ x α1AMPK^fl/fl mice and corresponding controls (LysM^Cre/wt). LyM-Cre⁺ mice were kept heterozygous for the Cre-recombinase (LysM^Cre/wt), since homozygous knock-in mice have compromised LysM gene expression and altered immune cell function. ATII at an infusion rate of 1 mg/kg/day was administered by subcutaneous implantation of osmotic minipumps (alzet model 1007D) under ketamine/xylazine anaesthesia (1 mL/kg intraperitoneal injection of a ready to use solution containing 80 mg/mL ketamine and 12 mg/mL xylazine, Sigma-Aldrich).

Systolic blood pressure was analysed by invasive recordings. For this purpose, telemetry transmitters were inserted under general anaesthesia (Midazolame 5 mg/kg, Medetomidine 0.5 mg/kg, and Fentanyl 0.05 mg/kg) in the carotid artery, and the tip of the catheter was advanced to the aortic arch.

Animals were sacrificed by exsanguination under isoflurane anaesthesia (5% inhalant in room air) and the thoracic/abdominal aorta, blood, and bone marrow were collected for further analyses as reported. All animal procedures were performed in accordance with the Institutional Animal Care and Use Committee (IACUC Guidelines).

Materials

Total-AMPK antibody was from Cell Signalling (Boston, MA, USA), total-eNOS and iNOS antibody were purchased from BD biosciences (San Jose, CA, USA), CCR-2 antibody was from Abcam (ab203128). For detection of tyrosine-nitration, a mouse monoclonal 3-nitrotyrosine antibody (Upstate Biotechnology, Waltham, MA, USA) was used (dil. 1:1000). Hydrogen peroxide was detected with an Amplex red Assay (Molecular Probes, Eugene, OR, USA). All other chemicals and reagents were of analytical grade purchased from Sigma-Aldrich, Fluka, or Merck. For isometric tension studies, glyceryltrinitrate (GTN) was used from a nitrolingual infusion solution (1 mg/mL) from G.Pohl-Boskamp (Hohenlockstedt, Germany). L-012 was purchased from Wako Pure Chemical Industries (Osaka, Japan). Polyethylene-glycolated superoxide dismutase (PEG-SOD) and PEG-catalase were obtained from Sigma.

Measurement of vascular reactivity and NO levels

Vascular reactivity of aortic rings was measured by isometric tension studies as described previously.¹⁹ Briefly, vasodilator responses to the endothelium dependent vasodilator acetylcholine (ACh), the endothelium-independent vasodilator GTN or the NO-donor DEA-NONOate were assessed with endothelium-intact isolated aortic rings (thoracic aorta, 3 mm length) mounted for isometric tension recordings in organ chambers. Aortic nitric oxide formation was measured in intact aortic rings by using electron paramagnetic resonance (EPR)-based spin trapping with iron-diethyldithiocarbamate (Fe(DETC)₂) colloid as described.²⁰

Blood and vascular reactive oxygen species (ROS) production

A complete blood count (white blood cells, platelets, and haemoglobin) was carried out by using Sysmex automated haematology diagnostics. The oxidative burst in response to phorbolester dibutyrate (PDBu, 10 μM) stimulation was analysed using L-012 (100 μM)-enhanced chemiluminescence in whole blood by a Centro plate reader (Berthold Technologies, Bad Wildbad, Germany). Aortic hydrogen peroxide formation was measured by an amplex red/peroxidase-based HPLC assay as described earlier.⁶ Briefly, aortic ring segments 4 mm in length were incubated with 150 μL of amplex red (100 μM) and HRP (0.2 μM) in PBS with Ca²⁺ and Mg²⁺ under stimulation with myxothiazol (20 μM) for 60 min at 37°C.

We used dot blot technique to analyse 3-nitrotyrosine levels as a surrogate for vascular peroxynitrite generation in aortic tissue.⁵ The detection of vascular superoxide production was performed in 3 mm aortic rings that were incubated in Krebs/Hepes solution for 15 min at 37°C, embedded in aluminium cups containing 1 mL of a polymeric resin (Tissue Tek^®, Sakura Finetek, Alphen aan den Rijn, Netherlands) and frozen in liquid nitrogen. Cryosections (6 μm) were stained with the superoxide-sensitive dye dihydroethidium (DHE, 1 μM in PBS) and incubated for 30 min at 37°C. Green and red fluorescence was detected using a Zeiss Axiovert 40 CFL camera (Zeiss, Oberkochen, Germany).

