Galectin-3 Is an Amplifier of Inflammation in Atherosclerotic Plaque Progression Through Macrophage Activation And Monocyte Chemoattraction
Objective— Galectin-3 (Gal-3) is a 26-kDa lectin known to regulate many aspects of inflammatory cell behavior. We assessed the hypothesis that increased levels of Gal-3 contribute to atherosclerotic plaque progression by enhancing monocyte chemoattraction through macrophage activation.
Methods and Results— Gal-3 was found to be upregulated in unstable plaque regions of carotid endarterectomy (CEA) specimens compared with stable regions from the same patient (3.2-fold, P<0.05) at the mRNA (n=12) and (2.3-fold, P<0.01) at the protein level (n=9). Analysis of aortic tissue from ApoE−/− mice on a high fat diet (n=14) and wild-type controls (n=9) showed that Gal-3 mRNA and protein levels are elevated by 16.3-fold (P<0.001) and 12.2-fold (P<0.01) and that Gal-3 staining colocalizes with macrophages. In vitro, conditioned media from Gal-3–treated human macrophages induced an up to 6-fold increase in human monocyte chemotaxis (P<0.01, ANOVA), an effect that was reduced by 66 and 60% by Pertussis Toxin (PTX) and the Vaccinia virus protein 35K, respectively. Microarray analysis of human macrophages and subsequent qPCR validation confirmed the upregulation of CC chemokines in response to Gal-3 treatment.
Conclusions— Our data suggest that Gal-3 is both a marker of atherosclerotic plaque progression and a central contributor to the pathology by amplification of key proinflammatory molecules.
Systemic and local inflammation underlies all phases of atherosclerotic plaque development, from initiation to plaque rupture leading to acute clinical events.1 Galectins are a family of highly conserved mammalian lectins that have recently been shown to function as novel regulators of inflammation.2,3 They are involved in cell adhesion, apoptosis, and chemotaxis and can also function as scavenger receptors.4 Recent work from our laboratories has identified at least 4 members of the Galectin family as present in advanced human atherosclerotic lesions, including Gal-3.5
Gal-3, a 26 kDa β-galactoside-binding protein, has been reported to regulate many inflammatory cell types including T and B lymphocytes,6,7 mast cells,8 neutrophils,9 and monocytes/macrophages.10,11 In vivo, targeted mutation of the Gal-3 gene results in an attenuated phagocytic clearance12 and inflammatory response13,14 after an immunologic challenge. In vitro, Gal-3 has been shown to bind to the cell surfaces of monocytes/ macrophages15 and cause superoxide release10 and Ca2+ influx.11 Gal-3 also induces the migration of monocytes and macrophages both in vivo and in vitro through a pertussis toxin (PTX)-sensitive pathway, which appears to differ from that used by the presently known monocyte attracting chemokines.11
The aim of this study was to assess the expression profile of Gal-3 in human and murine atherosclerotic lesions and to functionally characterize its potential role in plaque biology. We show that atherosclerotic lesions express increased levels of Gal-3 and that Gal-3 can exacerbate vascular inflammation by stimulating macrophages to express a range of chemokines and other proinflammatory molecules. Our data suggest that Gal-3 could serve as both a marker for macrophage-rich areas and a novel key upstream key target in antiinflammatory strategies for the treatment of atherosclerosis.
Collection of Human Specimens and Classification
Twenty-one human CEA specimens were collected from patients who underwent surgery at St Thomas’ Hospital. The study was approved by the Ethics Committee of St. Thomas’ Hospital and all patients gave informed consent (please see supplemental materials, available online at http://atvb.ahajournals.org). Twelve samples were used for mRNA and 9 for protein analysis. Plaques were segmented longitudinally and classified as either stable or unstable as previously described.5 Briefly, stable segments were covered by a smooth luminal surface indicative of an intact fibrous cap, whereas unstable segments had an ulcerated surface with or without thrombosis or hemorrhage.
