A Key Role for Orphan Nuclear Receptor Liver Receptor Homologue-1 in Activation of Fatty Acid Synthase Promoter by Liver X Receptor*

Liver X receptor (LXR) activates fatty acid synthase (FAS) gene expression through binding to a DR-4 element in the promoter. We show that a distinct nuclear receptor half-site 21 bases downstream of the DR-4 element is also critical for the response of FAS to LXR but is not involved in LXR binding to DNA. This half-site specifically binds liver receptor homologue-1 (LRH-1) in vitro and in vivo, and we show LRH-1 is required for maximal LXR responsiveness of the endogenous FAS gene as well as from promoter reporter constructs. We also demonstrate that LRH-1 stimulation of the FAS LXR response is blocked by the addition of small heterodimer partner (SHP) and that FAS mRNA is overexpressed in SHP knock-out animals, providing evidence that FAS is an in vivo target of SHP repression. Taken together, these findings identify the first direct lipogenic gene target of LRH-1/SHP repression and provide a mechanistic explanation for bile acid repression of FAS and lipogenesis recently reported by others.

Fatty acids subserve a large number of specialized cellular functions including cholesterol esterification, production of lung surfactant, mammary gland secretions, and signaling molecules. Fatty acids are also fundamental components of all biological membranes and the primary biochemical form of energy storage. Despite being a multistep enzymatic process, basic fatty acid biosynthesis in higher eukaryotes is accomplished through the catalytic activities of only two gene products, acetyl-CoA carboxylase (ACC) 3 and fatty acid synthase (FAS). ACC catalyzes the first step in which condensation of two acetyl-CoA molecules forms malonyl-CoA; FAS then catalyzes the remaining steps required to produce the fully saturated 16-carbon fatty acid palmitate.
Regulation of FAS by oxysterols was reported by Joseph et al. (6) who identified a DR-4-type nuclear receptor response element for the oxysterol-responsive RXR/LXR heterodimer in the FAS promoter. This DR-4 was identified in a region of the FAS promoter that was also shown to be required for TR regulation (13).
On further analysis of the FAS promoter, we identified an additional nuclear receptor half-site 21 bases downstream of the DR-4 element (Fig. 1). The region surrounding this putative nuclear receptor site did not suggest the existence of an additional DR-4-type element. However, because of its proximity to the DR-4, we hypothesized that it might play a role in the activation of FAS by either LXR or TR. In the study reported here, we show that the additional half-site binds the orphan nuclear receptor liver receptor homolog-1 (LRH-1)/fetoprotein transcription factor and that LRH-1 binding to this site specifically augments the LXR response. In contrast, LRH-1 has no effect on thyroid hormone activation of FAS.
Plasmids-The FAS-700/ϩ65 pGL2 luciferase reporter construct was subcloned by PCR from the rat FAS-1594/ϩ65 pGL2 and the FAS-700/ϩ65 MUT1 have been described previously (5,6). All other FAS mutant reporter constructs were generated using the QuikChange site-directed mutagenesis kit (Stratagene) by replacing the individual half-site sequences with an equal number of adenosine residues. The remaining reporter constructs were generated by PCR amplification and subcloning. pSynTATALuc is a reporter vector containing a minimal promoter region of the hydroxymethylglutaryl-CoA synthase promoter (Ϫ28/ϩ39) and has been described previously (19). The SREBP-1c promoter used here contains a sequence from Ϫ937/ϩ29 fused to the luciferase coding sequence in pGL2 basic. The following expression vectors were provided by other laboratories: CMX-hTR␤ (Barry Forman), CMX-mLXR␣ (Peter Tontonoz), CMX-hRXR␣ (Ron Evans and Bruce Blumberg), pCI-LRH-1 (Gregorio Gil), CMX-SHP (David Mangelsdorf).
Enzyme Assays-At the time of harvest cells were washed once with phosphate-buffered saline and then lysed in a reporter lysis buffer (25 mM Gly-Gly, 15 mM MgSO 4 , 4 mM EGTA, 0.25% Triton). Luciferase activity of the lysates was measured in an Analytical Luminescence MonoLight 2010 luminometer using 20 -40 l of cell extract plus 100 l of luciferase assay reagent (Promega) and expressed in relative light units. ␤-Galactosidase activity was measured by a standard colorimetric assay at 420 nM absorbance using 50 -100 l of cell lysate and 2-nitrophenyl ␤-galactopyranoside as the substrate. Luciferase activity for each sample was divided by the ␤-galactosidase activity to yield normalized relative light units. -Fold activation was determined by dividing the normalized relative light units for a given sample by the normalized relative light units for the control sample (no activators plus Me 2 SO). Each transfection was performed at least twice with similar results.
Electrophoretic Mobility Shift Assay-In vitro transcribed and translated LXR␣ and RXR␣ proteins were generated using the T7 TNT rabbit reticulocyte lysate (Promega). One or two microliters of each translation were added to each binding mixture (containing 10 mM Hepes, pH 7.6, 50 mM NaCl, 2.5 mM MgCl 2 , 0.1 g/l poly(dI:dC), 0.05% (v/v) Nonidet P-40, 10% glycerol) in a final volume of 20 l. For an electrophoretic mobility shift assay using GST-LRH-1, 20 ng of purified bacterially expressed GST-LRH-1 were added per reaction. A 57-base double-stranded oligo containing the wild type sequence of the three nuclear receptor half-sites was 5Ј endlabeled using T 4 polynucleotide kinase (U. S. Biochemical Corp.) and added to the binding mixtures (1-2 ng/reaction). Binding mixtures were incubated at 4°C for 1-2 h. Samples were then run on 5% polyacrylamide:bisacrylamide (19:1) gels at room temperature for 1.5 h, fixed in a solution of 10% methanol, 10% acetic acid, and dried onto Whatman No. 3MM chromatography paper at 80°C for 1 h. Dried gels were exposed to x-ray film for 12-48 h. DNA sequences (one strand from 5Ј-3Ј only) were as follows; note that mutant bases are underlined. DNA sequences at all three half-sites were: wild type, KM  Animal Studies-SHP knock-out mice described previously (20) and wild type C57B1/129sv mice were maintained on a normal chow diet. Adult animals were sacrificed, and liver RNA was extracted and analyzed for FAS expression by Q-PCR as described (21).
Chromatin Immunoprecipitation-B6/129 mice (6-week-old male) were purchased from Taconic and allowed to adapt for 2 weeks to a 12-h light/12-h dark cycle and sacrificed at the end of the dark cycle (8 a.m.). Livers from 4 mice were combined and placed in 40 ml of ice-cold phosphate-buffered saline containing a mixture of protease inhibitors (1 g/ml leupeptin, 1.4 g/ml pepstatin, and 2 g/ml phenylmethylsulfonyl fluoride) plus 1 mM EDTA and 1 mM EGTA. The tissue was disrupted in a Tekmar Tissumizer at the lowest setting. Formaldehyde was added from a 37% stock (v/v) to a final concentration of 1%, and samples were rotated on a shaker for 6 min followed by the addition of glycine to a final concentration of 0.125 M. Samples were returned to the shaker for an additional 5 min, and then cells were collected by centrifugation (2000 rpm in Sorvall RC3B) at 4°C. The cell pellet was washed once with homogenization buffer A (10 mM Hepes, pH 7.6), 25 mM KCl, 1 mM EDTA, 1 mM EGTA, 2 M sucrose, 10% glycerol, 0.15 mM spermine, plus protease inhibitors as above. The final pellet was resuspended in buffer A and homogenized in a Dounce homogenizer with a B pestle to release nuclei. The solution was layered over buffer A and centrifuged in a Beckman ultracentrifuge (1 h at 26,000 rpm, 4°C), and the nuclear pellet was resuspended in nuclei lysis buffer (1% SDS, 50 mM Tris, pH 7.6, 10 mM EDTA). Nuclei were disrupted using an Ultrasonic model W-220F sonicator for 5 ϫ 10 s to shear chromatin. Chromatin size was checked by agarose electrophoresis to ensure that the average size was between 200 and 500 bp. Aliquots were used in immunoprecipitation experiments with an antibody to LRH-1 (Santa Cruz; sc25389X) and processed for the rest of the chromatin immunoprecipitation protocol as described (22,23). Final DNA samples were analyzed by quantitative PCR in triplicate with a standard dilution curve of the input DNA performed in parallel. Oligo pairs for the FAS LRH-1 binding region or exon 4 from the YY1 gene used in the Q-PCR are as follows: FAS-700 (5Ј), ATCCTGGTCTCCAAGGTG; FAS-534 (3Ј), TAGGCAAT-AGGGTGATGGG; YY1, (5Ј) TCTGACGAGAGGATTGT-GTGGAC; YY1, (3Ј) CTGAAGGGCTTTTCTCCAGTATG.

