- Email: [email protected]
Copper induces the expression of cholesterogenic genes in human macrophages
Atherosclerosis 169 (2003) 71 /76 www.elsevier.com/locate/atherosclerosis
Copper induces the expression of cholesterogenic genes in human macrophages Per-Arne Svensson a,b,*, Mikael C.O. Englund b, Emilia Markstro¨m c, Bertil G. Ohlsson b, Margareta Jerna˚s a, Ha˚kan Billig c, Jarl S. Torgerson d, Olov Wiklund b, Lena M.S. Carlsson a, Bjo¨rn Carlsson a,d a
Department of Internal Medicine, Research Centre for Endocrinology and Metabolism (RCEM), Vita Stra˚ket 12, Sahlgrenska Academy, Go¨teborg University, S-41345 Gothenburg, Sweden b Wallenberg Laboratory for Cardiovascular Research, Bruna Stra˚ket 16, Sahlgrenska Academy, Go¨teborg University, S-41345 Gothenburg, Sweden c Department of Physiology, Medicinaregatan 9, Sahlgrenska Academy, Go¨teborg University, S-40530 Gothenburg, Sweden d Department of Body Composition and Metabolism, Vita Stra˚ket 15, Sahlgrenska Academy, Go¨teborg University, S-41345 Gothenburg, Sweden Received 2 December 2002; received in revised form 4 March 2003; accepted 31 March 2003
Abstract Accumulation of lipids and cholesterol by macrophages and subsequent transformation into foam cells are key features in development of atherosclerosis. Serum copper concentrations have been shown to be associated with cardiovascular disease. However, the mechanism behind the proatherogenic effect of copper is not clear. We used DNA microarrays to define the changes in gene expression profile in response to copper exposure of human macrophages. Expression monitoring by DNA microarray revealed 91 genes that were regulated. Copper increased the expression of seven cholesterogenic genes (3-hydroxy-3-methylglutaryl coenzyme A (HMG CoA) synthase, IPP isomerase, squalene synthase, squalene epoxidase, methyl sterol oxidase, H105e3 mRNA and sterolC5-desaturase) and low-density lipoprotein receptor (LDL-R), and decreased the expression of CD36 and lipid binding proteins. The expression of LDL-R and HMG CoA reductase was also investigated using real time PCR. The expression of both of these genes was increased after copper treatment of macrophages (P B/0.01 and P B/0.01, respectively). We conclude that copper activates cholesterogenic genes in macrophages, which may provide a mechanism for the association between copper and atherosclerosis. The effect of copper on cholesterogenic genes may also have implications for liver steatosis in early stages of Wilson’s disease. # 2003 Elsevier Science Ireland Ltd. All rights reserved. Keywords: DNA microarray; Copper; Macrophages; Lipid accumulation; Atherosclerosis; Wilson’s disease
1. Introduction Cardiovascular disease (CVD) is a major cause of morbidity and mortality and has a multifactoral pathogenesis including both genetic and environmental factors. Epidemiological research has identified several risk factors for CVD and reduction of several of these risk factors reduces the risk of disease. In addition to traditional risk factors such as high levels of low-density lipoprotein (LDL)-cholesterol, several prospective studies indicate that serum copper
* Corresponding author. Tel.: /46-31-342-6459; fax: /46-31-829426. E-mail address: [email protected] (P.-A. Svensson).
concentrations are associated with CVD [1]. In the Kuopio Ischaemic Heart Disease Risk Factor Study high serum copper concentrations were associated with more than a 3.5-fold increased risk of acute myocardial infarction [2]. An association between high serum copper and cardiovascular mortality has also been shown in a study of Dutch men [3]. However, an inverse correlation between serum cholesterol and copper levels has also been reported, suggesting that copper depletion can result in increased risk for CVD [4,5]. The development of atherosclerosis is characterized by an accumulation of lipids and fibrous elements in the large arteries. The lipids that accumulate in the atherosclerotic plaques are mainly derived from LDL, which can filter through the vascular endothelium into the
0021-9150/03/$ - see front matter # 2003 Elsevier Science Ireland Ltd. All rights reserved. doi:10.1016/S0021-9150(03)00145-X
72
P.-A. Svensson et al. / Atherosclerosis 169 (2003) 71 /76
artery wall. LDL entrapped in the artery wall can undergo modifications, such as oxidation [6]. Oxidized LDL (oxLDL) is recognized and taken up by macrophages in the artery wall. The presence of lipid-loaded macrophages (foam cells) in the artery wall is a hallmark of atherosclerosis. There is now considerable evidence that oxidative modification of the LDL particle is of fundamental importance in the development of atherosclerosis [6 /8]. Copper is prooxidative and stimulates oxidative modifications of LDL-cholesterol in vitro and may also be prooxidative in vivo. Thus, the atherogenic effect of copper may be explained by increased oxidation of LDL. However, copper is also an essential micronutrient that affects many biological processes [9] and it is possible that it influences the risk for CVD by other, so far unidentified, mechanisms. The effect of copper on the pathogenesis of atherosclerosis has been assessed in rats by implanting a silicon /copper cuff around rat carotid arteries. This local release of copper ions initiated and mimicked all morphological features of post-angioplasty restenotic and artherosclerotic lesions [10]. This indicates that copper may affect the cells in the artery directly. The aim of this study was to investigate if copper ions can have a direct effect on macrophages, which may contribute to the atherogenic effect of copper. We have used DNA microarray technology to determine gene expression profiles in human macrophages in response to copper.
(Affymetrix, Santa Clara, CA) [13]. Aliquots of the target preparation (15 mg cRNA) were hybridized to a HuGeneFL DNA microarray, washed, stained and scanned (Hewlett Packard, GeneArray scanner G2500A) according to procedures developed by the manufacturer (Affymetrix) [13]. 2.3. DNA microarray expression analysis
2. Materials and methods
Gene expression in macrophages exposed to copper or control medium for 6 h was analyzed on duplicate DNA microarrays. Scanned output files were visually inspected for hybridization artifacts and then analyzed with GENECHIP 3.1 software (Affymetrix). To allow comparison of gene expression, the DNA microarrays were globally scaled to an average intensity of 500. Comparisons were made between the results from the duplicate DNA microarrays used for analysis of the copper exposed macrophages, and the duplicate DNA microarrays used for analysis of the control macrophages, generating a total of four comparisons. Genes with different expression levels in copper exposed and control macrophages were identified by the ‘‘difference call’’ (Diff Call) algorithm (Affymetrix). With the Diff Call, a gene is classified as increased (I), marginally increased (MI), no change (NC), marginally decreased (MD), or decreased (D). Genes having a Diff Call of I, MI or MD, D in three out of the four comparisons were selected for further investigation. An average fold change was also calculated from the four comparisons. Down regulated transcripts are given a negative fold change.