Reverse transcription real-time PCR (qRT-PCR)

mRNA expression was analysed by quantitative real-time RT-PCR as previously described.⁸ Briefly, total RNA from mouse aorta was isolated (RNeasy Fibrous Tissue Mini Kit; Qiagen, Hilden, Germany), and 50 ng of total RNA was used for real-time RT-PCR analysis with the QuantiTect™ Probe RT-PCR kit (Qiagen). A TaqMan^® Gene Expression assay for TBP, CD11b, CD68, iNOS, IFN-gamma, TNF-alpha, IL-6, Nox2, CCR2, AGTR1 was purchased as probe-and-primer set (Applied Biosystems, Foster City, CA, USA). The comparative ΔΔCt method was used for relative mRNA quantification. Gene expression was normalized to the endogenous control, TATA box binding protein (TBP) mRNA, and the amount of target gene mRNA expression in each sample was expressed relative to that of untreated control animals.

Immunoblotting

Isolated aortic tissue was frozen and homogenized in liquid nitrogen. Proteins were separated by SDS-Page, blotted onto nitrocellulose membranes, and immunoblotting was performed as described previously.⁹

Bone marrow-derived macrophage isolation and culture

Bone marrow cells were isolated from mouse femur and tibia under sterile conditions as described previously.¹⁰ Briefly, bones were flushed with RPMI-1640 plus 10% calf serum. The bone marrow cells were passed through a 22-gauge needle followed by three passages through a 25-gauge needle to obtain a single-cell suspension. Bone marrow cells were cultured in RPMI-1640 medium supplemented with 10% foetal bovine serum, 1% glutamate, 20 μg/mL gentamycin, and 20 ng/mL M-CSF (Pepro-Tech, Hamburg, Germany) for 7 days to induce macrophage differentiation.

Isolation of peritoneal macrophages

Mice received an intraperitoneal injection of 2 mL thioglycolate (3%, Sigma Aldrich, St. Louis, MO, USA), vehicle solution was used as control. After 4 days, mice were euthanized and 5 mL phosphate-buffered saline was injected into the peritoneal cavity, followed by gentle massage of the mouse abdomen for 30 s. Peritoneal cells were collected by aspiration of the phosphate-buffered saline by using 21 gauge needles, immediately spun down at 1200 rpm for 10 min, resuspended in RPMI-1640 supplemented with 10% foetal bovine serum and cultured for 4 h. Non-adherent cells were discharged, adherent cells were considered as peritoneal macrophages.

Flow cytometry of aortic tissue

Aortic vessels were cleaned from adhering fat and minced with fine scissors. Mouse aortas were digested by using 1 mg/mL Liberase TM (Roche Applied Science, Penzberg, Germany) for 30 min at 37°C with constant agitation. Aortas were then passed through a 70 μm sterile cell strainer (Falcon; BD Biosciences, San Jose, CA, USA), yielding single-cell suspensions. Cells were washed twice with 1% BSA PBS buffer and additionally incubated for 30 min at 37°C with complete media (RPMI; 10% FCS), washed and counted. In order to block unspecific Fc receptor-mediated binding, single-cell suspensions were pre-incubated with unlabelled mAb against CD16/CD32. Cells were stained for 20 min with CD45 APC, CD11b PE-Cy7, F4/80 FITC, Gr-1 Pacific Blue, and Ly6c PerCP-Cy5.5 and for outgating of dead cells additionally with Fixable Viability Dye eFluor506 (CD11b from Pharmingen; all other antibodies from eBioscience, San Diego, CA, USA) by using multicolour flow cytometry as described in the previous paragraph. An initial gate was applied to exclude cell debris from further analysis, and CD45 staining was used to identify leucocytes within the aortic cell suspension. Viable cells were analysed with FACSDiva software.

Cell culture

The RAW 264.7 macrophage cell line was purchased from ATCC (ATCC-TIB-71) and cryopreserved in liquid nitrogen until required for the in vitro analyses. Cells were cultured in Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 10% (v/v) fetal bovine serum (FBS) and 1% (v/v) penicillin-streptomycin at 37°C in an incubator with a humidified atmosphere of 5% CO₂.

Transient transfection with small interfering RNA

RAW 264.7 macrophages were initially seeded and grown until confluence reached approximately 60% before transfection process was initiated. After overnight incubation, cells were transfected with scrambled control small interfering (siRNA) or siRNA targeting α1AMPK (Ambion, Catalog# AM16708, Silencer RNA) according to the manufacturer’s instructions (Thermo Fisher, Silencer Libraries: Transfection Protocol). Gene-silencing efficiency was assessed by qRT-PCR after 24 or 48 h of transfection.

Intravital microscopy

Mice were anaesthetized by intraperitoneal injection of Midazolame (5 mg/kg body weight), Medetomidine (0.5 mg/kg body weight), and Fentanyl (0.05 mg/kg body weight) and were fixed on a custom-built stage to maintain a physiological temperature. The right and left common carotid arteries were exposed for microscopic observation. In order to stain circulating leucocytes in vivo, 100 mL of acridine orange (0.5 mg/mL; Sigma Aldrich) was injected via a jugular vein catheter (inner diameter 0.28 mm; outer diameter 0.61 mm; Smiths Medical Deutschland GmbH, Grasbrunn, Germany). Assessment of leucocyte adhesion was performed by a high speed wide field Olympus BX51WI fluorescence microscope using a long-distance condenser and a 10x (numerical aperture, 0.3) water immersion objective with a monochromator (MT 20E; Olympus Deutschland GmbH, Germany) and a charge-coupled device camera (ORCA-R₂, Hamamatsu Photonics). For image acquisition and analysis, a real-time imaging system (eXcellence RT software, Olympus Deutschland GmbH, Germany) was used. Cell recruitment was quantified in four fields of view (100 × 150 μm) per carotid artery. Adherent cells were defined as cells that did not move or detach form the endothelial lining within an observation period of 10 c and were calculated per square millimetre. In part of the experiments, the CCR2 chemokine receptor antagonist RS504393 (5 mg/kg; Tocris Bioscience) was injected i.p. 1 h prior to recordings.

Statistical analysis

Results are expressed as mean ± SD. Statistical analysis was performed by two-way ANOVA with repeated measurements approach (with Tukey correction for comparison of multiple means) for concentration–relaxation curves, for whole blood oxidative burst kinetic traces, and for blood pressure measurements (Figures 2A–D and 3D,E).

Figure 2.

α1AMPK deletion in myelomonocytic cells enhances ATII-induced vascular dysfunction and decreases NO bioavailability. (A–C). Endothelial-dependent relaxation in response to acetylcholine (ACh, A) as well as endothelial independent relaxation in response to nitroglycerine (NTG, B) and the sGC stimulator BAY 41-8543 (C) were analysed by isometric tension studies in intact aortic rings from myelomonocytic cell-specific α1AMPK knockout mice (α1AMPK^fl/fl x LysM-Cre⁺) and corresponding wild-type littermates (LysM^Cre/wt) treated with ATII (1.0 mg/kg for 7 days; n = 8–10 mice/group). Statistical significance was calculated by two-way ANOVA and Tukey post hoc test (for details see Supplementary material online, Figures S7–S9). (D) Systolic blood pressure as assessed by invasive telemetry recordings during ATII infusion for 9 days (n = 9), two-way ANOVA was applied to calculate statistical significance. (E) Representative Western Blot showing total eNOS protein expression and ser1177 eNOS phosphorylation in aortic tissue. (F) Aortic NO production as measured by electron paramagnetic resonance (EPR) using Fe(DETC)₂. Summarized data (n = 9–10) are shown as mean ± SEM [arbitrary units (AU) per mg of dry weight]. Representative EPR spectra were generated from the mean values of all measurements per group. Statistical significance was calculated by one-way ANOVA and Tukey multiple comparison test. *P < 0.05 vs. untreated, ^#P < 0.05 vs. ATII-treated wild-type mice and ^§P < 0.05 vs. α1AMPK^fl/fl x LysM-Cre⁺ for all parts of the figure.

Figure 3.

Vascular oxidative stress is increased in mice lacking myelomonocytic cell α1AMPK. (A) Aortic 3-nitrotyrosine staining was used as a surrogate for peroxynitrite formation and assessed by dot blot technique. Pooled samples (n = 3) of n = 6 mice. (B) Vascular superoxide anion production was analysed by dihydroethidium staining of aortic cryosections. (C) Measurement of hydrogen peroxide formation in aortic tissue was performed using an amplex red-based HPLC assay (n = 4–7). Statistical significance was calculated by one-way ANOVA and Tukey multiple comparison test (A and C). (D) Aortic rings from myelomonocytic cell-specific α1AMPK knockout mice (α1AMPK^fl/fl x LysM-Cre⁺) and corresponding wild-type littermates (LysM^Cre/wt) treated with ATII (1.0 mg/kg for 7 days) were incubated with PEG-SOD ex vivo; endothelial-dependent relaxation in response to acetylcholine (ACh) was analysed by isometric tension studies (n = 8 animals/group). (E) Phorbol ester (PDBu) stimulated oxidative burst in whole blood was determined by L-012 enhanced chemiluminescence (n = 12). Statistical significance was calculated by two-way ANOVA and Tukey post hoc test (for detailed P-values see Supplementary material online, Figures S8 and S9). (F) Nox2 mRNA (n = 4–6) and protein expression in aortic tissue. Pooled samples (n = 3) of n = 6 mice. One-way ANOVA and Tukey multiple comparison test were used to calculate statistical significance. *P < 0.05 vs. untreated, ^§P < 0.05 vs. α1AMPK^fl/fl x LysM-Cre⁺, ^#P < 0.05 vs. ATII-treated wild-type mice for all parts of the figure.