To induce atherosclerosis, 3-week-old ApoE KO C57BL6 mice and wild-type controls were fed a Western-type diet (21% milk fat, 0.15% cholesterol; 100244 Dyets Inc) for 8 and 16 weeks, respectively. Aortic tissue from 18 ApoE−/− mice (11 and 19 weeks old) and 9 wild-type C57BL6 mice (19 weeks old) was used for mRNA, Western blot, and immunostaining analysis. All studies were conducted in accordance with UK Home Office Animals (Scientific Procedures) Act 1986. Freshly harvested thoracic aortas from ApoE−/− and wild-type C57BL6 were cleaned of adherent fat, cut to 1.5 cm in length, and snap frozen in liquid nitrogen (LN2) for mRNA and protein analysis. For immunostaining, tissues were perfusion-fixed in situ with 4% phosphate-buffered paraformaldehyde, excised and fixed in paraformaldehyde overnight, dehydrated in graded ethanols, and paraffin embedded.
Human Macrophage Isolation and Culture
Monocyte-derived macrophages were obtained from buffy-coat preparations (National Blood Service) by Ficoll density gradient centrifugation (Amersham), followed by adhesion-mediated purification on tissue culture plastic as previously described.16 Monocytes were seeded at a 2 million per well density (6-well plates, Costar) and left to differentiate into macrophages for 7 days by culturing in medium (X-vivo, Lonza) containing autologous fibrin-depleted plasma. On day 7, cells were treated with multiple batches of endotoxin-free recombinant human Gal-3 (R&D Systems) in plasma free media for 24 hours.
Human Monocyte Isolation and Chemotaxis Assay
Human monocytes were obtained from buffy coats (n=3) and healthy volunteers (n=2) by Optiprep density gradient centrifugation of whole blood in EDTA as previously described.17 Monocyte chemotaxis was performed using a 96-well chemotaxis plate (Neuroprobe). In brief, 25 μL of stimulant or chemotaxis buffer (25 mmol/L Hepes (Gibco) and 0.1% BSA (Sigma) in serum-free RPMI, Gibco) was added to the bottom wells of the plate. Stimulants included combinations of recombinant Gal-3 (R&D Systems), 200 ng/mL PTX (Sigma), and 1 μL of culture medium from Ad35K and AdGFP-infected 293 cells generated as previously described.18 A 5-μm membrane was placed in the assembly, and 75 000 monocytes in 30 μL of buffer were placed on the top of the membrane. The chemotaxis plate was incubated for 2 hours at 37°C with 5% CO2. The filters were removed, and the membranes were fixed in 4% formalin/PBS (BDH) and stained with DAPI (Molecular Probes). Each test group was assayed in quadruplicate. Three high power (×400) microscope fields were counted in each replicate well, and the results were expressed as a migration index (cell chemotaxis in response to stimulant divided by chemotaxis in response to buffer).
RNA Isolation and Quantification
RNA was isolated from cultured macrophages, from 12 CEA specimens (stable segments: ×12, unstable segments: ×12) and from murine thoracic aorta tissue using an adaptation of the Trizol method (Invitrogen) and the RNeasy mini column method (Qiagen) as previously described.5 Plaque segments were homogenized in Trizol reagent (Invitrogen) at a ratio of 1 mL Trizol per 80 mg tissue using an Ultra-Turrax T25 homogenizer (Ika Labortechnik). RNA quality was assessed using the RNA 6000 Nano LabChip Kit (Agilent Bioanalyser 2100, Agilent Technologies).
Microarray Analysis Using Illumina BeadChip
Microarray analysis of human macrophage gene expression in response to Gal-3 treatment was performed using the Illumina BeadChip system. Total RNA (200 ng) from Gal-3 treated and untreated human macrophages (n=3 donors) was amplified into biotin-labeled cRNA using the Illumina TotalPrep RNA amplification kit (Ambion) according to the manufacturer’s protocol. Amplified RNA (1.5 μg) was hybridized to the “whole genome” Sentrix Human-6 Expression BeadChips19 (Illumina, please see supplemental materials).
