Skip main navigation

Heme Induces Heme Oxygenase 1 via Nrf2

Role in the Homeostatic Macrophage Response to Intraplaque Hemorrhage
Originally published, Thrombosis, and Vascular Biology. 2011;31:2685–2691



Intraplaque hemorrhage (IPH) is an important progression event in advanced atherosclerosis, in large part because of the delivery of prooxidant hemoglobin in erythrocytes. We have previously defined a novel macrophage phenotype (hemorrhage-associated-mac) in human advanced plaques with IPH. These may be atheroprotective in view of raised heme oxygenase 1 (HO-1), CD163, and interleukin-10 expression and suppressed oxidative stress.

Methods and Results—

We have used a combination of small interfering RNA and pharmacological reagents, protein analysis, and oxidative stress measurements to dissect the pathway leading to the development of this phenotype. We found that erythrocytes, hemoglobin, or purified heme similarly induced CD163 and suppressed human leukocyte antigen and reactive oxygen species. HO-1 was required for the development of each of these features. Challenge of macrophages with purified heme provoked nuclear translocation of Nrf2, and Nrf2 small interfering RNA resulted in significant inhibition of the ability of heme to induce HO-1 protein. Furthermore, tert-butyl-hydroquinone, which activates Nrf2, upregulated CD163, suppressed human leukocyte antigen, and induced interleukin-10, further supporting a role for Nrf2-mediated signaling. However, an inducible protein transactivator is also probably necessary, as heme-induced HO-1 mRNA expression was fully inhibited by the protein synthesis inhibitor cycloheximide.


Our experiments define an Nrf2-mediated pathway by which heme induces a homeostatic macrophage response following IPH.


Intraplaque hemorrhage is a key determinant of atherosclerosis progression and plaque destabilization.1 Intraplaque hemorrhage results in the delivery of both cholesterol-enriched membrane lipids and hemoglobin (Hb) to the plaque interior,13 an environment already challenged by accumulated cholesterol and oxysterols.1,4 As the iron in heme is an effective peroxidant catalyst via H2O2 coordination and Fenton chemistry,5 Hb deposition enhances the lipid peroxidation of cholesterol.2 This in turn can activate macrophage release of H2O2 and free radicals.6 Because intraplaque hemorrhage induces monocyte recruitment,1 how monocytes respond to Hb is critical in determining the outcome of hemorrhage.

Heme is catabolized by cytoprotective heme oxygenases into free ferric ions, biliverdin, and carbon monoxide (CO).7 Low levels of CO, like nitric oxide, engage protective cell signaling pathways such as soluble guanylate cyclase.7 Free ferric ions are normally quickly chelated by coinduced ferritin, where they are safely locked out of Fenton activity.7 Biliverdin is further converted by biliverdin reductase to bilirubin, an endogenous antioxidant.7 Heme oxygenase (HO) has 2 isoforms, with HO-1 being inducible by heme and many other agonists and HO-2 being constitutively expressed.7 The inducibility of HO-1 by heme provides a key negative feedback loop, protecting phagocytes from oxidative stress associated with free iron.5 Surprisingly, there is relatively little published information on how heme induces HO-1 compared with the large investment of effort in inflammatory or pharmacological stimuli.7

We have recently identified a discrete phenotype of macrophages, designated hemorrhage-associated (HA)-mac, that are localized around hematoma in human plaques with fatal coronary thrombosis.6 Compared with more typical macrophage foam cells within the same plaque, these are characterized by suppressed oxidative stress, myeloperoxidase and activation marker human leukocyte antigen (HLA-DR), and increased HO-1 expression and interleukin-10.6 Culturing human blood-derived monocytes with Hb-haptoglobin (Hp) complexes (Hb:Hp) over 4 to 8 days reproduced these phenotypic features, which were prevented by inhibiting the antiinflammatory cytokine interleukin-10 (IL-10) or by inhibiting CD163 dependent endocytosis with cytochalasin-D or with CD163-small interfering RNA (siRNA).6

The basic biochemistry of the transcription factor Nrf2 has been extensively reviewed.8 Nrf2 is activated by oxidative stress and electrophiles because of conversion of cystine residues to cysteine (ie, disulfide cross-linking) in an inhibitory partner Keap-1.8 Activated Nrf2 translocates to the nucleus and binds to antioxidant response elements.8 The electrophile tert-butylhydroquinone (t-BHQ) is extensively used as a selective Nrf2 activator.8

CD163 is a macrophage scavenger receptor for Hb:Hp complexes, and it functions primarily as a means for their internalization.9 However, CD163 may also have signaling functions, as suggested by the effects of ligation by antibodies.10 If macrophage CD163 functions primarily as an endocytic rather than a signaling receptor, then other modes of entry of erythrocyte degradation products should induce HA-mac features. In this study, we demonstrated that this indeed occurs following exposure of monocytes to oxidatively damaged erythrocytes (OxRBCs) or free heme and that this involves a final common pathway mediated by activation of Nrf2.


A full description of the Methods is found in the supplemental materials, available online at These are previously described, with modification where indicated.6 Blood was collected from normal volunteers, and monocytes were aseptically purified and cultured as previously described. RNA was spin-column purified and analyzed by quantitative reverse transcription–polymerase chain reaction using standard molecular biology methods. Western blotting was by standard methods using a precast gel system (Novex, Invitrogen), commercial antibodies, and chemiluminescence (ECL Plus, GE Healthcare). Confocal microscopy was performed with a Zeiss LSM-510Meta using standard settings. RNA inhibition of human monocytes was by using a modification of our previously described methods. Flow cytometry was by modification of previously described methods, using standard live cell immunofluorescence staining and a Dako CyAN cytometer.


We tested whether CD163-independent entry of Hb would drive undifferentiated monocytes to the HA-mac phenotype. Oxidatively damaged erythrocytes (OxRBCs) contain Hb, are pathophysiologically relevant, and are known to be phagocytosed via macrophage scavenger receptor A (macrophage scavenger receptor 1, macrophage scavenger receptor class A, CD204) rather than by CD163.11,12 On the other hand, heme does not require a receptor to enter cells, as it is lipid soluble and freely membrane permeant. As shown in Figure 1, culturing monocytes for 6 days with OxRBCs or with purified heme increased surface CD163 and decreased surface HLA-DR in the same manner as Hb:Hp complexes. This result is therefore consistent with a common mechanism of monocyte response to Hb that is not dependent on CD163.

Figure 1.

Figure 1. Hemoglobin (Hb):haptoglobin (Hp) complexes, heme, and oxidatively damaged erythrocytes (OxRBCs) all induce a macrophage CD163hi human leukocyte antigen (HLA-DR)lo phenotype. Shown is flow cytometry for CD163, HLA-DR, or isotype control antibody as indicated, in human monocyte-derived macrophages (Mφ) after 6 days of culture. Monocytes were cultured from the start in the presence of heme (10 μmol/L) (A), Hb:Hp (100 nmol/L) (B), or OxRBCs (ratios as indicated) (C). Veh indicates control cells cultured with vehicle; Iso, staining with isotype-matched control antibody. Data are representative of 5 experiments on 5 subjects each. D, Human monocyte-derived macrophages were cultured for 48 hours, after which medium was changed to PBS containing aminophenylfluorescein and oxidized low-density lipoprotein (OxLDL) (10 μg · mL−1). Highly reactive oxygen species (hROS) were measured after a further 18 hours of incubation. E, Cell-free wells contained aminophenyl fluorescein in PBS were incubated with purified heme (1 mmol/L), hydrogen peroxide (1 mmol/L), or both, and hROS was measured after 18 hours of incubation. Values are mean+SE; P<0.05 with the Bonferroni's correction and ANOVA.

Because macrophage oxidants play an important role in driving atherosclerosis, we developed an assay using aminophenyl fluorescein, which has been previously validated to detect hypochlorite, hydroxyl free radical, and peroxynitrite (collectively termed highly reactive oxygen species [hROS]).13 As part of basic assay characterization, we found that oxidized low-density lipoprotein (OxLDL) evoked macrophage hROS (Supplemental Figure I) but other agonists (tumor necrosis factor, lipopolysaccharide, palmitoyl-oxo-valeryl-phosphatidylcholine, interferon-γ) had minimal effect under the assay conditions used (not shown). OxLDL-evoked hROS were at least partly dependent on myeloperoxidase, evidenced by suppression by the selective myeloperoxidase inhibitor 4-amino-benzoic acid hydrazide (Supplemental Figure IA) and were maximal with 30 μg · mL−1 OxLDL after 2 days of culture (Supplemental Figure IB and IC). OxLDL had no effect on aminophenyl fluorescein in the absence of macrophages (not shown). OxLDL was particulate with a fraction of the total in the 1 to 10 μm range (Supplemental Figure ID and IE), and OxLDL-evoked hROS were suppressed by cytochalasin-D, indicating that phagocytosis was required (Supplemental Figure IF). Adding 10 μm of polystyrene microbeads, which are phagocytosable but otherwise inert, to these cultures at a ratio of 1 bead/cell also evoked hROS (Supplemental Figure IG), indicating that phagocytosis itself may lead to hROS generation. Together, this set of experiments validated OxLDL as a plaque-relevant phagocytic stimulus for a macrophage oxidative burst and generation of pathogenic reactive species (hROS). This was then used as a standard assay for macrophage oxidative stress.

When human monocyte-derived macrophages were incubated with OxRBCs, heme, or Hb:Hp for 2 days and then challenged with OxLDL, OxRBCs and heme suppressed hROS to an even greater extent than Hb:Hp complexes (ANOVA, P<0.05, Figure 1D). Importantly, this effect required macrophages, because heme and hydrogen peroxide were strongly synergistic for aminophenyl fluorescein fluorescence in a cell-free system (Figure 1E). This is consistent with the known Fenton chemistry of heme in the presence of hydrogen peroxide and provides a positive control for the assay.

