OUP user menu

The absence of macrophage Nrf2 promotes early atherogenesis

Anna-Kaisa Ruotsalainen, Matias Inkala, Mervi E. Partanen, Jari P. Lappalainen, Emilia Kansanen, Petri I. Mäkinen, Suvi E. Heinonen, Heidi M. Laitinen, Janne Heikkilä, Tero Vatanen, Sohvi Hörkkö, Masayuki Yamamoto, Seppo Ylä-Herttuala, Matti Jauhiainen, Anna-Liisa Levonen
DOI: http://dx.doi.org/10.1093/cvr/cvt008 107-115 First published online: 22 January 2013

Abstract

Aims The loss of nuclear factor E2-related factor 2 (Nrf2) has been shown to protect against atherogenesis in apoE-deficient mice. The mechanism by which Nrf2 deficiency affords atheroprotection in this model is currently unknown, but combined systemic and local vascular effects on lesion macrophages have been proposed. We investigated the effect of bone marrow-specific loss of Nrf2 on early atherogenesis in low-density lipoprotein (LDL) receptor-deficient (LDLR−/−) mice, and assessed the effect of Nrf2 on cellular accumulation of modified LDLs and the expression of inflammatory markers in macrophages.

Methods and results The effect of bone marrow-specific loss of Nrf2 on atherogenesis was studied using bone marrow transplantation of wild-type (WT) or Nrf2−/− bone marrow to LDLR−/− mice. Mice transplanted with Nrf2−/− bone marrow and fed a high-fat diet for 6 weeks exhibited significantly larger atherosclerotic lesions than WT bone marrow transplanted mice. Moreover, in thioglycollate-elicited Nrf2−/− macrophages, the uptake of acetylated and malondialdehyde-modified LDLs was increased in comparison with WT controls, with the concomitant increase in the expression of scavenger receptor A and toll-like receptor 4. In addition, the expression of pro-inflammatory monocyte chemoattractant protein-1 and interleukin-6 were increased in Nrf2−/− vs. WT macrophages.

Conclusion Nrf2 deficiency specific to bone marrow-derived cells aggravates atherosclerosis in LDLR−/− mice. Furthermore, the loss of Nrf2 in macrophages enhances foam cell formation and promotes the pro-inflammatory phenotype.

  • Atherosclerosis
  • Nrf2
  • Macrophage
  • Bone marrow transplantation

1. Introduction

Atherosclerosis is a chronic inflammatory disease characterized by the accumulation of lipids and persistent inflammatory response within the arterial intima, leading to the formation of an atherosclerotic lesion. The contribution of oxidative stress, an imbalance between the production and disposal of reactive oxygen species (ROS), to the pathogenesis of atherosclerosis has been proposed. ROS have direct adverse effects on vascular function and integrity, but they can also oxidize low-density lipoprotein (LDL) entrapped in the arterial wall. Oxidized LDL (oxLDL) in turn can promote atherogenesis via multiple mechanisms, e.g. by serving as a ligand for scavenger receptors such as SR-A and CD36 resulting in the dysregulated uptake of oxLDL into intimal macrophages. These cells then transform into cholesterol-laden foam cells, a hallmark of both early as well as late advanced atherosclerotic lesions.1 In addition, oxLDL activates macrophages as well as endothelial and smooth muscle cells to produce pro-inflammatory mediators, thereby aggravating inflammation within the lesion.

