Portal de Periódicos da CAPES

Hyperglycemia-induced Production of Acute Phase Reactants in Adipose Tissue

2001; Elsevier BV; Volume: 276; Issue: 45 Linguagem: Inglês

10.1074/jbc.m107101200

1083-351X

Ying Lin, Michael W. Rajala, Joel P. Berger, David E. Moller, Nir Barzilai, Philipp E. Scherer,

Metabolism, Diabetes, and Cancer

Chronic elevation of systemic levels of acute phase reactants and inflammatory cytokines found in patients with diabetes and the often-associated metabolic syndrome X (hypertriglyceridemia, low serum high density lipoprotein cholesterol, hypertension, and accelerated atherosclerosis) may be responsible for the increased incidence of cardiovascular problems in this population. Here we examine the contribution of adipose tissue to the systemic elevation of acute phase reactants associated with chronic hyperglycemia. We demonstrate that adipose tissue expresses a number of acute phase reactants at high levels, including serum amyloid A3 (SAA3), αl-acid glycoprotein, the lipocalin 24p3 as well as plasminogen activator inhibitor-1 (PAI-1). Additionally, we show SAA3 is expressed at low levels under normal conditions but in the diabetic state is dramatically up-regulated in adipose tissue while down-regulated in liver. Furthermore, pro-inflammatory stimuli and high glucose can lead to the induction of SAA3 in adipose tissue in vivo as well as in the 3T3-L1 adipocyte cell line. Adipose tissue may therefore play a major role in the pathogenic sequelae of Type II diabetes, in particular the cardiovascular problems associated with prolonged hyperglycemia. Chronic elevation of systemic levels of acute phase reactants and inflammatory cytokines found in patients with diabetes and the often-associated metabolic syndrome X (hypertriglyceridemia, low serum high density lipoprotein cholesterol, hypertension, and accelerated atherosclerosis) may be responsible for the increased incidence of cardiovascular problems in this population. Here we examine the contribution of adipose tissue to the systemic elevation of acute phase reactants associated with chronic hyperglycemia. We demonstrate that adipose tissue expresses a number of acute phase reactants at high levels, including serum amyloid A3 (SAA3), αl-acid glycoprotein, the lipocalin 24p3 as well as plasminogen activator inhibitor-1 (PAI-1). Additionally, we show SAA3 is expressed at low levels under normal conditions but in the diabetic state is dramatically up-regulated in adipose tissue while down-regulated in liver. Furthermore, pro-inflammatory stimuli and high glucose can lead to the induction of SAA3 in adipose tissue in vivo as well as in the 3T3-L1 adipocyte cell line. Adipose tissue may therefore play a major role in the pathogenic sequelae of Type II diabetes, in particular the cardiovascular problems associated with prolonged hyperglycemia. serum amyloid A3 lipopolysaccharide tumor necrosis factor α interleukin-6 plasminogen activator inhibitor-1 CCAAT/enhancer-binding protein reverse transcription polymerase chain reaction α1-acid glycoprotein adipose glucose uptake high density lipoprotein The central regulatory role of the adipocyte in whole body energy homeostasis is well established. However, pre-adipocytes and adipocytes may also play an important physiological role in the regulation of both the innate and adaptive immune response. We have recently described the response of the adipocyte to various inflammatory stimuli including TNFα and IL-6, focusing primarily on the response to bacterial lipopolysaccharides (LPS), which are mediated through the newly identified Toll-like receptor family (TLRs) (1Lin Y. Lee H. Berg A.H. Lisanti M.P. Shapiro L. Scherer P.E. J. Biol. Chem. 2000; 275: 24255-24263Abstract Full Text Full Text PDF PubMed Scopus (276) Google Scholar). The expression of molecules involved in the innate immune response in adipose tissue, such as complement factors D (adipsin), B, and C3 (2White R.T. Damm D. Hancock N. Rosen B.S. Lowell B.B. Usher P. Flier J.S. Spiegelman B.M. J. Biol. Chem. 1992; 267: 9210-9213Abstract Full Text PDF PubMed Google Scholar, 3Choy L.N. Rosen B.S. Spiegelman B.M. J. Biol. Chem. 1992; 267: 12736-12741Abstract Full Text PDF PubMed Google Scholar, 4Cook K.S. Min H.Y. Johnson D. Chaplinsky R.J. Flier J.S. Hunt C.R. Spiegelman B.M. Science. 1987; 237: 402-405Crossref PubMed Scopus (286) Google Scholar) as well as acute phase reactant proteins (5Scherer P.E. Bickel P.E. Kotler M. Lodish H.F. Nat. Biotechnol. 1998; 16: 581-586Crossref PubMed Scopus (107) Google Scholar), has been demonstrated by our group as well as others. In addition, adipocytes actively secrete and respond to inflammatory cytokines such as TNFα, IL-1, and IL-6 (6Hotamisligil G.S. Shargill N.S. Spiegelman B.M. Science. 1993; 259: 87-91Crossref PubMed Scopus (6028) Google Scholar, 7Mattacks C.A. Pond C.M. Cytokine. 1999; 11: 334-346Crossref PubMed Scopus (72) Google Scholar, 8Mohamed-Ali V. Pinkney J.H. Coppack S.W. Int. J. Obes. Relat. Metab. Disord. 1998; 22: 1145-1158Crossref PubMed Scopus (794) Google Scholar). The latter two are primary cytokine mediators of the acute phase response. The transcription factors responsible for the downstream events of IL-1 and IL-6 include C/EBPβ as well as C/EBPδ, which are directly involved in the acute phase response of the liver (9Burgess-Beusse B.L. Darlington G.J. Mol. Cell. Biol. 1998; 18: 7269-7277Crossref PubMed Scopus (64) Google Scholar, 10Juan T.S. Wilson D.R. Wilde M.D. Darlington G.J. Proc. Natl. Acad. Sci. U. S. A. 1993; 90: 2584-2588Crossref PubMed Scopus (123) Google Scholar), typically considered the main contributor of circulating acute phase reactants. Because of the fact that adipocytes express significant levels of IL-1 and IL-6 receptors as well as high levels of C/EBPα, C/EBPβ, and C/EBPδ, it follows that adipocytes would be capable of producing high levels of acute phase reactant proteins in response to the proper stimuli (5Scherer P.E. Bickel P.E. Kotler M. Lodish H.F. Nat. Biotechnol. 