Because the effect of Hb:Hp, OxRBCs, and heme appeared similar, we next asked whether erythrocytes and Hb:Hp complexes liberated heme once within the macrophages. As porphyrins have characteristic visible-light absorption profiles, we first used visible spectrophotometry (Supplemental Figure II). Reference spectra for Hb and heme from 400 to 800 nm were consistent with those previously published.14 Hb had characteristic Soret peaks, whereas heme had a broader absorption profile, with a shoulder from 400 to 450 nm (Supplemental Figure IIA and IIB). We found that the absorption profile of Hb:Hp complexes became like that of heme after coincubation with macrophages (Supplemental Figure IIC). We then carried out the same procedure for OxRBCs and again found a change in the overall spectrum following incubation with macrophages, in a pattern consistent with release of heme (Supplemental Figure IID). Measurement of free heme using a commercially available chemical assay broadly corroborated the spectroscopy data, with a maximum heme release from Hb:Hp complexes at 3 days and generation of relatively large concentrations of intracellular heme in response to OxRBC treatment (Supplemental Figure IIE and IIF).

As heme is potentially prooxidant, we questioned whether there the heme-induced antioxidant state is preceded by oxidative stress (Supplemental Figure III). Indeed, treatment with purified heme (10 μmol/L) increased generation of hROS at 1 hour, and this returned to baseline at 4 hours and reversed to an antioxidant effect by 48 hours. However, oxidative stress alone is unlikely to explain the development of the HA-mac phenotype, as 2 other oxidative stress stimulants, H2O2 (100 nmol/L-10 μmol/L) and the purified synthetic oxylipid palmitoyl-oxo-valeryl-phosphatidylcholine,15 failed to lead to late suppression of hROS generation or HLA-DR expression or to an increase in CD163 (Supplemental Figure IV).

We next tested whether phagocytosed Hb:Hp complexes or erythrocytes traffic into lysosomes. Using live-cell confocal microscopy, Alexa 488–labeled Hb:Hp complexes were seen to internalize and traffic to lysosomes, as indicated by colocalization with the lysosomal marker LysoTracker Red. Thus, Figure 2A shows green-labeled Hb and red labeled lysosomes superimposed as yellow, indicating that Hb colocalizes with lysosomes as far as can be determined by optical microscopy. We then extended this analysis to calcein-labeled OxRBCs. A low-magnification image taken seconds after addition showed labeled erythrocytes surrounding the macrophages (Figure 2B). In contrast, images taken after 20 minutes showed OxRBCs (green-labeled) in lysosomes (red-labeled), with yellow colored areas in the overlay indicating colocalization (Figure 2C). In contrast, noninternalized erythrocytes were entirely green at this point, indicating that they were devoid of lysosomal coating (Figure 2D).

Figure 2.

Figure 2. Lysosomal colocalization of ingested erythrocytes or hemoglobin (Hb):haptoglobin (Hp) complexes. A, Hp was labeled with Alexa 488 (green), conjugated with Hb (1:1 molar ratio), and added to macrophages labeled with the lysosomal marker LysoTracker Red. The images were taken with a ×63 objective after 20 minutes. B to D, Autologous erythrocytes were oxidatively damaged (1 mmol/L H2O2 for 20 minutes), labeled with calcein (green) and added to LysoTracker Red–labeled macrophages. Images were acquired after 1 minute (B), 20 minutes (C), and 22 minutes (D) with a ×20 objective (B) or a ×63 objective (C and D). Red indicates lysosomes of varying sizes, some below optical resolution; larger lysosomes are pointed to by arrowheads (Ly); e, some of the calcein-labeled erythrocytes; Mφ, macrophage. Colocalization of green and red is indicated by yellow (open arrows indicate representative areas).

To ask whether generation of HA-mac features requires lysosomal function, we tested the effect of chloroquine.16 In unstimulated macrophages, chloroquine suppressed HLA-DR with minimal effect on surface CD163 in a concentration-dependent manner from 100 nmol/L to 10 μmol/L (Figure 3A). The lowest effective concentration was chosen for subsequent experiments (1 μmol/L). In contrast, chloroquine (1 μmol/L) instead increased HLA-DR in macrophages cultured in the presence of OxRBCs (Figure 3B) or with Hb:Hp (Figure 3C). In parallel, chloroquine (1 μmol/L) inhibited the induction of surface CD163 with OxRBCs or Hb:Hp complexes (Figure 3B and 3C). As a control, chloroquine (1 μmol/L) had minimal effect on either CD163 or on HLA-DR in heme-treated macrophages (Figure 3D). With the specificity provisos of a pharmacological inhibitor, this therefore suggests that inhibiting lysosomes prevents the CD163hi HLA-DRlo phenotype on challenge with complex hemorrhage products that are a source of heme but does not suppress equivalent responses to heme itself.

Figure 3.

Figure 3. Lysosomes are required for macrophage response to erythrocytes and hemoglobin (Hb):haptoglobin (Hp) complexes. Freshly isolated monocytes were cultured with chloroquine (10 μmol/L), a lysosomal inhibitor, either alone (A) or in combination with oxidatively damaged erythrocytes (OxRBCs) (10:1 erythrocytes:macrophages) (B), Hb:Hp complexes (100 nmol/L) (C), or purified heme (10 μmol/L) (D) for 6 days. Expression of CD163 and human leukocyte antigen (HLA-DR) was then assessed by flow cytometry. Data are representative of 4 donors. Iso indicates staining with isotype-matched control antibody.