The balance between the formation and disposal of ROS can be maintained by the induction of endogenous antioxidant enzymes as well as enzymes involved in the synthesis of small molecular weight antioxidants. Upon increased oxidative burden, such enzymes can be regulated in an integrated manner by the transcription factor Nuclear factor E2-related factor 2 (Nrf2), which is responsive to both environmental factors and endogenous stimuli. Interestingly, many Nrf2-regulated genes, such as haeme oxygenase-1 (HO-1), glutathione reductase, and peroxiredoxin-1 and -2, have been shown to be protective in mouse models of atherosclerosis.26 Furthermore, Nrf2 has been shown to mediate adaptive augmentation of antioxidant defences of vascular cells on exposure to oxLDL as well as oxidized phospholipids that can be found in atherosclerotic lesions at biologically relevant concentrations.79 Supporting the atheroprotective role of Nrf2, it is activated by laminar shear stress, a potent anti-atherogenic force, in vitro as well as in vivo.1014 Also, Nrf2 gene transfer inhibits vascular inflammation and oxidative stress in a rabbit balloon injury model,15 further strengthening the notion that Nrf2 protects against vascular diseases.

In contrast to these reported protective actions, Nrf2 has also been ascribed potentially pro-atherogenic functions. It has recently been reported by three independent research groups that Nrf2-deficient mice when crossed with apoE-null hypercholesterolaemic mice are protected against atherosclerosis.1618 Although the proposed mechanism of atheroprotection varies from one report to another, both systemic17 and local vascular effects1618 have been suggested. Specifically, Nrf2-dependent expression of macrophage scavenger receptor CD36 and enhancement of foam cell formation16,17 and the activation of Nrf2 by cholesterol crystals leading to the induction of pro-inflammatory IL-1 in macrophages18 suggested a seminal role of lesion macrophages in the observed phenotype. These findings prompted us to examine the role of macrophage Nrf2 expression on atherogenesis using bone marrow transplantation of Nrf2-deficient (Nrf2−/−) or wild-type (WT) bone marrow to lethally irradiated LDL receptor-deficient (LDLR−/−) mice.

Our results show that the absence of Nrf2 in bone marrow-derived cells aggravates early atherosclerosis in LDLR−/− mice. Furthermore, the loss of Nrf2 increases the uptake of modified LDL and the expression of inflammatory markers in thioglycollate-elicited peritoneal macrophages, suggesting that both macrophage lipid accumulation and inflammation are affected by Nrf2.

2. Methods

2.1 Bone marrow transplantation

To examine the effect of macrophage-specific loss of Nrf2 on atherogenesis, Nrf2−/− or WT bone marrow was transplanted to 12-week-old LDLR−/− female mice (n = 18) in C57BL/6 background (The Jackson Laboratory, Bar Harbor, ME, USA). Before bone marrow transplantation, mice were irradiated with a medical linear accelerator (Varian Clinac 600C, Varian medical systems, Palo Alto, CA, USA) to a dose of 5.5 Gy twice using 4 MV photon irradiation. WT C57BL/6J and Nrf2−/− mice19 were backcrossed to C57Bl/6J background and used as bone marrow donors. WT (n = 4) and Nrf2−/− (n = 4) mice were sacrificed with carbon dioxide at the age of 12 weeks, and their femurs and tibias were prepared for bone marrow isolation. Bones were removed and placed in ice-cold Hank's buffered salt solution (HBSS) in 10% foetal bovine serum (FBS). Bone marrow was flushed out with HBSS in 10% FBS. Recipient mice were anaesthetized by sc injection of ketamine (75 mg/kg) and medetomidine (1 mg/kg) for bone marrow transplantation. The response to toe pinch and the cornea reflex were monitored to ensure the adequacy of anaesthesia. Transplantation was performed with 5 × 106 cells per mouse 24 h after irradiation, followed by a 2-week recovery period on a chow diet, after which the mice were fed with a high-fat diet (TD 88137, Harlan Teklad: 42% of calories from fat and 0.15% cholesterol) for 6 weeks. The mice were then sacrificed with carbon dioxide for analyses. Mice were housed in sterilized cages under controlled conditions for temperature and humidity, using a 12 h light/dark cycle and given unlimited access to food and sterilized water. The experiments were approved by the National Animal Experiment Board Finland and conform to the Guide for the Care and Use of Laboratory Animals published by the United States National Institutes of Health. For details of the assessment of chimerism, see Supplementary material online, methods.