2.1. Subjects and samples
2.4. Real time PCR analysis of gene expression
Human mononuclear blood cells were isolated from buffy coats (Blood Bank, Sahlgrenska University Hospital, Go¨teborg, Sweden) with Ficoll-Paque (Amersham Pharmacia Biotech, Little Chalfont, UK) as previously described [11]. Monocyte-derived macrophages were prepared and differentiated in Macrophage-SFM media (GIBCO BRL, Grand Island, NY) as previously described [11] and denoted ‘‘macrophages’’.
Oligonucleotide primers and probes (primer and probe sequences available on request) were designed with PRIMER EXPRESS 1.5 software (Applied Biosystems, Foster City, CA). Primers and probes were purchased from Applied Biosystems. The probes consisted of oligonucleotides that were labeled at the 5? end with the reporter dye 5-carboxyfluorescein (FAM) and at the 3? end with the quencher N ,N ?,N ?-tetramethyl-6carboxyrhodamine (TAMRA). Reagents (TaqMan† Reverse Transcriptase reagents and TaqMan† Universal PCR Master mix, Applied Biosystems) and conditions were used according to the manufacturer’s protocol. Briefly, RNA was isolated using the RNeasy kit (Qiagen) and DNAse treated using the DNA-free kit (Ambion, Austin, TX). cDNA was synthesized 9/reverse transcriptase (RT) from RNA samples. Each amplification reaction consisted of diluted cDNA (corresponding to 20 ng RNA), 300 nM of each primer, 200 nM TaqMan probe and 1 / TaqMan† Universal PCR reaction mix. Amplification and detection of specific products was performed with the ABI Prism 7700
2.2. Preparation of cRNA and DNA microarray hybridization Macrophages were incubated in Macrophage-SFM media containing CuSO4 (0.4 mM) or control media for 6 h. This low copper concentration was selected to minimize oxidative stress responses in the macrophages [12]. RNA was isolated from the macrophages using the RNeasy kit (Qiagen, Hilden, Germany). Macrophage RNA from the four donors (2 mg each) were pooled and used for target preparation. Target preparation was performed according to the manufacturer’s instructions
P.-A. Svensson et al. / Atherosclerosis 169 (2003) 71 /76
sequence detection system (Applied Biosystems) using default cycle parameters. The critical threshold cycle (Ct) is defined as the cycle when the fluorescence is 10 times baseline sd and is inversely proportional to the logarithm of the initial number of template molecules in the sample. A standard curve was plotted for each primer-probe set with Ct values obtained from amplification of a serial dilution of pooled macrophage cDNA in the range of 0.625 /40 ng original RNA per reaction. Pre-developed assay reagents for human RPLP0 (large ribosomal protein) was obtained from Applied Biosystems and used as reference to normalize the expression levels between the samples. All standards and samples were analyzed in triplicate. Statistical analysis was performed using Wilcoxon signed-ranks test.
3. Results Expression monitoring of approximately 6800 genes in macrophages incubated in the absence or presence of copper (0.4 mM) for 6 h was analyzed using duplicate HuGeneFL DNA microarrays (Affymetrix). The data analysis revealed 91 genes (63 up regulated and 28 down regulated) with an altered expression in response to copper using the criteria described in Section 2 (see http://www.invmed.gu.se/rcem/svenssonCu/). The regulated genes were classified according to their assigned cellular role using information from the public human SDP database (http://www.proteome.com). 12 genes out of the 91 regulated genes had functions related to lipid, fatty acid and sterol metabolism and in this report we have focused on these genes. CuSO4 exposure of macrophages increased the expression of genes encoding several enzymes involved in de novo cholesterol biosynthesis pathway (Fig. 1). In addition, LDL-receptor (LDL-R) mRNA was increased (/3.3-fold) by CuSO4 exposure. Several of these genes are regulated by the transcription factor sterol regulatory element-binding protein (SREBP)-2, which showed a tendency to be increased (/1.3-fold). HMG CoA reductase, which is rate limiting for cholesterol synthesis, showed a tendency to be increased but did not fulfill the criteria for up regulation. The effect of copper on the expression of HMG CoA reductase mRNA was, therefore, investigated using real-time PCR on individual samples from eight donors. CuSO4 exposure of macrophages increased the expression of HMG CoA reductase (P B/0.01; Fig. 2). The increase in LDL-R expression after CuSO4 exposure was verified using real time PCR on individual samples from eight donors (P B/0.01; Fig. 2). CuSO4 exposure of macrophages down regulated the gene expression of the scavenger receptor CD36 (/2.0fold) and adipocyte lipid binding protein (/2.0-fold). The expression of the transcription factor, peroxisome
73
proliferator activated receptor gamma (PPARg) showed a tendency to be down regulated (/1.9-fold). Furthermore, CuSO4 exposure of macrophages also increased the expression of the CYP1B1 gene (/2.5fold) that oxidizes a variety of compounds, including steroids, fatty acids and xenobiotics and decreased the expression of the lysophospholipase homolog gene (/ 2.3-fold) that may regulate lysophospholipid levels.
4. Discussion Several investigators have addressed the possible connection between copper and CVD. However, this connection is still unclear, with reports indicating that copper may have both pro-atherogenic and anti-atherogenic properties. Epidemiological studies have suggested a positive relationship between coronary heart disease and serum copper levels [1 /3]. However, copper deficiency is associated with increased plasma cholesterol levels and has been proposed as a risk factor [4,5]. Several investigators have examined the relationship between copper status and atherogenesis in experimental animal models. Copper supplementation experiments in animals have shown increased resistance to injury following ischaemia-reperfusion [14] as well as decreased development of atherosclerosis [15]. It has been suggested that the relationship between copper status and CVD may be biphasic and that both copper depletion and copper excess may activate both pro-atherogenic and anti-atherogenic mechanisms [9]. We report here that copper exposure increased the expression of genes involved in LDL uptake and de novo cholesterol biosynthesis in human macrophages. This suggests a novel mechanism by which copper could influence cholesterol metabolism and macrophage function and may provide an explanation for the epidemiological association between serum copper levels and atherosclerosis. Serum concentration of copper may not be the best indicator of copper status in the body. Ceruloplasmin is a serum ferroxidase that contains greater than 95% of the copper found in plasma [16]. Only one of the copper ions in ceruloplasmin is considered to be bioavailable at neutral pH [17,18]. The levels of copper in serum may, therefore, not reflect the levels of bioavailable copper. In this study we have treated the macrophages with a copper dose that is in the range of the estimated freecopper levels in serum. Ceruloplasmin is an acute phase reactant protein exhibiting a 2/3-fold increase in response to inflammation. Another acute phase reactant, C-reactive protein, has recently been shown to be a strong predictor of the risk of cardiovascular events [19]. Elevated serum copper levels may, therefore, be a consequence of an acute phase reaction in response to the inflammation associated with atherosclerosis.