An important antiinflammatory and antioxidant gene induced by heme is HO-1.7 As shown in Supplemental Figure V, macrophages differentiated with Hb:Hp complexes had increased HO-1 as assessed by fluorescence image analysis (Supplemental Figure V). This increased HO-1 was in part colocalized with CD68 (macrosialin), a specific marker of macrophage lysosomes (Supplemental Figure V). HO-1 was not identified in unstimulated cells.

We then determined the contribution of HO-1 to the antioxidant profile of heme-stimulated cells, using OxLDL or serum deprivation (serum-free medium17,18) as stimuli for oxidative stress. Heme suppressed macrophage hROS (Figure 4A), and this protection was strongly reversed in the presence of the specific HO-1 inhibitor 10 μmol/L zinc protoporphyrin19,20 (Figure 4B). Zinc protoporphyrin also reduced macrophage survival in culture over 6 days, whereas both heme and the specific HO-1 activator cobalt protoporphyrin19,20 had the opposite effect, each in a concentration-dependent manner over 100 nmol/L to 10 μmol/L (Figure 4C). Furthermore, both the increase in CD163 and suppression of HLA-DR induced by heme were antagonized by zinc protoporphyrin (300 nmol/L-3 μmol/L) (Figure 4D). Conversely, surface levels of CD163 were increased and HLA-DR decreased by 3 μmol/L cobalt protoporphyrin (Figure 4E). Of note, 10 μmol/L cobalt protoporphyrin induced macrophage IL-10 after 48 hours of culture, consistent with our previous finding that HA-mac are mediated by autocrine IL-106 (Figure 4F).

Figure 4.

Figure 4. Heme oxygenase 1 (HO-1) is required for the macrophage response to heme. A, Monocytes were cultured in medium alone (open bars) or in the presence of heme (10 μmol/L) for 2 days. Then, 1 μmol/L aminophenyl fluorescein was added with oxidized low-density lipoprotein (OxLDL) (30 μg · mL−1), or medium was removed, and cells were washed twice in serum-free medium (SFM) and then incubated in SFM with 1 μmol/L aminophenyl fluorescein. B, Monocytes were cultured and challenged as in A, but in the presence or absence of zinc protoporphyrin (ZnPP) (10 μmol/L). C, Normalized cell number of monocytes cultured over 6 days in the presence of heme, ZnPP, or cobalt protoporphyrin (CoPP) (each over a concentration-effect curve 100 nmol/L to 100 μmol/L). D and E, human leukocyte antigen (HLA-DR) and CD163 expression in cells differentiated over 6 days with heme (10 μmol/L) and varying concentrations of ZnPP (D) or CoPP (10 μmol/L) (E). F, Flow cytometry histograms for intracellular interleukin-10 (IL-10) on monocytes incubated with CoPP (10 μmol/L) over 2 days. Values are mean±SE, with 5 subjects. P<0.05, ANOVA with the Bonferroni correction where appropriate. Iso indicates staining with isotype-matched control antibody.

Aware of the caveats of pharmacological reagents, we next used siRNA to further test the contribution of HO-1 to the development of HA-mac features (Figure 5). We found that freshly isolated monocytes took up labeled siRNA avidly (Figure 5A). Using the same protocol with unlabeled HO-1-specific or scrambled siRNA, we determined HO-1 expression by Western blotting (Figure 5B). This confirmed that heme strongly induces HO-1 and showed that this was specifically ablated by HO-1-siRNA (Figure 5B). Relative to control transfection, HO-1 siRNA reversed the antioxidant effect of Hb:Hp complexes on macrophage hROS triggered by OxLDL (Figure 5C). Similarly, heme-induced suppression of hROS was also reversed to a strongly prooxidant effect with HO-1 siRNA (Figure 5D). Furthermore, HO-1 siRNA reversed the suppression of HLA-DR and inhibited the induction of CD163 on day 6 macrophages by heme (Figure 5E and 5F). Finally, HO-1 siRNA inhibited heme-induced IL-10 secretion (Figure 5G). Thus, HO-1 induction is necessary for CD163 induction, HLA-DR suppression, and suppression of hROS induced by erythrocytes, Hb, and free heme.

Figure 5.

Figure 5. Heme oxygenase 1 (HO-1) is required for macrophage response to heme. A, We added transfection reagent polyethyleneimine (PEI) alone or transfection reagent plus fluorescein isothiocyanate (FITC)–labeled small interfering RNA (siRNA) immediately after monocyte purification and uptake assessed after 24 hours of culture. The x-axis shows uptake of FITC-labeled siRNA oligonucleotides into human monocytes at the start of culture, and the y-axis shows the cell number. Data are representative of 3 experiments. B, Western blotting for HO-1 27-kDa protein in extracts of monocytes cultured for 4 hours in the presence or absence of heme and siRNA, after 24 hours of siRNA added from the start of culture. Scr indicates scrambled control. C and D, Reversal by HO-1 siRNA of effects of hemoglobin (Hb):haptoglobin (Hp) (C) and heme (D) on highly reactive oxygen species (hROS) release by monocytes cultured for 2 days. Data are mean±SE, n=5 subjects. E to G show reversal by HO-1 siRNA of effects of heme on HLA-DR expression (E), CD163 expression (F), and interleukin-10 release (G) by monocytes cultured for 2 days. Data in histograms are mean±SE, with 5 donors. *P<0.05, Student t test. Iso indicates staining with isotype-matched control antibody.