Expression Analysis by qPCR
The expression of selected genes was assessed independently by qPCR. Total RNA was reverse-transcribed using Superscript III (Invitrogen) and qPCR amplification was performed using the QuantiTect SYBR Green kit (Qiagen) on the Rotor-Gene 3000 (Corbett Research). Target gene quantities were determined by comparison to a genomic DNA standard curve (Promega). Gene expression was normalized against GAPDH or β-actin. Dissociation curve and BLAST analysis ensured the specificity of all assays (please see supplemental materials). Genes were selected for confirmation according to their function and their potential role in regulating monocyte influx.
Protein lysates from CEA specimens (stable segments n=9, unstable segments n=9) and murine aortas (ApoE−/− n=4, wild-type n=4) were resolved using SDS-PAGE and transferred to nitrocellulose membranes. Western blotting was performed using a monoclonal antibody against human Gal-3 (R&D Systems) and a polyclonal goat antibody against murine Gal-3 (R&D Systems) followed by appropriate horseradish peroxidase–conjugated secondary antibodies (Promega, please see supplemental materials).
Immunohistochemistry was performed on serial sections of paraffin embedded ApoE−/−murine aortic root and brachiocephalic artery (BCA) lesions from 19-week-old mice after 16 weeks on a high-fat diet. Sections were stained for Gal-3 (R&D systems; 2 μg/mL), Mac-3 (BD Biosciences; 0.6 μg/mL) or α-SMA (Sigma-Aldrich; 1:50, please see supplemental materials).
Expression of Gal-3 in Stable and Unstable Segments of CEA Specimens
Our previous work, using microarray analysis, identified Gal-3 as an abundant transcript in CEA specimens.5 We assessed the expression of Gal-3 on a new panel of morphologically classified CEA specimens and found Gal-3 expression to be significantly higher in unstable regions of CEA specimens. Unstable segments of CEA specimens (n=12) contained 3.17-fold higher mRNA levels of Gal-3 than stable segments from the same patient (n=12), as determined by qPCR (paired t test, P<0.05, Figure 1A). Gal-3 mRNA expression correlated significantly with the expression of the macrophage marker CD68 (R2=0.62, P<0.0001, Figure 1B). At the protein level, unstable segments (n=9) contained 2.3-fold higher protein levels of Gal-3 as determined by Western blotting (paired t test, P<0.01) compared with their paired stable segments (n=9, Figure 1C and 1D). In vitro, human macrophages were found to be the main source of Gal-3 production at the mRNA level and the only source of Gal-3 at the protein level after examining expression in resting primary macrophages, smooth muscle cells, endothelial and T-cells (supplemental Figure I).
Expression of Gal-3 in Aortic Tissue From ApoE−/− Mice and Wild-Type Controls
To determine whether Gal-3 in expressed in murine atherosclerotic plaques, we analyzed aortic tissue from ApoE−/− mice and wild-type controls on a high-fat diet and found that Gal-3 also increased with age in ApoE−/− mice compared with wild-type controls. Aortic tissue from 19-week-old ApoE−/− mice (n=5) after 16 weeks on a high-fat diet contained 16.3-fold higher mRNA levels of Gal-3 compared with wild-type (P<0.001) and 5.3-fold higher levels compared with 11-week-old ApoE−/− mice (P<0.01, Tukey multiple comparison test, ANOVA, P<0.001). Gal-3 mRNA expression correlated significantly with the expression of the macrophage marker CD68 (R2=0.94, P<0.0001, Figure 2B). Nineteen-week-old ApoE−/− aortic tissue (n=4) also contained 12.2-fold higher protein levels of Gal-3 compared with wild-type (P<0.0001, unpaired Student t test, Figure 2C).
Gal-3, Mac-3, and α-actin staining of aortic root (Figure 3) and brachiocephalic artery (supplemental Figure IV) lesions of 19-week-old ApoE−/− mice after 16 weeks on a high-fat diet showed that Gal-3 colocalizes with macrophage-rich Mac-3 positive areas and not with smooth muscle cell-rich α-actin positive areas.
