Phagocytosis and Macrophage Activation Associated With Hemorrhagic Microvessels in Human Atherosclerosis
Objective— Previously, we demonstrated that activated inducible NO synthase (iNOS)-expressing foam cells in human carotid plaques often produce autofluorescent (per)oxidized lipids (ceroid). Here, we investigate whether intraplaque microvessels can provide foam cells with lipids and trigger macrophage activation.
Methods and Results— Microvessels (von Willebrand factor [vWf] immunoreactivity), activated macrophages (iNOS immunoreactivity), and ceroid were systematically mapped in longitudinal sections of 15 human carotid endarterectomy specimens. An unbiased hierarchical cluster analysis classified vascular regions into 2 categories. One type with normal vWf expression and without inflammatory cells was seen, and another type with cuboidal endothelial cells, perivascular vWf deposits, and iNOS and ceroid-containing foam cells was seen in 4 (27%) of 15 plaques. The perivascular foam cells frequently contained platelets (glycoprotein Ibα) and erythrocytes (hemoglobin, iron), pointing to microhemorrhage/thrombosis and subsequent phagocytosis. Similar lipid-containing cells, expressing both ceroid and iNOS, were generated in atherosclerosis-free settings by incubating murine J774 macrophages with platelets or oxidized erythrocytes and also in vivo in organizing thrombi in normocholesterolemic rabbits.
Conclusions— Focal intraplaque microhemorrhages initiate platelet and erythrocyte phagocytosis, leading to iron deposition, macrophage activation, ceroid production, and foam cell formation. Neovascularization, besides supplying plaques with leukocytes and lipoproteins, can thus promote focal plaque expansion when microvessels become thrombotic or rupture prone.
Intermittent growth is a characteristic of human atherosclerosis. This could be the consequence of recurrent rupture of the fibrous cap followed by thrombus organization into the plaque.1,2⇓ Other studies have suggested a causative role of hemorrhages of intraplaque microvessels in carotid plaque rupture.3–5⇓⇓ Paterson et al6 proposed the vascularization theory of plaque evolution by demonstrating hemosiderin deposition in early atheromatous plaques, and they related this to repeated intraplaque capillary rupture, but it remains unclear how hemorrhages contribute to lipid accumulation. Ceroid is one of these lipid components and consists of insoluble mixtures of oxidized lipids and proteins, which mark sites of previous oxidative events.7 Macrophages play a central role in the production of this fluorescent pigment, and extracellular ceroid deposits ultimately accumulate in the necrotic core of the plaque.7,8⇓ Furthermore, the upregulation of inducible NO synthase (iNOS), a major ancillary pathway of host defense by activated macrophages, is a characteristic feature of foam cell–rich plaque regions.9,10⇓ Recently, we showed that iNOS, which is predominantly expressed in macrophages,10,11⇓ often colocalizes with ceroid11 or platelet-derived amyloid β12 in advanced human plaques. Because the reasons
for these associations are unclear, we investigated the role of microvessels in plaque progression. To this end, the distribution of microvessels, ceroid and iNOS (as a marker of macrophage activation), was systematically mapped in human carotid artery plaques. The expression of von Willebrand factor (vWf) in the endothelial cells of intraplaque microvessels is highly variable, ranging from undetectable to thick perivascular deposits.5 The latter are due to increased vWf biosynthesis during atherogenesis.13,14⇓ Therefore, vWf was used as a marker of endothelial cell activation. For an unbiased identification of topographical associations, the results were subjected to a cluster analysis. Finally, the formation of iNOS-expressing ceroid-containing foam cells was demonstrated in normocholesterolemic settings on erythrophagocytosis by macrophages in experimental thrombi in rabbit carotid arteries and in murine J774 macrophages in culture.
The ethics committees of Middelheim Hospital and Antwerp University approved the studies.
Carotid artery endarterectomy specimens were obtained from 15 patients (stenosis >70% [digital subtraction angiography and duplex ultrasound], mean age 66 years), cut along their longitudinal axis, and fixed in 4% formaldehyde within 5 minutes after collection. Decalcification was performed in Bouin’s fixative (15 vol saturated picric acid, 5 vol of 40% formaldehyde, and 1 vol acetic acid) for 48 hours. Whole-mount longitudinal sections of paraffin-embedded specimens were mounted on 3-aminopropyltriethoxysilane–coated slides.11
Experimental Thrombosis in Rabbit Carotid Arteries
Male New Zealand White rabbits (n=38, weight 2.5 to 3.5 kg, normal diet) were anesthetized (sodium pentobarbital, 30 mg/kg IV). The right carotid artery was selected to induce thrombus by repeated inflation of an oversized (2.5-mm) angioplasty balloon.15 The uninjured contralateral artery served as the control. One animal was investigated immediately. The others were anesthetized again at 7 (n=6), 14 (n=13), or 21 (n=18) days after angioplasty, and both carotid arteries were removed and fixed in 60% methanol, 30% 1,1,1-trichloroethane, and 10% glacial acetic acid (Methacarn). Transversal sections of paraffin-embedded specimens were mounted on aminopropyltriethoxysilane–coated slides.
The following primary antibodies were used: rabbit polyclonal anti-human vWf (1:500, Dako), iNOS (1:1000, Biomol Research Laboratories), heme oxygenase-1 (1:100, Transduction Laboratories), hemoglobin (1:100, Lipshaw Immunon), and monoclonal anti-human glycoprotein Ibα (GPIbα, clone G28E5 1:3000, gift from Dr M.F. Hoylaerts, Center for Molecular and Vascular Biology, KUL, Leuven, Belgium) for human tissue and polyclonal sheep anti-human vWf (1:200, The Binding Site), monoclonal anti-human iNOS (1:25, Transduction Laboratories), and anti-rabbit macrophage RAM-11 (1:100, Dako) for rabbit tissue.
After inactivation of endogenous peroxidase (3% H2O2), monoclonal antibodies were detected with goat anti-mouse peroxidase antibody (Jackson); sheep polyclonal antibodies, with rabbit anti-sheep peroxidase (Jackson); and rabbit polyclonal antibodies, with the polyclonal Envision System (Dako). For demonstration of the complex, 0.1% H2O2 was used as a substrate, and 3-amino-9-ethylcarbazole was used as a chromogen. Controls without primary antibody were run for each protocol, resulting in consistently negative observations.
To detect iron deposits, Perls’ stain was used. To study colocalization, this was performed on the iNOS-immunostained slide by using neutral red counterstaining.
Two adjacent sections of the endarterectomy specimens were stained for iNOS and for vWf. To avoid bias, ceroid was assessed in the vWf section because ceroid often associates with iNOS.11 By use of a special x-y coordinate system of the object table, both sections were carefully aligned. Starting at a randomly selected position, the complete vWf section was systematically scanned by using 1-mm steps (Figure 1, region of interest [ROI] 0.4 mm2). A transparent grid containing 24 rectangular fields (0.0072 mm2) was superimposed on the ROI. In each ROI, fields with nuclei (ftotal), with microvessels without vWf, with intracellular vWf, with perivascular vWf deposits, and fields with intracellular ceroid (fceroid) were scored. Furthermore, all microvessels (without vWf, with intraendothelial vWf, or with subendothelial vWf deposits) were counted (nv). Using the x-y coordinates, the matching ROIs were relocated in the adjacent section, and ftotal and fiNOS were counted. The density of each variable was calculated for each ROI by dividing the number of positive fields by the number of fields containing nuclei (eg, fceroid/ftotal and fiNOS/ftotal). Vessel density (microvessels/mm2) was calculated by the following formula: nv/(ftotal×0.0072).
In rabbit carotid arteries, iron, RAM-11, iNOS, and vWf were quantified in adjacent serial sections, which were superimposed with an unbiased counting frame with a set of regularly spaced points.16 The points hitting immunoreactive structures were counted and expressed as percentage of the points in the intima.
Murine macrophages (5×105/800 μL, cell line J774A.1, American Type Culture Collection) were grown in RPMI and allowed to adhere in culture slides (Becton Dickinson Labware) at 37°C in 5% CO2/95% air as described.12 Thereafter, macrophages were incubated for 41 hours in DMEM with or without 108 washed human platelets12 or 4×107 oxidized (CuSO4+H2O2) erythrocytes (ox-RBCs).17 The cell-free supernatant was stored at −20°C for nitrite measurements (Griess reaction) as an index of iNOS activity.12 The cells were fixed with paraformaldehyde (1%, 2 minutes) followed by methanol (−20°C, 6 minutes), air-dried, and stained with oil red O.
Results are given as mean±SEM. For an unbiased classification, all ROIs were subjected to a hierarchical cluster analysis (procedure K-means cluster, SPSS release 10, SPSS Inc). Vessel density, the presence of subendothelial vWf deposits, and iNOS density were used as input variables after z transformations to obtain equal scaling (mean 0, SD 1). A minimum size of 10 ROIs was used as the criterion to select the number of categories. Differences among means were evaluated with an ANOVA (Bonferroni post hoc test); correlations were determined by the Spearman test. A value of P<0.05 was considered significant.
Each section showed American Heart Association (AHA) stages I to V.1 Complicated plaques (AHA VI) were present in 14 specimens, often at the dorsal wall of the carotid sinus. Thrombi were noted in 4 specimens (Table). All specimens contained iNOS-expressing macrophages (positive ROIs 7% to 50%, mean density 1.2% to 16.0%); ceroid was detected in 13 specimens (positive ROIs 2% to 52%, mean density 0.4% to 18.2%); and microvessels were detected in 14 specimens (positive ROIs 14% to 76%, mean density 2.1 to 25.8 vessels/mm2). Mean densities of iNOS and ceroid were not affected by sex, the maximum AHA score, or thrombus (P>0.05). However, vascular density was higher in specimens with thrombus (16±5 microvessels/mm2, n=4) than in specimens without thrombus (6±1 microvessels/mm2, n=11, P=0.027 by ANOVA). There was a strong positive correlation between the mean vascular density and iNOS expression in plaques (RSpearman=0.64, P=0.01).
The distribution of microvessels, intracellular ceroid, and iNOS-expressing macrophages was mapped for each plaque; Figure 1 shows an example. Microvessels were present in fibrous regions, often close to the media, or in foam cell–rich areas in the proximal shoulder. In the former vessels, vWf was either undetectable or present in flat endothelial cells; macrophages expressing iNOS or ceroid were rare. In contrast, microvessels in foam cell–rich regions often contained thick perivascular vWf deposits and were frequently associated with iNOS-expressing ceroid-containing macrophages. Finally, avascular clusters of iNOS-expressing foam cells frequently surrounded the necrotic core, commonly in association with ceroid.
These qualitative observations were substantiated by a cluster analysis of all ROIs (n=592). This unbiased assessment identified 4 relatively homogeneous groups (Figure 2). Category A was found in all sections and contained 413 ROIs in which microvessels and iNOS were rare or absent (Figures 2 and 3⇓A). Category B consisted of 69 ROIs in 14 sections (93%) with a high microvascular density but few iNOS-expressing cells or perivascular vWf deposits (Figures 2 and 3⇓B). Category C contained 49 ROIs in 4 sections (27%) with high vascular densities, cuboidal endothelial cells, subendothelial vWf deposits, and macrophages containing iNOS, ceroid, and lipid droplets (Figures 2 and 3⇓C). Finally, category D consisted of 61 ROIs in all plaques with many iNOS-containing macrophages but few microvessels. Ceroid was more abundant in categories C and D compared with categories A and B (Figure 2). Adding ceroid to the cluster analysis produced a very similar classification (results not shown).