74
P.-A. Svensson et al. / Atherosclerosis 169 (2003) 71 /76
Fig. 1. Effect of copper on the expression of genes encoding cholesterologenic enzymes in human macrophages as determined by DNA-microarray. Cells were exposed to CuSO4 (0.4 mM) for 6 h. Regulation is presented as fold change. The nomenclature is adopted from the KEGG database (http://www.genome.ad.jp/kegg).
Fig. 2. Real-time PCR analysis of LDL-R and HMG-CoA reductase mRNA levels in human macrophages exposed to CuSO4 for 6 h. Copper significantly increased the expression of the LDL-R mRNA (P B/0.01) and HMG-CoA reductase mRNA (P B/0.01) in macrophages. Expression was normalized to human RPLP0 mRNA expression and shown as mean9/S.E.M. (n /8). Statistical analysis was performed using Wilcoxon signed-ranks test.
Copper is an essential trace element, and can, therefore, affect many biological processes. Apart from the suggested role in LDL oxidation, the redox reactivity of copper may lead to risks of damage to cell and tissues. It has been shown that incubation of macrophages with
copper in the presence of the lipophilic chelating agent induces apoptosis [20], which may contribute to the formation of the lipid core in atherosclerotic lesions. Copper deficiency in humans is related to altered immune function [21]. Copper is also required for the catalytic activity of several enzymes, such as Cu /Znsuperoxide dismutase, cytochrome c oxidase and lysyl oxidase [9]. It has been shown that lysyl oxidase activity is reduced in Watanabe rabbits with severe atherosclerosis [22]. Lipid homeostasis in cells is regulated by a family of membrane-bound transcription factors called SREBPs. The mammalian genome encodes three SREBP isoforms (SREBP-1a, SREBP-1c, and SREBP-2). SREBP-1a is a potent activator of all SREBP-responsive genes, SREBP-1c preferentially activates genes of fatty acid and triglyceride metabolism, whereas SREBP-2 preferentially activates genes of cholesterol metabolism and the LDL-R gene [23,24]. This suggests that the activation of genes in de novo cholesterol biosynthesis and the LDL-R gene by copper exposure seen in the present study may be mediated by SREBP-2. In line with this, SREBP-2 showed a tendency to be up regulated in the present study (/1.3-fold). However, increased cholesterol content in the macrophage will lead to the
P.-A. Svensson et al. / Atherosclerosis 169 (2003) 71 /76
deactivation of SREBP-2 mediated transcription [23,24]. If copper can increase expression of LDL-R mRNA and cholesterogeneic genes also in cholesterol loaded macrophages remains to be investigated. Interestingly, copper deficiency also affects lipid homeostasis. In experimental animals, copper deficiency leads to an increase in hepatic triglyceride and cholesterol biosynthesis [25 /27]. Recently, copper deficiency was shown to involve an increased nuclear content of SREBP-1 [28] and increased expression of fatty acid synthase (FAS) [29]. However, copper deficiency had no effect on the expression of HMG-CoA synthase. There are indications that the effect of copper deficiency is mediated by an increased accumulation of iron in the liver which is known to occur as a consequence of copper deficiency [29,30]. In the present study, there were no effects of copper on the expression of FAS, a 6fold induction of HMG-CoA synthase and a tendency of induction of SREBP-2. Taken together, it may suggest that copper deficiency (iron accumulation) induces SREBP-1 and copper exposure induces SREBP-2 [23]. Wilson’s disease is an autosomal recessive disorder characterized by dramatic build-up of intracellular hepatic copper with subsequent hepatic and neurological abnormalities. The degenerative changes in the liver are proceeded by fatty depositions, which are not paralleled by increased serum lipid levels [31]. The phenotypic changes in early stages of Wilson’s disease resemble those seen in mice overexpressing SREBP-2 in the liver [32]. These mice have severe accumulation of cholesterol in the liver and relatively normal serum lipid levels probably due to the increased expression of LDLreceptors. This is similar to mice overexpressing SREBP-1a that have normal serum lipid levels despite severe accumulation of fat in the livers [33]. However, hyperlipidemia develops when SREBP1a is overexpressed in mice deficient in LDL-receptors [34]. We conclude that copper specifically induces the expression of cholesterogenic enzymes and the LDL-R, which may lead to lipid accumulation in macrophages. This provides a novel explanation for the epidemiological association between serum copper levels and atherosclerosis. Furthermore, if copper activates a similar set of genes in the liver, it also provides a possible mechanism for development of liver steatosis in early stages of Wilson’s disease.
Acknowledgements This work was supported by grants from Swedish Society for Medical Research, Go¨teborg Medical Association, LUA funds for clinical research, Swegene, the Swedish Heart-Lung Foundation and the Swedish
75
Medical Research Council (5239, 6816, 11285, 11502, 13488 and 13141).