Chemotaxis of Human Monocytes in Response to Gal-3 and Gal-3 Conditioned Media
Recombinant Gal-3 induced an increase in human monocyte chemotaxis in a dose-dependent manner (P<0.01, ANOVA) of up to 6.14-fold at 160 nmol/L. In the presence of 30 nmol/L lactose (a Gal-3 ligand that binds to and inactivates the lectin part of the molecule in vitro11) this effect was abolished at all concentrations (Figure 4A). Conditioned media from Gal-3 treated human macrophages also induced an increase in human monocyte chemotaxis of up to 10.9-fold at 160 nmol/L of Gal-3 treatment compared with chemotaxis buffer (in a dose-dependent manner, P<0.01, ANOVA). This effect was only partially blocked by lactose (up to 50% reduction) indicating that Gal-3 induced the release of other monocyte chemoattractants that do not require a lactose-inhibitable lectin domain to induce chemoattraction (Figure 4B). Chemotaxis induced by Gal-3 conditioned media in the presence of 30 nmol/L lactose was reduced by up to a further 66.1% in the presence of 200 ng/mL PTX (Figure 4C), indicating that a large percentage of the mediators released in response to Gal-3, exert their effects through G protein–coupled receptor signaling, Gαi/o subunit in particular.20 Chemotaxis was also reduced by 60.2% in the presence of the vaccinia virus CC-chemokine inhibitor 35K,18,21 indicating a major role for CC-chemokines in the secondary chemotaxis (Figure 4D). Each chemoattractant was assayed in quadruplicate. Data shown are representative data from 4 donors.
In pilot experiments we showed that macrophage expressed tumor necrosis factor (TNF)α mRNA and interleukin (IL)-6 mRNA levels are upregulated in response to Gal-3 in a dose-dependent manner (by 9.1- and 12.3-fold, respectively at 160 nmol/L; please see supplemental Figure II). A more comprehensive analysis of proinflammatory gene expression in response to Gal-3 was undertaken using microarray analysis. 576 unique mRNA species were found to be upregulated in response to Gal-3 treatment (supplemental Table II). Forty-eight of these genes were classified by Gene Ontology into groups relevant to inflammation (inflammatory response, immune response, GPCR binding, and chemotaxis; Table).
Confirmation of Candidate Genes
The expression of 5 genes selected for their potential to mediate inflammatory cell chemotaxis and their role in inflammation was confirmed by qPCR in a cDNA panel from Gal-3 treated and untreated human macrophages from 5 independent donors. Genes that were confirmed to be upregulated include the chemokines CCL5 (7.6-fold), CCL2 (27-fold), CCL8 (22.4-fold), CCL20 (21.3-fold), and CXCL8 (3.35-fold, P<0.001, ANOVA, Figure 5).
The expression of vascular noninflammatory molecule 2 (VNN2), found to be downregulated in response to Gal-3 treatment in the microarray analysis, was included as control and was also confirmed by qPCR (−2.76-fold, P<0.0001, ANOVA, Figure 5).
Correlation of Gal-3 mRNA With Gal-3 Induced Chemokines in Human CEA Specimens
The mRNA expression of 5 chemokines that were confirmed to be upregulated in response to Gal-3 treatment in vitro, correlated significantly with Gal-3 levels in stable and unstable segments of human CEA specimens. The expression of CCL2 (R2=0.26, P=0.01), CCL5 (R2=0.40, P=0.0009), CCL8 (R2=0.34, P=0.0026), CCL20 (R2=0.69, P<0.0001), and CXCL8 (R2=0.39, P=0.0012) had a similar expression pattern to Gal-3 in human CEA specimens with CXCL8 having the most significant correlation with Gal-3 (supplemental Figure III).
We assessed the expression of Gal-3 in human and murine atherosclerotic plaques and its potential role in plaque biology using primary monocytes and macrophages in vitro. This is the first report to show that human Gal-3 expression is upregulated in unstable regions of human atherosclerotic plaques and in murine ApoE−/− atherosclerotic plaques, compared with stable regions and wild-type controls, respectively. We have shown for the first time that Gal-3 localizes in macrophage-rich of ApoE−/− murine plaques and that it represents an excellent new marker of murine atherosclerotic macrophage-rich areas.