One-way ANOVA (with Tukey correction for comparison of multiple means) was used for comparisons of WBC, platelets, haemoglobin, aortic NO production, 3-nitrotyrosine, aortic hydrogen peroxide production, mRNA expression (except for Figure 6B,F, here we applied an unpaired t-test). An unpaired t-test was also used for the analysis of protein expression (Figure 1A,B).

Figure 6.

α1AMPK deletion in myelomonocytic cells increases CCR2 and AGTR1a expression. mRNA expression of CCR2 and the angiotensin II type 1a receptor (AGTR1a) was measured by real-time PCR from aortic tissue (A and E) and peritoneal macrophages (PM) (B and F) (n = 6). Effects of α1AMPK siRNA on markers of inflammation in RAW 264.7 macrophages (C). Statistical significance was calculated by one-way ANOVA and Tukey multiple comparison (A and E) and unpaired t-testing was applied for (B and F). Epifluorescence intravital microscopy was used to visualize adherent and rolling leucocytes in carotid arteries. Nucleated cells were stained with acridine orange (green fluorescence). Representative images of carotid arteries after 7 days of continuous ATII infusion (n = 8), before and 60 min after single intraperitoneal injection of anti-CCR2 antibody (CCR2ab) (C and D). For mRNA expression, P-values are shown in each figure.

Figure 1.

Assessment of macrophage α1AMPK protein expression and total blood count after myelomonocytic cell-specific α1AMPK deletion. (A, B) Immunoblotting showing α1AMPK protein expression in thioglycolate elicited peritoneal macrophages (PM) (A) and bone marrow-derived macrophages (B) from myelomonocytic cell-specific α1AMPK knockout mice (α1AMPK^fl/fl x LysM-Cre⁺) and wild-type littermates (LysM^Cre/wt). Representative western blot shown. Total of n = 10 mice, two animals pooled for one lane, unpaired t-test was used for statistical analysis. (C–E) Total white blood cell count, platelet count, and haemoglobin concentrations. n = 9–13 animals; statistical significance was calculated by one-way ANOVA and Tukey comparison test. P-values are displayed in each figure.

Kruskal–Wallis followed by Dunn’s method was used for the analysis of maximal aortic relaxation (NTG, see Supplementary material online, Figure S9). Friedman test was used for statistical analysis of EC50 (PEG-SOD and BAY) and maximal relaxation (BAY) calculation. P-values < 0.05 were considered statistically significant and are either provided in the figures or figure legends. All statistical calculations were performed by Prism for Windows, version 7.04, GraphPad Software Inc.

Results

Characterization of myelomonocytic cell-specific α1AMPK knockout mice

α1AMPK deletion in LysM⁺ cells resulted in a 60% decrease of α1AMPK protein expression in bone marrow-derived macrophages (BMDM) as compared to LysM^Cre/wt control mice (Figure 1B). In thioglycollate-elicited peritoneal macrophages, we found a similar extent of α1AMPK protein suppression (Figure 1A, Supplementary material online, Figure S1). Importantly, LysM⁺ cell-specific α1AMPK deletion caused a decrease in white blood cell count (Figure 1C) while platelet count (Figure 1D) and haemoglobin concentrations (Figure 1E) were not affected by myeloid cell-specific α1AMPK deletion.

Myelomonocytic cell-specific α1AMPK deletion aggravates ATII-induced vascular dysfunction

ATII induced endothelial dysfunction was significantly impaired in LysM-Cre⁺ x α1AMPK^fl/fl mice as compared to LysM^Cre/wt control mice, while ATII already induced a significant degree of endothelial dysfunction in wild-type mice (Figure 2A). Similarly, endothelial-independent relaxation was reduced in these mice on top of an existing loss of NTG potency in wild-type mice (Figure 2B). Blood pressure measurements during chronic ATII treatment (1 mg/kg/day) showed a similar increase of systolic blood pressure in mice lacking α1AMPK in myelomonocytic cells (Figure 2D) as compared to controls. Since altered NO bioavailability may account for the observed changes in endothelial function, we next assessed aortic NO levels by EPR. Although LysM⁺ cell-specific α1AMPK deletion did not cause changes in endothelial NO-synthase (eNOS) expression or phosporylation at ser1177 (Figures 2E, Supplementary material online, Figure S5), we observed a significant decrease of NO-Fe(DETC)₂ adduct formation (Figure 2F), indicating diminished vascular NO bioavailability in these animals. This observation was supported by a similar vascular relaxation to the sGC stimulator BAY 41-8543 in both ATII-treated groups (Figure 2C), suggesting that an appreciable part of the observed impairment of endothelial function in ATII-treated LysM-Cre⁺ x α1AMPK^fl/fl mice must be due to inefficient sGC activation (e.g. by less bioavailable NO) or impaired sGC activity. The remaining impairment of BAY 41-8543 potency may be related to direct structural changes in the vascular wall caused by ATII.

