|
|
||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Thematic Review |






* Division of Endocrinology and Metabolism, Department of Medicine, Faculty of Medicine, Chulalongkorn University and King Chulalongkorn Memorial Hospital, Bangkok, Thailand
Department of Medicine, University of California, San Francisco, and Metabolism Section, Medical Service, Department of Veterans Affairs Medical Center, San Francisco, CA
Published, JLR Papers in Press, April 21, 2004. DOI 10.1194/jlr.R300019-JLR200
1 This paper is dedicated to the memory of Dr. Riaz A. Memon, deceased. ![]()
2 To whom correspondence should be addressed. e-mail: grunfld{at}itsa.ucsf.edu (C.G.); kfngld{at}itsa.ucsf.edu (K.R.F.)
| ABSTRACT |
|---|
|
|
|---|
APR-induced alterations initially protect the host from the harmful effects of bacteria, viruses, and parasites. However, if prolonged, these changes in the structure and function of lipoproteins will contribute to atherogenesis.
Abbreviations: ABC, ATP binding cassette; ACC, acetyl CoA carboxylase; ACS, acyl-CoA synthetase; AIDS, acquired immune deficiency syndrome; AP2, adipocyte P2; apoE, apolipoprotein E; APR, acute-phase response; BSEP, bile salt export pump; CAD, coronary artery disease; CAR, constitutive androstane receptor; CETP, cholesteryl ester transfer protein; CNTF, ciliary neurotropic factor; CPT, carnitine palmitoyl transferase; CRP, C-reactive protein; CYP7A1, cholesterol 7
-hydroxylase; CYP7B1, oxysterol 7
-hydroxylase; CYP8B1, sterol 12
-hydroxylase; CYP27A1, sterol 27-hydroxylase; DR, direct repeat; ERK, extracellular signal-related kinase; EL, endothelial lipase; FABP, fatty acid binding protein; FAS, fatty acid synthase; FAT, fatty acid translocase; FATP, fatty acid transport protein; FXR, farnesoid X receptor; GlcCer, glucosylceramide; GSL, glycosphingolipid; HIV, human immunodeficiency virus; HNF, hepatocyte nuclear factor; HSL, hormone-sensitive lipase; IL, interleukin; KB, ketone body; KGF, keratinocyte growth factor; LBP, lipopolysaccharide binding protein; LIF, leukemia inhibitory factor; Lp[a], lipoprotein [a]; LPC, lysophosphatidylcholine; LPL, lipoprotein lipase; LPS, lipopolysaccharide; LRH-1, liver receptor homolog-1; LTA, lipoteichoic acid; LXR, liver X receptor; MDR3, multidrug resistance-3; MEK, mitogen-activated protein kinase kinase; MRP2, multidrug resistance-associated protein-2; NF-
B, nuclear factor
B; NF-IL-6, nuclear factor interleukin-6; NGF, nerve growth factor; OATP, organic anion-transporting protein; oxPAPC, oxidized 1-palmitoyl-2-arachidonoyl-sn-glycero-3-phosphorylcholine; PAF, platelet-activating factor; PAF-AH, platelet-activating factor acetylhydrolase; PKA, protein kinase A; PLTP, phospholipid transfer protein; PON1, paraoxonase 1; PPAR, peroxisome proliferator-activated receptor; PTHrP, parathyroid hormone-related protein; PXR, pregnane X receptor; RCT, reverse cholesterol transport; RXR, retinoid X receptor; SAA, serum amyloid A; SHP, small heterodimer partner; sPLA2, secretory phospholipase A2; SPT, serine palmitoyltransferase; SR-BI, scavenger receptor class B type I; TG, triglyceride; TLF, trypanosome lytic factor; TNF, tumor necrosis factor; TR, thyroid hormone receptor
Supplementary key words acute-phase response endotoxin lipopolysaccharide cytokine atherosclerosis
| INTRODUCTION |
|---|
|
|
|---|
-fetoprotein) (1). Changes in acute-phase protein concentrations are largely attributable to alterations in their rate of synthesis in the liver, although similar changes occur in extrahepatic tissues. Microarrays of mouse liver after endotoxin treatment demonstrate that
7% of the genes respond to endotoxin challenge (2). These changes in acute-phase proteins are often species specific with regard to the magnitude and direction of change. The APR induced during infection/inflammation protects the host from further injury (1). Changes in acute-phase proteins neutralize invading microorganisms, minimize the extent of tissue damage, participate in the local immune response and tissue regeneration, and replenish proteins used in the inflammatory process. These changes, if present for a prolonged period of time, can lead to detrimental consequences to the host, such as the development of systemic amyloidosis after chronic infection or inflammation.
Changes in acute-phase protein synthesis are mediated by cytokines produced in response to a variety of stimuli in multiple cell types, including macrophages, monocytes, T-lymphocytes, and endothelial cells (1). Key cytokines responsible for the coordination of both immune and inflammatory responses include tumor necrosis factors (TNF-
and TNF-ß), interleukins (ILs), and interferons (IFN-
, -ß, and -
) (1). Redundancy classically occurs in essential parts of the host response, as several structurally different cytokines may exert similar biological effects even though they bind to different receptors. Combinations of certain cytokines produce additive or synergistic effects, whereas other cytokines may have inhibitory effects, indicating the complex nature of the host response (35).
Infection and inflammation are accompanied by similar cytokine-induced alterations in lipid and lipoprotein metabolism. Of note, inflammatory cytokines are increased and play a pathogenic role in a variety of very common disorders, such as diabetes, obesity, metabolic syndrome, hypertension, chronic heart failure, chronic renal failure, and atherosclerosis (612). Many of these disorders display abnormalities in lipid metabolism that are similar to those that occur during infection and inflammation.