References [1] Ford ES. Serum copper concentration and coronary heart disease among US adults. Am J Epidemiol 2000;151:1182 /8. [2] Salonen JT, Salonen R, Korpela H, Suntioinen S, Tuomilehto J. Serum copper and the risk of acute myocardial infarction: a prospective population study in men in eastern Finland. Am J Epidemiol 1991;134:268 /76. [3] Kok FJ, Van Duijn CM, Hofman A, et al. Serum copper and zinc and the risk of death from cancer and cardiovascular disease. Am J Epidemiol 1988;128:352 /9. [4] Klevay LM, Inman L, Johnson LK, et al. Increased cholesterol in plasma in a young man during experimental copper depletion. Metabolism 1984;33:1112 /8. [5] He JA, Tell GS, Tang YC, Mo PS, He GQ. Relation of serum zinc and copper to lipids and lipoproteins: the Yi People Study. J Am Coll Nutr 1992;11:74 /8. [6] Steinberg D. Low density lipoprotein oxidation and its pathobiological significance. J Biol Chem 1997;272:20963 /6. [7] Carew TE, Schwenke DC, Steinberg D. Antiatherogenic effect of probucol unrelated to its hypocholesterolemic effect: evidence that antioxidants in vivo can selectively inhibit low density lipoprotein degradation in macrophage-rich fatty streaks and slow the progression of atherosclerosis in the Watanabe heritable hyperlipidemic rabbit. Proc Natl Acad Sci USA 1987;84:7725 /9. [8] Prasad K, Kalra J. Oxygen free radicals and hypercholesterolemic atherosclerosis: effect of vitamin E. Am Heart J 1993;125:958 /73. [9] Ferns GA, Lamb DJ, Taylor A. The possible role of copper ions in atherogenesis: the Blue Janus. Atherosclerosis 1997;133:139 / 52. [10] Volker W, Dorszewski A, Unruh V, Robenek H, Breithardt G, Buddecke E. Copper-induced inflammatory reactions of rat carotid arteries mimic restenosis/arteriosclerosis-like neointima formation. Atherosclerosis 1997;130:29 /36. [11] Ohlsson BG, Englund MC, Karlsson AL, et al. Oxidized low density lipoprotein inhibits lipopolysaccharide-induced binding of nuclear factor-kappaB to DNA and the subsequent expression of tumor necrosis factor-alpha and interleukin-1beta in macrophages. J Clin Invest 1996;98:78 /89. [12] Hultberg B, Andersson A, Isaksson A. The cell-damaging effects of low amounts of homocysteine and copper ions in human cell line cultures are caused by oxidative stress. Toxicology 1997;123:33 /40. [13] Benson M, Svensson PA, Carlsson B, et al. DNA microarrays to study gene expression in allergic airways. Clin Exp Allergy 2002;32:301 /8. [14] Pucheu S, Coudray C, Tresallet N, Favier A, de Leiris J. Effect of dietary antioxidant trace element supply on cardiac tolerance to ischemia-reperfusion in the rat. J Mol Cell Cardiol 1995;27:2303 / 14. [15] Lamb DJ, Reeves GL, Taylor A, Ferns GA. Dietary copper supplementation reduces atherosclerosis in the cholesterol-fed rabbit. Atherosclerosis 1999;146:33 /43. [16] Hellman NE, Gitlin JD. Ceruloplasmin metabolism and function. Annu Rev Nutr 2002;22:439 /58. [17] Lamb DJ, Leake DS. Acidic pH enables caeruloplasmin to catalyse the modification of low-density lipoprotein. FEBS Lett 1994;338:122 /6. [18] Ehrenwald E, Chisolm GM, Fox PL. Intact human ceruloplasmin oxidatively modifies low density lipoprotein. J Clin Invest 1994;93:1493 /501.
76
P.-A. Svensson et al. / Atherosclerosis 169 (2003) 71 /76
[19] Ridker PM. Clinical application of C-reactive protein for cardiovascular disease detection and prevention. Circulation 2003;107:363 /9. [20] Pang JH, Chau LY. Copper-induced apoptosis and immediate early gene expression in macrophages. Atherosclerosis 1999;146:45 /52. [21] Licastro F, Morini MC, Chiricolo M, et al. Plasmic levels of trace elements and immune functions in the healthy elderly. J Trace Elem Electrolytes Health Dis 1993;7:234 /6. [22] Dalessandri KM, Eisele P, Wong TS, Parker J, Rucker RB. Elastin degradation in the aorta of Watanabe hereditary hyperlipidemic rabbits. Mech Ageing Dev 1994;74:117 /20. [23] Osborne TF. Sterol regulatory element-binding proteins (SREBPs): key regulators of nutritional homeostasis and insulin action. J Biol Chem 2000;275:32379 /82. [24] Horton JD, Goldstein JL, Brown MS. SREBPs: activators of the complete program of cholesterol and fatty acid synthesis in the liver. J Clin Invest 2002;109:1125 /31. [25] al-Othman AA, Rosenstein F, Lei KY. Copper deficiency increases in vivo hepatic synthesis of fatty acids, triacylglycerols, and phospholipids in rats. Proc Soc Exp Biol Med 1993;204:97 / 103. [26] Kim S, Chao PY, Allen KG. Inhibition of elevated hepatic glutathione abolishes copper deficiency cholesterolemia. FASEB J 1992;6:2467 /71. [27] Shao MJ, Lei KY. Conversion of [2-14C] mevalonate into cholesterol, lanosterol and squalene in copper-deficient rats. J Nutr 1980;110:859 /67.
[28] Tang Z, Gasperkova D, Xu J, Baillie R, Lee JH, Clarke SD. Copper deficiency induces hepatic fatty acid synthase gene transcription in rats by increasing the nuclear content of mature sterol regulatory element binding protein 1. J Nutr 2000;130:2915 /21. [29] Wilson J, Kim S, Allen KG, Baillie R, Clarke SD. Hepatic fatty acid synthase gene transcription is induced by a dietary copper deficiency. Am J Physiol 1997;272:E1124 /9. [30] Fields M, Lewis CG. Hepatic iron overload may contribute to hypertriglyceridemia and hypercholesterolemia in copper-deficient rats. Metabolism 1997;46:377 /81. [31] Scheinberg IH, Sternlieb I. Wilson disease and idiopathic copper toxicosis. Am J Clin Nutr 1996;63:842S /5S. [32] Horton JD, Shimomura I, Brown MS, Hammer RE, Goldstein JL, Shimano H. Activation of cholesterol synthesis in preference to fatty acid synthesis in liver and adipose tissue of transgenic mice overproducing sterol regulatory element-binding protein-2. J Clin Invest 1998;101:2331 /9. [33] Shimano H, Horton JD, Hammer RE, Shimomura I, Brown MS, Goldstein JL. Overproduction of cholesterol and fatty acids causes massive liver enlargement in transgenic mice expressing truncated SREBP-1a. J Clin Invest 1996;98:1575 /84. [34] Horton JD, Shimano H, Hamilton RL, Brown MS, Goldstein JL. Disruption of LDL receptor gene in transgenic SREBP-1a mice unmasks hyperlipidemia resulting from production of lipid-rich VLDL. J Clin Invest 1999;103:1067 /76.