We have found that Gal-3 amplifies inflammation by inducing the expression of a series of well-known proinflammatory molecules in plaque pathology. Conditioned media from Gal-3 treated primary macrophages induced an up to 11-fold increase in monocyte chemoattaction. This effect shown to be largely independent of Gal-3 (using lactose to bind to and inactivate Gal-3), indicating that induced proinflammatory molecules are responsible for these functional effects. These molecules appear to act via Gαi linked G protein–coupled receptors (GPCRs) as PTX caused a 66% inhibition of chemotaxis. A microarray experiment, followed by real-time polymerase chain reaction (PCR) validation, indeed confirmed the upregulation of a range of monocyte chemoattractant molecules that act via the Gαi domain of G protein–coupled receptors in response to Gal-3 treatment, including several members of the chemokine family. CC chemokines, in particular, appear to be mainly responsible for the secondary chemotaxis induced by Gal-3 as 35K—a vaccinia virus protein that is known to bind to and inactivate all CC chemokines18,21—was able to inhibit the chemotactic effects of these mediators by 60%.
Galectin-3 was not only found to be upregulated in advanced lesions, but was an abundant molecule at both mRNA and protein level. Gal-3 mRNA levels correlated significantly with CD68 levels, which we have previously shown to be elevated in unstable regions of CEA specimens and not with a-actin levels (smooth muscle cell–SMC marker) which are low in unstable plaque areas.5 These findings, in combination with the notion that high macrophage numbers and low smooth muscle cell content are associated with plaque vulnerability,22 suggest that Gal-3 levels could potentially be used as a novel biomarker of advanced plaques. In addition, further correlations of Gal-3 expression in human CEA specimens with proinflammatory mediators identified in in vitro experiments, indicate that chemokines such as CCL2, -5, -8, and -20 and CXCL2 are upregulated in regions of human atherosclerotic plaque where Gal-3 levels are also elevated. This observation provides further evidence that Gal-3 is responsible for the amplification of the expression of such inflammatory mediators within the atherosclerotic plaque.
Gal-3 has been previously shown to be expressed in aortas of hypercholesterolemic rabbits and in aortas of rats after balloon injury, but not in normal rabbit and rat aorta, respectively.23 The authors suggested that despite the lack of Gal-3 expression by quiescient SMCs in vitro, activated SMCs could be the main source of Gal-3 production in experimental atherosclerotic lesions.23 In our study, using immunolocalization analysis of ApoE−/− mice atherosclerotic lesions and correlations of Gal-3 mRNA expression with macrophage marker expression, we demonstrate that macrophages are, in fact, the main source of Gal-3 production in human and murine ApoE−/− lesions. In addition, we confirm these findings by assessing the expression of Gal-3 at protein and mRNA level in all main relevant human cell types in vitro (supplemental Figure I). Macrophages appear to be the only source of Gal-3 production at both protein and mRNA level after assessment of Gal-3 expression in primary human macrophages, smooth muscle cells, endothelial cells, and T-cells.
As a secreted protein, Gal-3 can act in an autocrine manner to activate macrophages to secrete higher levels of proinflammatory and chemoattractant molecules. The structural properties of Gal-3 are ideal in allowing this molecule to activate cells in a proinflammatory manner, by cross-linking of surface proteins. As a multivalent lectin, Gal-3 is able to cross-link glycosylated membrane receptors by binding their glycan parts with its C-terminal domain.24 This can trigger multiple signal transduction cascades, which in turn results in cell activation. By activating macrophages and possibly other cell types within the atherosclerotic plaque, Gal-3 may propagate the release of proinflammatory mediators that ensure a continuous influx of monocytes within the vessel wall.
On a systemic level, it is interesting to note that recently, Gal-3 levels in sera and synovial fluid from rheumatoid arthritis (RA) patients were found to correlate with C-reactive protein levels25 suggesting that Gal-3 could also be used as a prognostic circulating marker. In preliminary investigations, we have been able to detect Gal-3 in the plasma of a cohort of 30 patients undergoing coronary artery bypass graft surgery at an average concentration of 3 ng/mL (±SD 1.2 ng/mL, data not shown), however larger clinical studies of patients with well-defined coronary artery or peripheral vascular disease and suitable controls would be needed to properly assess Gal-3 as a circulating biomarker of atherosclerosis.