We next assessed changes in levels of HO-1 mRNA in response to heme and found that these were near maximally elevated by 10 μmol/L heme (Supplemental Figure VI). In a time-course analysis, we found that there was minimal induction at 1 hour but near-maximal expression by 4 hours (Supplemental Figure VI). We therefore chose 10 μmol/L and 4 hours as standard conditions to test the role of Nrf2 in induction of HO-1. Purified heme (Figure 6A) or Hb:Hp (not shown) evoked clear-cut nuclear translocation of Nrf2 to the nucleus when added to monocytes. Moreover, Nrf2-siRNA, which suppressed Nrf2 protein by more than 89%, decreased the ability of heme to induce HO-1 protein by ≈50% (Figure 6B). Further support for a role for Nrf2 signaling came from the use of the known Nrf2 activator t-BHQ,21 which suppressed heme-induced hROS in a concentration-dependent manner, maximal by 10 μmol/L (Figure 6C). Similarly, 50 μmol/L t-BHQ evoked IL-10 secretion (control 71.8±21.8; t-BHQ 477±40 pg · mL−1) and induced a subpopulation with upregulated CD163 and reciprocally suppressed HLA-DR (Figure 6D).

Figure 6.

Figure 6. Heme induces heme oxygenase 1 (HO-1) via Nrf2. Human monocytes were rested overnight before challenging with heme. A, Confocal immunofluorescence staining of Nrf2 localization at 4 hours, with Nrf2 shown in green and nuclear dye (TOPRO) in red. Yellow staining implies nuclear colocalization of Nrf2. Inset indicates low magnification (vehicle) or staining control (heme-treated). Shown is a representative of n=3 experiments. B, Western blots of Nrf2, HO-1, and β-actin (loading control) in lysates of cells treated with small interfering RNA (siRNA) for Nrf2 or control siRNA as indicated. Lower panels show densitometry quantification of Nrf2 or HO-1 (mean±SE, n=3). C and D, Monocytes were cultured with the Nrf2 activator tert-butylhydroquinone (t-BHQ) over the concentration effect curve 1 to 100 μmol/L for highly reactive oxygen species (hROS) at day 2 (C) and expression of HLA-DR and CD163 at day 6 (D). *ANOVA with the Bonferroni adjustment; maximum effect of t-BHQ is at 10 μmol/L. Arrow in D points to a population within the ellipse with increased CD163 and reduced HLA-DR. Suppression of hROS by t-BHQ was also seen at culture day 6 (not shown).

As established above, oxidant stress alone does not induce the HA-mac phenotype, suggesting that Nrf2 activation is not sufficient. Consistent with this, we found that heme induction of HO-1 mRNA was completely suppressed in monocytes from 12 separate donors by the protein synthesis inhibitor cycloheximide (100 μmol/L), without loss of cell viability (Supplemental Figure VI). As Nrf2-mediated transcriptional activation does not require de novo protein synthesis, these observations therefore indicate that the induction of HO-1 mRNA in monocytes by heme also requires synthesis of an induced intermediary protein transactivator.


The data presented in this article provide fresh insights into the origins of the HA-mac phenotype in hemorrhaged plaques6 by showing that the phenotype is driven by internalization of Hb:Hp or erythrocytes complexes followed by the release of heme. We have established for the first time that heme induces HO-1 expression, in part via Nrf2. HO-1 is then required for the increase in CD163, suppression of HLA-DR and suppression of oxidative stress. We therefore propose that a heme→Nrf2→HO-1 pathway is part of a final common pathway for HA-mac differentiation. We have demonstrated that Hb:Hp complexes and oxidized erythrocytes are internalized by macrophages, sent to lysosomes, and then liberate heme. This induces HO-1, which is itself trafficked to lysosomes. Using HO-1 inhibitors and inducers, we have found that the induced HO-1 is necessary for hemorrhage products to evoke the features of HA-mac. Taking these data together, we suggest a unifying schema whereby hemorrhage-related degradation products enter macrophages (Supplemental Figure VII). Heme is lipophilic and freely passes cell membranes. On the other hand, Hb:Hp complexes undergo receptor-mediated endocytosis via CD163, and OxRBCs are internalized via CD204/macrophage scavenger receptor class A.11,12 As far as can be determined with state-of-the-art reagents and microscopy, internalized products are passed to lysosomes, where the globin scaffold is proteolysed, liberating heme. This framework goes against the concept that CD163 ligation by Hb:Hp complexes initiates a classical protein phosphorylation signaling cascade that directly causes the HA-mac transcriptional response.