2.2 Measurements of atherosclerosis

For the analysis of atherosclerosis, tissues were fixed in 4% paraformaldehyde (PFA). Aortic sinuses were embedded in paraffin for cross-sectional analysis. The aortic sinus level cross-sections were stained with haematoxylin-eosin, and the average lesion area was quantified from every fifth serial cross-section (n = 5) per mouse. Similarly, for the analysis of the morphology and macrophage content of atherosclerotic lesions, every fifth serial cross-section (n = 5) from the aortic sinus level was stained with a modified Movat's pentachrome stain or immunostained with mMQ (mMQ AIA31240, 1 : 6500, Accurate Chemical & Scientific Corp.) or MCP-1 (ab21667, 1:500, Abcam) antibodies. Photographs of histological sections were taken using an Olympus AX70 microscope (Olympus Optical) and analyses were performed in a blinded fashion with the AnalySIS software (Soft Imaging System GmbH).

2.3 Preparation of LDL

LDL was isolated from normal healthy human PLASMA (Finnish Red Cross BLOOD SERVICE) by sequential ultracentrifugation in the presence of EDTA.20 The study was approved by the Ethics Committee of the Finnish Red Cross Blood Service (approval reference number 42/2010). Isolation conformed to the principles of Declaration of Helsinki for the use of human tissues. For details of oxidation, acetylation, and malondialdehyde (MDA) modification of LDL, see Supplementary material online, methods.

2.4 Assessment of macrophage uptake of modified LDL particles

Nrf2−/− and WT mice were treated with 1 mL 3% thioglycollate broth (Sigma-Aldrich) ip injection and sacrificed with carbon dioxide for peritoneal macrophage isolation after a 4-day treatment. Macrophages were grown in RPMI 1640 medium (Sigma-Aldrich) supplemented with 5% FBS in 24-well culture plates and the medium was changed to Optimem (Gibco®) supplemented with 10% lipoprotein-deficient serum (LPDS) for 4 h. To compare the ability of macrophages to take up modified LDL, macrophages were incubated for 12 h at 37°C in the absence or presence of acLDL (50 µg protein/mL). Cellular lipids were extracted twice with 3:2 (v/v) n-hexane:isopropanol and fractions were pooled. Solvent from extracted lipids was evaporated under nitrogen and the extract re-dissolved in 2:1 (v/v) chloroform:methanol. The cholesterol concentration was analysed by using the enzymatic method (total cholesterol, CHOD-PAP 1489232 kit, Roche Diagnostics GmbH, Mannheim, Germany). For the total protein analysis, cells were lysed with 0.2 M NaOH, incubated overnight +4°C and then 1 h at room temperature on a shaking platform, and the protein concentrations were determined using the DC Protein Assay (BioRad). In addition, macrophages grown on chamber slides (BD FalconTM) and grown for 4 h in Optimem (Gibco®) with 10% LPDS before incubating at + 37°C in the absence or presence of acLDL or MDA-modified LDL (50 µg protein/mL) for 12 h. Cells were fixed with 4% PFA and stained with Oil Red O. The total lipid area per cell was calculated by the AnalySIS software (Soft Imaging System GmbH).

2.5 Reverse cholesterol transport

Peritoneal macrophages from Nrf2−/− and WT mice were maintained in RPMI 1640 medium (5% FBS) in 24-well culture plates and grown in Optimem (Gibco®) with 10% LPDS for 4 h before incubating for 16 h at + 37°C in the presence of 3H-labelled-acLDL (10 µg protein/mL). After 3H-labelled-acLDL loading, the cells were incubated with human HDL2 (25 μg/mL) or apoAI (10 μg/mL) for 4 h at + 37°C. The cell medium was removed and the radioactivity was measured with a beta counter (Microbeta, Perkin Elmer). Cells were lysed with 0.2 M NaOH, and protein concentrations of the cell lysates were determined using a protein assay (DC Protein Assay, BioRad). The fractional cholesterol efflux in each well was expressed as the percentage radioactivity in the medium relative to the total radioactivity contained in the medium and the cells.