This review summarizes the changes in lipid and lipoprotein during infection/inflammation and their molecular mechanisms. Most mechanistic studies were carried out in animal models of infection using endotoxin [lipopolysaccharide (LPS)], a well-characterized inducer of cytokines and the APR, or the proinflammatory cytokines (TNF and IL-1), which mediate the APR. We describe the role of transcription factors in regulating lipid metabolism during infection/inflammation. Finally, we discuss both the beneficial effects and deleterious consequences to the host of APR-induced changes in lipid and lipoprotein metabolism.
| CHANGES IN LIPID AND LIPOPROTEIN METABOLISM DURING INFECTION AND INFLAMMATION |
|---|
|
|
|---|
The effects of infection and inflammation on TG metabolism are similar in all species, whereas changes in cholesterol metabolism differ between rodents and primates. In rodents, there is an increase in serum total cholesterol levels and hepatic cholesterol synthesis, whereas nonhuman primates and humans have either no change or a decrease in serum cholesterol and LDL levels (13). The mechanisms underlying this species difference is not known. HDL levels are decreased in both rodents and primates during the APR, and there are marked changes in proteins associated with HDL metabolism (14). Finally, infection produces alterations in the composition and function of lipoproteins, including changes in sphingolipid concentrations, decreased reverse cholesterol transport (RCT), and increased oxidation of lipids.
TG metabolism
Patients with gram-negative or gram-positive bacterial infections and viral infections have increased serum TG levels (1517). In animals, administration of LPS, a major component of the cell wall of gram-negative bacteria, or lipoteichoic acid (LTA), a component of the cell wall of gram-positive bacteria, produces hypertriglyceridemia (1828) (Table 1). Multiple cytokines increase serum TG levels in rodents and in humans (2940). The hypertriglyceridemic effect of LPS and cytokines is rapid, occurring within 2 h after administration, and is sustained for at least 24 h (26, 29). The doses of LPS or cytokines that produce hypertriglyceridemia in rodents are similar to those that produce fever, anorexia, and changes in acute-phase protein synthesis, suggesting that hypertriglyceridemia is a very sensitive, physiological part of the host response to infection rather than a manifestation of toxicity (26).
|
Serum TG levels are increased by multiple cytokines, including TNF, IL-1, IL-2, IL-6, leukemia inhibitory factor (LIF), ciliary neurotropic factor (CNTF), nerve growth factor (NGF), keratinocyte growth factor (KGF), platelet-activating factor (PAF), and parathyroid hormone-related protein (PTHrP) (30, 3239, 4147) (Table 1), suggesting redundancy. IL-4, an anti-inflammatory cytokine, opposes the action of some, but not all, of these cytokines (48). The effects of cytokines on TG metabolism are likely direct and not mediated by hormones such as insulin, cortisol, or catecholamines, as TNF increases serum TG levels in insulinopenic diabetic animals and adrenalectomized rats (49, 50). Moreover, TNF also increases serum TG levels under various dietary conditions from high sucrose, which stimulates endogenous FA synthesis, to high fat, which suppresses endogenous FA synthesis (51, 52).
Increased VLDL production
Increased de novo FA and TG synthesis
LPS and several cytokines, including TNF-
, TNF-ß (lymphotoxin), IL-1, IL-6, IFN-
, LIF, CNTF, NGF, PAF, and PTHrP, rapidly induce de novo FA synthesis and hepatic TG synthesis in rodents (26, 29, 31, 32, 34, 43, 44, 47, 50, 53) (Table 1). Hepatic secretion of apoB also increases (54), resulting in an increased number of VLDL particles secreted. In contrast, other cytokines, such as IL-2, IL-4, and IFN-
, do not stimulate hepatic FA synthesis (31, 48).
TNF rapidly increases hepatic FA synthesis within 1 h after administration, which is sustained for at least 17 h (29). The time course for stimulation of hepatic FA synthesis and VLDL secretion is consistent with the time course for TNF-induced hypertriglyceridemia (29, 55). However, TNF does not acutely increase the total activity of the rate-limiting enzymes of FA synthesis [i.e., acetyl CoA carboxylase (ACC) and FA synthase (FAS)] or alter the phosphorylation state of ACC, a mechanism that regulates ACC activity (56). Instead, TNF acutely increases intracellular concentrations of citrate, an allosteric activator of ACC (56) (Fig. 1)
. IL-1 and IL-6 increase hepatic FA synthesis by increasing hepatic citrate levels, whereas IFN-
, which also increases hepatic FA synthesis, has no effect on citrate levels, suggesting a different mechanism (53). The stimulatory effects of TNF or IL-1 and IFN-
on hepatic FA synthesis are additive or synergistic, whereas there is no such synergy between TNF and IL-1 or TNF and IL-6 (53). Finally, IL-4, an anti-inflammatory cytokine, inhibits the stimulatory effects of TNF, IL-1, and IL-6 on hepatic FA synthesis by blocking the increase in hepatic citrate levels (48). In contrast, IL-4 does not block the stimulatory effect of IFN-
on FA synthesis in liver (48). Thus, analogous to cytokine regulation of the immune response, there are complex interactions among the metabolic effects of cytokines that may be additive, synergistic, or antagonistic.
|
Increased adipose tissue lipolysis
Adipose tissue lipolysis also provides FAs for increased hepatic TG synthesis during infection. The mobilized FAs are delivered to the liver and, instead of being oxidized, become reesterified into TGs and secreted into the circulation as VLDL.
LPS, LTA, and several cytokines induce adipose tissue lipolysis in both intact animals and 3T3-L1 adipocytes (26, 28, 34, 37, 43, 45, 52, 5863). The effects of different cytokines are specific and dependent upon the nutritional status of the host (Table 1). TNF acutely induces lipolysis in chow-fed but not in sucrose-fed animals (52). IL-1 does not stimulate lipolysis; its effect on serum TG levels is attributable to enhanced hepatic FA synthesis and TG secretion (32). IL-6, LIF, and CNTF, which act through the same receptor transducer (gp130), stimulate both hepatic FA synthesis and adipose tissue lipolysis (34, 43). On the other hand, KGF stimulates lipolysis but has no effect on hepatic FA synthesis (45). Finally, both IFN-
and IFN-
stimulate lipolysis, but those peripherally derived FAs do not contribute to increased TG synthesis in the liver because they are oxidized, producing ketone bodies (KBs) (63).