Copper induces the expression of cholesterogenic genes in human macrophages Per-Arne Svensson a,b,*, Mikael C.O. Englund b, Emilia Markstro¨m c, Bertil G. Ohlsson b, Margareta Jerna˚s a, Ha˚kan Billig c, Jarl S. Torgerson d, Olov Wiklund b, Lena M.S. Carlsson a, Bjo¨rn Carlsson a,d a
Department of Internal Medicine, Research Centre for Endocrinology and Metabolism (RCEM), Vita Stra˚ket 12, Sahlgrenska Academy, Go¨teborg University, S-41345 Gothenburg, Sweden b Wallenberg Laboratory for Cardiovascular Research, Bruna Stra˚ket 16, Sahlgrenska Academy, Go¨teborg University, S-41345 Gothenburg, Sweden c Department of Physiology, Medicinaregatan 9, Sahlgrenska Academy, Go¨teborg University, S-40530 Gothenburg, Sweden d Department of Body Composition and Metabolism, Vita Stra˚ket 15, Sahlgrenska Academy, Go¨teborg University, S-41345 Gothenburg, Sweden Received 2 December 2002; received in revised form 4 March 2003; accepted 31 March 2003
Abstract Accumulation of lipids and cholesterol by macrophages and subsequent transformation into foam cells are key features in development of atherosclerosis. Serum copper concentrations have been shown to be associated with cardiovascular disease. However, the mechanism behind the proatherogenic effect of copper is not clear. We used DNA microarrays to define the changes in gene expression profile in response to copper exposure of human macrophages. Expression monitoring by DNA microarray revealed 91 genes that were regulated. Copper increased the expression of seven cholesterogenic genes (3-hydroxy-3-methylglutaryl coenzyme A (HMG CoA) synthase, IPP isomerase, squalene synthase, squalene epoxidase, methyl sterol oxidase, H105e3 mRNA and sterolC5-desaturase) and low-density lipoprotein receptor (LDL-R), and decreased the expression of CD36 and lipid binding proteins. The expression of LDL-R and HMG CoA reductase was also investigated using real time PCR. The expression of both of these genes was increased after copper treatment of macrophages (P B/0.01 and P B/0.01, respectively). We conclude that copper activates cholesterogenic genes in macrophages, which may provide a mechanism for the association between copper and atherosclerosis. The effect of copper on cholesterogenic genes may also have implications for liver steatosis in early stages of Wilson’s disease. # 2003 Elsevier Science Ireland Ltd. All rights reserved. Keywords: DNA microarray; Copper; Macrophages; Lipid accumulation; Atherosclerosis; Wilson’s disease
1. Introduction Cardiovascular disease (CVD) is a major cause of morbidity and mortality and has a multifactoral pathogenesis including both genetic and environmental factors. Epidemiological research has identified several risk factors for CVD and reduction of several of these risk factors reduces the risk of disease. In addition to traditional risk factors such as high levels of low-density lipoprotein (LDL)-cholesterol, several prospective studies indicate that serum copper
* Corresponding author. Tel.: /46-31-342-6459; fax: /46-31-829426. E-mail address: [email protected] (P.-A. Svensson).
concentrations are associated with CVD [1]. In the Kuopio Ischaemic Heart Disease Risk Factor Study high serum copper concentrations were associated with more than a 3.5-fold increased risk of acute myocardial infarction [2]. An association between high serum copper and cardiovascular mortality has also been shown in a study of Dutch men [3]. However, an inverse correlation between serum cholesterol and copper levels has also been reported, suggesting that copper depletion can result in increased risk for CVD [4,5]. The development of atherosclerosis is characterized by an accumulation of lipids and fibrous elements in the large arteries. The lipids that accumulate in the atherosclerotic plaques are mainly derived from LDL, which can filter through the vascular endothelium into the
0021-9150/03/$ - see front matter # 2003 Elsevier Science Ireland Ltd. All rights reserved. doi:10.1016/S0021-9150(03)00145-X
72
P.-A. Svensson et al. / Atherosclerosis 169 (2003) 71 /76
artery wall. LDL entrapped in the artery wall can undergo modifications, such as oxidation [6]. Oxidized LDL (oxLDL) is recognized and taken up by macrophages in the artery wall. The presence of lipid-loaded macrophages (foam cells) in the artery wall is a hallmark of atherosclerosis. There is now considerable evidence that oxidative modification of the LDL particle is of fundamental importance in the development of atherosclerosis [6 /8]. Copper is prooxidative and stimulates oxidative modifications of LDL-cholesterol in vitro and may also be prooxidative in vivo. Thus, the atherogenic effect of copper may be explained by increased oxidation of LDL. However, copper is also an essential micronutrient that affects many biological processes [9] and it is possible that it influences the risk for CVD by other, so far unidentified, mechanisms. The effect of copper on the pathogenesis of atherosclerosis has been assessed in rats by implanting a silicon /copper cuff around rat carotid arteries. This local release of copper ions initiated and mimicked all morphological features of post-angioplasty restenotic and artherosclerotic lesions [10]. This indicates that copper may affect the cells in the artery directly. The aim of this study was to investigate if copper ions can have a direct effect on macrophages, which may contribute to the atherogenic effect of copper. We have used DNA microarray technology to determine gene expression profiles in human macrophages in response to copper.
(Affymetrix, Santa Clara, CA) [13]. Aliquots of the target preparation (15 mg cRNA) were hybridized to a HuGeneFL DNA microarray, washed, stained and scanned (Hewlett Packard, GeneArray scanner G2500A) according to procedures developed by the manufacturer (Affymetrix) [13]. 2.3. DNA microarray expression analysis
2. Materials and methods
Gene expression in macrophages exposed to copper or control medium for 6 h was analyzed on duplicate DNA microarrays. Scanned output files were visually inspected for hybridization artifacts and then analyzed with GENECHIP 3.1 software (Affymetrix). To allow comparison of gene expression, the DNA microarrays were globally scaled to an average intensity of 500. Comparisons were made between the results from the duplicate DNA microarrays used for analysis of the copper exposed macrophages, and the duplicate DNA microarrays used for analysis of the control macrophages, generating a total of four comparisons. Genes with different expression levels in copper exposed and control macrophages were identified by the ‘‘difference call’’ (Diff Call) algorithm (Affymetrix). With the Diff Call, a gene is classified as increased (I), marginally increased (MI), no change (NC), marginally decreased (MD), or decreased (D). Genes having a Diff Call of I, MI or MD, D in three out of the four comparisons were selected for further investigation. An average fold change was also calculated from the four comparisons. Down regulated transcripts are given a negative fold change.
