Identification of PLTP as an LXR target gene and apoE as an FXR target gene reveals overlapping targets for the two nuclear receptors.

Affymetrix microarray data and Northern blot assays demonstrated that phospholipid transfer protein (PLTP) was induced 6-fold when either murine or human macrophages were incubated in the presence of ligands for the liver X receptor (LXR) and the retinoid X receptor. Two functional LXR response elements (LXREs) were identified and characterized in the proximal promoter of the human PLTP gene. One LXRE corresponds to a traditional direct repeat separated by 4 bp. However, the second LXRE is novel in that it corresponds to an inverted repeat separated by 1 bp, and is identical to the farnesoid X receptor response element. These studies demonstrate that PLTP is a direct target for activated LXR and farnesoid X receptor (FXR). In addition, apolipoprotein E (apoE), a known LXR target gene in macrophages, was shown to be activated in liver cells by FXR ligands. Taken together, the current data suggest that a small number of genes that currently include PLTP, apoE, and apoC-II, are induced in macrophages by activated LXR and in liver by activated FXR.

Phospholipid transfer protein (PLTP) is a member of the lipid transfer/lipopolysaccharide binding protein gene family that includes CETP, the lipopolysaccharide binding protein, and the bactericidal/permeability increasing protein (1). PLTP has broad substrate specificity and functions to transfer phospholipids from triglyceride rich particles (VLDL and chylomicrons) to HDL during lipoprotein lipolysis (2). PLTP mRNA is expressed in a number of tissues, including liver, ovary, thymus, and placenta (3). Recent studies have demonstrated that PLTP is a tar-get gene for activated farnesoid X receptor (FXR) (4,5). This conclusion was based on the findings that i ) FXR ligands induced PLTP mRNA levels in cultured human hepatocytes (4) and in the livers of wild type but not FXR null mice (6), and ii ) the proximal promoter of the human gene contains a functional FXR response element (FXRE) (4,5).
In the current report, we utilized murine and human macrophages to demonstrate that PLTP is also an liver X receptor (LXR) target gene. Since macrophages do not express FXR, the results demonstrate that PLTP can be regulated independently by either receptor. Apolipoprotein E (apoE) is a known LXR target gene in macrophages (7,8). We now demonstrate that the hepatic expression of apoE is induced by FXR ligands. Taken together, these studies provide evidence that a subset of genes, which currently includes PLTP, apoE, and apoC-II (4)(5)(6)(7)(8), are induced following activation of LXR or FXR and control lipid metabolism. 11). HUH7 cells were maintained in DMEM containing 10% calf serum, and were incubated for 24 h in DMEM containing 10% super stripped FBS prior to addition of ligands or vehicle.

Identification of PLTP as an LXR target gene in primary macrophages
Murine peritoneal macrophages were isolated from C57BL/6 mice 4 days after thioglycollate injection, allowed to adhere to tissue culture dishes for 15 h, and then incubated for an additional 36 h either in a cholesterolpoor medium or in medium supplemented with 1 M T0901317 and 100 nM LG, ligands that are specific for LXR and retinoid X receptor (RXR), respectively. RNA was isolated and processed for hybridization to Affymetrix microarrays (MG-U74Av2) as described (7).
The Affymetrix data suggested that PLTP mRNA levels were induced 7-fold in response to T0901317 and LG. Consequently, we isolated RNA from murine peritoneal macrophages that had been incubated in the presence of ligands for LXR (22( R )-hydroxycholesterol or T0901317), RXR (LG), or FXR. As shown in Fig. 1A , PLTP mRNA levels were induced р 7.6-fold in response to ligands for LXR and/or RXR. In contrast to the liver (4-6), macrophages do not express FXR (data not shown) and thus FXR ligands alone do not induce PLTP mRNA levels (data not shown). Figure 1B shows that the human PLTP gene was also induced when primary human macrophages were incubated with specific LXR/RXR ligands. PLTP mRNA levels were undetectable in sterol-deprived human macrophages (Fig. 1B, lane 1), were induced significantly in response to T0901317 (lanes 2-6), and were further increased in the presence of both LG and T0901317 (lanes 7-10). PLTP mRNA levels were induced within 7 h and remained elevated ( Ͼ 17-fold) after 60 h in the presence of both ligands (data not shown).