Lipolysis in adipose tissue is primarily driven by hormone-sensitive lipase (HSL), which is regulated either by alteration in its phosphorylation state or by induction of gene expression. Several cytokines that induce lipolysis, including TNF, IFN-
, and IFN-
, produce a marked decrease in HSL mRNA (64), indicating that gene regulation of HSL does not play a role in cytokine-induced lipolysis. Rather, lipolysis is likely attributable to phosphorylation of HSL or its associated proteins. TNF-induced lipolysis in cultured human adipocytes is associated with the activation of mitogen-activated protein kinase kinase (MEK)-extracellular signal-related kinase (ERK) (65), leading to decreases in cyclic nucleotide phosphodiesterase 3B, an enzyme that hydrolyzes cAMP. Increased intracellular cAMP consequently activates cAMP-dependent protein kinase A (PKA), which phosphorylates perilipins, phosphoproteins located at the surface of lipid droplets in adipocytes. Phosphorylation of perilipin A or B modifies lipid surfaces, allowing access of lipases to the lipid droplets, promoting lipolysis. Activation of the MEK-ERK pathway and PKA has also been shown to phosphorylate HSL and increase its lipolytic activity (65, 66).
LPS and cytokines may also induce lipolysis by decreasing the expression of acyl-CoA synthetase (ACS) in adipose tissue (64). ACS catalyzes the activation of long-chain FAs to acyl-CoA esters that are subsequently metabolized in anabolic or catabolic pathways depending on the type of tissue, the nutritional status, and the hormonal milieu of the host. Although FA transport across biological membranes is a bidirectional process, activation of FAs to acyl-CoA esters prevents the efflux of FAs from cells and hence renders FA transport unidirectional. In adipose tissue, ACS is primarily associated with microsomes to support the synthesis of TG for storage of energy. During the APR, there is a coordinated decrease in the mRNA expression of fatty acid transport proteins (FATPs) and ACS mRNA and activity in adipose tissue (67, 68) that likely prevents the activation and storage of FAs and may promote the mobilization of FAs.
Decreased hepatic FA oxidation and ketogenesis
Bacterial infections are accompanied by the suppression of hepatic FA oxidation (69, 70). Increased FA substrate provided by increased hepatic FA synthesis and adipose tissue lipolysis is then directed away from oxidation and channeled toward reesterification. This concept is supported by the demonstration that LPS, TNF, and IL-1 decrease mitochondrial but increase microsomal ACS activity in the liver (68). Decreased mitochondrial ACS prevents the activation of FA for entry into mitochondria for oxidation, whereas increased microsomal ACS enhances the reesterification of FAs for TG synthesis.
LPS and cytokines differentially regulate the hepatic mRNA expression of membrane-associated FATPs involved in the uptake of peripherally derived FAs. LPS and cytokines increase the expression of CD36/fatty acid translocase (FAT) while decreasing the mRNA levels of FATP in the liver, suggesting that these proteins may be involved in directing FAs to different intracellular locations (67) (Fig. 1). We propose that CD36/FAT transports FAs to cytosol for reesterification, which is enhanced during the APR, whereas FATP transports FAs toward mitochondria for oxidation, which is suppressed during the APR. LPS also decreases the mRNA and protein levels of cytosolic fatty acid binding protein (FABP) in liver, heart, and muscle (71). Because FABPs are thought to facilitate the transport of FAs to the site of utilization in the cell, the decrease in FABP may also contribute to decreased FA oxidation during infection. The fact that TNF does not acutely increase the activities of regulatory enzymes of TG synthesis, such as glycerol phosphate acyltransferase and diacylglycerol acyltransferase (52), also suggests that the acute increase in TG synthesis is driven by increased FA substrate.
Mitochondrial ACS converts FA into fatty acyl-CoA, which is subsequently metabolized by mitochondrial carnitine palmitoyl transferase-I (CPT-I) into acylcarnitine. CPT-II subsequently metabolizes acylcarnitine into acyl-CoA, which allows FA entrance into the mitochondria, where it undergoes ß-oxidation. Hepatic expression of both CPT-I, the rate-limiting enzyme for mitochondrial FA oxidation, and CPT-II is decreased during sepsis (72, 73) (Fig. 1). LPS, IL-1, and TNF increase levels of hepatic malonyl-CoA, an allosteric inhibitor of CPT-I, which inhibits the remaining CPT-I, decreasing FA oxidation (74) (Table 1). IFN-
at high doses increases hepatic malonyl-CoA levels (63), whereas IFN-
does not affect hepatic malonyl-CoA levels (63).
Given the decrease in FA oxidation, infection is associated with the suppression of hepatic KB production (69, 70). Serum KB levels are regulated by their rates of synthesis in the liver and utilization in peripheral tissues. Infection decreases KB production through the inhibition of FA oxidation but also likely by increased peripheral KB utilization.
Various cytokines have different effects on KB metabolism (Table 1). Both TNF and IL-1 acutely decrease serum KB levels in mice (30, 74). In the fed state, IL-1 increases hepatic malonyl-CoA levels, inhibiting CPT-I and preventing KB production. During fasting, IL-1 inhibits lipolysis, reducing FA substrate to the liver for KB synthesis (74). Although TNF increases hepatic malonyl-CoA levels, it stimulates peripheral lipolysis, increasing the flux of FA substrate to the liver, with no net effect on hepatic KB levels (74), suggesting that TNF decreases serum KB through changes in KB catabolism. IL-6 has no effect on serum KB levels (34). IFN-
has biphasic effects: low doses of IFN-
increase serum KB levels by mobilization of FA substrate, whereas higher doses have no effect (63). IFN-
stimulates adipose tissue lipolysis, increasing serum and hepatic KB levels (63).
FA uptake and oxidation decrease in heart and skeletal muscle during the APR, shifting their metabolism from FA as the preferred fuel substrate to glucose, whose uptake and utilization are increased (7577). This makes more FA available to liver and other tissues, such as those of the immune system. IL-1, but not TNF, decreases LPL activity in the heart (7880). LPS, TNF, and IL-1 decrease the mRNA expression of FA transport and binding proteins and ACS in heart and muscle (67, 71). It is likely that this coordinated decrease in FA transport and binding proteins and ACS is the mechanism for the decreased uptake and utilization of FA in heart and muscle during infection/inflammation.