2.1. Subjects and samples
2.4. Real time PCR analysis of gene expression
Human mononuclear blood cells were isolated from buffy coats (Blood Bank, Sahlgrenska University Hospital, Go¨teborg, Sweden) with Ficoll-Paque (Amersham Pharmacia Biotech, Little Chalfont, UK) as previously described [11]. Monocyte-derived macrophages were prepared and differentiated in Macrophage-SFM media (GIBCO BRL, Grand Island, NY) as previously described [11] and denoted ‘‘macrophages’’.
Oligonucleotide primers and probes (primer and probe sequences available on request) were designed with PRIMER EXPRESS 1.5 software (Applied Biosystems, Foster City, CA). Primers and probes were purchased from Applied Biosystems. The probes consisted of oligonucleotides that were labeled at the 5? end with the reporter dye 5-carboxyfluorescein (FAM) and at the 3? end with the quencher N ,N ?,N ?-tetramethyl-6carboxyrhodamine (TAMRA). Reagents (TaqMan† Reverse Transcriptase reagents and TaqMan† Universal PCR Master mix, Applied Biosystems) and conditions were used according to the manufacturer’s protocol. Briefly, RNA was isolated using the RNeasy kit (Qiagen) and DNAse treated using the DNA-free kit (Ambion, Austin, TX). cDNA was synthesized 9/reverse transcriptase (RT) from RNA samples. Each amplification reaction consisted of diluted cDNA (corresponding to 20 ng RNA), 300 nM of each primer, 200 nM TaqMan probe and 1 / TaqMan† Universal PCR reaction mix. Amplification and detection of specific products was performed with the ABI Prism 7700
2.2. Preparation of cRNA and DNA microarray hybridization Macrophages were incubated in Macrophage-SFM media containing CuSO4 (0.4 mM) or control media for 6 h. This low copper concentration was selected to minimize oxidative stress responses in the macrophages [12]. RNA was isolated from the macrophages using the RNeasy kit (Qiagen, Hilden, Germany). Macrophage RNA from the four donors (2 mg each) were pooled and used for target preparation. Target preparation was performed according to the manufacturer’s instructions
P.-A. Svensson et al. / Atherosclerosis 169 (2003) 71 /76
sequence detection system (Applied Biosystems) using default cycle parameters. The critical threshold cycle (Ct) is defined as the cycle when the fluorescence is 10 times baseline sd and is inversely proportional to the logarithm of the initial number of template molecules in the sample. A standard curve was plotted for each primer-probe set with Ct values obtained from amplification of a serial dilution of pooled macrophage cDNA in the range of 0.625 /40 ng original RNA per reaction. Pre-developed assay reagents for human RPLP0 (large ribosomal protein) was obtained from Applied Biosystems and used as reference to normalize the expression levels between the samples. All standards and samples were analyzed in triplicate. Statistical analysis was performed using Wilcoxon signed-ranks test.
3. Results Expression monitoring of approximately 6800 genes in macrophages incubated in the absence or presence of copper (0.4 mM) for 6 h was analyzed using duplicate HuGeneFL DNA microarrays (Affymetrix). The data analysis revealed 91 genes (63 up regulated and 28 down regulated) with an altered expression in response to copper using the criteria described in Section 2 (see http://www.invmed.gu.se/rcem/svenssonCu/). The regulated genes were classified according to their assigned cellular role using information from the public human SDP database (http://www.proteome.com). 12 genes out of the 91 regulated genes had functions related to lipid, fatty acid and sterol metabolism and in this report we have focused on these genes. CuSO4 exposure of macrophages increased the expression of genes encoding several enzymes involved in de novo cholesterol biosynthesis pathway (Fig. 1). In addition, LDL-receptor (LDL-R) mRNA was increased (/3.3-fold) by CuSO4 exposure. Several of these genes are regulated by the transcription factor sterol regulatory element-binding protein (SREBP)-2, which showed a tendency to be increased (/1.3-fold). HMG CoA reductase, which is rate limiting for cholesterol synthesis, showed a tendency to be increased but did not fulfill the criteria for up regulation. The effect of copper on the expression of HMG CoA reductase mRNA was, therefore, investigated using real-time PCR on individual samples from eight donors. CuSO4 exposure of macrophages increased the expression of HMG CoA reductase (P B/0.01; Fig. 2). The increase in LDL-R expression after CuSO4 exposure was verified using real time PCR on individual samples from eight donors (P B/0.01; Fig. 2). CuSO4 exposure of macrophages down regulated the gene expression of the scavenger receptor CD36 (/2.0fold) and adipocyte lipid binding protein (/2.0-fold). The expression of the transcription factor, peroxisome
73
proliferator activated receptor gamma (PPARg) showed a tendency to be down regulated (/1.9-fold). Furthermore, CuSO4 exposure of macrophages also increased the expression of the CYP1B1 gene (/2.5fold) that oxidizes a variety of compounds, including steroids, fatty acids and xenobiotics and decreased the expression of the lysophospholipase homolog gene (/ 2.3-fold) that may regulate lysophospholipid levels.
4. Discussion Several investigators have addressed the possible connection between copper and CVD. However, this connection is still unclear, with reports indicating that copper may have both pro-atherogenic and anti-atherogenic properties. Epidemiological studies have suggested a positive relationship between coronary heart disease and serum copper levels [1 /3]. However, copper deficiency is associated with increased plasma cholesterol levels and has been proposed as a risk factor [4,5]. Several investigators have examined the relationship between copper status and atherogenesis in experimental animal models. Copper supplementation experiments in animals have shown increased resistance to injury following ischaemia-reperfusion [14] as well as decreased development of atherosclerosis [15]. It has been suggested that the relationship between copper status and CVD may be biphasic and that both copper depletion and copper excess may activate both pro-atherogenic and anti-atherogenic mechanisms [9]. We report here that copper exposure increased the expression of genes involved in LDL uptake and de novo cholesterol biosynthesis in human macrophages. This suggests a novel mechanism by which copper could influence cholesterol metabolism and macrophage function and may provide an explanation for the epidemiological association between serum copper levels and atherosclerosis. Serum concentration of copper may not be the best indicator of copper status in the body. Ceruloplasmin is a serum ferroxidase that contains greater than 95% of the copper found in plasma [16]. Only one of the copper ions in ceruloplasmin is considered to be bioavailable at neutral pH [17,18]. The levels of copper in serum may, therefore, not reflect the levels of bioavailable copper. In this study we have treated the macrophages with a copper dose that is in the range of the estimated freecopper levels in serum. Ceruloplasmin is an acute phase reactant protein exhibiting a 2/3-fold increase in response to inflammation. Another acute phase reactant, C-reactive protein, has recently been shown to be a strong predictor of the risk of cardiovascular events [19]. Elevated serum copper levels may, therefore, be a consequence of an acute phase reaction in response to the inflammation associated with atherosclerosis.