Increased reactive oxygen species production in the vasculature of mice lacking myelomonocytic α1AMPK

Since vascular NO may be inactivated by an increased production of superoxide anions, we next assessed aortic 3-nitrotyrosine levels as a marker of peroxynitrite formation. ATII treatment caused a significant stronger increase of aortic 3-nitrotyrosine in LysM-Cre⁺ x α1AMPK^fl/fl mice than in controls (Figure 3A), indicating enhanced NO consumption by superoxide anions. Accordingly, we observed increased superoxide anion production in all layers of the vascular wall by using DHE staining of aortic cryosections (Figure 3B). This finding was paralleled by increased hydrogen peroxide formation assessed by amplex red assay (Figure 3C), suggesting that the global oxidative stress in aortic tissue from LysM-Cre⁺ x α1AMPK^fl/fl mice was significantly augmented. Mechanistically, ROS scavenging by in vitro treatment of intact aortic rings with PEG-SOD (but not with PEG-catalase) partially restored endothelial dependent relaxation, indicating that increased superoxide generation in these animals contributes to the observed impairment of vascular function (Figure 3D, Supplementary material online, Figure S2). Besides the vascular NADPH oxidase, phagocytic NADPH oxidases may contribute significantly to vascular ROS production as these cells infiltrate the vascular wall under pro-inflammatory stimuli such as ATII. As α1AMPK deletion in myelomonocytic cells may specifically affect their ROS-producing capacity, we analysed the phagocytic oxidative burst after in vitro stimulation with the protein kinase C activator PDBu and found an increased susceptibility towards pro-oxidant stimuli in myelomonocytic cells from LysM-Cre⁺ x α1AMPK^fl/fl mice (Figure 3E). This was supported by the observation that myelomonocytic α1AMPK deletion lead to an increased abundance of aortic Nox2 mRNA and protein expression (Figure 3F), which may originate from myelomonocytic cells infiltrating the vascular wall.

α1AMPK deletion in myelomonocytic cells enhances their accumulation in the vascular wall

In order to prove an α1AMPK-dependent recruitment of myelomonocytic cells to the vascular wall, we next characterized number and subgroups of these cells in aortic tissue. After chronic ATII treatment, we observed a higher abundance of aortic CD11b mRNA and CD68 mRNA in aortas from myelomonocytic cell-specific α1AMPK knockout mice, suggesting an increased presence of innate immune cells in the vasculature (Figure 4A,B).

Figure 4.

α1AMPK deletion causes increased vascular infiltration of myelomonocytic cells during chronic ATII infusion. (A, B) Aortic mRNA expression of CD11b (A) and CD68 (B) was assessed in myelomonocytic cell-specific α1AMPK knockout mice (α1AMPK^fl/fl x LysM-Cre⁺) and corresponding wild-type littermates (LysM^Cre/wt) by real-time PCR (n = 6). Statistical significance was calculated by one-way ANOVA and Tukey multiple comparison test. (C) Flow cytometry from aortic tissue showing a representative plot for CD45+ leucocyte infiltration and CD11b+/F4/80+ macrophages (D). Flow cytometry event plots from aortic tissue showing infiltration of neutrophils (E) and inflammatory monocytes (Ly6c high/intermediate/low) (F) after 7 days of ATII infusion (n = 6). P-values are displayed in A and B.

To study the number and nature of these inflammatory cells infiltrating the vasculature, we analysed aortic single-cell suspensions by flow cytometry. We found a significant vascular accumulation of CD45⁺ leucocytes in response to ATII treatment in LysM^Cre/wt control mice, that was 2.5-fold increase in aortas from LysM-Cre⁺ x α1AMPK^fl/fl mice. A further characterization of these cells revealed that CD11b⁺/F4/80⁺ macrophages as well as CD11b⁺/Gr1⁺ neutrophils were particularly prominent in aortic tissue (Figure 4C–E). Since CD11b⁺/F4/80⁺ macrophages play a pivotal role in the development of ATII-induced hypertension and vascular dysfunction, we investigated monocyte subsets (Ly6C high/intermediate/low) and found a predominant increase of Ly6C^int positive cells in the vessel wall of LysM-Cre⁺ x α1AMPK^fl/fl mice (Figure 4F). In order to prove the concept that a significant portion of Nox2-derived oxidative stress originates from these infiltrating macrophages, we performed aortic immune fluorescence histology showing an increased Nox2/F4/80 double staining in ATII-treated LysM-Cre+ x α1AMPKfl/fl mice (Supplementary material online, Figure S3).