Decreased VLDL clearance Infection may also increase serum TG levels by decreasing VLDL clearance. Early in vitro studies showed that TNF decreases LPL expression in cultured adipocytes (81, 82). In vivo, however, there is little evidence that hypertriglyceridemia is attributable to decreased LPL activity. First, although TNF reduces LPL activity in epididymal fat pads in rodents (80, 83), this decrease requires many hours, whereas the TNF-induced increase in serum TG levels occurs very rapidly (29). Second, TNF administration does not decrease LPL activity in other adipose tissue sites or in muscle (41, 83). Third, TNF-neutralizing antibodies block the LPS-induced increase in serum TG levels in mice but they do not block LPS-induced inhibition of LPL in mouse adipose tissue, again dissociating the LPS-induced increase in serum TGs from changes in LPL activity (33). Finally, TNF does not decrease the clearance of chylomicrons or VLDL from the circulation, the mechanism by which changes in LPL might influence TG levels (41, 49, 84).
Like TNF, IL-1, IL-6, and LIF also require several hours to decrease LPL activity in vivo in mouse adipose tissue (80). IFN-
and IFN-
increase serum TG levels in humans (38, 39) but do not increase TG levels in rodents, despite decreasing LPL activity in cultured murine 3T3-L1 fat cells (64, 82), again showing discordance between LPL activity and TG levels.
There may be a role for the decreased clearance of TG with high doses of LPS. Low doses of LPS enhance hepatic VLDL secretion and increase serum TG levels without affecting TG clearance in rats. In contrast, high doses of LPS inhibit the clearance of TG-rich lipoproteins (26). Moreover, high doses of LPS decrease postheparin plasma LPL activity and LPL activity in adipose tissue and muscle (80).
LPS and cytokines also decrease apoE mRNA in many tissues, including the liver, and VLDL has lower amounts of apoE during infection (54, 85, 86). Because apoE is required for the clearance of TG-rich lipoproteins, decreased apoE could contribute to the delayed clearance observed in rats with infection (87).
Cholesterol metabolism
There are marked alterations in the metabolism of cholesterol, LDL, HDL, and RCT during infection. LPS and cytokines decrease total serum cholesterol levels in primates, whereas in rodents they increase cholesterol levels by stimulating de novo cholesterol synthesis, decreasing lipoprotein clearance, and decreasing the conversion of cholesterol into bile acids. Such species-specific responses in the APR are common, but the underlying mechanisms responsible for these differences are not yet understood. There are baseline differences in serum cholesterol levels among species, with rodents having low LDL levels and primates having relatively high LDL levels. Baseline levels are often related to the direction of changes in the APR. There are classic positive acute-phase proteins that are expressed at baseline in some species, and they do not increase during the APR in those species.
Hepatic cholesterol synthesis In rodents, LPS stimulates hepatic cholesterol synthesis (27) (Table 2). In contrast to the acute effect of LPS on de novo FA synthesis, the effect of LPS on hepatic cholesterol synthesis is delayed, occurring 16 h after administration (27). LPS stimulates hepatic cholesterol synthesis by increasing the transcription rate, mRNA expression, protein mass, and activity of HMG-CoA reductase, the rate-limiting enzyme in the biosynthetic pathway of cholesterol liver (27, 88). The effect of LPS on HMG-CoA reductase is specific, as the mRNA expression of several other enzymes in the cholesterol synthetic pathway, including HMG-CoA synthase and farnesyl pyrophosphate synthase, which are usually coordinately regulated with HMG-CoA reductase under nutritional or pharmacological manipulations, is not altered by LPS treatment (2, 88) (Fig. 2) . Moreover, LPS still upregulates HMG-CoA reductase mRNA expression when its basal expression is increased by treatment with bile acid binding resins or decreased by feeding a high-cholesterol diet (88). Thus, the stimulatory effect of LPS on HMG-CoA reductase is independent of dietary regulation and persists over a wide range of basal expression.
|
|
Like LPS, several cytokines, including TNF, IL-1, IL-6, KGF, and NGF, produce a delayed increase in serum cholesterol levels in rodents (29, 32, 34, 44, 45) (Table 2). TNF-
, TNF-ß, IL-1, and IFN-
stimulate hepatic cholesterol synthesis in mice, whereas IFN-
and IL-2 have no such effect (31). Like LPS, both TNF and IL-1 stimulate de novo hepatic cholesterol by increasing the activity and mRNA expression of HMG-CoA reductase (88, 92). TNF and IL-1 decrease squalene synthase activity and mRNA expression (89); they may also divert the flux of mevalonate metabolites into nonsterol pathways during the APR.
In primates, including humans, infection/inflammation decreases serum cholesterol as a result of decreases in both LDL and HDL cholesterol (16, 17, 24, 25). LPS, TNF, IL-2, IFN-ß, granulocyte-macrophage colony-stimulating factor, and macrophage colony-stimulating factor decrease serum cholesterol, whereas IL-1 has no effect (24, 25, 36, 42, 9397). The decrease in cholesterol is accompanied by a reduction in serum apoB levels.
The mechanism by which infection/inflammation decreases cholesterol levels has not been thoroughly studied in intact primates. Most of the mechanistic studies were performed in vitro using human hepatoma HepG2 cells. IL-1 inhibits cholesterol synthesis and decreases cholesterol and apoB secretion, whereas IL-6 increases cholesterol synthesis but decreases cholesterol secretion (98). IFN-ß also decreases apoB synthesis (99).
LDL clearance In rats, LPS significantly inhibits the clearance of LDL from the circulation (100). LPS decreases the expression of hepatic LDL receptor protein (Table 2), but the decrease in protein levels could not be explained by changes in mRNA levels, suggesting that posttranscriptional regulation occurs during the APR (101). In a rat model of gram-negative sepsis, the rate of apoB degradation is decreased (87). In hamsters, however, LPS, IL-1, and TNF either have no effect or produce a slight increase in hepatic LDL receptor mRNA and protein levels (27). In human HepG2 cells, IL-1 and TNF increase LDL receptor activity (102, 103). The differences may explain the species-specific response in cholesterol metabolism commonly seen during the APR.
