Regulation of Lipoprotein Lipase by the Oxysterol Receptors, LXRα and LXRβ*

Lipoprotein lipase (LPL) is a key enzyme for lipoprotein metabolism and is responsible for hydrolysis of triglycerides in circulating lipoproteins, releasing free fatty acids to peripheral tissues. In liver, LPL is also believed to promote uptake of high density lipoprotein (HDL)-cholesterol and thereby facilitate reverse cholesterol transport. In this study we show that the Lpl gene is a direct target of the oxysterol liver X receptor, LXRα. Mice fed diets containing high cholesterol or an LXR-selective agonist exhibited a significant increase in LPL expression in the liver and macrophages, but not in other tissues (e.g. adipose and muscle). Studies inLxr-deficient mice confirmed that this response was dependent more on the presence of LXRα than LXRβ. Analysis of theLpl gene revealed the presence of a functional DR4 LXR response element in the intronic region between exons 1 and 2. This response element directly binds rexinoid receptor (RXR)/LXR heterodimers and is sufficient for rexinoid- and LXR agonist-induced transcription of the Lpl gene. Together, these studies further distinguish the roles of LXRα and β and support a growing body of evidence that LXRs function as key regulators of lipid metabolism and are anti-atherogenic.

enzymatic activity, LPL has been proposed to be a bridging factor that facilitates the uptake of HDL-associated cholesterol esters into the liver (3).
LPL is a member of a family of lipase enzymes that also includes hepatic lipase and pancreatic lipase (4,5). Although hepatic lipase and pancreatic lipase expression are limited to liver and pancreas, respectively, LPL is expressed at high levels in heart, adipose tissue, skeletal muscle, kidney, and mammary gland, and at lower levels in liver, adrenal, and brain (4, 6 -9). LPL activity is regulated by a variety of nutritional factors, as well as insulin, glucocorticoids, catecholamines, pro-inflammatory cytokines, and thiazolidinediones (10,11). Depending on the stimulus, LPL activity can also vary greatly between tissues. For example, in response to cytokines LPL expression is decreased in adipose tissue and increased in liver (12)(13)(14). LPL gene expression is also regulated by a number of transcription factors, including SP1, SREBP-1, and the peroxisome proliferator-activated receptors, PPAR␣ and PPAR␥ (11,(15)(16)(17)(18)(19)(20).
LXR␣ and LXR␤ are members of the nuclear hormone receptor superfamily that form heterodimers with RXRs and can be activated by both RXR and LXR ligands (21). RXR/LXR heterodimers activate their target genes by binding to specific response elements (termed LXREs) that contain a hexameric nucleotide direct repeat spaced by four bases (DR4). LXRs are activated by naturally occurring oxysterol ligands (22)(23)(24) and regulate the expression of a number of genes involved in cholesterol metabolism (24 -33). Studies in Lxr-deficient mice have helped elucidate the role of LXRs as one of the body's main sensors for regulating sterol homeostasis (25,27,34,35). Mice harboring Lxr-null mutations lack the ability to sense dietary cholesterol and as a consequence fail to regulate a number of lipid metabolic pathways, including cholesterol absorption and transport, and bile acid synthesis. Several lines of evidence suggest that, in addition to sterol metabolism, LXRs reciprocally regulate fatty acid metabolism. For example, in wild-type mice dietary cholesterol induces a marked increase in fatty acid synthesis and accumulation of hepatic triglycerides, whereas cholesterol levels are maintained at normal levels due to increased catabolism. This phenotype is reversed in Lxr␣ knockout animals, which accumulate high levels of cholesterol but show no increase in hepatic fatty acid synthesis and have normal triglyceride levels (25). Recently, we have shown that the cholesterol-induced elevation in fatty acid synthesis is due to the direct activation by LXRs of the gene encoding SREBP-1c (36,37). SREBP-1c is the primary transcription factor responsible for regulating fatty acid synthesis in the liver and peripheral tissues (16,38). We have suggested that the coordinate regulation of fatty acid and cholesterol metabolism by LXRs provides a means by which the body may protect itself from elevated levels of cholesterol. These findings prompted us to investigate other lipid-regulating genes as potential targets of LXR␣ and LXR␤ action.