1998; 16: 581-586Crossref PubMed Scopus (107) Google Scholar). However, the relative contribution of adipose tissue to systemic acute phase reactant levels, particularly in the diabetic state, has not been studied to date.Here we investigate the link between the inflammatory/acute phase reactant response and hyperglycemia/hyperinsulinemia at the level of the adipocyte. We show that hyperglycemia, but not hyperinsulinemia, leads to the induction and secretion of SAA3,1 an acute phase reactant in the adipocyte. Adipose tissue, and not the liver, may therefore be responsible for the increased SAA levels in the diabetic state reported in a number of clinical studies.DISCUSSIONThere is an increasing body of evidence that correlates the diabetic phenotype with chronically elevated systemic levels of acute phase reactants and inflammatory cytokines. It has been suggested that these elevated levels may contribute or be partially causative to some of the pathologies of the disease, in particular the increased incidence of cardiovascular problems. Pickup et al. (20Pickup J.C. Mattock M.B. Chusney G.D. Burt D. Diabetologia. 1997; 40: 1286-1292Crossref PubMed Scopus (1040) Google Scholar) show a correlation between the increased levels of several acute phase reactants and IL-6 and the metabolic syndrome X. These authors report elevated levels of circulating α1-acid glycoprotein and serum amyloid A. In addition, other groups show that fibrinogen, another acute phase reactant protein associated with coronary heart disease, is increased in Type II diabetes (21Kannel W.B. D'Agostino R.B. Wilson P.W. Belanger A.J. Gagnon D.R. Am. Heart J. 1990; 120: 672-676Crossref PubMed Scopus (318) Google Scholar, 22Ganda O.P. Arkin C.F. Diabetes Care. 1992; 15: 1245-1250Crossref PubMed Scopus (153) Google Scholar). Furthermore, elevated levels of complement factor C3 have also been reported in diabetic patients (23Figueredo A. Ibarra J.L. Bagazgoitia J. Rodriguez A. Molino A.M. Fernandez-Cruz A. Patino R. Diabetes Care. 1993; 16: 445-449Crossref PubMed Scopus (25) Google Scholar) as well as elevated levels of PAI-1 (24Yudkin J.S. J. Intern. Med. 1995; 238: 21-30Crossref PubMed Scopus (43) Google Scholar). Yokoyama et al.(25Yokoyama H. Jensen J.S. Jensen T. Deckert T. J. Intern. Med. 1995; 237: 519-523Crossref PubMed Scopus (49) Google Scholar) find that acute phase reactants are elevated in Type I diabetes, especially in albuminuric patients. Serum leptin concentrations are responsive to an acute phase or stress response, independent of body mass index (26Pickup J.C. Chusney G.D. Mattock M.B. Clin. Endocrinol. (Oxf). 2000; 52: 107-112Crossref PubMed Scopus (38) Google Scholar).It is theoretically possible that the increase in acute phase reactant proteins, some of which have shown to directly affect lipid metabolism, may contribute to the dyslipidemia associated with diabetes. For example, serum amyloid A displaces apolipoprotein A1 from HDL3, thereby increasing HDL binding to macrophages with the net effect of redirecting HDL cholesterol from the liver to the macrophage for tissue repair (27Cabana V.G. Siegel J.N. Sabesin S.M. J. Lipid Res. 1989; 30: 39-49Abstract Full Text PDF PubMed Google Scholar, 28Steel D.M. Whitehead A.S. Immunol. Today. 1994; 15: 81-88Abstract Full Text PDF PubMed Scopus (867) Google Scholar). A series of prospective cohort studies demonstrate that inflammatory parameters (such as serum amyloid A) and cytokines (such as interleukin-6) are all elevated at base line among patients at risk for future coronary occlusion. Furthermore, data derived from randomized clinical trials suggest that the efficacy of common preventive agents for coronary heart disease, such as aspirin and hydroxymethylglutaryl-CoA reductase inhibitors, may be derived in part from interactions with the inflammatory system (for review, see Ref. 29Ridker P.M. Blood Coagul. Fibrinolysis. 1999; 10 Suppl. 1: 9-12Google Scholar). A more recent study determined that elevated levels of C-reactive protein, even in the absence of hyperlipidemia, are associated with an increased risk of coronary events (30Ridker P.M. Rifai N. Clearfield M. Downs J.R. Weis S.E. Miles J.S. Gotto Jr., A.M. N. Engl. J. Med. 2001; 344: 1959-1965Crossref PubMed Scopus (1474) Google Scholar). Statin therapy reduces the level of C-reactive protein independent of its effect on lipid levels, potentially through an anti-inflammatory mechanism that these compounds can also exert. Statin therapy may therefore be effective in the primary prevention of coronary events among subjects with relatively low lipid levels but with elevated levels of the acute reactant C-reactive protein.Very little is known about the potential diagnostic and therapeutic significance of the subclinical inflammatory state and its relation to hyperglycemia. However, taken together, our results show that therapies targeting chronic low grade inflammation in the hyperglycemic state for cardiovascular disease prevention should evaluate their effects on the adipocyte as well as the liver.Although the liver