Nrf2 is a cap'n'collar transcription factor activated by oxidation of cysteines on its inhibitory partner, Keap.22 Thus, Nrf2 is activated by cellular oxidative stress or electrophiles, and its target genes include many that decrease oxidative stress (eg, superoxide dismutase, HO-1) or induce phase II metabolism (which catabolises exogenous electrophiles, eg, xenobiotics).22 Given this background, Nrf2 was a priori a strong candidate to mediate adaptive macrophage responses to heme,23 and this was supported both by experimental activation with t-BHQ and by gene silencing with siRNA. We cannot find a previous mechanistic dissection of the role of Nrf2 in HO-1 induction by heme, the physiological substrate for this enzyme.

Interestingly, Kadl et al have recently published a landmark article on Nrf2 and HO-1in mouse macrophages.24 They stimulated mouse macrophages with the synthetic lipid 1-palmitoyl-2-arachidonoyl-sn-3-glycero-phosphorylcholine, which they had purchased and oxidized in-house by passive exposure to atmospheric oxygen. Oxidized 1-palmitoyl-2-arachidonoyl-sn-3-glycero-phosphorylcholine and palmitoyl-oxo-valeryl-phosphatidylcholine are defined oxylipids related to phosphatidylcholine; they are found in minimally modified low-density lipoprotein and have been extensively characterized as being bioactive.15 Oxidized 1-palmitoyl-2-arachidonoyl-sn-3-glycero-phosphorylcholine evoked a macrophage transcriptome pattern, designated by the authors as Mox, that included induction of HO-1, along with sulfiredoxin and thioredoxin genes. This did not occur in Nrf2-null mice, indicating that Nrf2 was required for HO-1 induction in this model.24 Nrf2-mediated induction of HO-1 via the antioxidant response element has been previously described in response to a wide variety of stimuli, including toxic electrophiles, dietary (vegetable) indirect antioxidants, and cigarette smoke.23,2532 Although the Mox transcriptome analysis sets the oxidized 1-palmitoyl-2-arachidonoyl-sn-3-glycero-phosphorylcholine–responding gene set apart from M1 and M2 subsets in the mouse, it remains to be seen whether this represents a genuinely committed stable lineage or a plastic state. Notwithstanding, our own treanscriptome analysis of HA-Mac clearly distinguishes this putative atheroprotective phenotype from Mox, as well as M1 and M2 cells (J.J. Boyle et al, unpublished data, 2011).

A strength of our study has been the use of freshly isolated cells. Transcriptome analyses of plaque macrophages and pooled cell line microarray data indicate that gene expression in freshly isolated blood-derived monocytes is relevant to plaques.33 Moreover, a metaanalysis of microarray data giving a global map of human gene expression has revealed enormous differences between transformed cell lines and primary tissues.34 Thus, although human blood-derived monocytes are difficult to transfect, they are likely to be to be far more informative about human pathophysiology. However, the use of fresh human cells has limitations, not least the difficulty of obtaining a complete knockdown using siRNA. We corroborated data with siRNA and extensively characterized pharmacological inhibitors, establishing concentration-dependent action with effects at appropriate concentrations.16,1921 We cannot be sure to what degree the relatively modest reduction in HO-1 seen in cells treated with Nrf2 siRNA was due to gene silencing being only partial (89%) or to what degree it reflects the status of Nrf2 as just one of the transcription factors necessary for eliciting HA-mac features. The latter seems likely, as an experiment with cycloheximide showed that induction of HO-1 mRNA requires protein synthesis as an intermediary step. The identification of transactivators that operate in this system and that may deviate the stress response toward the HA-mac rather than Mox phenotype24 is the subject of a separate study (J.J. Boyle et al, unpublished data).

In conclusion, uncommitted monocyte-derived macrophages respond to intraplaque hemorrhage by internalizing Hb:Hp complexes, oxidized red blood cells, or both into lysosomes, resulting in the intracellular release of free heme. This in turn induces HO-1, in part via Nrf2, which leads to a phenotype similar the one we have previously identified in the vicinity of hemorrhage in human plaques.

Sources of Funding

We thank the British Heart Foundation (Gerry Turner Intermediate FellowshipFS07/010 to J.J.B. and Professorial Core Support to D.O.H.) and the Imperial College Biomedical Research Centre for financial support.




Correspondence to Joseph J. Boyle,
Cardiovascular Sciences, NHLI, Imperial College London, B Block, Hammersmith Hospital, Du Cane Rd, London, United Kingdom
. E-mail