In this work we show that cholesterol-induced LPL gene expression in the liver is directly regulated by RXR/LXR heterodimers in a tissue-specific manner and that in vivo this regulation is mediated predominantly by LXR␣. These studies define a new mechanism for governing tissue-specific LPL expression and further expand the role of the LXRs as key regulators of lipid metabolism.
Animal Studies-Animal experiments were approved by the Institution Animal Care and Research Advisory Committee of the University of Texas Southwestern Medical Center and were conducted as described previously (27). Mice were fed ad libitum Teklad 7001 rodent diet supplemented with cholesterol and receptor agonists or vehicle. The rexinoid LG268 was solubilized at 4.5 mg/ml in a vehicle containing 0.9% carboxymethylcellulose, 9% polyethylene glycol 400, and 0.05% Tween 80 and provided in the diet to give a final concentration of 30 mg/kg body weight. The LXR agonist T0901317 was solubilized in a vehicle containing 1% methylcellulose and 1% Tween 80 and administered by oral gavage or provided in the diet at a dose of 50 mg/kg body weight.
Northern Analysis-RNA was extracted and used in Northern analysis as described previously (25). For these experiments poly(A) ϩ RNA (5 g/lane) or total RNA (10 g/lane) was separated on 1% formaldehyde agarose gels, transferred to nylon membranes, and probed with 32 P-labeled human LPL or mouse actin or cyclophilin cDNAs. The human LPL cDNA probe has been described previously (40).
Macrophage Experiments-Peritoneal macrophages were obtained from thioglycolate-injected male mice of wild-type or Lxr␣/␤Ϫ/Ϫ genotype as described in a previous study (31). Cells were pooled from four wild-type or five Lxr␣/␤Ϫ/Ϫ mice and distributed on two plates, one for each treatment condition. Cells were allowed to adhere for 7 h in DMEM containing10% fetal bovine serum and penicillin/streptomycin. The medium was then replaced with DMEM supplemented with 10% lipoprotein-deficient serum, penicillin/streptomycin, and either vehicle (Me 2 SO) or 10 M LXR agonist (T0901317) and incubated for 42 h.
Cell Culture and Cotransfection Assays-The human embryonic kidney cell line, HEK 293, was maintained at 37°C, 5% CO 2 in DMEM containing 10% fetal bovine serum. Transfections were performed in 96-well plates in media containing 10% dextran-charcoal-stripped fetal bovine serum using the calcium phosphate coprecipitation technique (41). Ligands were added at final concentrations of 0.1 M LG268 and 5 M 22(R)-hydroxycholesterol. All transfection data were normalized using an internal ␤-galactosidase marker and represent the mean of triplicate assays.
Electrophoretic Mobility Shift Assay-Electrophoretic mobility shift assay was performed as described previously (42). Sequences of the Lpl DR4.1, DR4.2, and DR4.2m elements are shown in Figs. 4 and 5. Competitor DR4 containing the perfect tandem repeat of AGGTCA was described previously (42). After electrophoresis, the gel was dried at 80°C for 1.5 h and autoradiographed with intensifying screens at Ϫ80°C overnight.