LysM-Cre⁺ x α1AMPK^fl/fl mice display a pro-inflammatory aortic cytokine pattern in response to ATII

The inflammatory response triggered by macrophages involves large amounts of NO produced by their inducible NO-synthase (iNOS). In accordance with the enhanced macrophage infiltration in ATII treated LysM-Cre⁺ x α1AMPK^fl/fl mice, we observed a significantly elevated aortic iNOS mRNA expression (Figure 5A).

Figure 5.

Vascular cytokine patterns reveal a pro-inflammatory phenotype in myelomonocytic cell-specific α1AMPK knockout mice treated with ATII. Aortic iNOS (A) mRNA expression (n = 11–12) and pro-inflammatory cytokine levels [TNF-α (n = 4) (B), IFN-γ (n = 10) (C), and IL-6 (n = 4) (D)] were measured by real-time PCR. Statistical significance was calculated by one-way ANOVA and Tukey multiple comparison test, P-values are shown in each figure.

The release of cytokines is necessary to initiate and propagate the inflammatory response and can be used as signature of infiltrating immune cells. In this respect, recent publications have highlighted the diversity of macrophage polarization reflecting a distinctive activation phenotype of these heterogeneous immune cells.¹¹ Since macrophages lacking the α1AMPK subunit are disturbed in switching from a M1 polarized phenotype towards a M2 polarization in the setting of inflammatory resolution,¹² we investigated aortic expression of interleukin-6, TNF-alpha, and IFN-gamma as indicators of an M1-polarization macrophage phenotype.¹³ Aortic mRNA levels of all these cytokines were significantly increased in ATII treated LysM-Cre⁺ x α1AMPK^fl/fl mice, supporting a pro-inflammatory M1 macrophage phenotype in these animals (Figure 5B–D).

α1AMPK deletion in myelomonocytic cells increases CCR2 and AGTR1a dependent vascular inflammation

Our previous results show that myelomonocytic cell-specific α1AMPK deletion leads to a pro-inflammatory milieu in the vasculature, while the mechanisms underlying increased vascular infiltration and ROS-producing capacity of these cells remain uncertain. We therefore analysed C-C chemokine receptor type 2 and angiotensin II receptor subtype 1a (AGTR1a) abundance in aortic tissue and isolated macrophages, as they are important regulators of immune cell invasion and their oxidative burst.

CCR2 is expressed in myelomonocytic cells and constitutes a receptor for monocyte chemoattractant protein-1 (MCP-1), which mediates monocyte chemotaxis. We found that CCR2 mRNA expression was elevated in aortas of ATII-treated control mice, but significantly further increased in LysM-Cre⁺ x α1AMPK^fl/fl mice (Figure 6A). Since F4/80+ macrophages were the main infiltrating immune cell lineage in our model, we analysed CCR2 expression specifically in these cells and observed a significantly higher CCR2 expression in macrophages from LysM-Cre⁺ x α1AMPK^fl/fl mice than from corresponding wild-type mice (Figure 6B). In order to dissect whether this change in macrophage phenotype is directly related to α1AMPK abundance or caused by changes in the extracellular milieu, cultured RAW 264.7 macrophages were pretreated with α1AMPK siRNA and challenged with ATII in vitro. α1AMPK gene silencing resulted in an increased expression of iNOS and CCR2 protein, indicating that basal levels of myelomonocytic α1AMPK are required to suppress a pro-inflammatory response in these cells (Figure 6C). In accordance with this observation, pharmacological inhibition of CCR2 by a selective antagonist prevented leucocyte adhesion as shown by intravital microscopy of carotid arteries (Figure 6D). In summary, our data support a causal involvement of CCR2 upregulation for the enhanced vascular infiltration of myelomonocytic cells (as shown by flow cytometry experiments, see Figure 4).

LysM⁺ cells deficient in α1AMPK displayed a pro-oxidant phenotype that may be a consequence of enhanced signalling pathways leading to their NADPH oxidase activation. A major signalling event to activate vascular or phagocytic NADPH oxidases involves AT-1 receptor dependent activation of protein kinase C. Intriguingly, we found an increased expression of the AT-1 receptor subtype AGTR1a in vascular tissue of ATII treated LysM-Cre⁺ x α1AMPK^fl/fl mice (Figure 6E) and macrophages isolated from LysM-Cre⁺ x α1AMPK^fl/fl mice (Figure 6F). Consistent with the enhanced oxidative burst shown in Figure 3E, AGTR1a upregulation may cause facilitated ROS production in these cells.