is conventionally viewed as the major site of synthesis of many acute phase reactants, it has become increasingly apparent that adipose tissue may play a significant role in the overall systemic levels of these proteins in response to appropriate stimuli. In addition to our work here focusing on SAA3, adipocytes have been proven to express large amounts of complement factor C3 (3Choy L.N. Rosen B.S. Spiegelman B.M. J. Biol. Chem. 1992; 267: 12736-12741Abstract Full Text PDF PubMed Google Scholar) and PAI-1 (for review, see Ref. 31Loskutoff D.J. Samad F. Arterioscler. Thromb. Vasc. Biol. 1998; 18: 1-6Crossref PubMed Scopus (250) Google Scholar).This is the first report that systematically looks at the presence of acute phase reactant proteins in adipose tissue and describes the transcriptional phenomena associated with pro-inflammatory stimuli on these cells. Our data indicate that α1-acid glycoprotein is abundantly expressed in adipocytes, as judged by the high abundance of the mRNA (this paper) and protein. 2Y. Lin and P. E. Scherer, unpublished observations. AGP, also known as orosomucoid (ORM), is a 41–43-kDa glycoprotein and is one of the major acute phase proteins in humans and mice. As with most acute phase proteins, its serum concentration increases in response to systemic tissue injury, inflammation, or infection, and these changes in serum protein concentrations have been correlated with increases in hepatic synthesis (for review, see Ref. 32Fournier T. Medjoubi N.N. Porquet D. Biochim. Biophys. Acta. 2000; 1482: 157-171Crossref PubMed Scopus (778) Google Scholar). Even though the levels of α1-acid glycoprotein and serum amyloid A are mostly governed by the same factors in liver, we found strikingly different regulation in adipose tissue. Although α1-acid glycoprotein is expressed at high levels, its synthesis cannot be further induced by pro-inflammatory stimuli, and its expression is repressed in the diabetic state. This demonstrates a different regulation from SAA3 and may indicate a need for constitutive high level expression in adipocytes. Future studies will have to address which, if any, additional proteinaceous and non-proteinaceous factors may be associated with α1-acid glycoprotein released from adipocytes. Interestingly, a very recent report by Cierniewski and co-workers (33Boncela J. Papiewska I. Fijalkowska I. Walkowiak B. Cierniewski C. J. Biol. Chem. 2001; 276: 35305-35311Abstract Full Text Full Text PDF PubMed Scopus (45) Google Scholar) demonstrates that AGP interacts with plasminogen activator inhibitor type 1 and stabilizes its inhibitory activity. The authors propose that the complex of PAI-1 with AGP could play a role as an alternative reservoir of the physiologically active form of the inhibitor, particularly during inflammation or other acute phase reactions. The fact that both AGP and PAI-1 are co-expressed in adipose tissue is therefore not surprising, and it seems likely that AGP would associate with PAI-1 during biogenesis and potentially influence PAI-1 activity.We found that the lipocalin 24p3 is also abundantly expressed in adipose tissue and repressed in the diabetic state. In contrast to α1-acid glycoprotein, basal levels of 24p3 are very low in 3T3-L1 adipocytes. Similar to SAA3, 24p3 can be induced by LPS. However, unlike SAA3, 24p3 is not induced by TNFα in 3T3-L1 adipocytes. 24p3 has many postulated functions (16Flower D.R. North A.C. Attwood T.K. Biochem. Biophys. Res. Commun. 1991; 180: 69-74Crossref PubMed Scopus (91) Google Scholar, 34Chu S.T. Huang H.L. Chen J.M. Chen Y.H. Biochem. J. 1996; 316: 545-550Crossref PubMed Scopus (36) Google Scholar, 35Chu S.T. Lin H.J. Huang H.L. Chen Y.H. J. Pept. Res. 1998; 52: 390-397Crossref PubMed Scopus (60) Google Scholar, 36Chu S.T. Lee Y.C. Nein K.M. Chen Y.H. Mol. Reprod Dev. 2000; 57: 26-36Crossref PubMed Scopus (36) Google Scholar, 37Garay-Rojas E. Harper M. Hraba-Renevey S. Kress M. Gene. 1996; 170: 173-180Crossref PubMed Scopus (73) Google Scholar, 38Orabona C. Dumoutier L. Renauld J.C. Eur. Cytokine Netw. 2001; 12: 154-161PubMed Google Scholar), among them the potential to serve as a carrier protein, similar to the function proposed for AGP, since the protein has been shown to have binding sites for hydrophobic ligands (35Chu S.T. Lin H.J. Huang H.L. Chen Y.H. J. Pept. Res. 1998; 52: 390-397Crossref PubMed Scopus (60) Google Scholar).Interestingly, Friedman and co-workers (39Soukas A. Cohen P. Socci N.D. Friedman J.M. Genes Dev. 2000; 14: 963-980PubMed Google Scholar) report microarray data showing an increase in SAA3 levels in the white adipose tissue of ob/ob mice when compared with their lean littermates (39Soukas A. Cohen P. Socci N.D. Friedman J.M. Genes Dev. 2000; 14: 963-980PubMed Google Scholar). In addition, they also report 24p3 to be up-regulated to approximately the same extent as SAA3. Although our data fully support the reported up-regulation in SAA3, we failed to observe the up-regulation of 24p3 but actually observed a significant decrease in 24p3 in the diabetic model. Although we cannot explain the observed differences with respect to this specific EST comprising 24p3, there are many interesting conclusions that can be drawn from Friedman's comprehensive microarray study. As the authors point out, many of the most highly up-regulated mRNAs in adipose tissue in the diabetic state fall into the category of inflammatory molecules, including the induction of acute phase reactants and several macrophage marker proteins that are induced within the adipocytes