Decreased hepatic cholesterol catabolism and excretion
Equipped with a number of enzymes and transporters, hepatocytes secrete bile salts, phospholipids, cholesterol, organic anions, and cations into the bile. Cholesterol returned to the liver is primarily metabolized into bile acids, representing the major pathway for the elimination of cholesterol from the body. There are two distinct pathways of bile acid synthesis in mammalian liver (104, 105). The classic or neutral pathway is initiated by microsomal cholesterol 7
-hydroxylase (CYP7A1) that converts cholesterol into 7
-hydroxycholesterol, which is subsequently converted into primary bile acids. The alternative or acidic pathway is initiated by mitochondrial sterol 27-hydroxylase (CYP27A1) that converts cholesterol into 27-hydroxycholesterol, which is then converted into 7
,27-dihydroxycholesterol by oxysterol 7
-hydroxylase (CYP7B1) and subsequently metabolized into primary bile acids. The alternative pathway may contribute as much as 50% to total bile acid synthesis (104, 105). Primary bile acids synthesized in hepatocytes are conjugated with taurine and glycine. At physiological pH, these conjugates exist in the anionic salt form; therefore, they are called bile salts. Secretion of bile salts mediates the solubilization of lipids from the canalicular membrane, resulting in the secretion of biliary phospholipids and cholesterol.
As polarized cells, hepatocytes contain multiple transporters at the basolateral (sinusoidal) and apical (canalicular) surfaces (106). Basolateral bile salt uptake from the portal circulation is primarily mediated by sodium taurocholate-cotransporting protein. Several organic anion-transporting proteins (OATPs), including OATP1, OATP2, and OATP4, are also involved in sodium-independent bile salt uptake. At the canalicular surface, bile salt secretion into the bile duct is mediated by members of the ATP binding cassette (ABC) superfamily. An ABC transporter hydrolyzes intracellular ATP to transport biliary components against the concentration gradient into the bile. The canalicular bile salt export pump (BSEP or ABCB11) secretes monovalent bile salts, whereas multidrug resistance-associated protein-2 (MRP2 or ABCC2) secretes divalent bile salts. Once secreted into the bile, bile salts stimulate the secretion of phospholipids and cholesterol from the canalicular membrane, forming micelles. Multidrug resistance-3 (MDR3 or ABCB4 in humans or MDR2 in rodents) is a phospholipid transporter. Secretion of intact cholesterol into bile is mediated by a heterodimer of two ABC transporters, ABCG5 and ABCG8 (107, 108). These transporters are transcriptionally regulated by a variety of nuclear hormone receptors (106).
LPS and cytokines decrease the catabolism and excretion of cholesterol. In the liver, LPS markedly decreases the mRNA expression and activity of CYP7A1, the rate-limiting enzyme in the classic pathway of bile acid synthesis (109) (Fig. 3) . This effect is very rapid, occurring within 90 min of LPS administration, and is sustained for at least 16 h (109). LPS also decreases the mRNA expression and activity of CYP27A1, the rate-limiting enzyme in the alternative pathway of bile acid synthesis, and mRNA levels of CYP7B1 in the liver (110) (Fig. 3). The decreases in CYP27A1 and CYP7B1 occur 816 h after LPS administration and persist for at least 24 h, suggesting that both the classic and alternative pathways of bile acid synthesis are sequentially downregulated during infection and inflammation. Like LPS, both TNF and IL-1 also decrease hepatic CYP27A1 and CYP7B1 mRNA expression (110).
|
The coordinated downregulation of both pathways of bile acid synthesis during the APR is in contrast to most other situations, including studies in knockout animals, in which during the suppression or absence of one pathway of bile acid synthesis the enzymes of the other pathway are upregulated to compensate for the deficiency. The decreases in the regulatory enzymes of both the classic and alternative pathways of bile acid synthesis as well as the decrease in ABCG5 and ABCG8 induced by LPS and cytokines suggest that during infection the body's need to conserve cholesterol is so essential that all of these pathways are downregulated to limit the elimination of cholesterol from the body. A decrease in cholesterol catabolism would make cholesterol more available for hepatic lipoprotein production.
Lipoprotein [a] Lipoprotein [a] (Lp[a]) is a distinct lipoprotein consisting of an LDL particle attached to apo[a] that is present in primates but not in rodents and most other species (118). Lp[a] is cholesterol-rich; increased serum levels have been associated with a higher risk for atherosclerosis. The physiological role of Lp[a] is not known, but it is thought to be involved in wound healing. The structure of apo[a] resembles plasminogen, and apo[a] has been found in the lesions during early stages of wound healing. Alternatively, Lp[a] may act as a scavenger of oxidized lipids, as Lp[a] contains platelet-activating factor acetylhydrolase (PAF-AH) (119), an enzyme that inactivates PAF and oxidized lipids.
Whether Lp[a] is an acute-phase reactant is unclear. Some studies showed that levels of Lp[a] are increased during stress (120, 121), whereas others reported no changes or a reduction (122, 123). These conflicting data may be attributable to the specificity of the assays used to measure Lp[a] levels or to interindividual variation in plasma Lp[a] levels in the population.
HDL metabolism and decreased RCT During infection and inflammation, there is a marked decrease in serum levels of HDL and apoA-I (16, 17, 27, 124). Furthermore, circulating HDL during infection, known as acute-phase HDL, has different characteristics from normal HDL. Acute-phase HDL is larger than normal HDL3, its radius extending into the HDL2 range, but it has a density comparable to that of HDL3 (125). Acute-phase HDL is depleted in cholesteryl ester but enriched in free cholesterol, TG, and free FAs (24, 25, 27, 125127). The phospholipid content of acute-phase HDL was increased in some studies (24, 27) but decreased in others (124, 125). In patients who underwent bypass surgery, acute-phase HDL had the same phospholipid-neutral lipid ratio, a decrease in phosphatidylethanolamine and phosphatidylinositol, and an increase in isoprostane-containing phosphatidylcholine and lysophosphatidylcholine (LPC) (127). In humans, there was a decrease in HDL sphingomyelin content (127), but an increase was observed in hamsters (128).