The third indication of the association between activated macrophages and microvessels was obtained by selecting ROIs containing microvessels, irrespective of their density. Macrophages expressing iNOS were more abundant in ROIs with perivascular vWf deposits (mean density 0.21±0.04, n=50, P<0.001 versus other ROIs) than in ROIs with microvessels without (0.07±0.02, n=40) or with (0.10±0.02, n=112) intraendothelial vWf. Also, ceroid was more profuse in ROIs with perivascular vWf (mean density 0.14±0.03, n=50, P<0.05 versus other ROIs) than in regions without (0.08±0.03, n=40) or with (0.07±0.02, n=112) intraendothelial vWf.
Because we were interested in associations of microvessels, activated macrophages, and ceroid, the regions (classified in category C) were studied in greater detail. These microvessels were found not only between the plaque shoulder and necrotic core (Figure 1) but also in the fibrous cap at sites of plaque rupture. Macrophages around microvessels with cuboidal endothelial cells and perivascular vWf deposits often showed signs of platelet and erythrophagocytosis. The platelet marker GPIbα was detected as a cytoplasmic granular stain, and hemoglobin, heme oxygenase-1, and iron deposits were noted in the macrophages (Figure 3E through 3H). In these regions, iNOS and heme oxygenase-1 were demonstrated as granular staining patterns in the cytoplasm. Finally, a strong colocalization of iron deposits and iNOS expression in the macrophages was found (Figure 3H). The presence of erythrocytes, iron deposits, and hemoglobin in plaque regions with microvessels points to microhemorrhages of these microvessels.
Macrophage Activation and Ceroid Formation in Experimental Thrombus
To prove that erythrophagocytosis evokes macrophage activation and foam cell formation in an atherosclerosis-free setting, mural thrombi were induced in the rabbit carotid artery.15 Neither macrophages, iNOS, neovessels, nor iron was found immediately after thrombus formation or in uninjured vessels (not shown). Seven days after injury, scattered RAM-11 immunoreactive macrophages were present between the erythrocytes, and they increased until week 3 (Figure 4). A significant fraction of these macrophages expressed iNOS. From day 14, microvessels appeared, and the macrophages formed multinucleated giant cell foam cells and showed iron deposits, which increased until week 3. At that time, autofluorescent ceroid was clearly visible in the macrophages (please see online Figure I, available at http//: www.atvb.ahajournals.org). Strong positive correlations (P<0.001) existed between RAM-11 and iron deposition (RSpearman=0.68), iNOS and iron deposition (RSpearman=0.73), microvessels and RAM-11 (RSpearman=0.70), and microvessels and iNOS (RSpearman=0.62).
Macrophage Activation and Ceroid Formation In Vitro
J774 macrophages transformed into foam cells containing cytoplasmic oil red O–positive lipid droplets on incubation with ox-RBCs or platelets for 41 hours (Figure 5). Fluorescent ceroid pigment was seen in macrophages exposed to ox-RBCs; this became more evident 2 days later (not shown). Increased nitrite concentrations were produced by macrophages incubated with ox-RBCs (2.0±0.3 μmol/L, control 0.2±0.1 μmol/L, n=4, P=0.027) or platelets (0.8±0.3 μmol/L, control 0.4±0.2 μmol/L, n=4, P=0.029 by paired t tests).
This quantitative study of human carotid artery plaques demonstrated (1) a highly significant association between the overall densities of microvessels and activated iNOS-expressing macrophages; (2) that macrophages were particularly abundant around a subset of microvessels with activated endothelial cells, as indicated by perivascular vWf deposits; (3) that in those regions, phagocytosis of erythrocytes and blood platelets may occur; and (4) that this phagocytic response evokes macrophage activation and foam cell formation.