RESULTS
In our analysis of the FAS promoter, we noted a highly conserved sequence 21 bases downstream of the RXR/LXR response element bearing a strong resemblance to other nuclear receptor half-sites (Fig. 1). To determine whether this sequence played a role in the FAS LXR response, we first compared LXR activation of the wild type FAS promoter to LXR activation of similar reporter constructs containing mutations in each of the three half-sites ( Fig. 2A). As expected, mutations of either half-site 1 or 2 (comprising the known LXR DR-4 element) resulted in a significant decrease in LXR activation. Surprisingly, however, mutation of the third half-site alone also resulted in a substantial defect in LXR activation, suggesting that the third half-site played an important and previously unappreciated role in LXR activation of FAS. Furthermore, FIGURE 1. DNA sequence alignment for the region from ؊669 to ؊615 (relative to the mRNA start site) for rat, chicken, and human FAS promoters. The three nuclear receptor half-sites in this region are capitalized and denoted by arrows. The DR-4 and putative LRH-1/fetoprotein transcription factor (FTF) sites are indicated.

FIGURE 2.
A, all three nuclear receptor half-sites of the FAS promoter are required for the maximal LXR response. HepG2 cells were transfected with CMX-LXR␣, CMX-RXR␣, and the wild type rat FAS (FASWT)(Ϫ700/ϩ65) luciferase reporter or mutant versions containing base-pair changes in the three nuclear receptor half-sites as indicated in the figure (denoted by ϫ's) and CMV ␤-galactosidase. Cells were treated with either Me 2 SO (as a vehicle control) or GW3965, a synthetic LXR agonist (1 M) for 48 h and then harvested and assayed for luciferase and ␤-galactosidase activity. B, LXR␣/RXR␣ preferentially bind the FAS DR-4 (half-sites 1 and2) and not half-site 3. In vitro translated LXR␣ and RXR␣ proteins were incubated with a 32 P-labeled oligonucleotide probe containing the rat FAS three half-sites and processed for an electrophoretic mobility shift assay as described under "Materials and Methods." In lanes 2-9, 200ϫ unlabeled competitor DNA, containing either the wild type sequence or altered versions containing mutations in the various half-sites, was added to the binding mixture as indicated.
mutation of all three half-sites simultaneously resulted in the most significant defect. