Predominance of a Proinflammatory Phenotype in Monocyte-Derived Macrophages From Subjects With Low Plasma HDL-Cholesterol
Objective— Reduced plasma concentrations of high-density lipoprotein-cholesterol (HDL-C) are a significant risk factor for cardiovascular disease. Mechanisms that regulate HDL-C concentrations represent an important area of investigation.
Methods and Results— Comparative transcriptome analyses of monocyte-derived macrophages (MDM) from a large population of low HDL-C subjects and age- and sex-matched controls revealed a cluster of inflammatory genes highly expressed in low HDL-C subjects. The expression levels of peroxisome proliferator activated receptor (PPAR) γ and several antioxidant metallothionein genes were decreased in MDM from all low HDL-C groups compared with controls, as was the expression of other genes regulated by PPARγ, including CD36, adipocyte fatty acid binding protein (FABP4), and adipophilin (ADFP). In contrast, PPARδ expression was increased in MDM from low HDL-C groups. Quantitative RT-PCR corroborated all major findings from the microarray analysis in two separate patient cohorts. Expression of several inflammatory cytokine genes including interleukin 1β, interleukin 8, and tumor necrosis factor α were highly increased in low HDL-C subjects.
Conclusions— The activated proinflammatory state of monocytes and MDM in low HDL-C subjects constitutes a novel parameter of risk associated with HDL deficiency, related to altered expression of metallothionein genes and the reciprocal regulation of PPARγ and PPARδ.
Reduced plasma concentrations of high-density lipoprotein- cholesterol (HDL-C) are a significant risk factor for cardiovascular disease, epidemiologically documented in numerous population studies and further supported through interventional randomized controlled trials.1 HDL mitigates against atherosclerosis through multiple mechanisms including enhancement of reverse cholesterol transport,2 regulation of eNOS and endothelial function,3,4 suppression of the induction of cell-adhesion molecules,5 antiinflammatory effects,6 and antioxidative properties.7,8 The metabolic and genetic etiologies of low HDL-C (hypoalphalipoproteinemia) are varied. Overproduction and/or impaired clearance of triglyceride (TG)-rich lipoproteins results in TG enrichment of HDL and accelerated apoA-I catabolism and may account for low HDL-C levels in many patients with other features of the metabolic syndrome. Low HDL-C can also result from mutations in a number of genes regulating HDL production or catabolism: ApoA-I,9,10 lecithin: cholesterol acyltransferase (LCAT),11,12 phospholipid transfer protein (PLTP),13–15 and the ATP-binding cassette protein, ABCA1.16
Here we have studied a previously defined cohort of subjects with low HDL and controls that was extensively characterized in terms of incidence of nonsynonymous mutations in apoA-I, ABCA1, LCAT, and PLTP17 and in whom the incidence of cholesterol efflux defects from monocyte-derived macrophages (MDM) had been defined.17a This study has established the high prevalence of efflux defects in low HDL syndromes that are not attributable to nonsynonymous mutations in ABCA1. MDM from the low HDL and control subjects were used to evaluate gene expression by microarray analysis as a function of efflux defects, functional mutations in ABCA1, and absence of mutations in the other candidate genes. Here, we demonstrate that cholesterol loaded MDM from low HDL-C subjects exhibit a complex inflammatory gene response, independent of efflux defects, but related to altered expression of PPARγ, PPARδ, and a cluster of metallothionein genes. Our results suggest that a heightened macrophage inflammatory response may contribute to the pathophysiological consequences of HDL deficiency.
For details, please consult the method section in the supplemental materials, available online at http://atvb.ahajournals.org.
Selection of Low HDL-C Subjects and Controls
Low HDL-C subjects were selected from consecutive referrals to the Heart Institute Lipid Clinic based on the following criteria: White, untreated values for HDL-C <5th%ile, TG <95th%ile (most <75th%ile), LDL-C <75th%ile adjusted for age and sex. Exclusion criteria included diabetes and clinical conditions or medications causative of low HDL. Control subjects were healthy normolipemic volunteers of the same ethnic background recruited from Ottawa region. The study was approved by the University of Ottawa Heart Institute Human Research Ethics Committee, and written informed consent was obtained from all participants. Further details are provided in supplemental Table I and supplemental methods.