and not within the stromal cell fraction. Additional mRNAs that are highly induced correspond to messages encoding proteins such as heme oxygenase and superoxide dismutase that help cells cope with the increased oxidative stress seen under hyperglycemic conditions. In agreement with the notion that reactive oxygen species are present at elevated levels during hyperglycemia, we found that hyperglycemia per se without the concomitant increase of insulin is at the core of the observed increase of SAA3 in white adipose tissue.Another perhaps unexpected finding is that in adipocytes, not all members of the acute phase reactant family of proteins are regulated in the same manner as they are in hepatocytes. It is clear, however, that the constitutive expression of some of these proteins reflects a functional requirement for their presence under normal conditions to sustain homeostasis as part of an autocrine, paracrine, or endocrine loop. Future studies will have to address whether any bioactive molecules are associated with these proteins as part of their well described carrier functions. Furthermore, the signal transduction pathways that lead to the highly specific transcriptional induction patterns observed for the various acute phase reactants described in this paper will have to be worked out. The central regulatory role of the adipocyte in whole body energy homeostasis is well established. However, pre-adipocytes and adipocytes may also play an important physiological role in the regulation of both the innate and adaptive immune response. We have recently described the response of the adipocyte to various inflammatory stimuli including TNFα and IL-6, focusing primarily on the response to bacterial lipopolysaccharides (LPS), which are mediated through the newly identified Toll-like receptor family (TLRs) (1Lin Y. Lee H. Berg A.H. Lisanti M.P. Shapiro L. Scherer P.E. J. Biol. Chem. 2000; 275: 24255-24263Abstract Full Text Full Text PDF PubMed Scopus (276) Google Scholar). The expression of molecules involved in the innate immune response in adipose tissue, such as complement factors D (adipsin), B, and C3 (2White R.T. Damm D. Hancock N. Rosen B.S. Lowell B.B. Usher P. Flier J.S. Spiegelman B.M. J. Biol. Chem. 1992; 267: 9210-9213Abstract Full Text PDF PubMed Google Scholar, 3Choy L.N. Rosen B.S. Spiegelman B.M. J. Biol. Chem. 1992; 267: 12736-12741Abstract Full Text PDF PubMed Google Scholar, 4Cook K.S. Min H.Y. Johnson D. Chaplinsky R.J. Flier J.S. Hunt C.R. Spiegelman B.M. Science. 1987; 237: 402-405Crossref PubMed Scopus (286) Google Scholar) as well as acute phase reactant proteins (5Scherer P.E. Bickel P.E. Kotler M. Lodish H.F. Nat. Biotechnol. 1998; 16: 581-586Crossref PubMed Scopus (107) Google Scholar), has been demonstrated by our group as well as others. In addition, adipocytes actively secrete and respond to inflammatory cytokines such as TNFα, IL-1, and IL-6 (6Hotamisligil G.S. Shargill N.S. Spiegelman B.M. Science. 1993; 259: 87-91Crossref PubMed Scopus (6028) Google Scholar, 7Mattacks C.A. Pond C.M. Cytokine. 1999; 11: 334-346Crossref PubMed Scopus (72) Google Scholar, 8Mohamed-Ali V. Pinkney J.H. Coppack S.W. Int. J. Obes. Relat. Metab. Disord. 1998; 22: 1145-1158Crossref PubMed Scopus (794) Google Scholar). The latter two are primary cytokine mediators of the acute phase response. The transcription factors responsible for the downstream events of IL-1 and IL-6 include C/EBPβ as well as C/EBPδ, which are directly involved in the acute phase response of the liver (9Burgess-Beusse B.L. Darlington G.J. Mol. Cell. Biol. 1998; 18: 7269-7277Crossref PubMed Scopus (64) Google Scholar, 10Juan T.S. Wilson D.R. Wilde M.D. Darlington G.J. Proc. Natl. Acad. Sci. U. S. A. 1993; 90: 2584-2588Crossref PubMed Scopus (123) Google Scholar), typically considered the main contributor of circulating acute phase reactants. Because of the fact that adipocytes express significant levels of IL-1 and IL-6 receptors as well as high levels of C/EBPα, C/EBPβ, and C/EBPδ, it follows that adipocytes would be capable of producing high levels of acute phase reactant proteins in response to the proper stimuli (5Scherer P.E. Bickel P.E. Kotler M. Lodish H.F. Nat. Biotechnol. 1998; 16: 581-586Crossref PubMed Scopus (107) Google Scholar). However, the relative contribution of adipose tissue to systemic acute phase reactant levels, particularly in the diabetic state, has not been studied to date. Here we investigate the link between the inflammatory/acute phase reactant response and hyperglycemia/hyperinsulinemia at the level of the adipocyte. We show that hyperglycemia, but not hyperinsulinemia, leads to the induction and secretion of SAA3,1 an acute phase reactant in the adipocyte. Adipose tissue, and not the liver, may therefore be responsible for the increased SAA levels in the diabetic state reported in a number of clinical studies. DISCUSSIONThere is an increasing body of evidence that correlates the diabetic phenotype with chronically elevated systemic levels of acute phase reactants and inflammatory cytokines. It has been suggested that these elevated levels may contribute or be partially causative to some of the pathologies of the disease, in particular the increased incidence of cardiovascular problems. Pickup et al. (20Pickup J.C. Mattock M.B. Chusney G.D. Burt D. Diabetologia. 1997; 40: 1286-1292Crossref PubMed Scopus (1040) Google Scholar) show a correlation between the increased levels of several acute phase reactants and IL-6 and the metabolic syndrome X. These authors report elevated levels of circulating α1-acid glycoprotein and serum amyloid A. In addition, other groups show that fibrinogen, another acute phase reactant protein associated with coronary heart disease, is increased in Type II diabetes (21Kannel W.B. D'Agostino R.B. Wilson P.W. Belanger A.J. Gagnon D.R. Am. Heart J. 1990; 120: 672-676Crossref PubMed Scopus (318) Google Scholar, 22Ganda O.P. Arkin C.F. Diabetes Care. 1992; 15: 1245-1250Crossref PubMed Scopus (153) Google Scholar). Furthermore, elevated levels of complement factor C3 have also been reported in diabetic patients (23Figueredo A. Ibarra J.L. Bagazgoitia J. Rodriguez A. Molino A.M. Fernandez-Cruz A. Patino R. Diabetes Care. 1993; 16: 445-449Crossref PubMed Scopus (25) Google Scholar) as well as elevated levels of PAI-1 (24Yudkin J.S. J. Intern. Med. 1995; 238: 21-30Crossref PubMed Scopus (43) Google Scholar). Yokoyama et al.(25Yokoyama H. Jensen J.S. Jensen T. Deckert T. J. Intern. Med. 1995; 237: 519-523Crossref PubMed Scopus (49) Google Scholar) find that acute phase reactants are elevated in Type I diabetes, especially in albuminuric patients. Serum leptin concentrations are responsive to an acute phase or stress response, independent of body mass index (26Pickup J.C. Chusney G.D. Mattock M.B. Clin. Endocrinol. (Oxf). 2000; 52: 107-112Crossref PubMed Scopus (38) Google Scholar).It is theoretically possible that the increase in acute phase reactant proteins, some of which have shown to directly affect lipid metabolism, may contribute to the dyslipidemia associated with diabetes. For example, serum amyloid A displaces apolipoprotein A1 from HDL3, thereby increasing HDL binding to macrophages with the net effect of redirecting HDL cholesterol from the liver to the macrophage for tissue repair (27Cabana V.G. Siegel J.N. Sabesin S.M. J. Lipid Res. 1989; 30: 39-49Abstract Full Text PDF PubMed Google Scholar, 28Steel D.M. Whitehead A.S. Immunol. Today. 1994; 15: 81-88Abstract Full Text PDF PubMed Scopus (867) Google Scholar). A series of prospective cohort studies demonstrate that inflammatory parameters (such as serum amyloid A) and cytokines (such as interleukin-6) are all elevated at base line among patients at risk for future coronary occlusion. Furthermore, data derived from randomized clinical trials suggest that the efficacy of common preventive agents for coronary heart disease, such as aspirin and hydroxymethylglutaryl-CoA reductase inhibitors, may be derived in part from interactions with the inflammatory system (for review, see Ref. 29Ridker P.M. Blood Coagul. Fibrinolysis. 1999; 10 Suppl. 1: 9-12Google Scholar). A more recent study determined that elevated levels of C-reactive protein, even in the absence of hyperlipidemia, are associated with an increased risk of coronary events (30Ridker P.M. Rifai N. Clearfield M. Downs J.R. Weis S.E. Miles J.S. Gotto Jr., A.M. N. Engl. J. Med. 2001; 344: 1959-1965Crossref PubMed Scopus (1474) Google Scholar). Statin therapy reduces the level of C-reactive protein independent of its effect on lipid levels, potentially through an anti-inflammatory mechanism that these compounds can also exert. Statin therapy may therefore be effective in the primary prevention of coronary events among subjects with relatively low lipid levels but with elevated levels of the acute reactant C-reactive protein.Very little is known about the potential diagnostic and therapeutic significance of the subclinical inflammatory state and its relation to hyperglycemia. However, taken together, our results show that therapies targeting chronic low grade inflammation in the hyperglycemic state for cardiovascular disease prevention should evaluate their effects on the adipocyte as well as the liver.Although the liver is conventionally viewed as the major site of synthesis of many acute phase reactants, it has become increasingly apparent that adipose tissue may play a significant role in the overall systemic levels of these proteins in response to appropriate stimuli. In addition to our work here focusing on SAA3, adipocytes have been proven to express large amounts of complement factor C3 (3Choy L.N. Rosen B.S. Spiegelman B.M. J. Biol. Chem. 1992; 267: 12736-12741Abstract Full Text PDF PubMed Google Scholar) and PAI-1 (for review, see Ref. 31Loskutoff D.J. Samad F. Arterioscler. Thromb. Vasc. Biol. 1998; 18: 1-6Crossref PubMed Scopus (250) Google Scholar).This is the first report that systematically looks at the presence of acute phase reactant proteins in adipose tissue and describes the transcriptional phenomena associated with pro-inflammatory stimuli on these cells. Our data indicate that α1-acid glycoprotein is abundantly expressed in adipocytes, as judged by the high abundance of the mRNA (this paper) and protein. 