The hallmark of acute-phase HDL is an increase in apoSAA (24, 124, 125, 129, 130) and a decrease in apoA-I content (24, 124, 127, 130) (Table 3). The content of apoA-II and apoCs is decreased (24, 124, 130, 131), whereas apoE is found to be increased in some studies (24, 132) but decreased in others (130). HDL-associated apoJ is increased during inflammation and infection in rodents and humans (133135). In contrast, several other proteins, including LCAT (24, 25, 136), cholesteryl ester transfer protein (CETP) (137, 138), hepatic lipase (HL) (139), and paraoxonase 1 (PON1) (134, 140), are decreased during the APR. The activity of HDL-associated plasma PAF-AH is acutely increased during inflammation in several rodent species (141), but a late decrease has also been reported in rabbits and mice (134, 135). Phospholipid transfer protein (PLTP) is decreased in rats injected with LPS (142), but data in humans are conflicting (132, 143). Finally, secretory phospholipase A2 (sPLA2), a phospholipase enzyme that hydrolyzes phospholipids in HDL, and LPS-binding protein (LBP) are markedly induced during infection and inflammation (144). SAA-rich HDL particles that are devoid of apoA-I have also been reported (145). We recently found that apoA-IV and apoA-V levels are increased in acute-phase HDL (our unpublished observations).
|
An increase in sPLA2 has also been proposed to contribute to the reduction in HDL during infection/inflammation. Mice overexpressing sPLA2 have reduced HDL concentrations (150), and HDL from these mice is catabolized more rapidly than HDL from normal mice (151). Although apoSAA is known to activate sPLA2, overexpression of SAA in addition to sPLA2 does not cause a greater reduction in the levels of HDL or apoA-I (152), further suggesting that the reduction of HDL during infection is not caused by an increase in apoSAA.
Endothelial lipase (EL) has been shown to regulate HDL metabolism (153155). EL is synthesized by the endothelial cells and possesses phospholipase A-I activity. Overexpression of EL reduces HDL cholesterol levels (153), whereas inhibition of EL increases HDL levels (156). Treatment of cultured endothelial cells with TNF-
or IL-1ß has been shown to increase the expression of EL (157). If similar effects occur in vivo, it may provide another mechanism for the reduction in HDL levels during infection.
The decrease in LCAT activity during infection may decrease HDL cholesterol levels caused by impaired esterification, similar to what is found in humans or animals with mutations in the LCAT gene (158). The decrease in HL may reduce pre-ß HDL generation. Moreover, TG enrichment of HDL during infection may lead to the rapid clearance of apoA-I (159). Which of these changes contributes to the reduction of HDL and apoA-I during the APR is not yet established, but none accounts for the early decrease.
HDL metabolism is tightly linked to RCT, a process by which cholesterol is removed from peripheral cells and transported to the liver for metabolism and/or excretion (160, 161). Several HDL-associated proteins and a number of cell surface receptors play a key role in RCT (Fig. 4)
. ApoA-I on HDL and ABCA1 in the plasma membrane are required for apolipoprotein-mediated cholesterol efflux. Subsequently, LCAT, which converts free cholesterol on HDL into cholesteryl ester, assists in cholesterol efflux by an aqueous diffusion mechanism. CETP then mediates the exchange of cholesteryl ester in HDL for TG in TG-rich lipoproteins. PLTP transfers phospholipids from TG-rich lipoproteins into HDL and promotes the remodeling of HDL. HL hydrolyzes TG and phospholipids in large
-HDL, generating small pre-ß HDL particles that are efficient acceptors of cholesterol from plasma membrane. In the liver, scavenger receptor class B type I (SR-BI) plays a key role in the selective uptake of cholesteryl ester, whereas the ß-chain of ATP synthase mediates endocytosis of HDL particles.
|
Sphingolipid metabolism
Sphingolipids such as ceramide and sphingomyelin are important constituents of plasma membranes. Glycosphingolipids (GSLs) are complex sphingolipids that contain a hydrophobic ceramide moiety and a hydrophilic oligosaccharide residue. Both sphingolipids and GSLs are components of plasma lipoproteins and are involved in several biological processes, including cell recognition and proliferation, signal transduction, interaction with bacterial toxins, and modulation of the immune response.
The metabolism of sphingolipids and GSLs is altered during infection and inflammation. LPS stimulates hepatic ceramide and sphingomyelin synthesis by increasing the mRNA expression and activity of serine palmitoyltransferase (SPT), the first and rate-limiting enzyme in sphingolipid synthesis that catalyzes the condensation of serine with palmitoyl-CoA (128) (Fig. 5) . LPS increases the transcription rate, mRNA expression, and activity of glucosylceramide (GlcCer) synthase, the first committed enzyme in the GSL synthesis pathway, in the liver (165). GlcCer is the precursor of all neutral GSLs as well as sialic acid-containing acidic GSLs or gangliosides. The LPS-induced increase in GlcCer expression occurs earlier than the increase in SPT mRNA levels. It is possible that the increase in hepatic GlcCer production during the APR is the primary event, which then signals for more substrate, resulting in the induction of SPT and subsequent increase in ceramide synthesis. This hypothesis is supported by the fact that steady-state levels of GlcCer and its distal metabolites, including ceramide trihexoside and ganglioside GM3, are increased in the liver after LPS treatment (165), whereas in contrast, the content of ceramide, the substrate for GlcCer synthesis, is decreased in the liver despite the increase in SPT (165). Like LPS, TNF and IL-1 also increase both SPT and GlcCer mRNA expression in the liver, suggesting that these cytokines mediate the LPS effect (128, 165).
|
LPS also upregulates the mRNA expression and activities of SPT and GlcCer synthase in extrahepatic tissues, including spleen and kidney (166). The content of ceramide in spleen or kidney, however, is not increased, suggesting that newly synthesized ceramide is used as a substrate to increase GlcCer synthesis (166). Specific GSLs are ligands for a T-cell receptor expressed on natural killer T-lymphocytes, and GSLs stimulate the proliferation of specific subsets of lymphocytes (167). One can speculate that the LPS-induced increase in GSL content of these tissues is used to regulate cellular proliferation and modulate the immune response.