Monocytes were isolated over Histopaque 1077 and plated at 4.4×106 cells/mL in 48-well plates (efflux assays) or 12-well plates (mRNA isolation). Peripheral blood mononuclear cells (PBMCs) were plated for 48 hours before harvesting of mRNA. MDM were fed every 48 hours and cultured for 12 days to differentiate monocytes into macrophages.
Macrophage Cholesterol Loading and Efflux Assay
Cells were incubated for 36 hours with acetylated LDL (AcLDL) (37.5 μg AcLDL/mL) labeled with 3H-cholesterol, equilibrated in media with 2 mg/mL BSA for 12 hours and mRNA was isolated. For efflux assay, cells were incubated without and with apoA-I (50 μg/mL) in RPMI/BSA for 2 hours.
RNA was assessed with an Agilent 2100 Bioanalyzer (Agilent Technologies) and RiboGreen RNA quantitation reagent (Molecular Probes). Affymetrix human U133A GeneChips (Affymetrix Inc) were used.
Taq-Man Quantitative Real-Time PCR
Total RNA (1 mg) was converted to cDNA using a High Capacity cDNA Archive Kit (Applied Biosystems) and the equivalent of 10 ng per well was arrayed into high-density 384-well plates using a Biomek FX robot (Beckman Coulter). Quantitative RT-PCR was carried out using a 7900HT Sequence Detector System (Applied Biosystems).
The Resolver error model18 was used to select probesets differing between experimental groups (Figure 1). Each group was compared with the control group using a cutoff of P<0.05 and xDEV greater than 2.5 in at least one comparison. Average linkage clustering, using the Pearson correlation, was applied and the data were viewed by Tree view (Treeview 1.60; Michael Eisen, Stanford.edu).
Subject Phenotyping and Genotyping
After extensive characterization of the MDM from low HDL subjects and after exclusion of all subjects with sequence variants in apoA-I, LCAT, or PLTP, subjects were divided into age- and sex-matched groups: A (low HDL-C, low efflux, no ABCA1 mutation); B (low HDL-C, low efflux, ABCA1 mutation); C (low HDL-C, normal efflux); D (controls with normal HDL-C and normal efflux). The 3 groups with low HDL-C demonstrated similar plasma lipoprotein values, and all groups were well matched for BMI.17a Only a few participants had clinically apparent vascular disease (ie, prior myocardial infarction, angiographic evidence of CAD, or positive stress test; supplemental Material and Table I and data not shown).
Microarray Transcriptome Analysis
MDM from 6 to 8 subjects in each group (all age- and sex- matched) were loaded with AcLDL. mRNA was isolated from cholesterol loaded MDM and analyzed using Affymetrix GeneChips (see Methods). Hierarchical clustering of significantly differing gene expression (P<0.05) for all subjects in all groups demonstrated significant homogeneity within each of the groups, with the exception of 2 subjects in group A (low efflux, normal ABCA1), who demonstrated distinct profiles with extremely altered expression of liver X receptor (LXR) and LXR-regulated genes. Sequencing of coding regions and 2 kb of the 5′flanking sequence of the LXR gene in these 2 subjects failed to identify any unique sequence variant. They were not included with the rest of group A (low efflux, normal ABCA1) (identified as A ) in the subsequent analyses because of possible skewing of the correlation.
The first microarray study of the 3 low HDL-C groups compared with controls demonstrated the relative homogeneity within groups and consistent differences between groups, as indicated by several highlights emerging from the hierarchical cluster analysis (Figure 1). First, a striking cluster consisting of the metallothionein-1 (MT1) genes was uniformly lower in MDM from all low HDL-C subjects (Figure 1, bottom panel). Secondly, the cluster analysis revealed increased expression of a group of cholesterol regulated genes in the two efflux defective groups, consistent with impaired intracellular trafficking of cholesterol (Figure 1, top panel).