2Y. Lin and P. E. Scherer, unpublished observations. AGP, also known as orosomucoid (ORM), is a 41–43-kDa glycoprotein and is one of the major acute phase proteins in humans and mice. As with most acute phase proteins, its serum concentration increases in response to systemic tissue injury, inflammation, or infection, and these changes in serum protein concentrations have been correlated with increases in hepatic synthesis (for review, see Ref. 32Fournier T. Medjoubi N.N. Porquet D. Biochim. Biophys. Acta. 2000; 1482: 157-171Crossref PubMed Scopus (778) Google Scholar). Even though the levels of α1-acid glycoprotein and serum amyloid A are mostly governed by the same factors in liver, we found strikingly different regulation in adipose tissue. Although α1-acid glycoprotein is expressed at high levels, its synthesis cannot be further induced by pro-inflammatory stimuli, and its expression is repressed in the diabetic state. This demonstrates a different regulation from SAA3 and may indicate a need for constitutive high level expression in adipocytes. Future studies will have to address which, if any, additional proteinaceous and non-proteinaceous factors may be associated with α1-acid glycoprotein released from adipocytes. Interestingly, a very recent report by Cierniewski and co-workers (33Boncela J. Papiewska I. Fijalkowska I. Walkowiak B. Cierniewski C. J. Biol. Chem. 2001; 276: 35305-35311Abstract Full Text Full Text PDF PubMed Scopus (45) Google Scholar) demonstrates that AGP interacts with plasminogen activator inhibitor type 1 and stabilizes its inhibitory activity. The authors propose that the complex of PAI-1 with AGP could play a role as an alternative reservoir of the physiologically active form of the inhibitor, particularly during inflammation or other acute phase reactions. The fact that both AGP and PAI-1 are co-expressed in adipose tissue is therefore not surprising, and it seems likely that AGP would associate with PAI-1 during biogenesis and potentially influence PAI-1 activity.We found that the lipocalin 24p3 is also abundantly expressed in adipose tissue and repressed in the diabetic state. In contrast to α1-acid glycoprotein, basal levels of 24p3 are very low in 3T3-L1 adipocytes. Similar to SAA3, 24p3 can be induced by LPS. However, unlike SAA3, 24p3 is not induced by TNFα in 3T3-L1 adipocytes. 24p3 has many postulated functions (16Flower D.R. North A.C. Attwood T.K. Biochem. Biophys. Res. Commun. 1991; 180: 69-74Crossref PubMed Scopus (91) Google Scholar, 34Chu S.T. Huang H.L. Chen J.M. Chen Y.H. Biochem. J. 1996; 316: 545-550Crossref PubMed Scopus (36) Google Scholar, 35Chu S.T. Lin H.J. Huang H.L. Chen Y.H. J. Pept. Res. 1998; 52: 390-397Crossref PubMed Scopus (60) Google Scholar, 36Chu S.T. Lee Y.C. Nein K.M. Chen Y.H. Mol. Reprod Dev. 2000; 57: 26-36Crossref PubMed Scopus (36) Google Scholar, 37Garay-Rojas E. Harper M. Hraba-Renevey S. Kress M. Gene. 1996; 170: 173-180Crossref PubMed Scopus (73) Google Scholar, 38Orabona C. Dumoutier L. Renauld J.C. Eur. Cytokine Netw. 2001; 12: 154-161PubMed Google Scholar), among them the potential to serve as a carrier protein, similar to the function proposed for AGP, since the protein has been shown to have binding sites for hydrophobic ligands (35Chu S.T. Lin H.J. Huang H.L. Chen Y.H. J. Pept. Res. 1998; 52: 390-397Crossref PubMed Scopus (60) Google Scholar).Interestingly, Friedman and co-workers (39Soukas A. Cohen P. Socci N.D. Friedman J.M. Genes Dev. 2000; 14: 963-980PubMed Google Scholar) report microarray data showing an increase in SAA3 levels in the white adipose tissue of ob/ob mice when compared with their lean littermates (39Soukas A. Cohen P. Socci N.D. Friedman J.M. Genes Dev. 2000; 14: 963-980PubMed Google Scholar). In addition, they also report 24p3 to be up-regulated to approximately the same extent as SAA3. Although our data fully support the reported up-regulation in SAA3, we failed to observe the up-regulation of 24p3 but actually observed a significant decrease in 24p3 in the diabetic model. Although we cannot explain the observed differences with respect to this specific EST comprising 24p3, there are many interesting conclusions that can be drawn from Friedman's comprehensive microarray study. As the authors point out, many of the most highly up-regulated mRNAs in adipose tissue in the diabetic state fall into the category of inflammatory molecules, including the induction of acute phase reactants and several macrophage marker proteins that are induced within the adipocytes and not within the stromal cell fraction. Additional mRNAs that are highly induced correspond to messages encoding proteins such as heme oxygenase and superoxide dismutase that help cells cope with the increased oxidative stress seen under hyperglycemic conditions. In agreement with the notion that reactive oxygen species are present at elevated levels during hyperglycemia, we found that hyperglycemia per se without the concomitant increase of insulin is at the core of the observed increase of SAA3 in white adipose tissue.Another perhaps unexpected finding is that in adipocytes, not all members of the acute phase reactant family of proteins are regulated in the same manner as they are in hepatocytes. It is clear, however, that the constitutive expression of some of these proteins reflects a functional requirement for their presence under normal conditions to sustain homeostasis as part of an autocrine, paracrine, or endocrine loop. Future studies will have to address whether any bioactive molecules are associated with these proteins as part of their well described carrier functions. Furthermore, the signal transduction pathways that lead to the highly specific transcriptional induction patterns observed for the various acute phase reactants described in this paper will have to be worked out. There is an increasing body of evidence that correlates the diabetic phenotype with chronically elevated systemic levels of acute phase reactants and inflammatory cytokines. It has been suggested that these elevated levels may contribute or be partially causative to some of the pathologies of the disease, in particular the increased incidence of cardiovascular problems. Pickup et al. (20Pickup J.C. Mattock M.B. Chusney G.D. Burt D. Diabetologia. 