Because endothelial cells raise the biosynthesis and deposition of vWf during atherogenesis,13,14⇓ vWf was selected as a vascular marker. To facilitate detection of perivascular vWf deposits, simultaneous staining of CD-31 and CD-345 was omitted. This could predispose to underestimation of vWf-negative microvessels. Yet, microvessels were detected in more cases (93%) than in triple-stained cross sections (60%) of endarterectomy specimens,5 presumably because a much greater plaque area is sampled in longitudinal sections. Another advantage of whole-mount preparations is the occurrence of several stages of atherosclerosis in each specimen,11 although assigning a single AHA score to the specimen becomes impossible. Because the strength of the AHA classification is also debated,2 an unbiased hierarchical cluster analysis was used to categorize the data. This algorithm takes not only the presence or absence but also the abundance of each variable into account and classifies the microvessels into 2 categories. One category (B, present in all but 1 plaque) was lined by flat endothelial cells without vWf deposits and contained few perivascular macrophages. Those microvessels were present in fibrous regions and at the base of the plaque. The other category (C) was lined by cuboidal endothelial cells showing increased expression and perivascular deposits of vWf. Ceroid-containing foam cells of macrophage origin, which were activated as demonstrated by their iNOS expression, surrounded the microvessels of the second category. Those microvessels were detected in 27% of the specimens, but their incidence may increase when more sections per specimen are analyzed.
By use of different approaches, a higher microvessel density has been demonstrated in lipid-rich plaques compared with fibrous plaques,18 and microvessels in shoulder regions are often associated with mast cells and T cells as well.5,19⇓ Furthermore, endothelial cells in lipid-rich plaques express increased levels of intercellular adhesion molecule-1, vascular adhesion molecule-1, and E-selectin.18,20⇓ Those findings indicate that microvessels in lipid-rich plaques are important in the continuous recruitment of macrophages in atherosclerotic plaques. In this concept, the recruited macrophages transform into foam cells as a consequence of uptake of plasma-derived LDL.
The present study suggests an additional mechanism of foam cell formation. First, we demonstrated immunoreactivity of the platelet marker GPIbα, as well as hemoglobin, heme oxygenase-1, and iron, in the cytoplasm of macrophages around microvessels in certain lipid-rich plaque regions, pointing to microhemorrhages and subsequent phagocytosis of platelets and erythrocytes. Furthermore, microvessels, located around GPIbα-immunoreactive macrophages, often contained platelet-rich microthrombi. Erythrophagocytosis and processing of hemoglobin lead to the accumulation of heme and iron in the macrophage. This could explain the expression of heme oxygenase-1 in these macrophages, inasmuch as heme is a potent inducer of heme oxygenase-1.21 Heme oxygenase-1 protects the macrophages against oxidative stress and cell death by regulating cellular iron.22 Interestingly, iron and ceroid colocalization has been demonstrated in human plaques.8 Moreover, erythrophagocytosis increases the capacity of the macrophage to oxidize LDL,23 and iron promotes ceroid formation.8 Collectively, these data suggest that hemorrhages occur in certain foam cell–rich areas and that macrophages surrounding those microvessels transform into activated foam cells producing ceroid pigment and iron as a result of repeated phagocytosis of platelets and erythrocytes.
It is unclear whether quiescent microvessels in fibrous regions (category B) will eventually become fragile and prone to rupture because of endothelial cell activation and recruitment of inflammatory cells. Ongoing inflammation, matrix-degrading metalloproteinases released from macrophages, or proteases secreted by accumulating mast cells could invoke proteolytic damage to the microvessels and facilitate intraplaque microhemorrhages.5,19⇓ Previously, we reported that foam cell accumulation is associated with smooth muscle cell death.24 Therefore, it would be interesting to investigate whether apoptosis of vascular cells leads to the transition of the foam cell–rich vascular regions (category C) to avascular areas (category D) around the necrotic core. Furthermore, erythrophagocytosis could also promote expansion of the necrotic core, inasmuch as hemoglobin-derived iron and phagocyte-generated oxidants collaborate to cause macrophage dysfunction.25