Importantly, a remarkable cluster of highly expressed inflammatory genes was evident in MDM of low HDL-C subjects without ABCA1 mutations (groups A and C) compared with controls (Figure 1, middle panel). The same groups compared with controls had decreased expression of PPARγ, a nuclear receptor known to repress transcriptional activation of inflammatory response genes in macrophages.19,20 In contrast, PPARδ was increased in low HDL-C subjects compared with controls (Figure 1, top panel). We then examined the relationships of PPARγ and PPARδ mRNA expression to that of other genes tested in the microarray. Based on a Pearson correlation coefficient of r>0.6, PPARγ correlated positively with 105 genes and negatively with 21. Of these, data for genes associated with cholesterol metabolism or free radical scavenging are shown in Table 1. In this set, the metallothionein genes, adipophilin (ADFP), CD36, adipocyte fatty acid binding protein (FABP4), ATP binding cassette protein A1 (ABCA1), and liver X receptor (LXRα) all emerged as strongly positively correlated with PPARγ expression. Most of these same genes were negatively correlated with PPARδ (Table 1), suggesting a reciprocal regulatory relationship between PPARγ and PPARδ. On the other hand, PPARδ positively correlated with 138 genes (Table 1), which included a large number of cholesterol synthesis and cholesterol metabolism related genes, including low density lipoprotein receptor (LDLR), 3-hydroxy-3-methylglutaryl (HMG)-coenzyme A (CoA) reductase (HMGCR), and INSIG1.
The expression of selected genes, representative of the different clusters, was verified by TaqMan real-time PCR, and the results corroborated the major findings from this first series of microarray experiments (Table 2). MT1E expression was low in all low HDL-C groups compared with control. PPARγ expression was reduced in MDM from low HDL-C subjects, as were FABP4 and ADFP, genes regulated by PPARγ. In contrast, PPARδ was expressed at higher levels in low HDL-C groups as compared with controls. Again, the inflammatory cytokines, interleukin 1β (IL-1β), IL-8, tumor necrosis factor α (TNF-α), as well as superoxide dismutase 2 (SOD2) were significantly increased in low HDL-C subjects (Table 2).
To validate these observations, a second series of duplicate transcriptome analyses were carried out on MDM from an additional 21 low HDL-C subjects and 18 controls, genotyped and phenotyped as before (group A′: low HDL-C, low efflux, normal ABCA1 [10 subjects], group C′ low HDL-C, normal efflux [11 subjects], and group D′ normal controls [18 subjects]). Overall, the microarray results of this second series corroborated those of series 1 and revealed the same clusters of overexpressed and downregulated genes in the different groups. A heat map of the composite microarray for both series 1 and series 2 cohorts vividly illustrated the prominence of the inflammation phenotype in low HDL groups and its independence of efflux phenotype (Figure 2).
Unsupervised methods were also used to further analyze the complete dataset. Small but consistent changes between the two datasets led to adoption of an approach where the datasets were analyzed singly, and in concert, using K-means clustering. Probesets, which were consistently assigned as members of similarly behaving clusters in each of the 3 comparisons, were assigned into one of 5 patterns (supplemental Table II). Only 2 of these patterns show clear discrimination between low-HDL and normal groups, and these correspond to those previously identified from the supervised analysis of set 1. Comparisons to GO categories using Fisher exact test (supplemental Table II) confirmed our interpretations: (1) mRNA expression of many inflammation genes is increased in low-HDL subjects in the A and C groups; (2) a general gene cluster corresponding to energy metabolism with PPARγ at its center presented with reduced expression in most low HDL subjects. For further analysis, see the online data supplements for supplemental Results to Microarray Analysis.