2.6 Macrophage gene expression

Total RNA was extracted using TRI® reagent (Sigma, USA) from Nrf2−/− and WT peritoneal macrophages maintained in RPMI 1640 medium (5% FBS) in 24-well culture plates and grown in Optimem (Gibco®) with 10% LPDS for 4 h before incubating at + 37°C in the absence or presence of native LDL (natLDL), oxLDL, or acLDL (50 µg protein/mL) for 12 h. For real-time quantitative RT–PCR, total RNA was reverse transcribed into cDNA using random hexamers (Promega) and M-MuLV reverse transcriptase (MBI Fermentas). Quantitative measurements of gene expression were done using appropriate Taqman® Assays-on-demand gene expression products (Applied Biosystems) with the StepOnePlusTM Real-Time PCR System (Applied Biosystems). Measurements were done as triplicates. The expression levels were normalized to β2-microglobulin expression and presented as a fold change in the expression vs. control.

2.7 ELISA

Peritoneal macrophages from Nrf2−/− and WT mice were maintained in RPMI 1640 medium (5% FBS) in 24-well culture plates and grown in Optimem (Gibco®) with 10% LPDS for 4 h before incubating at +37°C in the presence of natLDL, oxLDL, or acLDL (50 µg protein/mL) for 12 h. After incubation, the cell medium was collected and MCP-1 levels were detected with Mouse MCP-1 ELISA Set (BD OptEIATM, Cat.No. 555260) according to the manufacturer's protocol.

2.8 Western blotting

Nrf2−/− and WT peritoneal macrophages treated with natLDL, oxLDL, or acLDL (50 µg protein/mL) for 12 h, were lysed, and protein concentrations were measured with the BCA kit (Pierce). An equal amount of protein in each lane (8 μg) was used for electrophoresis. Proteins were transferred to nitrocellulose membrane, blocked for 1 h at room temperature with 5% horse serum in TBS-Tween, and incubated overnight at +4°C with primary goat polyclonal CD36 antibody (CD36/SR-B3 antibody R&D, AF2519) and β-actin antibody (Cell Signaling). Blots were detected by incubating the blots with HRP-conjugated secondary antibody (Thermo Scientific) and by using ECL Plus Western blotting Detection System (GE Healthcare) with Typhoon 9400 (GE Healthcare). Images were processed with ImageQuantTM TL (GE Healthcare).

2.9 Analysis of plasma cholesterol and triglycerides

After an overnight fasting period, mice were anaesthetized by sc injection of ketamine (75 mg/kg) and medetomidine (1 mg/kg), and ∼100 μL of blood was drawn from each individual mouse by tail bleeding before and after the high-fat diet. The concentrations of total cholesterol and triglycerides in EDTA plasma were determined by Konelab TM/T Series—assays.

2.10 Statistical analyses

To evaluate statistical significance (P < 0.05), independent samples two-tailed t-test was used. Numerical values for each measurement are shown as mean ± SD. All statistical analyses were performed using GraphPad Prism version 4.00 (GraphPad Software).

3. Results

3.1 Nrf2 deficiency specific to bone marrow-derived cells increases the aortic root cross-sectional lesion area in LDLR−/− mice