1997; 40: 1286-1292Crossref PubMed Scopus (1040) Google Scholar) show a correlation between the increased levels of several acute phase reactants and IL-6 and the metabolic syndrome X. These authors report elevated levels of circulating α1-acid glycoprotein and serum amyloid A. In addition, other groups show that fibrinogen, another acute phase reactant protein associated with coronary heart disease, is increased in Type II diabetes (21Kannel W.B. D'Agostino R.B. Wilson P.W. Belanger A.J. Gagnon D.R. Am. Heart J. 1990; 120: 672-676Crossref PubMed Scopus (318) Google Scholar, 22Ganda O.P. Arkin C.F. Diabetes Care. 1992; 15: 1245-1250Crossref PubMed Scopus (153) Google Scholar). Furthermore, elevated levels of complement factor C3 have also been reported in diabetic patients (23Figueredo A. Ibarra J.L. Bagazgoitia J. Rodriguez A. Molino A.M. Fernandez-Cruz A. Patino R. Diabetes Care. 1993; 16: 445-449Crossref PubMed Scopus (25) Google Scholar) as well as elevated levels of PAI-1 (24Yudkin J.S. J. Intern. Med. 1995; 238: 21-30Crossref PubMed Scopus (43) Google Scholar). Yokoyama et al.(25Yokoyama H. Jensen J.S. Jensen T. Deckert T. J. Intern. Med. 1995; 237: 519-523Crossref PubMed Scopus (49) Google Scholar) find that acute phase reactants are elevated in Type I diabetes, especially in albuminuric patients. Serum leptin concentrations are responsive to an acute phase or stress response, independent of body mass index (26Pickup J.C. Chusney G.D. Mattock M.B. Clin. Endocrinol. (Oxf). 2000; 52: 107-112Crossref PubMed Scopus (38) Google Scholar). It is theoretically possible that the increase in acute phase reactant proteins, some of which have shown to directly affect lipid metabolism, may contribute to the dyslipidemia associated with diabetes. For example, serum amyloid A displaces apolipoprotein A1 from HDL3, thereby increasing HDL binding to macrophages with the net effect of redirecting HDL cholesterol from the liver to the macrophage for tissue repair (27Cabana V.G. Siegel J.N. Sabesin S.M. J. Lipid Res. 1989; 30: 39-49Abstract Full Text PDF PubMed Google Scholar, 28Steel D.M. Whitehead A.S. Immunol. Today. 1994; 15: 81-88Abstract Full Text PDF PubMed Scopus (867) Google Scholar). A series of prospective cohort studies demonstrate that inflammatory parameters (such as serum amyloid A) and cytokines (such as interleukin-6) are all elevated at base line among patients at risk for future coronary occlusion. Furthermore, data derived from randomized clinical trials suggest that the efficacy of common preventive agents for coronary heart disease, such as aspirin and hydroxymethylglutaryl-CoA reductase inhibitors, may be derived in part from interactions with the inflammatory system (for review, see Ref. 29Ridker P.M. Blood Coagul. Fibrinolysis. 1999; 10 Suppl. 1: 9-12Google Scholar). A more recent study determined that elevated levels of C-reactive protein, even in the absence of hyperlipidemia, are associated with an increased risk of coronary events (30Ridker P.M. Rifai N. Clearfield M. Downs J.R. Weis S.E. Miles J.S. Gotto Jr., A.M. N. Engl. J. Med. 2001; 344: 1959-1965Crossref PubMed Scopus (1474) Google Scholar). Statin therapy reduces the level of C-reactive protein independent of its effect on lipid levels, potentially through an anti-inflammatory mechanism that these compounds can also exert. Statin therapy may therefore be effective in the primary prevention of coronary events among subjects with relatively low lipid levels but with elevated levels of the acute reactant C-reactive protein. Very little is known about the potential diagnostic and therapeutic significance of the subclinical inflammatory state and its relation to hyperglycemia. However, taken together, our results show that therapies targeting chronic low grade inflammation in the hyperglycemic state for cardiovascular disease prevention should evaluate their effects on the adipocyte as well as the liver. Although the liver is conventionally viewed as the major site of synthesis of many acute phase reactants, it has become increasingly apparent that adipose tissue may play a significant role in the overall systemic levels of these proteins in response to appropriate stimuli. In addition to our work here focusing on SAA3, adipocytes have been proven to express large amounts of complement factor C3 (3Choy L.N. Rosen B.S. Spiegelman B.M. J. Biol. Chem. 1992; 267: 12736-12741Abstract Full Text PDF PubMed Google Scholar) and PAI-1 (for review, see Ref. 31Loskutoff D.J. Samad F. Arterioscler. Thromb. Vasc. Biol. 1998; 18: 1-6Crossref PubMed Scopus (250) Google Scholar). This is the first report that systematically looks at the presence of acute phase reactant proteins in adipose tissue and describes the transcriptional phenomena associated with pro-inflammatory stimuli on these cells. Our data indicate that α1-acid glycoprotein is abundantly expressed in adipocytes, as judged by the high abundance of the mRNA (this paper) and protein. 