TaqMan real-time PCR studies of the second series of low HDL and control subjects were consistent with the initial results and showed that inflammation-related genes were significantly increased in MDM derived from low HDL-C subjects with normal efflux. (ie, Group C; Table 2). The expression of PPARγ and certain PPARγ regulated genes including FABP4 were decreased in group C subjects (low HDL-C, normal efflux), whereas expression of PPARδ was increased in all low HDL-C subjects. The expression of inflammatory cytokines IL-1β and IL-8 was significantly higher in low HDL-C subjects in group C (low HDL-C, normal efflux) versus controls. To evaluate whether the proinflammatory phenotype might precede differentiation of monocytes and loading with cholesterol, low HDL-C subjects of group A (low HDL-C, low efflux) and controls were restudied. Peripheral blood mononuclear cells (PBMCs; including T-cells and monocytes) were isolated and cultured for 2 days in normal growth medium. Compared with controls, non–cholesterol loaded PBMCs from low HDL-C subjects demonstrated significantly higher expression of both IL-1β and IL-8 by quantitative RT-PCR (3.0- and 2.6-fold, respectively, P<0.01). Separately isolated and cultured MDM from low HDL-C subjects responded to cholesterol loading as expected by increasing ABCA1 and ABCG1 and decreasing HMG-CoA reductase, but the expression of MTE1 and proinflammatory genes did not increase with cholesterol loading (supplemental Table III). This indicates that reduced expression of metallothionein and elevated expression of inflammatory genes in MDM from low HDL-C subjects is not a direct consequence of cholesterol loading (supplemental Table II). Thus, the proinflammatory expression profile appears to be a constitutive feature of PBMC from low HDL-C subjects.
The gene expression profiles reflect the differing response of control and low HDL subjects to AcLDL loading and are likely dependent on the genotype of the monocyte/macrophage as well as potentially the epigenetic conditioning of monocyte precursors in the original low or normal HDL environments. Interestingly, Group B subjects who presented with efflux defects of a defined nature (ABCA1 functional mutations) clearly did not have an obvious proinflammatory phenotype (Figure 1), and neither did a group of low HDL subjects with defined functional mutations affecting LCAT function (data not shown). This result argues against low HDL by itself causing the inflammation. Close to 20% of patients with premature CAD have isolated hypoalphalipoproteinemia.21 However, low HDL is not consistently associated with premature CAD. Subjects with Tangier disease are only moderately more susceptible to heart disease than controls, even though their HDL levels are barely detectable.22 Thus, other properties of HDL are likely important, including its pro- and antiinflammatory effects.23,24 Keeping in mind that the vast majority of subjects with low HDL have no mutation in apoA-I, LCAT, PLTP, or ABCA1 (88% of low HDL subjects17a), the subjects in Group A and C are representative of the typical hypoalphalipoproteinemic clinical population. In view of this novel relationship between HDL and inflammation in macrophages, we are left with the question: Which comes first? Plasma HDL production is predominantly regulated by hepatic expression of apoA-I (the main protein component of HDL) and ABCA1 (essential for the initial lipidation of apoA-I). Monocytes with a proinflammatory signature could be recruited to the liver and macrophage inflammation could affect hepatic expression of apoA-I or ABCA1, but we have no evidence to support such hypotheses. Alternatively, these findings may reflect epigenetic conditioning. There are many subtypes of monocytes, and it is possible that a specific subtype of inflammation-prone monocyte may be selected for in bone marrow exposed to a low HDL environment. There is only scant evidence to support such a model, but we are currently examining gene expression in CD14(+)-monocytes and MDM from control and low HDL subjects to evaluate this possibility.
We consistently observed decreased expression of a cluster of metallothionein genes, which encode for a family of small proteins characterized by a high metal [Zn(II), Cu(I)] content that contribute to cellular protection from reactive oxygen species.25,26 They are free radical scavengers and have both antioxidant and antiapoptotic functions.27 Metallothionein gene expression was closely correlated with that of PPARγ (r=0.83 for MTE1), whereas expression of PPARδ was negatively correlated with that of MTE1 (r=−0.67 for MTE1). The basal transcription of PPARγ is negatively regulated by cellular cholesterol28 and was found to be decreased in AcLDL-treated macrophages.29 That its expression is further reduced in all low HDL-C groups compared with controls, as observed here, is significant. The lower PPARγ expression is consistent with the observed decrease of PPARγ-regulated genes, ADFP, CD36, ABCG1, and ABCA1.30,31 PPARγ is a general antiinflammatory regulator; it increases the expression of proteins involved in reverse cholesterol transport32 and suppresses proinflammatory signals.32,33 Metallothioneins have also been shown to decrease the expression of proinflammatory cytokines in mice with experimental autoimmune encephalitis.34 Thus, the low expression of both metallothioneins and PPARγ in monocyte-derived macrophages from low HDL-C subjects may be causally linked to the proinflammatory transcriptome profile in these cells.