To examine whether Nrf2 deficiency confined to bone marrow-derived cells impacts early atherogenesis, bone marrow transplantation of WT or Nrf2−/− bone marrow to LDLR−/− mice was performed. The mice were allowed to recover for 2 weeks after which the mice were fed a high-fat diet for 6 weeks. The assessment of chimerism from the leucocytes collected at the time of sacrifice showed that bone marrow was repopulated with the donor bone marrow (Supplementary material online, Figure S1). There were no significant differences between serum total cholesterol in mice transplanted with WT or Nrf2−/− bone marrow before or after 6 weeks on a high-fat diet [7.0 ± 1.7 mmol/L in WT vs. 9.1 ± 2.5 mmol/L in Nrf2−/− before (P = 0.16) and 14.8 ± 3.0 mmol/L in WT and 18.9 ± 5.0 mmol/L in Nrf2−/− after the high-fat diet (P = 0.05)]. Triglyceride levels did not differ between the groups before the high-fat diet (1.8 ± 0.5 mmol/L in WT vs. 2.2 ± 0.4 mmol/L in Nrf2−/−, P = 0.21), but the mice that were transplanted with Nrf2−/− bone marrow had increased triglyceride levels (1.5 ± 0.5 mmol/L in WT vs. 2.2 ± 0.6 mmol/L in Nrf2−/−, P = 0.02) after the high-fat diet. The mice that were transplanted with Nrf2−/− bone marrow exhibited significantly larger lesions, assessed by the cross-sectional lesion area at the sinus level (Figure 1A). Characterization of plaque morphology assessed by Movat's pentachrome staining showed no difference in the necrotic lesion area between the groups (Figure 1B). The mean macrophage content within lesions was higher in the Nrf2−/− group, but the difference between the two groups did not reach statistical significance (P = 0.08, Figure 1C).

Figure 1

Macrophage-specific loss of Nrf2 increases atherosclerosis in LDL-receptor-deficient mice after 6 weeks on a high-fat diet. (A) Quantification of the cross-sectional lesion area from the aortic arch of WT (dots) and Nrf2−/− (rectangles) bone marrow transplanted mice with representative pictures of haematoxylin-eosin-stained sections. (B) Quantification of the necrotic lesion area from the aortic arch with representative pictures of MOVAT Pentachrome-stained sections. Asterisks denote necrotic areas (C) Quantification of the macrophage positive lesion area from the aortic arch with representative pictures stained with—mouse macrophage antibody. Representative photomicrographs are shown with original magnification × 40, scale bar 100 μm. Values are presented as the total surface area. Each symbol represents one mouse and the bar represents the mean, *P < 0.05.

3.2 The effect of Nrf2 on macrophage lipid content, scavenger receptor expression, and reverse cholesterol transport in thioglycollate-elicited peritoneal macrophages

Inasmuch as the lesion size in mice transplanted with Nrf2−/− bone marrow was increased, we next determined the lipid uptake of peritoneal macrophages of WT and Nrf2−/− mice incubated with modified LDL. The cellular neutral lipid content was significantly increased in Nrf2−/− macrophages in comparison with WT controls upon incubation with either acLDL or MDA-LDL, assessed by Oil Red O staining (Figure 2A and B). Also the cholesterol content of Nrf2−/− macrophages was significantly increased in acLDL-loaded macrophages in comparison with WT controls, measured by the enzymatic cholesterol assay (Figure 2C).

Figure 2

The effect of Nrf2 on the macrophage lipid content. (A) Representative images of acLDL (50 μg/mL) and MDA-LDL (50 μg/mL)-treated peritoneal macrophages after Oil red O staining. Representative photomicrographs are shown with original magnification × 200, scale bar 25 μm. (B) Quantification of Oil Red O-stained peritoneal macrophages. (C) Measurement of the total cholesterol from peritoneal macrophages incubated with 50 μg/mL acLDL. The data are depicted as mean ± SD, *P < 0.05, **P < 0.01, ***P < 0.001.