  • 1. Kolodgie FD, Gold HK, Burke AP, Fowler DR, Kruth HS, Weber DK, Farb A, Guerrero LJ, Hayase M, Kutys R, Narula J, Finn AV, Virmani R. Intraplaque hemorrhage and progression of coronary atheroma. N Engl J Med. 2003; 349:2316–2325.CrossrefMedlineGoogle Scholar
  • 2. Levy AP, Levy JE, Kalet-Litman S, Miller-Lotan R, Levy NS, Asaf R, Guetta J, Yang C, Purushothaman KR, Fuster V, Moreno PR. Haptoglobin genotype is a determinant of iron, lipid peroxidation, and macrophage accumulation in the atherosclerotic plaque. Arterioscler Thromb Vasc Biol. 2007; 27:134–140.LinkGoogle Scholar
  • 3. Lin HL, Xu XS, Lu HX, Zhang L, Li CJ, Tang MX, Sun HW, Liu Y, Zhang Y. Pathological mechanisms and dose dependency of erythrocyte-induced vulnerability of atherosclerotic plaques. J Mol Cell Cardiol. 2007; 43:272–280.CrossrefMedlineGoogle Scholar
  • 4. Daugherty A, Rateri DL, Lu H. As macrophages indulge, atherosclerotic lesions bulge. Circ Res. 2008; 102:1445–1447.LinkGoogle Scholar
  • 5. Sadrzadeh SM, Graf E, Panter SS, Hallaway PE, Eaton JW. Hemoglobin: a biologic fenton reagent. J Biol Chem. 1984; 259:14354–14356.CrossrefMedlineGoogle Scholar
  • 6. Boyle JJ, Harrington HA, Piper E, Elderfield K, Stark J, Landis RC, Haskard DO. Coronary intraplaque hemorrhage evokes a novel atheroprotective macrophage phenotype. Am J Pathol. 2009; 174:1097–1108.CrossrefMedlineGoogle Scholar
  • 7. Otterbein LE, Soares MP, Yamashita K, Bach FH. Heme oxygenase-1: unleashing the protective properties of heme. Trends Immunol. 2003; 24:449–455.CrossrefMedlineGoogle Scholar
  • 8. Singh S, Vrishni S, Singh BK, Rahman I, Kakkar P. Nrf2-ARE stress response mechanism: a control point in oxidative stress-mediated dysfunctions and chronic inflammatory diseases. Free Radic Res. 2010; 44:1267–1288.CrossrefMedlineGoogle Scholar
  • 9. Schaer CA, Schoedon G, Imhof A, Kurrer MO, Schaer DJ. Constitutive endocytosis of CD163 mediates hemoglobin-heme uptake and determines the noninflammatory and protective transcriptional response of macrophages to hemoglobin. Circ Res. 2006; 99:943–950.LinkGoogle Scholar
  • 10. Polfliet MM, Fabriek BO, Daniels WP, Dijkstra CD, van den Berg TK. The rat macrophage scavenger receptor CD163: expression, regulation and role in inflammatory mediator production. Immunobiology. 2006; 211:419–425.CrossrefMedlineGoogle Scholar
  • 11. Sambrano GR, Terpstra V, Steinberg D. Independent mechanisms for macrophage binding and macrophage phagocytosis of damaged erythrocytes. Evidence of receptor cooperativity. Arterioscler Thromb Vasc Biol. 1997; 17:3442–3448.LinkGoogle Scholar
  • 12. Sambrano GR, Parthasarathy S, Steinberg D. Recognition of oxidatively damaged erythrocytes by a macrophage receptor with specificity for oxidized low density lipoprotein. Proc Natl Acad Sci U S A. 1994; 91:3265–3269.CrossrefMedlineGoogle Scholar
  • 13. Setsukinai K, Urano Y, Kakinuma K, Majima HJ, Nagano T. Development of novel fluorescence probes that can reliably detect reactive oxygen species and distinguish specific species. J Biol Chem. 2003; 278:3170–3175.CrossrefMedlineGoogle Scholar
  • 14. Vasudevan G, McDonald MJ. Soret spectral and bioinformatic approaches provide evidence for a critical role of the α-subunit in assembly of tetrameric hemoglobin. Protein J. 2006; 25:45–56.CrossrefMedlineGoogle Scholar
  • 15. Watson AD, Leitinger N, Navab M, Faull KF, Horkko S, Witztum JL, Palinski W, Schwenke D, Salomon RG, Sha W, Subbanagounder G, Fogelman AM, Berliner JA. Structural identification by mass spectrometry of oxidized phospholipids in minimally oxidized low density lipoprotein that induce monocyte/endothelial interactions and evidence for their presence in vivo. J Biol Chem. 1997; 272:13597–13607.CrossrefMedlineGoogle Scholar
  • 16. Wen Y, Leake DS. Low density lipoprotein undergoes oxidation within lysosomes in cells. Circ Res. 2007; 100:1337–1343.LinkGoogle Scholar
  • 17. Maeda Y, Hirano K, Nishimura J, Sasaki T, Kanaide H. Endothelial dysfunction and altered bradykinin response due to oxidative stress induced by serum deprivation in the bovine cerebral artery. Eur J Pharmacol. 2004; 491:53–60.CrossrefMedlineGoogle Scholar
  • 18. Liu W, Kato M, Itoigawa M, Murakami H, Yajima M, Wu J, Ishikawa N, Nakashima I. Distinct involvement of NF-κB and p38 mitogen-activated protein kinase pathways in serum deprivation-mediated stimulation of inducible nitric oxide synthase and its inhibition by 4-hydroxynonenal. J Cell Biochem. 2001; 83:271–280.CrossrefMedlineGoogle Scholar