FABP4, a PPARγ regulated gene, was the most prominently downregulated gene in all low HDL-C groups compared with control. FABP4 facilitates transport of fatty acid ligands to the nucleus,35,36 where it interacts directly with PPARγ.37,38 Various data support the concept that FABP4 is at the crossroads of 2 central pathways in macrophages, where it coordinates cholesterol metabolism and inflammatory response.39 In mice, it appears that FABP4 is required for expression of inflammatory cytokines and/or chemoattractants in macrophages and adipose tissue. Our findings that macrophages from low HDL-C subjects have decreased expression of both PPARγ and FABP4 and elevated expression of TNF-α and IL-1β suggest that the observed differences in cytokine expression in low HDL-C versus control macrophages are related to altered expression of MTE1 and/or PPARγ and FABP4.
In contrast to PPARγ and FABP4, PPARδ expression was increased in cholesterol-loaded macrophages from low HDL-C subjects compared with controls (Table 2). PPARδ agonists significantly increase HDL-C levels in both insulin-resistant rhesus monkeys40 and obese mice,41 whereas PPARδ agonists reduce cholesterol efflux and promote lipid accumulation in human macrophages.42 Human and mouse macrophages respond differently to PPARδ, the function of which in mice may overlap with that of PPARγ in the regulation of genes related to cholesterol efflux. The role of PPARδ in human macrophages remains unresolved.43 In our study, PPARδ expression strongly and positively correlated with the expression level of many cholesterol synthesis and metabolism genes, an effect also noted recently by others.44 The relationship of PPARδ to inflammation has been previously established in mice: macrophages overexpressing PPARδ display increased production of inflammatory proteins, while genetic ablation of PPARδ increases the availability of transcriptional repressors, resulting in decreased macrophage expression of MCP-1, IL-1β, matrix metalloproteinase (MMP)-9, and decreased lesion area.45 In the present study, we observed higher expression of both PPARδ and inflammatory cytokines in cholesterol-loaded macrophages of low HDL-C subjects compared with controls. Considered in the context of the observations in PPARδ KO mice, elevated PPARδ expression may be partly responsible for the observed increase in inflammatory cytokine expression in macrophages from subjects with low HDL-C.
Interestingly, the mRNA expression of neither PPARγ nor PPARδ correlated with that of individual inflammation genes, indicating that neither has an independent influence on inflammatory gene expression, but implying that they are major coregulators of lipid homeostasis in MDM: PPARγ regulates genes influencing lipid uptake (CD36), storage (ADFP), and efflux (LXRα),32 which directly upregulates ABCA1 and ABCG1. PPARδ negatively regulates those same genes and positively controls cholesterol metabolism. We hypothesize that the expression of the antioxidant metallothionein genes and the reciprocal relationship between PPARγ and PPARδ are major determinants of the MDM inflammatory response. It is worth noting that although unliganded PPARδ levels correlate with inflammatory gene expression in mouse macrophages, activation of PPARδ in mouse macrophages by the synthetic agonist GW501516 attenuates inflammatory gene expression.45
In conclusion, we have demonstrated the increased expression of proinflammatory genes in MDM from duplicate cohorts of well characterized subjects with low HDL-C as compared with age and sex matched controls. The proinflammatory expression profile appears to be a constitutive feature of PBMCs from low HDL-C subjects which may be established in response to genetic and nongenetic factors during the differentiation of monocyte precursors in bone marrow. Notably, neither cholesterol loading (supplemental Table II) nor differentiation of PBMCs into MDM is a prerequisite. Furthermore, this proinflammatory phenotype is independent of cholesterol efflux defects and of mutations in specific candidate genes. We propose that the activated proinflammatory state of monocytes and the heightened macrophage inflammatory response in low HDL-C subjects may contribute to the pathophysiological consequences of low HDL. This constitutes a novel parameter of risk associated with HDL deficiency.
Sources of Funding
This work was supported by a Grant from the Heart and Stroke Foundation of Ontario (T5911) to Y.L.M. and R.M.P., a CIHR Grant(44359) to Y.L.M., and a CIHR-Industry grant to Y.L.M. This work was also supported in part by GlaxoSmithKline Canada.