To delineate if the increase in the macrophage lipid content is due to increased uptake of modified LDL or a decrease in reverse cholesterol transport, we first assessed the mRNA expression of scavenger receptors as well the expression of genes involved in the cholesterol efflux. We used HO-1, glutamate-cysteine ligase, and NAD(P)H:quinone oxidoreductase-1, known antioxidative target genes of Nrf2, as positive controls. The expression of these genes were suppressed by Nrf2 deficiency (Supplementary material online, Figure SII). In line with previous reports,17,21 the expression of CD36 was significantly attenuated on the mRNA and protein level in Nrf2−/− macrophages (Figure 3A and B). The expression of scavenger receptors SR-A, LOX-1, and CXCL16 were increased in Nrf2−/− peritoneal macrophages, as was TLR4, which has been implicated to enhance macrophage lipid accumulation via macropinocytosis (Figure 3C–F).22 With respect to genes involved in reverse cholesterol transport, ABCA1 mRNA expression was significantly increased, whereas ABCG1 was decreased in Nrf2−/− macrophages in comparison with WT controls (Figure 4A and B). However, the early step in the reverse cholesterol transport process, the cholesterol efflux from macrophages using either apoAI or HDL2 as an acceptor, did not differ between Nrf2−/− and WT macrophages (Figure 4C), indicating that the increase in the macrophage lipid content in Nrf2−/− macrophages was due to increased lipid uptake in the form of modified LDLs.

Figure 3

The expression of scavenger receptors CD36, SR-A, LOX-1, CXCL16, and TLR4 in Nrf2−/− macrophages. (A) The mRNA expression and (B) protein level of scavenger receptor CD36 detected by western blot, (C) SR-A, (D), LOX-1, (E), CXCL16, and (F), TLR4 in peritoneal macrophages incubated with 50 μg/mL natLDL, 50 μg/mL oxLDL, or 50 μg/mL acLDL. The data are depicted as mean ± SD, *P < 0.05, **P < 0.01, ***P < 0.001. Western blot is representative of three independent experiments.

Figure 4

The effect of Nrf2 on the expression of ABC transport proteins and reverse cholesterol transport. (A) The mRNA expression of ABCA1 and (B), ABCG1 in peritoneal macrophages incubated with 50 μg/mL natLDL, 50 μg/mL oxLDL, or 50 μg/mL acLDL. (C) Measurement of reverse cholesterol transport using ApoAI (10 μg/mL) or HDL2 (25 μg/mL) as cholesterol acceptors. The data are expressed as mean ± SD, *P < 0.05, **P < 0.01, ***P < 0.001.

3.3 The effect of Nrf2 deficiency on inflammatory gene expression

Nrf2 has been implicated to play a role in the regulation of inflammation in mouse peritoneal macrophages.17,23 Next, the expression of MCP-1, IL-6, and TNF-α, markers of the pro-inflammatory M1 phenotype,24 were examined in peritoneal macrophages exposed to modified LDLs. In keeping with a previous report,17 the MCP-1 mRNA expression and protein content measured from the media were significantly higher in macrophages isolated from Nrf2−/− mice in comparison with WT controls (Figure 5A and B). Also, the intensity of MCP-1 immunostaining in atherosclerotic lesions of mice transplanted with Nrf2−/− bone marrow was stronger than in control mice (Figure 5C). In addition, the mRNA expression of IL-6 was induced by modified LDL and was higher in Nrf2−/− macrophages compared with WT (Figure 6A), and TNF-α mRNA levels were higher in Nrf2−/− macrophages under all conditions except for oxLDL (Figure 6B). There was no difference between WT and Nrf2−/− mice in the mRNA expression of arginase (Arg) I, a marker of anti-inflammatory M2 phenotype (Figure 6C). These results support the notion that the loss of Nrf2 promotes a macrophage pro-inflammatory phenotype.

Figure 5

Increased expression of inflammatory cytokine MCP-1 in Nrf2−/− macrophages and atherosclerotic lesions. (A) The mRNA expression of MCP-1, (B), MCP-1 protein in the media in peritoneal macrophages incubated with 50 μg/mL natLDL, 50 μg/mL oxLDL, or 50 μg/mL acLDL. (C) Representative images of MCP-1 immunohistochemical stainings of aortic cross sections from WT or Nrf2−/− bone marrow-transplanted LDLR−/− mice fed a high-fat diet for 6 weeks. Representative photomicrographs are shown with original magnification × 100, scale bar 100 μm. The data are shown as mean ± SD, *P < 0.05, **P < 0.01, ***P < 0.001.