Liver LPL Gene Expression Is Induced by Cholesterol and
Synthetic RXR/LXR Agonists-Previous work has shown that the hepatic accumulation of triglycerides in cholesterol-fed mice is dependent on the expression of LXR␣ (25). To explore the possibility that changes in hepatic LPL expression are associated with this response, wild-type, Lxr␣Ϫ/Ϫ, Lxr␤Ϫ/Ϫ, and Lxr␣/␤Ϫ/Ϫ mice were treated with various dietary regimens for 12 h or 7-10 days and then sacrificed, and hepatic LPL expression was examined by Northern analysis. On a low cholesterol diet in the absence of added agonists, LPL expression in the liver was barely detectable regardless of genotype (Vehicle lanes in Fig. 1). After feeding a diet rich in cholesterol for 7 days, which is known to generate endogenous oxysterol LXR ligands (43), LPL mRNA expression was increased 2.5-to 5-fold in wild-type mice (Fig. 1, A and B). In parallel experiments, treatment with a potent synthetic LXR agonist (27) resulted in a 12-to 17-fold increase in LPL mRNA in wild-type mice (Fig. 1, A and B). Elevation of LPL mRNA was also evident after just one 12-h treatment with LXR agonist (Fig.  1C), suggesting that Lpl is a direct LXR target gene. Treatment of wild-type mice for 10 days (Fig. 1C) or 12 h (Fig. 1D) with the RXR agonist LG268 (27) also resulted in activation of hepatic LPL gene expression, consistent with the notion that RXR works as a permissive heterodimer with LXR (42). Significantly, hepatic LPL expression induced by either LXR or RXR agonists was virtually absent in mice lacking both LXRs (Fig. 1, C and D). The RXR/LXR-induced expression of LPL was still observed in the Lxr␤ knockout mice (Fig. 1B) but was not present in the Lxr␣ knockout mice (Fig. 1A). These data indicate that LPL expression in the liver is regulated by the RXR/ LXR heterodimer. In addition, because the liver expresses both LXR␣ and ␤ subtypes, these data suggest that Lpl is predominantly an LXR␣ target gene. It is of interest that in these experiments high cholesterol diets did not induce LPL expression in either the Lxr␣ or ␤ knockout animals (Fig. 1, A and B), even though the potent synthetic agonist did. At present, the significance of this finding is unknown but could reflect differences in the pharmacology of the synthetic agonists versus endogenous ligands, or differences in the strain background of the Lxr␣ knockout (which is in an A129 background) versus Lxr␤ knockout (which is in a C57BL/6 and A129 mixed background) animals. In any case, these experiments show unequivocally that the induction of LPL expression requires the expression of at least one LXR subtype.
LXR-dependent Regulation of LPL Expression in Other Tissues-We next looked at regulation of LPL expression in other known LXR target tissues. Macrophages and adipose tissue express high levels of LXR proteins and are known to regulate several LXR target genes (27,31,36). LPL expression in peritoneal macrophages was stimulated 2-fold after a 42-h treatment with LXR agonist (Fig. 2A). Although the induction was not as high as that seen in liver (Fig. 1), the effect in macrophages was consistently reproducible (in four independent experiments). It is also of significance to note that the basal expression in macrophages was higher, which affected the fold induction but not the maximal level of expression induced by the LXR agonist. This induction was absent in macrophages isolated from Lxr␣/␤ double-knockout mice. In contrast, in adipose tissue, which expresses higher basal levels of LPL, the LXR agonist exhibited no significant effect on LPL expression in either mouse genotype after a 12-h or 10-day oral administration of the drug (Fig. 2, B and C, respectively). We also looked at expression of LPL in heart, muscle, kidney, intestine, and adrenals, tissues that are known to express LPL. Similar to the results found in adipose tissue, none of these other tissues exhibited LXR-dependent regulation of LPL (Fig. 2D). These data support the conclusion that LPL expression is differentially regulated by LXR in a tissue-specific manner.