L.S.-B. and R.S.K. contributed equally to this study.
Original received October 16, 2006; final version accepted February 4, 2007.
Brewer HB Jr, Remaley AT, Neufeld EB, Basso F, Joyce C. Regulation of plasma high-density lipoprotein levels by the ABCA1 transporter and the emerging role of high-density lipoprotein in the treatment of cardiovascular disease. Arterioscler Thromb Vasc Biol. 2004; 24: 1755–1760.
Uittenbogaard A, Shaul PW, Yuhanna IS, Blair A, Smart EJ. High density lipoprotein prevents oxidized low density lipoprotein-induced inhibition of endothelial nitric-oxide synthase localization and activation in caveolae. J Biol Chem. 2000; 275: 11278–11283.
Yuhanna IS, Zhu Y, Cox BE, Hahner LD, Osborne-Lawrence S, Marcel YL, Anderson RGW, Mendelsohn ME, Hobbs HH, Shaul PW. High-density lipoprotein binding to scavenger receptor-BI activates endothelial nitric oxide synthase. Nature Med. 2001; 7: 855–859.
Cockerill GW, Rye KA, Gamble JR, Vadas MA, Barter PJ. High-density lipoproteins inhibit cytokine-induced expression of endothelial cell adhesion molecules. Arterioscler Thromb Vasc Biol. 1995; 15: 1987–1994.
Barter PJ, Nicholls S, Rye KA, Anantharamaiah GM, Navab M, Fogelman AM. Antiinflammatory properties of HDL. Circ Res. 2004; 95: 764–772.
Navab M, Hama SY, Anantharamaiah GM, Hassan K, Hough GP, Watson AD, Reddy ST, Sevanian A, Fonarow GC, Fogelman AM. Normal high density lipoprotein inhibits three steps in the formation of mildly oxidized low density lipoprotein: steps 2 and 3. J Lipid Res. 2000; 41: 1495–1508.
Navab M, Hama SY, Cooke CJ, Anantharamaiah GM, Chaddha M, Jin L, Subbanagounder G, Faull KF, Reddy ST, Miller NE, Fogelman AM. Normal high density lipoprotein inhibits three steps in the formation of mildly oxidized low density lipoprotein: step 1. J Lipid Res. 2000; 41: 1481–1494.
Lambert G, Sakai N, Vaisman BL, Neufeld EB, Marteyn B, Chan CC, Paigen B, Lupia E, Thomas A, Striker LJ, Blanchette-Mackie J, Csako G, Brady JN, Costello R, Striker GE, Remaley AT, Brewer HB Jr, Santamarina-Fojo S. Analysis of glomerulosclerosis and atherosclerosis in lecithin cholesterol acyltransferase-deficient mice. J Biol Chem. 2001; 276: 15090–15098.
Qin SC, Kawano K, Bruce C, Lin M, Bisgaier C, Tall AR, Jiang XC. Phospholipid transfer protein gene knock-out mice have low high density lipoprotein levels, due to hypercatabolism, and accumulate apoA-IV-rich lamellar lipoproteins. J Lipid Res. 2000; 41: 269–276.
Aouizerat BE, Engler MB, Natanzon Y, Kulkarni M, Song J, Eng C, Huuskonen J, Rivera C, Poon A, Bensley M, Sehnert A, Zellner C, Malloy M, Kane J, Pullinger CR. Genetic variation of PLTP modulates lipoprotein profiles in hypoalphalipoproteinemia. J Lipid Res. 2006; 47: 787–793.
Brunham LR, Singaraja RR, Hayden MR. Variations on a Gene: Rare and Common Variants in ABCA1 and Their Impact on HDL Cholesterol Levels and Atherosclerosis. Annu Rev Nutr. 2006.
Cohen JC, Kiss RS, Pertsemlidis A, Marcel YL, McPherson R, Hobbs HH. Multiple rare alleles contribute to low plasma levels of HDL cholesterol. Science. 2004; 305: 869–872.