Previous studies have suggested that Gal-3 is a “macrophage activation” marker mainly because of its elevated expression levels in phagocytic macrophages.26 These findings in combination with our new in vitro data and the demonstration of Gal-3 upregulation in advanced lesions support the role of Gal-3 as an amplifier of inflammation and suggest that Gal-3 may find application as a novel biomarker of advanced atherosclerotic lesions. A selective, sensitive, and high-affinity ligand and inhibitor of Gal-3 has recently emerged as a promising novel tool that could be used to visualize Gal-3 activity in vivo.27 If Gal-3 levels and activity detection could be coupled with imaging technologies, they could provide accurate information not just on the size of the atherosclerotic lesion but potentially its inflammatory state.
In summary, we have shown that Gal-3 mRNA and protein expression is elevated in unstable areas of human atherosclerotic plaques and that Gal-3 expression also increases with age and plaque size in murine atherosclerotic plaques, whereas the source of Gal-3 production in atherosclerotic plaques appears to be the macrophage cell. Gal-3 appears to play an active role in the propagation of inflammation as it induces the expression of a variety of well characterized proinflammatory mediators by human macrophages. Gal-3 may represent a promising novel biomarker and therapeutic target in antiinflammatory strategies in atherosclerotic disease.
Sources of Funding
This work was supported by the British Heart Foundation.
DRG has received lecture fees from ChemoCentryx Inc. and Takeda.
Original received August 3, 2007; final version accepted December 3, 2007.
Ross R. Atherosclerosis–an inflammatory disease. N Engl J Med. 1999; 14: 115–126.
Papaspyridonos M, Smith A, Burnand KG, Taylor P, Padayachee S, Suckling KE, James CH, Greaves DR, Patel L. Novel candidate genes in unstable areas of human atherosclerotic plaques. Arterioscler Thromb Vasc Biol. 2006; 26: 1837–1844.
Yamaoka A, Kuwabara I, Frigeri LG, Liu FT. A human lectin, galectin-3 (epsilon bp/Mac-2), stimulates superoxide production by neutrophils. J Immunol. 1995; 154: 3479–3487.
Sano H, Hsu DK, Yu L, Apgar JR, Kuwabara I, Yamanaka T, Hirashima M, Liu FT. Human galectin-3 is a novel chemoattractant for monocytes and macrophages. J Immunol. 2000; 165: 2156–2164.
Frigeri LG, Liu FT. Surface expression of functional IgE binding protein, an endogenous lectin, on mast cells and macrophages. J Immunol. 1992; 148: 861–867.
Kuhn K, Baker SC, Chudin E, Lieu MH, Oeser S, Bennett H, Rigault P, Barker D, McDaniel TK, Chee MS. A novel, high-performance random array platform for quantitative gene expression profiling. Genome Res. 2004; 14: 2347–2356.
Fairley-Grenot K, Assmann SM. Evidence for G-protein regulation of inward K+ channel current in guard cells of fava bean. Plant Cell. 1991; 3: 1037–1044.
Bursill CA, Choudhury RP, Ali Z, Greaves DR, Channon KM. Broad-spectrum CC-chemokine blockade by gene transfer inhibits macrophage recruitment and atherosclerotic plaque formation in apolipoprotein E-knockout mice. Circulation. 2004; 110: 2460–2466.
Davies MJ, Richardson PD, Woolf N, Katz DR, Mann J. Risk of thrombosis in human atherosclerotic plaques: role of extracellular lipid, macrophage, and smooth muscle cell content. Br Heart J. 1993; 69: 377–381.
Elliott MJ, Strasser A, Metcalf D. Selective up-regulation of macrophage function in granulocyte-macrophage colony-stimulating factor transgenic mice. J Immunol. 1991; 147: 2957–2963.