In addition to activating the enzymes that synthesize sphingolipids and GSLs, LPS and cytokines also induce enzymes involved in the hydrolysis of sphingolipids (Fig. 5). Treatment with LPS, TNF, or IL-1 acutely increases the serum activity of secretory sphingomyelinase (168). Serum ceramide levels are increased in animals treated with LPS and in patients with sepsis (128, 169, 170). The APR also activates ceramide-metabolizing enzymes. IL-1 activates both neutral and acid ceramidases in cultured rat hepatocytes, resulting in increased formation of sphingosine (171), whereas in cultured endothelial cells, TNF induces sphingosine kinase activity and increases the formation of sphingosine-1-phosphate (172). These studies suggest that several enzymes involved either in the de novo synthesis of ceramide and its downstream metabolites or in the hydrolysis of ceramide are induced by LPS and cytokines. Because ceramide and its metabolites are involved in signal transduction and cellular regulation, particularly in cells of the immune system, it makes sense that several anabolic and catabolic pathways of sphingolipid metabolism are induced during infection and inflammation to maintain a delicate balance between ceramide and its metabolites in the cell. Figure 5 summarizes the effects of LPS and APR-inducing cytokines on sphingolipid and GSL metabolism.
| ROLE OF NUCLEAR HORMONE RECEPTORS IN THE REGULATION OF LIPID METABOLISM DURING INFECTION AND INFLAMMATION |
|---|
|
|
|---|
B (NF-
B) and nuclear factor interleukin-6 (NF-IL-6) mediates IL-1-stimulated increases in acute-phase protein transcription, whereas activation of NF-IL-6 and the Janus kinase-signal transducers and activators of transcription pathway mediates IL-6 family stimulation of acute-phase protein transcription (174). Much less is understood regarding the mechanism of the downregulation of transcription of negative acute-phase proteins during the APR, and many of the changes in lipid metabolism seen in infection and inflammation are mediated by decreases in proteins and their transcription (13). Nuclear hormone receptors are a large family of transcription factors, characterized by a central DNA binding domain that targets the receptor to specific DNA sequences (response elements) and a C-terminal portion that includes a ligand binding domain, which recognizes specific hormones, vitamins, drugs, or other lipophilic compounds (175178). Several nuclear hormone receptors, including the peroxisome proliferator-activated receptors (PPARs), liver X receptors (LXRs), and farnesoid X receptor (FXR), bind and are activated by lipids (176181). Furthermore, the increased activity of these receptors regulates the transcription of a large number of genes involved in multiple aspects of lipid and lipoprotein metabolism (182). Because of their abilities to sense intracellular lipid levels and orchestrate changes in lipid metabolism, these nuclear hormone receptors have been recognized as liposensors (178). Finally, these liposensors (PPARs, LXRs, and FXR) heterodimerize with retinoid X receptors (RXRs) for efficient gene regulation (175). As discussed in detail below, most of the genes of lipid metabolism that decrease during the APR are regulated by these liposensors and related transcription factors, and the downregulation of these liposensors plays a key role in those changes.
Regulation of liposensors during infection and inflammation
In hamsters and mice, LPS administration decreases both protein and mRNA levels of RXR-
, -ß, and -
in the liver (183) (Table 4). The decrease in RXR occurs rapidly (within 4 h) and is sustained. Administering TNF and IL-1 reproduces these LPS effects. Similar reductions in RXR isoforms are seen in Hep3B cells treated with TNF and IL-1 but not IL-6, indicating that the decreases are directly induced by the cytokines (M-S. Kim, J. K. Shigenaga, A. H. Moser et al., unpublished observations). Furthermore, LPS administration also significantly reduces the hepatic nuclear DNA-binding activity of RXR homodimers to an RXR response element (183).
|
and -
, LXR-
, FXR, pregnane X receptor (PXR), and constitutive androstane receptor (CAR) (183185). These decreases were associated with reductions in nuclear binding activity to a direct repeat-1 (DR-1) PPAR response element, a DR-4 LXR response element, and an inverted repeat-1 FXR response element (183, 184). In contrast, mRNA levels of PPAR-ß/
and LXR-ß were not significantly altered in the liver after LPS treatment.
In adipose tissue, PPAR-
levels decrease after the administration of LPS or TNF (186) (Table 4). Treatment of adipocytes in vitro with TNF, IFN-
, and IL-11 decreases mRNA levels of PPAR-
(187191). The effect of LPS and cytokines on RXR isoforms and other liposensors in adipose tissue remains to be determined. In cardiac muscle, our laboratory recently reported that LPS administration decreases RXR-
, -ß, and -
and PPAR-
and -ß/
expression (192) (Table 4). To our knowledge, studies of the effect of inflammation and infection on the expression of RXR, PPAR, and other liposensors in skeletal muscle have not been carried out. Lastly, although the levels of liposensors are regulated in tissues that play a major role in the alterations of lipid metabolism during the APR, recent studies by our laboratory have shown that changes in the levels of RXR, PPARs, and LXRs were not found in the small intestine, an organ in which lipid metabolism is not significantly altered during infection and inflammation (117). Thus, liposensor levels specifically change in the tissues that exhibit changes in lipid metabolism during the APR.
Consequences of decreased expression of liposensors
Although it is likely that many factors influence the diverse changes in lipid and lipoprotein metabolism that occur in response to infection/inflammation, alterations in the activity of nuclear hormone receptor liposensors are likely to play a pivotal role in the coordinated regulation of FA and cholesterol metabolism that occurs during the APR, as can be seen by examining the effects on genes that liposensors are known to regulate.
FA and TG metabolism
As discussed earlier, infection/inflammation is characterized by an increase in lipolysis and a decrease in FA oxidation in adipose tissue, contributing to hypertriglyceridemia (26). PPAR-
has been shown to directly regulate genes that promote the storage of fat in adipose tissue, including adipocyte P2, LPL, FATP, CD36/FAT, and ACS (179, 193, 194). As discussed above, during infection and inflammation the expression of these genes is decreased, and it is likely that the reduction in PPAR-
activity in adipose tissue contributes to the changes in these proteins that would reduce fat storage and enhance lipolysis.