Identification of a Functional RXR/LXR Binding Site in the
First Intron of the Mouse LPL Gene-The mouse Lpl gene promoter and partial coding sequences were obtained from GenBank. A computer-assisted search for potential LXR response elements was performed. The search revealed two DR4like sequences (Fig. 3A), one at Ϫ274 to Ϫ259 (DR4.1, 5Ј-TAAATCagtgTAAACC-3Ј) and another at ϩ635 to ϩ650 (DR4.2, 5Ј-TGACCGgtggTGACCT-3Ј). Electrophoresis mobility gel shift assays were performed to investigate the direct binding of receptors to each sequence. An oligonucleotide containing the DR4.2 sequence was radiolabeled and used in the experiments shown in Fig. 3 (B and C). The DR4.2 oligo produced a significant band shift when incubated with in vitro translated RXR/LXR␣ protein (Fig. 3B, lane 4) but not when incubated with either receptor alone (lanes 2 and 3). Binding of the RXR/LXR␣ heterodimer was completely inhibited by a 10-and 50-fold molar excess of either unlabeled DR4.2 (Fig. 3B, lanes 5  and 6) or the canonical DR4 LXRE (lanes 11 and 12) oligonucleotide identified previously (42). This inhibition was specific, because a mutated version of the DR4.2 element (see Fig. 5A for sequence) was unable to compete for binding, even at a 50-fold molar excess (Fig. 3B, lanes 7 and 8). In contrast to the DR4.2 element, the DR4.1 element was unable to compete for binding to the RXR/LXR␣ heterodimer (Fig. 3B, lanes 9 and 10), suggesting that this site does not function as an LXRE.
We also looked at the ability of LXR␤ to bind to the two sites. LXR␤ protein was radiolabeled by in vitro translation to a specific activity that was similar to LXR␣ and tested in the band-shift assay. In contrast to LXR␣ (Figs. 3B and 3C, lane 6), the RXR/LXR␤ heterodimer bound only weakly (ϳ10-fold less) to the DR4.2 oligonucleotide (Fig. 3C, lane 4). As expected, LXR␤ exhibited no binding to the DR4.1 site (data not shown). Similar results were seen using either human or mouse receptor proteins.
To test the ability of the DR4-like elements to function as LXR response elements, one copy of either the DR4.1 or DR4.2 elements was cloned into the luciferase reporter gene TK-Luc and cotransfected into HEK 293 cells with or without expression plasmids for mouse LXR␣, LXR␤, and RXR␣ receptors. After treatment with ethanol vehicle, rexinoid LG268, LXR ligand 22(R)-hydroxycholesterol, or both ligands, the cells were harvested and assayed for luciferase activity (Fig. 4, A and B). Consistent with the band-shift results in Fig. 3, only the DR4.2 element was able to mediate RXR/LXR-dependent transactivation of the reporter gene. In addition, LXR␣-mediated transcription was significantly higher than that of LXR␤, further supporting the notion that the Lpl gene is more selectively activated by LXR␣ than LXR␤.
The DR4.2 Sequence Is a Functional LXRE in the Full-length LPL Promoter-To demonstrate that the DR4.2 element iden-

FIG. 1. Liver LPL mRNA expression is induced by cholesterol and synthetic RXR/LXR agonists.
A, male A129 wild-type and Lxr␣Ϫ/Ϫ mice were fed chow diets containing vehicle (Veh) or 2% cholesterol (2% Chol). A third group of mice was fed a chow diet and gavaged once a day with 50 mg/kg body weight LXR agonist T0901317 (LXR Ag). On day 7, mRNA was isolated from livers of five different animals per group, and each mRNA preparation was pooled for Northern analysis. Results were quantified by phosphorimaging, standardized against actin, and mathematically adjusted to yield a unit of 1 for the wild-type group receiving the control diet. B, male mixedstrain (A129/C57BL/6) wild-type and Lxr␤Ϫ/Ϫ mice were treated the same as described in A. C, male mixed-strain (A129/C57BL/6) wild-type and Lxr␣/ ␤Ϫ/Ϫ mice were fed 0.2% cholesterol diets containing vehicle (Veh), 30 mg/kg body weight LG268 (RXR Ag), or 50 mg/kg body weight T0901317 (LXR Ag) for 10 days. Liver mRNA was isolated and pooled as above (n ϭ 6) for Northern analysis. D, male mixed-strain (A129/ C57BL/6) wild-type and Lxr␣/␤Ϫ/Ϫ mice were treated as in C for 12 h. Liver mRNA was