These results indicate that α1AMPK in LysM⁺ cells affects vascular ROS production by two different mechanisms, namely enhanced recruitment of myelomonocytic cells to the vasculature and an increased susceptibility towards pro-oxidant stimuli.

Discussion

The data presented here indicate that selective modulation of the myelomonocytic cell phenotype has profound implications on vascular function in a pro-oxidant milieu. During chronic ATII treatment, deletion of α1AMPK in LysM⁺ cells increased their capacity to produce ROS and to invade the vascular wall. Increased responsiveness to ATII stimulation (by upregulation of AGTR1a) and facilitated chemotaxis to the vascular wall (by upregulation of CCR2) were key events resulting from α1AMPK deletion in myelomonocytic cells. In vitro ROS scavenging by PEG-SOD was able to improve endothelial dysfunction in ATII-treated LysM-Cre⁺ x α1AMPK^fl/fl mice, indicating a mechanistic role of oxidative stress in this setting. In contrast to PEG-SOD, PEG-catalase had no effect on ATII-induced endothelial dysfunction in LysM-Cre⁺ x α1AMPK^fl/fl mice (Supplementary material online, Figure S2). Of note, H₂O₂ plays a dual role for vasodilation. While high concentrations are clearly associated with vascular dysfunction (e.g. by S-glutathionylation dependent uncoupling of eNOS or activation of redox-sensitive kinases such as protein kinase C), low concentrations were reported to cause vasodilation and play a role in endothelial-derived growth factor signalling. Since H₂O₂ changes were small in our model (see Amplex Red Assay; Figure 3C), this could explain why PEG-catalase was not able to correct endothelial dysfunction in this setting.

The vascular effects of ATII treatment result in endothelial dysfunction, mainly driven by an increased production of ROS and in particular superoxide anions. The activation of the NADPH oxidase complex with its subunits plays a crucial role in this process, while the relative importance of vascular vs. phagocytic NADPH oxidases is still a matter of controversy. Our study clearly supports the importance of phagocytic derived ROS for the development of ATII-induced vascular dysfunction.

Previous work highlighted the central role of a phagocytic NADPH oxidase for the development of ATII-induced hypertension and vascular dysfunction.⁸ In that study, ablation of LysM-positive cells largely prevented blood pressure increase in response to ATII. Interestingly, reconstitution with wild-type monocytes re-established the hypertensive phenotype, while monocytes lacking the AGTR1a receptor or Nox2 did not.⁸ With respect to vascular function, our observations support the importance of ROS-producing capacities in myelomonocytic cells, since we observed an upregulation of AT-1 receptor expression as well as an enhanced oxidative burst in myelomonocytic cells lacking α1AMPK. The facilitated ROS generation had important functional consequences, as ROS scavenging by PEG-SOD resulted in a significant improvement of endothelial function. In addition, the observed changes in cytokine expression patterns (IL-6, TNF-alpha, IFN-gamma) may also contribute to the increased responsiveness towards pro-oxidant stimuli in these cells. Functional immune cell modulation resulting from increased AMPK activity has recently been demonstrated since AMPK activation by AICAR promoted innate immune defence in an in vivo model of bacterial endophthalmitis.²¹ However, our observations were generated in a loss-of-function model, and it remains to be established whether a gain of α1AMPK activity has the desired opposite (beneficial) effects on vascular function.

Although α1AMPK is critically involved in the regulation of vascular endothelial function, its specific deletion in myelomonocytic cells or ATII infusion did not affect overall vascular α1AMPK protein expression (Supplementary material online, Figure S4). However, we observed an increased AMPK activity and α1AMPK protein expression in raw macrophages after in vitro ATII stimulation (Supplementary material online, Figure S6). This discrepancy might be due to tissue-specific actions of ATII or differences for short-term vs. continuous ATII administration.

In our study, myelomonocytic cell-specific α1AMPK deletion did not affect blood pressure increase in response to ATII. Since the development of hypertension requires vascular infiltration of myelomonocytic cells with an intact ROS-producing machinery (AT-1 receptor, Nox2)⁸, it is tempting to speculate that an enhanced vascular recruitment of these cells and responsiveness towards pro-oxidant stimuli could increase blood pressure in mice lacking myelomonocytic cell α1AMPK. However, the used dose of ATII in the current study might be too high in order to unmask this difference. In addition, the efficiency of myelomononcytic α1AMPK gene silencing was approximately 60% in our animal model, and it remains uncertain whether a complete suppression of α1AMPK activity may have affected blood pressure in response to ATII.