74
P.-A. Svensson et al. / Atherosclerosis 169 (2003) 71 /76
Fig. 1. Effect of copper on the expression of genes encoding cholesterologenic enzymes in human macrophages as determined by DNA-microarray. Cells were exposed to CuSO4 (0.4 mM) for 6 h. Regulation is presented as fold change. The nomenclature is adopted from the KEGG database (http://www.genome.ad.jp/kegg).
Fig. 2. Real-time PCR analysis of LDL-R and HMG-CoA reductase mRNA levels in human macrophages exposed to CuSO4 for 6 h. Copper significantly increased the expression of the LDL-R mRNA (P B/0.01) and HMG-CoA reductase mRNA (P B/0.01) in macrophages. Expression was normalized to human RPLP0 mRNA expression and shown as mean9/S.E.M. (n /8). Statistical analysis was performed using Wilcoxon signed-ranks test.
Copper is an essential trace element, and can, therefore, affect many biological processes. Apart from the suggested role in LDL oxidation, the redox reactivity of copper may lead to risks of damage to cell and tissues. It has been shown that incubation of macrophages with
copper in the presence of the lipophilic chelating agent induces apoptosis [20], which may contribute to the formation of the lipid core in atherosclerotic lesions. Copper deficiency in humans is related to altered immune function [21]. Copper is also required for the catalytic activity of several enzymes, such as Cu /Znsuperoxide dismutase, cytochrome c oxidase and lysyl oxidase [9]. It has been shown that lysyl oxidase activity is reduced in Watanabe rabbits with severe atherosclerosis [22]. Lipid homeostasis in cells is regulated by a family of membrane-bound transcription factors called SREBPs. The mammalian genome encodes three SREBP isoforms (SREBP-1a, SREBP-1c, and SREBP-2). SREBP-1a is a potent activator of all SREBP-responsive genes, SREBP-1c preferentially activates genes of fatty acid and triglyceride metabolism, whereas SREBP-2 preferentially activates genes of cholesterol metabolism and the LDL-R gene [23,24]. This suggests that the activation of genes in de novo cholesterol biosynthesis and the LDL-R gene by copper exposure seen in the present study may be mediated by SREBP-2. In line with this, SREBP-2 showed a tendency to be up regulated in the present study (/1.3-fold). However, increased cholesterol content in the macrophage will lead to the
P.-A. Svensson et al. / Atherosclerosis 169 (2003) 71 /76
deactivation of SREBP-2 mediated transcription [23,24]. If copper can increase expression of LDL-R mRNA and cholesterogeneic genes also in cholesterol loaded macrophages remains to be investigated. Interestingly, copper deficiency also affects lipid homeostasis. In experimental animals, copper deficiency leads to an increase in hepatic triglyceride and cholesterol biosynthesis [25 /27]. Recently, copper deficiency was shown to involve an increased nuclear content of SREBP-1 [28] and increased expression of fatty acid synthase (FAS) [29]. However, copper deficiency had no effect on the expression of HMG-CoA synthase. There are indications that the effect of copper deficiency is mediated by an increased accumulation of iron in the liver which is known to occur as a consequence of copper deficiency [29,30]. In the present study, there were no effects of copper on the expression of FAS, a 6fold induction of HMG-CoA synthase and a tendency of induction of SREBP-2. Taken together, it may suggest that copper deficiency (iron accumulation) induces SREBP-1 and copper exposure induces SREBP-2 [23]. Wilson’s disease is an autosomal recessive disorder characterized by dramatic build-up of intracellular hepatic copper with subsequent hepatic and neurological abnormalities. The degenerative changes in the liver are proceeded by fatty depositions, which are not paralleled by increased serum lipid levels [31]. The phenotypic changes in early stages of Wilson’s disease resemble those seen in mice overexpressing SREBP-2 in the liver [32]. These mice have severe accumulation of cholesterol in the liver and relatively normal serum lipid levels probably due to the increased expression of LDLreceptors. This is similar to mice overexpressing SREBP-1a that have normal serum lipid levels despite severe accumulation of fat in the livers [33]. However, hyperlipidemia develops when SREBP1a is overexpressed in mice deficient in LDL-receptors [34]. We conclude that copper specifically induces the expression of cholesterogenic enzymes and the LDL-R, which may lead to lipid accumulation in macrophages. This provides a novel explanation for the epidemiological association between serum copper levels and atherosclerosis. Furthermore, if copper activates a similar set of genes in the liver, it also provides a possible mechanism for development of liver steatosis in early stages of Wilson’s disease.
Acknowledgements This work was supported by grants from Swedish Society for Medical Research, Go¨teborg Medical Association, LUA funds for clinical research, Swegene, the Swedish Heart-Lung Foundation and the Swedish
75
Medical Research Council (5239, 6816, 11285, 11502, 13488 and 13141).