isolated, and Northern analysis was performed as above on individual mice. Results shown are representative of three independent experiments. The mouse LPL mRNA migrates as a doublet of 3.4 and 3.6 kb (4, 50). tified above functions as an LXRE in the context of the Lpl gene promoter, the mouse Lpl gene from Ϫ289 to ϩ752 was cloned into the luciferase vector pGL2 and the resultant reporter gene (pLPL wt -Luc) was tested for LXR-dependent transactivation. This portion of the Lpl gene contains both the DR4.1 and DR4.2 sequences (Fig. 5A). A similar construct in which the DR4.2 site was mutated to destroy LXR binding (see Fig. 3B, lanes 7 and 8) was also tested. As shown in Fig. 5B, the wild-type promoter conferred significant RXR/LXR␣-dependent transactivation to the Lpl promoter, whereas the mutated version in which the DR4.2 site was altered (pLPL mut -Luc) did not. We conclude from these data that the Lpl promoter contains a functional LXRE located within the first intron of the gene. In a similar experiment LXR␤ was also able to mediate activation of the wild-type promoter, albeit at a significantly reduced level (data not shown). Taken together, these results support the conclusion that the Lpl gene is a direct target of LXR-mediated regulation by oxysterols. Furthermore, these data suggest that this regulation is governed primarily by LXR␣. DISCUSSION In this study we show that the mouse Lpl gene is a direct target of the oxysterol receptor, LXR. The activation of Lpl gene transcription by LXR is mediated via an LXRE in the first intron of the mouse gene. This LXRE is conserved in the human Lpl gene as well, indicating that LXR responsiveness is also conserved in humans. Furthermore, we show that LXR regulation of LPL expression is tissue-specific. LPL is markedly up-regulated by RXR/LXR agonists in liver and to a lesser extent in macrophages but not in adipose, muscle, kidney, adrenal, intestine, or heart.
One of the questions raised by these findings is what is the physiological basis of the tissue-specific regulation of LPL by LXRs. LPL is the rate-limiting enzyme that catalyzes the hydrolysis of lipoprotein triglycerides for uptake of fatty acids into adjacent tissues. Thus, in tissues such as adipose and muscle, LPL activity is required to meet the high demand these tissues have for free fatty acids. Despite the fact that LXRs are expressed abundantly in adipose and are known to regulate other FIG. 2. LPL mRNA expression in other tissues. A, peritoneal macrophages were isolated from male A129/C57BL/6 mice as described (31) and incubated with 10% lipoprotein-deficient serum in the presence of vehicle (Veh) or 10 M T0901317 (LXR Ag) for 42 h. Total RNA was isolated, and Northern analysis was performed as in Fig. 1. Results shown are representative of four independent experiments. B and C, white adipose tissue mRNA from male mixed strain mice treated for 12 h (B) or 10 days (C) was prepared as described in Fig. 1D and used for Northern analysis (n ϭ 3 for 12-h and n ϭ 5 for 10-day experiments). D, male A129 wild-type mice were fed chow diets containing vehicle (Veh) or 50 mg/kg body weight T0901317 (LXR Ag) for 7 days. Tissues were obtained and total RNA was isolated from 10 mice of each group. mRNA was pooled for Northern analysis as described in Fig. 1 (n ϭ 5 for each lane). Fold induction was calculated after standardizing to cyclophilin, using the average of both groups for each treatment.
adipose target genes, such as SREBP-1c, it is of significance that Lpl is not regulated by LXRs in this tissue. One explanation for this lack of regulation may be that subtle changes in the expression of LPL are masked by the already high basal level of LPL expression in adipose. However, it is worth noting that other transcription factors, in particular PPAR␥ (see below), are able to regulate Lpl in this tissue. A more likely explanation is that tissue-specific regulation of Lpl by LXRs in liver and macrophages is directly linked to the role LXRs play in maintaining whole body cholesterol balance.