2Y. Lin and P. E. Scherer, unpublished observations. AGP, also known as orosomucoid (ORM), is a 41–43-kDa glycoprotein and is one of the major acute phase proteins in humans and mice. As with most acute phase proteins, its serum concentration increases in response to systemic tissue injury, inflammation, or infection, and these changes in serum protein concentrations have been correlated with increases in hepatic synthesis (for review, see Ref. 32Fournier T. Medjoubi N.N. Porquet D. Biochim. Biophys. Acta. 2000; 1482: 157-171Crossref PubMed Scopus (778) Google Scholar). Even though the levels of α1-acid glycoprotein and serum amyloid A are mostly governed by the same factors in liver, we found strikingly different regulation in adipose tissue. Although α1-acid glycoprotein is expressed at high levels, its synthesis cannot be further induced by pro-inflammatory stimuli, and its expression is repressed in the diabetic state. This demonstrates a different regulation from SAA3 and may indicate a need for constitutive high level expression in adipocytes. Future studies will have to address which, if any, additional proteinaceous and non-proteinaceous factors may be associated with α1-acid glycoprotein released from adipocytes. Interestingly, a very recent report by Cierniewski and co-workers (33Boncela J. Papiewska I. Fijalkowska I. Walkowiak B. Cierniewski C. J. Biol. Chem. 2001; 276: 35305-35311Abstract Full Text Full Text PDF PubMed Scopus (45) Google Scholar) demonstrates that AGP interacts with plasminogen activator inhibitor type 1 and stabilizes its inhibitory activity. The authors propose that the complex of PAI-1 with AGP could play a role as an alternative reservoir of the physiologically active form of the inhibitor, particularly during inflammation or other acute phase reactions. The fact that both AGP and PAI-1 are co-expressed in adipose tissue is therefore not surprising, and it seems likely that AGP would associate with PAI-1 during biogenesis and potentially influence PAI-1 activity. We found that the lipocalin 24p3 is also abundantly expressed in adipose tissue and repressed in the diabetic state. In contrast to α1-acid glycoprotein, basal levels of 24p3 are very low in 3T3-L1 adipocytes. Similar to SAA3, 24p3 can be induced by LPS. However, unlike SAA3, 24p3 is not induced by TNFα in 3T3-L1 adipocytes. 24p3 has many postulated functions (16Flower D.R. North A.C. Attwood T.K. Biochem. Biophys. Res. Commun. 1991; 180: 69-74Crossref PubMed Scopus (91) Google Scholar, 34Chu S.T. Huang H.L. Chen J.M. Chen Y.H. Biochem. J. 1996; 316: 545-550Crossref PubMed Scopus (36) Google Scholar, 35Chu S.T. Lin H.J. Huang H.L. Chen Y.H. J. Pept. Res. 1998; 52: 390-397Crossref PubMed Scopus (60) Google Scholar, 36Chu S.T. Lee Y.C. Nein K.M. Chen Y.H. Mol. Reprod Dev. 2000; 57: 26-36Crossref PubMed Scopus (36) Google Scholar, 37Garay-Rojas E. Harper M. Hraba-Renevey S. Kress M. Gene. 1996; 170: 173-180Crossref PubMed Scopus (73) Google Scholar, 38Orabona C. Dumoutier L. Renauld J.C. Eur. Cytokine Netw. 2001; 12: 154-161PubMed Google Scholar), among them the potential to serve as a carrier protein, similar to the function proposed for AGP, since the protein has been shown to have binding sites for hydrophobic ligands (35Chu S.T. Lin H.J. Huang H.L. Chen Y.H. J. Pept. Res. 1998; 52: 390-397Crossref PubMed Scopus (60) Google Scholar). Interestingly, Friedman and co-workers (39Soukas A. Cohen P. Socci N.D. Friedman J.M. Genes Dev. 2000; 14: 963-980PubMed Google Scholar) report microarray data showing an increase in SAA3 levels in the white adipose tissue of ob/ob mice when compared with their lean littermates (39Soukas A. Cohen P. Socci N.D. Friedman J.M. Genes Dev. 2000; 14: 963-980PubMed Google Scholar). In addition, they also report 24p3 to be up-regulated to approximately the same extent as SAA3. Although our data fully support the reported up-regulation in SAA3, we failed to observe the up-regulation of 24p3 but actually observed a significant decrease in 24p3 in the diabetic model. Although we cannot explain the observed differences with respect to this specific EST comprising 24p3, there are many interesting conclusions that can be drawn from Friedman's comprehensive microarray study. As the authors point out, many of the most highly up-regulated mRNAs in adipose tissue in the diabetic state fall into the category of inflammatory molecules, including the induction of acute phase reactants and several macrophage marker proteins that are induced within the adipocytes and not within the stromal cell fraction. Additional mRNAs that are highly induced correspond to messages encoding proteins such as heme oxygenase and superoxide dismutase that help cells cope with the increased oxidative stress seen under hyperglycemic conditions. In agreement with the notion that reactive oxygen species are present at elevated levels during hyperglycemia, we found that hyperglycemia per se without the concomitant increase of insulin is at the core of the observed increase of SAA3 in white adipose tissue. Another perhaps unexpected finding is that in adipocytes, not all members of the acute phase reactant family of proteins are regulated in the same manner as they are in hepatocytes. It is clear, however, that the constitutive expression of some of these proteins reflects a functional requirement for their presence under normal conditions to sustain homeostasis as part of an autocrine, paracrine, or endocrine loop. Future studies will have to address whether any bioactive molecules are associated with these proteins as part of their well described carrier functions. Furthermore, the signal transduction pathways that lead to the highly specific transcriptional induction patterns observed for the various acute phase reactants described in this paper will have to be worked out. We thank members of the Scherer laboratory, especially Dr. Terry Combs for valuable assistance with the in vivo experiments. We also thank Dr. Michael Brownlee for helpful discussions, Dr. David Cohen and Sonya VanPatten for assistance with the diabetic mouse models, and Xiao Man Yang and Marie Thompson for technical assistance.