  • 19. Kinderlerer AR, Pombo GI, Hamdulay SS, Ali F, Steinberg R, Silva G, Ali N, Wang B, Haskard DO, Soares MP, Mason JC. Heme oxygenase-1 expression enhances vascular endothelial resistance to complement-mediated injury through induction of decay-accelerating factor: a role for increased bilirubin and ferritin. Blood. 2009; 113:1598–1607.CrossrefMedlineGoogle Scholar
  • 20. Bussolati B, Ahmed A, Pemberton H, Landis RC, Di CF, Haskard DO, Mason JC. Bifunctional role for VEGF-induced heme oxygenase-1 in vivo: induction of angiogenesis and inhibition of leukocytic infiltration. Blood. 2004; 103:761–766.CrossrefMedlineGoogle Scholar
  • 21. Bloom D, Dhakshinamoorthy S, Jaiswal AK. Site-directed mutagenesis of cysteine to serine in the DNA binding region of Nrf2 decreases its capacity to upregulate antioxidant response element-mediated expression and antioxidant induction of NAD(P)H:quinone oxidoreductase1 gene. Oncogene. 2002; 21:2191–2200.CrossrefMedlineGoogle Scholar
  • 22. Kang KW, Lee SJ, Kim SG. Molecular mechanism of nrf2 activation by oxidative stress. Antioxid Redox Signal. 2005; 7:1664–1673.CrossrefMedlineGoogle Scholar
  • 23. Ishii T, Itoh K, Takahashi S, Sato H, Yanagawa T, Katoh Y, Bannai S, Yamamoto M. Transcription factor Nrf2 coordinately regulates a group of oxidative stress-inducible genes in macrophages. J Biol Chem. 2000; 275:16023–16029.CrossrefMedlineGoogle Scholar
  • 24. Kadl A, Meher AK, Sharma PR, Lee MY, Doran AC, Johnstone SR, Elliott MR, Gruber F, Han J, Chen W, Kensler T, Ravichandran KS, Isakson BE, Wamhoff BR, Leitinger N. Identification of a novel macrophage phenotype that develops in response to atherogenic phospholipids via Nrf2. Circ Res. 2010; 107:737–746.LinkGoogle Scholar
  • 25. Alam J, Stewart D, Touchard C, Boinapally S, Choi AM, Cook JL. Nrf2, a Cap‘n’Collar transcription factor, regulates induction of the heme oxygenase-1 gene. J Biol Chem. 1999; 274:26071–26078.CrossrefMedlineGoogle Scholar
  • 26. Venugopal R, Jaiswal AK. Nrf2 and Nrf1 in association with Jun proteins regulate antioxidant response element-mediated expression and coordinated induction of genes encoding detoxifying enzymes. Oncogene. 1998; 17:3145–3156.CrossrefMedlineGoogle Scholar
  • 27. Venugopal R, Jaiswal AK. Nrf1 and Nrf2 positively and c-Fos and Fra1 negatively regulate the human antioxidant response element-mediated expression of NAD(P)H:quinone oxidoreductase1 gene. Proc Natl Acad Sci U S A. 1996; 93:14960–14965.CrossrefMedlineGoogle Scholar
  • 28. McWalter GK, Higgins LG, McLellan LI, Henderson CJ, Song L, Thornalley PJ, Itoh K, Yamamoto M, Hayes JD. Transcription factor Nrf2 is essential for induction of NAD(P)H:quinone oxidoreductase 1, glutathione S-transferases, and glutamate cysteine ligase by broccoli seeds and isothiocyanates. 203. J Nutr. 2004; 134:3499S–506S.CrossrefMedlineGoogle Scholar
  • 29. Andreadi CK, Howells LM, Atherfold PA, Manson MM. Involvement of Nrf2, p38, B-Raf, and nuclear factor-κB, but not phosphatidylinositol 3-kinase, in induction of hemeoxygenase-1 by dietary polyphenols. Mol Pharmacol. 2006; 69:1033–1040.CrossrefMedlineGoogle Scholar
  • 30. Balogun E, Hoque M, Gong P, Killeen E, Green CJ, Foresti R, Alam J, Motterlini R. Curcumin activates the haem oxygenase-1 gene via regulation of Nrf2 and the antioxidant-responsive element. Biochem J. 2003; 371:887–895.CrossrefMedlineGoogle Scholar
  • 31. Jeong WS, Jun M, Kong AN. Nrf2: a potential molecular target for cancer chemoprevention by natural compounds. Antioxid Redox Signal. 2006; 8:99–106.CrossrefMedlineGoogle Scholar
  • 32. Knorr-Wittmann C, Hengstermann A, Gebel S, Alam J, Muller T. Characterization of Nrf2 activation and heme oxygenase-1 expression in NIH3T3 cells exposed to aqueous extracts of cigarette smoke. Free Radic Biol Med. 2005; 39:1438–1448.CrossrefMedlineGoogle Scholar
  • 33. Eijgelaar WJ, Horrevoets AJ, Bijnens AP, Daemen MJ, Verhaegh WF. Equivalence testing in microarray analysis: similarities in the transcriptome of human atherosclerotic and non-atherosclerotic macrophages [Epub ahead of print Jan. 12, 2010]. Physiol Genomics. 2010.CrossrefMedlineGoogle Scholar
  • 34. Lukk M, Kapushesky M, Nikkila J, Parkinson H, Goncalves A, Huber W, Ukkonen E, Brazma A. A global map of human gene expression. Nat Biotechnol. 2010; 28:322–324.CrossrefMedlineGoogle Scholar