One attractive hypothesis is that LXR up-regulates LPL to help the body clear the serum of cholesterol-rich lipoproteins (via macrophages) and transport this cholesterol (via HDL) back to the liver for catabolism and elimination. At present the function of LPL in macrophages and liver is controversial and just beginning to be elucidated. LPL has been suggested to have both pro-and anti-atherogenic properties in these tissues (44). For example, in macrophages LPL has been shown to facilitate uptake of cholesterol esters, presumably by remodel-ing triglyceride-rich chylomicrons and very low density lipoprotein into chylomicron remnants and low density lipoprotein that are then taken up by macrophages (45). Under pathogenic conditions, the excess cholesterol build-up in macrophages would lead to the development of foam cells. However, under non-pathogenic conditions, this cholesterol ester may be converted to oxysterols and free cholesterol that are effluxed out of the macrophage into HDL and transported to the liver. In the liver, LPL expression is normally undetectable in adult animals (9,46), but as shown by this study hepatic LPL expression is up-regulated markedly by high cholesterol diets and LXR agonists (Fig. 1). Several recent studies implicate the role of hepatic LPL expression as being anti-atherogenic by helping to facilitate reverse cholesterol transport through the HDL pathway. Strauss et al. (2) showed that adenoviral-mediated expression of LPL in Lpl-deficient mice is necessary and sufficient to promote maturation of HDL. It has also been shown that LPL can mediate the selective uptake of HDL-cholesterol into liver cells by a mechanism that evidently does not involve the enzy- matic activity of the lipase (3). Our observation that LPL is induced by LXR and its agonists and that cholesterol induction of LPL is LXR-dependent is consistent with the recent identification of other LXR target genes in macrophages and liver that are in the reverse cholesterol transport pathway. These genes include the ABC sterol transport proteins ABCA1, ABCG1, ABCG5, and ABCG8; apolipoprotein E; cholesterol ester transfer protein; and cholesterol 7␣-hydroxylase (24 -33). 2 Together these findings support a growing body of evidence that demonstrates LXRs are key sensors of sterol metabolism and maintain normal cholesterol balance by promoting sterol efflux from peripheral cells, increasing circulating HDLcholesterol, increasing hepatic sterol catabolism and excretion, and inhibiting further sterol absorption (25,27,34,35).
Another conclusion that can be extrapolated from this and other recent studies is that LPL expression is coordinately regulated by multiple dietary factors via nuclear receptors.
Here we have shown that sterol-mediated regulation of Lpl requires the LXRs. Previous work has shown that, in macrophages, high glucose induces LPL expression, in part through enhanced expression of PPAR␣ (20). Fatty acids and other PPAR␣ agonists also induce LPL expression in the liver through a pathway that may involve LXR activation (11,47). TNF has also been shown to enhance both fatty acid synthesis and LPL activity in liver (12,14). The mechanism of TNF stimulation in these cases is also believed to be indirect, suggesting that either PPAR or LXR may be modulating TNF induction of LPL. Finally, in adipocytes, LPL expression is specifically induced by PPAR␥ agonists (11), but not by LXR agonists (Fig. 2), even though LXR␣ is abundantly expressed in adipose tissue and can potently up-regulate other adipose target genes such as SREBP-1c (36). These data support the notion that the LXRs (oxysterol receptors) and PPARs (fatty acid receptors) define two distinct, but overlapping metabolic pathways that govern fatty acid metabolism in response to different dietary lipids (i.e. cholesterol and fatty acids).
A final intriguing observation from this work is the finding that LXR␣ is a more selective regulator of Lpl than LXR␤. In liver, LPL synthesis has been reported to be confined to Kupffer cells (48), although there is evidence that in newborn rats LPL is expressed in hepatocytes (49). Although it is possible that part of the LXR␣ selectivity for LPL expression is attributable to differential cell type expression of the two LXR genes, we note that the in vitro data on Lpl promoter binding and activation by the LXRs (Figs. 3 and 4) support the idea that LXR␣ has a higher affinity than LXR␤ for the Lpl LXRE. In either case, the identification of Lpl as an LXR␣-selective target gene is in keeping with the previous notion that LXR␣ and LXR␤ have distinct targets in vivo (25,34). Because LXR agonists have pharmacologic effects that are both desirable (e.g. increased reverse cholesterol transport) (27,35) and undesirable (e.g. hypertriglyceridemia) (37), the identification of specific LXR␣ or LXR␤ agonists or antagonists may have considerable therapeutic potential.