Myelomonocytic cells are an integral part of the immune system and their absence has serious consequences for host defence. Therefore, strategies that target immunological processes in order to prevent vascular disease should not compromise the protection against infectious diseases. This may be achieved by reversible inhibition of immune cell function or a modulation of myelomonocytic cell function rather than their complete ablation, although we did not assess the immunological consequences of α1AMPK deletion in LysM⁺ cells in detail. Interestingly, myeloid cell-specific α1AMPK deletion resulted in decreased circulating white blood cells. Since previous work by Dai et al.²² demonstrated a crucial role of the α1AMPK subunit for the prevention of apoptosis in myeloid cells, our observation might be related to pro-apoptotic signalling in myeloid cells. This finding is in line with previous work of our group showing that AMPK activation suppresses ROS-mediated c-Jun N-terminal kinase (JNK) activation and endothelial cell death. Interestingly, attenuation of JNK-signalling was absent in mice lacking α1AMPK.²³ However, apoptosis in response to ATII involves increased ROS production and stimulation of redox-dependent signalling cascades including mitogen-activated protein kinases (MAPK)^24.25 Therefore, the modulation of ATII triggered oxidative stress—e.g. by myeloid cell α1AMPK deletion—will also affect apoptosis.

Today it is our common understanding that macrophages can switch from a pro-inflammatory M1 to an anti-inflammatory M2 phenotype and vice versa, a phenomenon called macrophage plasticity or polarization. Previous work by Mounier et al.¹⁸ demonstrated that deletion of myelomonocytic α1AMPK shifted macrophages to a M1 phenotype in a model of muscle injury, resulting in impaired phagocytosis of necrotic and apoptotic cells at the time of resolution of inflammation. Another study showed that the α1AMPK subunit in macrophages was associated with an anti-inflammatory immune cell phenotype.²⁶ Our results are in line with these observations, since α1AMPK deletion in myelomonocytic cells resulted in polarization to a M1 phenotype indicated by surface markers (Ly6c^int) and cytokine patterns. Initially, macrophage polarization was thought to be a consequence of environmental stimuli including cytokines such as IFN-γ or TNF-α for development of a M1 phenotype. Besides that, intracellular signalling events in macrophages itself may determine their plasticity. Several mediators have been identified that affect macrophage transition from a M1 to a M2 phenotype including cAMP response element-binding protein, mitogen-activated protein kinase phosphatase-1, Akt, hypoxia-inducible factor and others.²⁷ Among these signalling events, α1AMPK may be unique since it is closely associated with macrophage metabolism, and metabolic reprogramming is an essential prerequisite for macrophage polarization as discussed below.

Since changes of AMPK activity and cellular metabolism often go in parallel, it remains a major difficulty to assess whether the macrophage phenotype is primarily caused by an altered cell metabolism or directly by AMPK activity. The inflammatory response triggered by tissue injury or infection is an energy demanding process that requires enhanced glucose uptake and glycolysis in activated macrophages. In contrast, anti-inflammatory M2 macrophages have a higher rate of oxidative metabolism. These adaptations are necessary in order to allow an inflammatory response and are governed by AMPK.²⁸ On the other hand, an interesting hypothesis by O’Neill and Hardie suggests that metabolism may also modulate inflammation via AMPK. With respect to energy regulation of the whole body, they speculate that an underfed state will promote AMPK activation through release of hormones such as adiponectin, while hormones associated with an overfed state (e.g. insulin, leptin) are known to inhibit AMPK.²⁹ These changes may explain AMPK suppression in overweight subjects and could therefore constitute a link between obesity and inflammation.

In addition to its effects on metabolism, AMPK modulates inflammation by the release of pro-inflammatory cytokines (e.g. TNF-alpha, interleukin-6) or enhanced CCR2 expression in macrophages.^30,31 Our data extend these findings, as altered CCR2 expression in α1AMPK deficient LysM⁺ cells augmented their recruitment to the vascular wall in vivo. The causal role of CCR2 upregulation was proven by intravital microscopy, as a highly specific CCR2 antagonist completely prevented the increased leucocyte adhesion observed in ATII-treated LysM-Cre⁺ x α1AMPK^fl/fl mice.

In addition, we identified an increased responsiveness towards pro-oxidant stimuli (by upregulation of AGTR1a) as a new factor that contributes to the pro-inflammatory phenotype seen during α1AMPK deficiency. Since most effects of ATII are mediated by angiotensin II type 1 receptor (AT-1) binding, its expression determines the biological efficacy of ATII. AT-1 receptor gene expression is predominantly regulated by its transcription rate and AT-1 mRNA stability.³² However, the precise mechanism how α1AMPK modulates AT-1 receptor expression remains unclear at present and needs further investigation. Importantly, pharmacological interventions in order to suppress undesired ATII signalling events exist today and future studies will reveal whether co-treatment with angiotensin receptor blockers may reverse some of the pro-inflammatory effects caused by myelomonocytic α1AMPK deletion.

Time for primary review: 44 days