The proximal promoter of the human PLTP gene contains two functional LXREs
To date, all functional LXREs identified in LXR target genes correspond to a DR-4 (12,13). Analysis of the proximal promoter of the human PLTP gene, using both visual scanning and FINDPATTERNS algorithm from the Genetic Computing Group package, identified a DR-4 element ( Ϫ 2669 AGGTTActagAGTTCA Ϫ 2654). In order to determine whether the LXRE motif is important in the transcriptional activation of the human PLTP gene, we generated a number of human PLTP promoter-luciferase constructs ( Fig. 2A ). HepG2 cells were transiently transfected with the indicated reporter plasmids in the presence or absence of the plasmids encoding LXR and RXR, and the cells were then incubated for 24 h in the presence of specific ligands. Figure 2A shows that pPLTP2.8 ( Ϫ 2776/ ϩ 86) was activated by ligands for LXR and/or RXR in the presence of both nuclear receptors. However, we consistently noted that a shorter construct (pPLTP0.5), which does not contain the DR-4, was modestly activated (Fig. 2A). These data suggest that there may be an additional LXRE within the Ϫ 539/ ϩ 86 region. The only hexanucleotide consensus sequence present within this region corresponds to an IR-1 ( Ϫ 249 GGGTCAgTGACCC Ϫ 236) that is known to function as an FXRE (4,5). In order to investigate the importance of the DR-4 and IR-1 motifs, named LXRE1 and LXRE2 respectively, we introduced a 4-bp mutation into each of the response elements. The data of Fig. 2B show that the wild type promoter was induced 2.3-fold by T0901317 in the presence of LXR/RXR, and that the induction was attenuated, but not abolished, when mutations were incorporated into either LXRE1 or LXRE2. In contrast, mutation of both response elements (pPLTP2.8-mutLXRE1 ϩ 2) completely abolished the induction by T0901317 (Fig. 2B).
In order to demonstrate that the IR-1 element functions as an LXRE, we transfected cells with a TK-reporter gene under the control of two copies of the IR-1 from the PLTP promoter. This reporter gene was activated 5.3-fold by T0901317 in the presence of LXR/RXR, while the empty TK-luciferase construct was unresponsive (Fig. 2C). Taken together, these data identify human PLTP as a direct target gene of LXR and suggest that both the LXRE1 and LXRE2 are necessary for maximal induction in response to activated LXR/RXR. Since LXR ligands also ac- tivate murine PLTP (Fig. 1) (14), it is likely that the murine gene is also a direct LXR target, although to date a functional LXRE has not been identified.

LXR/RXR binds to LXRE1 and LXRE2
EMSAs utilized radiolabeled probes corresponding to wild-type LXREs or mutated LXREs (containing the 4-bp mutation utilized in the promoter-luciferase constructs) and nuclear receptors derived from in vitro transcription/ translation reactions. The data indicate that LXR and RXR bind to both the DR-4 (LXRE1) and IR-1 (LXRE2) ( Fig. 3A , lanes 4 and 14). Antibody to RXR had two effects: i ) the total amount of radiolabeled probe that was shifted decreased significantly, consistent with the antibody interfering with the complex formation and ii ) a minor supershifted band was observed (Fig. 3A, lanes 5 and  15). In contrast, no protein/DNA complex was observed when the probe was incubated with either LXR or RXR alone (lanes 2, 3, 12, and 13) or when the probe con- Competition studies demonstrate that the formation of the LXR/RXR complex with either LXRE1 or LXRE2 is reduced in the presence of excess unlabelled DNA containing sequences corresponding to the LXRE1 and the DR-4 of rat Cyp7A1 (Fig. 3B, left panel), or LXRE2 and the IR-1 of I-BABP (Fig. 3B, right panel), respectively. In contrast, the formation of the LXR/RXR/DNA complexes was unaffected when the competitor DNA contained four mutations in the LXRE ( Fig. 3B , lanes 7-10 and 21-24).   4. Induction of apolipoprotein E mRNA in response to farnesoid X receptor ligands. HuH7 or HepG2-FXR (6) liver cells were maintained in DMEM ϩ 10% super stripped FBS and treated for 24 h with either vehicle or ligands for FXR (GW4064 or chenodeoxycholic acid) and/or RXR (LG). Northern blots were performed as described (6). Together with the transfection data, these studies demonstrate that PLTP is an LXR-target gene that contains two functional LXREs (a DR-4 and an IR-1) in the proximal promoter. To our knowledge, an IR-1 has not previously been shown to function as an LXRE and transactivate a target gene. Since the IR-1 can also function as an FXRE (4, 5), we conclude that either FXR/RXR or LXR/RXR can bind to this sequence. Additional studies will be required to determine what parameters regulate the preferential binding of distinct heterodimers to the same element in the liver.

Identification of apoE as a new FXR target gene
The findings that PLTP is transcriptionally activated in macrophages by LXR (Fig. 1) and in liver by FXR (4,5,15) and that the apoC-II gene is induced in macrophages by LXR (7) and in liver by FXR (6) suggest that there may be other genes involved in lipid metabolism that are also activated by both of these nuclear receptors.
The data of Fig. 4 demonstrate that apoE, a known LXR target gene in macrophages (8), is induced in liver cells by FXR ligands (CDCA and GW). ApoE is a member of a gene cluster (apoE/C-I/C-IV/C-II) that is known to be regulated in the liver by two distal enhancers, termed hepatic control regions (16)(17)(18)(19). We recently demonstrated that FXREs, contained within these enhancers, regulate the expression of hepatic apoC-II in response to activated FXR (6). Consequently, we hypothesize that the same FXREs are involved in the induction of hepatic apoE mRNAs in response to FXR ligands (Fig. 4). Since apoE is induced by ligands for FXR (Fig. 4) and LXR (8), these data suggest that a subset of genes, currently limited to apoE, PLTP, and apoC-II, can each be independently regulated by either FXR (Fig. 4) or LXR, thus providing alternative mechanisms that control lipid homeostasis in different tissues.
The function of macrophage derived PLTP remains to be established. Our studies suggest that PLTP expression and secretion would be highly induced as macrophages convert to lipid-loaded foam cells in the fatty streak of the artery wall. Since treatment of macrophages with LXR ligands also activates LPL (20), its obligate cofactor apoC-II, as well as apo E, C-I, and C-IV (7), we hypothesize that this milieu of proteins will alter lipoprotein metabolism in the artery wall and possibly affect the development of the atherosclerotic plaque.