Figure 6

Increased expression of pro-atherogenic inflammatory cytokines in Nrf2−/− macrophages. (A) IL6, (B), TNF-α, and (C), ArgI in peritoneal macrophages incubated with 50 μg/mL natLDL, 50 μg/mL oxLDL, or 50 μg/mL acLDL. The data are shown as mean ± SD, *P < 0.05, **P < 0.01, ***P < 0.001.

4. Discussion

Oxidative stress has been implicated to play a role in atherogenesis via direct adverse effects on the vasculature as well as indirectly via oxidation of LDL and subsequent foam cell formation.25 Moreover, a number of antioxidant enzyme deficiencies aggravate atherosclerosis in hypercholesterolaemic mouse models.26 Nevertheless, the loss of Nrf2, the key transcription factor involved in the regulation of antioxidant genes protects against atherogenesis in apoE−/− mice, but the role of systemic vs. local vascular effects remains unclear. In this study, we provide evidence that Nrf2 deficiency in bone marrow-derived cells is pro-atherogenic in LDLR−/− mice, and that Nrf2 increases both lipid uptake and inflammation in mouse peritoneal macrophages in culture.

It has previously been demonstrated in three separate studies conducted in apoE null mice that Nrf2 promotes atherogenesis.1618 It is notable that these studies have come to a different conclusion as to what is the underlying mechanism behind the observed phenotype, some suggesting a localized, specific effect of Nrf2 on macrophages either via increased expression of CD36 and subsequent enhancement of foam cell formation,16 or Nrf2-dependent induction of IL-1 by cholesterol crystals,18 or in contrast, a combination of local and systemic effects.17 Barajas et al.17 observed a sex-dependent reduction in atherosclerosis in Nrf2 null mice, the male mice exhibiting 53% decrease in the degree of aortic atherosclerosis in comparison with WT littermates, with a concomitant reduction in plasma non-HDL levels. There are notable differences in the design and execution of these studies: in reports by Sussan et al.16 and Freigang et al.18, both of which were conducted with the same Nrf2-deficient mouse strain19 backcrossed to C57BL/6, the mice were fed an atherogenic diet (0.515 or 1.25%13 cholesterol) for 10 or 20 weeks, whereas in the study by Barajas et al.17, the mice that were of different strains26 were kept on a chow diet. This may explain the differences in lipid profiles observed in these studies, as in the studies where cholesterol-fed mice were used, the plasma total cholesterol and the non-HDL cholesterol were similar between the groups.16,18 With respect to foam cell formation, Freigang et al.18 observed no differences in oxLDL uptake between Nrf2+/− and Nrf2−/− mice in apoE−/− background, whereas both Sussan et al.16 and Barajas et al.17 noted that the uptake of oxidized or acetylated LDL was less in peritoneal macrophages derived from Nrf2-deficient mice in comparison with WT controls. In contrast to the latter studies, our results show that the uptake of modified LDL was significantly greater in Nrf2-deficient mice, assessed by two different experimental methods. Similar to the previous studies,16,17,21,27 we did see a significant reduction in the expression of CD36 in Nrf2-deficient peritoneal macrophages, but we note here that the other pathways important for the uptake of LDLs such as those mediated by SR-A and TLR4 outweigh the effect of CD36.