Because human plaques already contain plasma-derived lipoproteins, further studies were performed in normocholesterolemic rabbits and with macrophages in culture to prove that the lipids in perivascular foam cells can originate from phagocytosed membranes. The in vivo experiment confirmed the formation of macrophage-derived foam cells in organizing thrombi after platelet or erythrocyte phagocytosis in the absence of elevated LDL levels.26,27⇓
To exclude the possibility that increased oxidation of LDL after erythrophagocytosis23 might explain the accumulation of lipids and ceroid during thrombus organization, murine J774 macrophages were incubated with human blood platelets or ox-RBCs. Platelet phagocytosis evoked foam cell formation, confirming older reports,27,28⇓ and macrophage activation,12 as indicated by the iNOS metabolite nitrite. Recently, we reported that platelet phagocytosis also stimulates tumor necrosis factor-α and cyclooxygenase-2 expression, that priming with interferon-γ drastically raises J774 activation, and that the activation occurs in human macrophages as well.12 In the present study, we show that macrophages after uptake of ox-RBCs, presumably via scavenger receptors,17 (1) transform to foam cells as well, (2) become activated, as indicated by the iNOS activity, and (3) produce autofluorescent pigments with characteristics of ceroid in foam cells around microvessels in human plaques in spite of the serum-free environment. Thus, both experiments directly support the hypothesis that lipids and ceroid in foam cells around microvessels could be derived from platelets and erythrocytes. Previously, we demonstrated that ceroid and iNOS expression are frequently associated in macrophages in human plaques.11 The present study could explain this phenomenon because we now show that erythrophagocytosis, similar to platelet phagocytosis,12 stimulates iNOS expression and because iron is known to promote lipid peroxidation25 and ceroid formation.8 The function of iNOS expression is not clear, but possible roles in atherogenesis and thrombus vascularization have to be considered because plaque progression29 and angiogenesis during wound healing30 are reduced in iNOS-deficient mice.
In summary, the cluster analysis provided an unbiased confirmation of the abundance of activated microvessels in certain lipid-rich atherosclerotic plaques reported by others.5,18,20⇓⇓ Their findings were extended by demonstrating platelet and erythrocyte phagocytosis by the perivascular foam cells. Moreover, these foam cells were activated (as indicated by their iNOS expression) and produced autofluorescent pigments that were compatible with ceroid. During thrombus organization and also in vitro, the same foam cells, ceroid formation, and iNOS upregulation could be induced in macrophages by platelet or erythrophagocytosis in atherosclerosis-free settings. Therefore, we conclude that in some plaques, microhemorrhages or thrombi initiate the phagocytosis of platelets and erythrocytes. This not only explains iron deposition,6 but it also invokes nonimmune macrophage activation, foam cell formation, and ceroid production. These focal processes may contribute to plaque expansion3 and may eventually promote plaque vulnerability.4,5⇓
The financial support by the Fund for Scientific Research-Flanders (F.W.O. grants G.0180.01, G.0427.02) and IUAP P5/02 is greatly appreciated. Dr Kockx is a recipient of the FWO fund for fundamental clinical research. The authors wish to thank Anneliese Van Hoydonck, Hermine Fret and Martine De Bie for technical assistance and Liliane Van den Eynde for secretarial assistance.
Received August 20, 2002; revision accepted December 17, 2002.
- ↵Stary HC, Chandler AB, Dinsmore RE, Fuster V, Glagov S, Insull W Jr, Rosenfeld ME, Schwartz CJ, Wagner WD, Wissler RW. A definition of advanced types of atherosclerotic lesions and a histological classification of atherosclerosis. Circulation. 1995; 92: 1355–1374.
- ↵Virmani R, Kolodgie FD, Burke AP, Farb A, Schwartz SM. Lessons from sudden coronary death: a comprehensive morphological classification scheme for atherosclerotic lesions. Arterioscler Thromb Vasc Biol. 2000; 20: 1262–1275.
- ↵Paterson JC. Capillary rupture with intimal haemorrhages as a causative factor in coronary thrombosis. Arch Pathol. 1938; 25: 474–479.
- ↵Paterson JC, Moffat T, Mills J. Haemosiderin deposition in early atherosclerotic plaques. Arch Pathol. 1956; 61: 496–502.
- ↵Wilcox JN, Subramanian RR, Sundell CL, Tracey WR, Pollock JS, Harrison DG, Marsden PA. Expression of multiple isoforms of nitric oxide synthase in normal and atherosclerotic vessels. Arterioscler Thromb Vasc Biol. 1997; 17: 2479–2488.