Likewise, downregulation of RXR-
, -ß, and -
and PPAR-
and -ß/
in cardiac muscle would be expected to reduce FA oxidation. Activation of PPAR-
and -ß/
induces the expression of many key enzymes required for FA oxidation, including LPL, FATP, CD36/FAT, heart-FABP (H-FABP), CPT-Iß, and ACS (179, 195198). One can postulate that a reduction in PPAR-
and -ß/
activity in the heart during the APR contributes to the decreased expression of these genes (67, 68, 71, 199) (Table 4). In skeletal muscle, there is also a decrease in FA oxidation, which is associated with a decrease in LPL, FATP, CD36/FAT, H-FABP, and ACS (67, 68, 71, 200). Whether levels of RXR-
, -ß, and -
and PPAR-
and -ß/
change in skeletal muscle during the APR remains to be determined.
Downregulation of RXR-
, -ß, and -
and PPAR-
and -
in the liver during the APR could also reduce hepatic FA oxidation, as a number of key PPAR-regulated proteins required for FA oxidation are decreased, including FATP, CD36/FAT, liver-FABP, and CPT-I
(ACS is decreased in mitochondria but not in endoplasmic reticulum) (67, 68, 70, 71) (Table 4). In contrast, many proteins involved in the reesterification of FA and the secretion of VLDL from the liver are not decreased and are not regulated by the PPARs.
Decreased hepatic FXR activity could also contribute to the increase in serum TGs during infection (184). FXR-deficient mice have increased serum TG levels (201), and FXR has been shown to regulate the hepatic expression of apoC-II and apoE (202, 203), both of which are decreased during the APR (85, 184).
Regulation of gene transcription is complex, involving multiple transcription factors. Therefore, changes in PPARs, FXR, and RXR are unlikely to be the only transcription factors that regulate the genes of interest during the APR. For example, Berg, Calnek, and Grinnell (204, 205) have shown that IL-1- and IL-6-induced decreases in apoE mRNA levels in HepG2 cells are associated with the phosphorylation of BK virus enhancer factor-1, a member of the NF-1 family of nuclear factors, to its isoform B1. An increase in B1 is associated, by unknown mechanisms, with decreases in apoE mRNA levels (205). Thus, an increase in the B1 isoform coupled with the reduction in FXR activity during infection and inflammation may together result in the decrease in apoE expression. Likewise, we recently found that PPAR-
coactivating factor-1 (PGC-1), which interacts with PPAR-
, PPAR-
, hepatocyte nuclear factor-4 (HNF-4), and other nuclear hormone receptors, is reduced during the APR (M-S. Kim, J. K. Shigenaga, A. H. Moser, et al., unpublished observations).
Thus, decreases in RXR, PPARs, LXR, and related transcription factors in adipose tissue, muscle, and liver could be mechanisms by which the characteristic changes in TG and FA metabolism that occur during infection and inflammation are induced.
RCT RCT is a complex process that involves transporters in peripheral tissues and liver, enzymes and transfer proteins in the serum, receptors in the liver, the synthesis of bile acids in the liver, and the secretion of cholesterol and bile acids into the bile (160, 161). Many of the proteins essential for RCT are regulated by liposensors (181), whose changes could mediate the reduction in RCT that occurs during infection and inflammation.
Peripheral tissues
ABCA1 transporters play a dominant role in the movement of cholesterol from cells to HDL and are regulated by LXR (206208). Treatment of macrophages with LPS or cytokines decreases ABCA1 (117, 162). However, no change in RXR or LXR that could account for the reduction in ABCA1 expression was found in macrophages (117, 209) (Table 5). Recently, bacterial infection was found to activate Toll-like receptor 4, inhibiting the induction of LXR target genes, including ABCA1 (209). This cross-talk between LXR and Toll-like receptor signaling decreases cholesterol efflux from macrophages (209). In addition, LPS-induced decreases in the expression of CYP27A1 would decrease the production of 27-hydroxycholesterol, a likely endogenous ligand of LXR, further explaining the effect of LPS on LXR target genes (110).
|
Receptors in the liver
SR-BI mediates the selective uptake of cholesteryl esters from HDL into the liver (213). PPARs and FXR regulate the expression of SR-BI (214, 215). Therefore, the decrease in PPAR/RXR and FXR/RXR activity in the liver could mediate the decrease in SR-BI expression during the APR (Table 5).
Hepatic synthesis of bile acids
CYP7A1 is the key rate-limiting enzyme in the neutral bile acid synthetic pathway (105). CYP7A1 is regulated by LXR in rodents and FXR in rodents and humans (216219). In contrast, increases in FXR/RXR activation reduce CYP7A1 activity by increasing small heterodimer partner (SHP), which in turn blocks the ability of the transcription factor liver receptor homolog-1 (LRH-1), to stimulate CYP7A1 expression (220).
During infection/inflammation, RXR, LXR, FXR, SHP, LRH-1, and other transcription factors decrease, with the net result being a decrease in CYP7A1 activity, despite the decreases in FXR and SHP (183, 184). There are several possible explanations for the decrease in CYP7A1 activity during the APR. First, LXR/RXR activation may be a dominant factor in regulating the transcription of CYP7A1 (221); hence, the reduction in LXR/RXR activity may result in decreased CYP7A1 expression. Second, the decrease in FXR/RXR activity and SHP may not be crucial in the complex setting of inflammation; their decrease would normally result in an increase in the activity of LRH-1, but during the APR LRH-1 is independently reduced (184), thereby decreasing CYP7A1. Lastly, expression of CYP7A1 is regulated by a number of other transcription factors, such as HNF-4, PXR, and thyroid hormone receptor (TR) (222); our laboratory and others have shown that these transcription factors are also downregulated during the APR (185, 223, 224). Thus, multiple factors may produce the decrease in CYP7A1 expression (Table 5).
Sterol 12
-hydroxylase (CYP8B1) is an e