During the revision of our manuscript, two studies on the role of Nrf2 in bone marrow-derived cells in advanced atherosclerosis have been published. In a study by Harada et al.,28 it was shown that the transfer of ApoE−/−Nrf2−/− bone marrow to ApoE−/− mice fed an atherogenic diet for 12 weeks reduced atherosclerosis compared with ApoE−/−Nrf2+/+ bone marrow. This is in stark contrast with a study by Collins et al.29, in which LDLR−/− mice when transferred with Nrf2−/− bone marrow and fed a prolonged obesogenic high-fat, high-cholesterol diet for 7 months had a significant increase in the atherosclerotic lesion area in comparison with LDLR−/− mice receiving WT bone marrow. The experimental approach of the latter study was very different from ours, as the length of the diet in our study was only 6 weeks, resulting in lesions consisting mainly of foamy macrophages vs. complex lesions found in the study by Collins et al.29 Also, prolonged feeding on a high-fat diet resulted in extensive liver inflammation and fibrosis that was aggravated in Nrf2−/− mice.29 Thus the two studies complement each other, showing that Nrf2 deficiency in bone marrow-derived cells aggravates both early as well as late stages of atherosclerosis. These results also emphasize the necessity of studying different genetic models of atherosclerosis, as widely discordant results may be achieved.

Despite the popularity of apoE−/− as a mouse model of atherosclerosis, it has certain drawbacks as a model of human atherosclerosis as it has a rather non-physiological lipoprotein profile consisting mainly of very low density lipoproteins (VLDLs) and chylomicron remnants that have apoB48 as the major apolipoprotein.30 In contrast, LDL receptor-deficient mice have a relatively isolated elevation of LDLs and the corresponding apolipoprotein, apoB100.30 Furthermore, apoE has direct functions within atherosclerotic lesions via facilitating cholesterol efflux from macrophage foam cells.31 Importantly, macrophage-derived apoE facilitates the cholesterol efflux from peritoneal macrophages, even in the absence of cholesterol acceptors.3234 As the absence of apoE thus alters the balance of influx and efflux of cholesterol, it may explain the differences in foam cell formation in this study in comparison with earlier reports, in which macrophages devoid of both Nrf2 and apoE were used.1618 In addition, apoE directly modifies macrophage- and T lymphocyte-mediated immune responses, thereby affecting the vascular inflammation present in atherosclerosis.31 Given the pleiotropic anti-atherogenic effects of apoE and the aberrant lipoprotein profile of apoE-deficient mice, confirmation of the results with more physiological models of atherosclerosis is warranted.

Macrophages have a central role in the chronic inflammatory process characteristic of atherosclerosis.35 Macrophages present in atherosclerotic lesions are phenotypically diverse and can have both pro-inflammatory and anti-inflammatory features.24,35,36 Although a somewhat controversial issue, pro-inflammatory M1 and anti-inflammatory M2 phenotypes represent the two extremes of the spectrum of phenotypes that can be found in atherosclerotic lesions.24,35,36 Interestingly, Kadl et al.37 recently characterized a novel macrophage phenotype (Mox) within atherosclerotic lesions that develops in response to oxidized phospholipids. In this study, the authors showed that the mRNA expression signature characteristic of Mox macrophages is largely dependent on Nrf2, but did not study the functional role of Nrf2 in this phenotype. In our study, the peritoneal macrophages isolated from Nrf2−/− mice show an increased expression of M1 markers MCP-1, IL-6, and TNF-α, in line with previous reports.17,23 The anti-inflammatory effect of Nrf2 in macrophages is likely due to improved antioxidant protection, as the production of ROS in Nrf2−/− peritoneal macrophages challenged with oxLDL or LPS is increased with concomitant induction of pro-inflammatory genes.17,38,39

In summary, our findings indicate that the macrophage-specific loss of Nrf2 promotes atherogenesis in the early phases of lesion development. Further efforts will be directed at fully characterizing the role of Nrf2 in macrophage gene expression and function in relation to atherosclerosis.

Funding

This work was supported by the Academy of Finland, the Sigrid Juselius Foundation, and the Foundation of Cardiovascular Research in Finland.

Acknowledgements

Arja Korhonen, Seija Sahrio, Anne Martikainen, and Sari Nuutinen are acknowledged for their excellent technical assistance. We thank the personnel at the National Laboratory Animal Centre (NLAC) for the animal care.

Conflict of interest: none declared.

References

View Abstract