- ↵Cromheeke KM, Kockx MM, De Meyer GRY, Bosmans JM, Bult H, Vrints CJ, Herman AG. Inducible nitric oxide synthase colocalizes with signs of lipid oxidation/peroxidation in human atherosclerotic plaques. Cardiovasc Res. 1999; 43: 744–754.
- ↵De Meyer GRY, De Cleen DM, Cooper S, Knaapen MWM, Jans DM, Martinet W, Herman AG, Bult H, Kockx MM. Platelet phagocytosis and processing of beta-amyloid precursor protein as a mechanism of macrophage activation in atherosclerosis. Circ Res. 2002; 90: 1197–1204.
- ↵Mannucci PM. von Willebrand factor: a marker of endothelial damage? Arterioscler Thromb Vasc Biol. 1998; 18: 1359–1362.
- ↵De Meyer GRY, Hoylaerts MF, Kockx MM, Yamamoto H, Herman AG, Bult H. Intimal deposition of functional von Willebrand factor in atherogenesis. Arterioscler Thromb Vasc Biol. 1999; 19: 2524–2534.
- ↵Bosmans JM, Kockx MM, Vrints CJ, Bult H, De Meyer GRY, Herman AG. Fibrin(ogen) and von Willebrand factor deposition are associated with intimal thickening after balloon angioplasty of the rabbit carotid artery. Arterioscler Thromb Vasc Biol. 1997; 17: 634–645.
- ↵Gundersen HJG, Bendtsen TF, Korbo L, Marcussen N, Moller A, Nielsen K, Nyengaard JR, Pakkenberg B, Sorensen FB, Vesterby A, West MJ. Some new, simple and efficient stereological methods and their use in pathological research and diagnosis. Acta Pathol Microbiol Immunol Scan. 1988; 96: 379–394.
- ↵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.
- ↵de Boer OJ, van der Wal AC, Teeling P, Becker AE. Leukocyte recruitment in rupture prone regions of lipid-rich plaques: a prominent role of neovascularization? Cardiovasc Res. 1999; 41: 443–449.
- ↵O’Brien KD, McDonald TO, Chait A, Allen MD, Alpers CE. Neovascular expression of E-selectin, intercellular adhesion molecule-1, and vascular adhesion molecule-1 in human atherosclerosis and their relation to intimal leukocyte content. Circulation. 1996; 93: 672–682.
- ↵Alam J, Shibahara S, Smith A. Transcriptional activation of the heme oxygenase gene by heme and cadmium in mouse hepatoma cells. J Biol Chem. 1989; 264: 6371–6375.
- ↵Kockx MM, De Meyer GRY, Bortier H, de Meyere N, Muhring J, Bakker A, Jacob W, Van Vaeck L, Herman A. Luminal foam cell accumulation is associated with smooth muscle cell death in the intimal thickening of human saphenous vein grafts. Circulation. 1996; 94: 1255–1262.
- ↵Loegering DJ, Raley MJ, Reho TA, Eaton JW. Macrophage dysfunction following the phagocytosis of IgG-coated erythrocytes: production of lipid peroxidation products. J Leukoc Biol. 1996; 59: 357–362.
- ↵Chandler AB, Hand RA. Phagocytized platelets: a source of lipids in human thrombi and atherosclerotic plaques. Science. 1961; 134: 946–947.
- ↵Curtiss LK, Black AS, Takagi Y, Plow EF. New mechanism for foam cell generation in atherosclerotic lesions. J Clin Invest. 1987; 80: 367–373.
- ↵Detmers PA, Hernandez M, Mudgett J, Hassing H, Burton C, Mundt S, Chun S, Fletcher D, Card DJ, Lisnock J, Weikel R, Bergstrom JD, Shevell DE, Hermanowski-Vosatka A, Sparrow CP, Chao YS, Rader DJ, Wright SD, Pure E. Deficiency in inducible nitric oxide synthase results in reduced atherosclerosis in apolipoprotein E-deficient mice. J Immunol. 2000; 165: 3430–3435.