References [1] Ford ES. Serum copper concentration and coronary heart disease among US adults. Am J Epidemiol 2000;151:1182 /8. [2] Salonen JT, Salonen R, Korpela H, Suntioinen S, Tuomilehto J. Serum copper and the risk of acute myocardial infarction: a prospective population study in men in eastern Finland. Am J Epidemiol 1991;134:268 /76. [3] Kok FJ, Van Duijn CM, Hofman A, et al. Serum copper and zinc and the risk of death from cancer and cardiovascular disease. Am J Epidemiol 1988;128:352 /9. [4] Klevay LM, Inman L, Johnson LK, et al. Increased cholesterol in plasma in a young man during experimental copper depletion. Metabolism 1984;33:1112 /8. [5] He JA, Tell GS, Tang YC, Mo PS, He GQ. Relation of serum zinc and copper to lipids and lipoproteins: the Yi People Study. J Am Coll Nutr 1992;11:74 /8. [6] Steinberg D. Low density lipoprotein oxidation and its pathobiological significance. J Biol Chem 1997;272:20963 /6. [7] Carew TE, Schwenke DC, Steinberg D. Antiatherogenic effect of probucol unrelated to its hypocholesterolemic effect: evidence that antioxidants in vivo can selectively inhibit low density lipoprotein degradation in macrophage-rich fatty streaks and slow the progression of atherosclerosis in the Watanabe heritable hyperlipidemic rabbit. Proc Natl Acad Sci USA 1987;84:7725 /9. [8] Prasad K, Kalra J. Oxygen free radicals and hypercholesterolemic atherosclerosis: effect of vitamin E. Am Heart J 1993;125:958 /73. [9] Ferns GA, Lamb DJ, Taylor A. The possible role of copper ions in atherogenesis: the Blue Janus. Atherosclerosis 1997;133:139 / 52. [10] Volker W, Dorszewski A, Unruh V, Robenek H, Breithardt G, Buddecke E. Copper-induced inflammatory reactions of rat carotid arteries mimic restenosis/arteriosclerosis-like neointima formation. Atherosclerosis 1997;130:29 /36. [11] Ohlsson BG, Englund MC, Karlsson AL, et al. Oxidized low density lipoprotein inhibits lipopolysaccharide-induced binding of nuclear factor-kappaB to DNA and the subsequent expression of tumor necrosis factor-alpha and interleukin-1beta in macrophages. J Clin Invest 1996;98:78 /89. [12] Hultberg B, Andersson A, Isaksson A. The cell-damaging effects of low amounts of homocysteine and copper ions in human cell line cultures are caused by oxidative stress. Toxicology 1997;123:33 /40. [13] Benson M, Svensson PA, Carlsson B, et al. DNA microarrays to study gene expression in allergic airways. Clin Exp Allergy 2002;32:301 /8. [14] Pucheu S, Coudray C, Tresallet N, Favier A, de Leiris J. Effect of dietary antioxidant trace element supply on cardiac tolerance to ischemia-reperfusion in the rat. J Mol Cell Cardiol 1995;27:2303 / 14. [15] Lamb DJ, Reeves GL, Taylor A, Ferns GA. Dietary copper supplementation reduces atherosclerosis in the cholesterol-fed rabbit. Atherosclerosis 1999;146:33 /43. [16] Hellman NE, Gitlin JD. Ceruloplasmin metabolism and function. Annu Rev Nutr 2002;22:439 /58. [17] Lamb DJ, Leake DS. Acidic pH enables caeruloplasmin to catalyse the modification of low-density lipoprotein. FEBS Lett 1994;338:122 /6. [18] Ehrenwald E, Chisolm GM, Fox PL. Intact human ceruloplasmin oxidatively modifies low density lipoprotein. J Clin Invest 1994;93:1493 /501.
76
P.-A. Svensson et al. / Atherosclerosis 169 (2003) 71 /76
[19] Ridker PM. Clinical application of C-reactive protein for cardiovascular disease detection and prevention. Circulation 2003;107:363 /9. [20] Pang JH, Chau LY. Copper-induced apoptosis and immediate early gene expression in macrophages. Atherosclerosis 1999;146:45 /52. [21] Licastro F, Morini MC, Chiricolo M, et al. Plasmic levels of trace elements and immune functions in the healthy elderly. J Trace Elem Electrolytes Health Dis 1993;7:234 /6. [22] Dalessandri KM, Eisele P, Wong TS, Parker J, Rucker RB. Elastin degradation in the aorta of Watanabe hereditary hyperlipidemic rabbits. Mech Ageing Dev 1994;74:117 /20. [23] Osborne TF. Sterol regulatory element-binding proteins (SREBPs): key regulators of nutritional homeostasis and insulin action. J Biol Chem 2000;275:32379 /82. [24] Horton JD, Goldstein JL, Brown MS. SREBPs: activators of the complete program of cholesterol and fatty acid synthesis in the liver. J Clin Invest 2002;109:1125 /31. [25] al-Othman AA, Rosenstein F, Lei KY. Copper deficiency increases in vivo hepatic synthesis of fatty acids, triacylglycerols, and phospholipids in rats. Proc Soc Exp Biol Med 1993;204:97 / 103. [26] Kim S, Chao PY, Allen KG. Inhibition of elevated hepatic glutathione abolishes copper deficiency cholesterolemia. FASEB J 1992;6:2467 /71. [27] Shao MJ, Lei KY. Conversion of [2-14C] mevalonate into cholesterol, lanosterol and squalene in copper-deficient rats. J Nutr 1980;110:859 /67.
[28] Tang Z, Gasperkova D, Xu J, Baillie R, Lee JH, Clarke SD. Copper deficiency induces hepatic fatty acid synthase gene transcription in rats by increasing the nuclear content of mature sterol regulatory element binding protein 1. J Nutr 2000;130:2915 /21. [29] Wilson J, Kim S, Allen KG, Baillie R, Clarke SD. Hepatic fatty acid synthase gene transcription is induced by a dietary copper deficiency. Am J Physiol 1997;272:E1124 /9. [30] Fields M, Lewis CG. Hepatic iron overload may contribute to hypertriglyceridemia and hypercholesterolemia in copper-deficient rats. Metabolism 1997;46:377 /81. [31] Scheinberg IH, Sternlieb I. Wilson disease and idiopathic copper toxicosis. Am J Clin Nutr 1996;63:842S /5S. [32] Horton JD, Shimomura I, Brown MS, Hammer RE, Goldstein JL, Shimano H. Activation of cholesterol synthesis in preference to fatty acid synthesis in liver and adipose tissue of transgenic mice overproducing sterol regulatory element-binding protein-2. J Clin Invest 1998;101:2331 /9. [33] Shimano H, Horton JD, Hammer RE, Shimomura I, Brown MS, Goldstein JL. Overproduction of cholesterol and fatty acids causes massive liver enlargement in transgenic mice expressing truncated SREBP-1a. J Clin Invest 1996;98:1575 /84. [34] Horton JD, Shimano H, Hamilton RL, Brown MS, Goldstein JL. Disruption of LDL receptor gene in transgenic SREBP-1a mice unmasks hyperlipidemia resulting from production of lipid-rich VLDL. J Clin Invest 1999;103:1067 /76.