Kiss RS, Kavaslar N, Okuhira KI, Freeman MW, Walter S, Milne RW, McPherson R, Marcel YL. Genetic etiology of isolated low HDL syndrome: incidence and heterogeneity of efflux defects. Arterioscler Thromb Vasc Biol. 2007; 27: 1138–1144.
Weng L, Dai H, Zhan Y, He Y, Stepaniants SB, Bassett DE. Rosetta error model for gene expression analysis. Bioinformatics. 2006; 22: 1111–1121.
Genest JJ Jr, Martin-Munley SS, McNamara JR, Ordovas JM, Jenner J, Myers RH, Silberman SR, Wilson PW, Salem DN, Schaefer EJ. Familial lipoprotein disorders in patients with premature coronary artery disease. Circulation. 1992; 85: 2025–2033.
Singaraja RR, Brunham LR, Visscher H, Kastelein JJ, Hayden MR. Efflux and atherosclerosis: the clinical and biochemical impact of variations in the ABCA1 gene. Arterioscler Thromb Vasc Biol. 2003; 23: 1322–1332.
Heinecke JW. Lipoprotein oxidation in cardiovascular disease: chief culprit or innocent bystander? J Exp Med. 2006; 203: 813–816.
De SK, McMaster MT, Andrews GK. Endotoxin induction of murine metallothionein gene expression. J Biol Chem. 1990; 265: 15267–15274.
Schroeder JJ, Cousins RJ. Interleukin 6 regulates metallothionein gene expression and zinc metabolism in hepatocyte monolayer cultures. Proc Natl Acad Sci U S A. 1990; 87: 3137–3141.
Fajas L, Schoonjans K, Gelman L, Kim JB, Najib J, Martin G, Fruchart JC, Briggs M, Spiegelman BM, Auwerx J. Regulation of peroxisome proliferator-activated receptor gamma expression by adipocyte differentiation and determination factor 1/sterol regulatory element binding protein 1: implications for adipocyte differentiation and metabolism. Mol Cell Biol. 1999; 19: 5495–5503.
Han JH, Hajjar DP, Zhou XY, Gotto AM Jr, Nicholson AC. Regulation of peroxisome proliferator-activated receptor-gamma-mediated gene expression - A new mechanism of action for high density lipoprotein. J Biol Chem. 2002; 277: 23582–23586.
Tan NS, Shaw NS, Vinckenbosch N, Liu P, Yasmin R, Desvergne B, Wahli W, Noy N. Selective cooperation between fatty acid binding proteins and peroxisome proliferator-activated receptors in regulating transcription. Mol Cell Biol. 2002; 22: 5114–5127.
Rival Y, Stennevin A, Puech L, Rouquette A, Cathala C, Lestienne F, Dupont-Passelaigue E, Patoiseau JF, Wurch T, Junquero D. Human adipocyte fatty acid-binding protein (aP2) gene promoter-driven reporter assay discriminates nonlipogenic peroxisome proliferator-activated receptor gamma ligands. J Pharmacol Exp Ther. 2004; 311: 467–475.
Makowski L, Hotamisligil GS. Fatty acid binding proteins–the evolutionary crossroads of inflammatory and metabolic responses. J Nutr. 2004; 134: 2464S–2468S.
Oliver WR Jr, Shenk JL, Snaith MR, Russell CS, Plunket KD, Bodkin NL, Lewis MC, Winegar DA, Sznaidman ML, Lambert MH, Xu HE, Sternbach DD, Kliewer SA, Hansen BC, Willson TM. A selective peroxisome proliferator-activated receptor delta agonist promotes reverse cholesterol transport. Proc Natl Acad Sci U S A. 2001; 98: 5306–5311.
Vosper H, Patel L, Graham TL, Khoudoli GA, Hill A, Macphee CH, Pinto I, Smith SA, Suckling KE, Wolf CR, Palmer CN. The peroxisome proliferator-activated receptor delta promotes lipid accumulation in human macrophages. J Biol Chem. 2001; 276: 44258–44265.
Lee CH, Chawla A, Urbiztondo N, Liao D, Boisvert WA, Evans RM, Curtiss LK. Transcriptional repression of atherogenic inflammation: modulation by PPARdelta. Science. 2003; 302: 453–457.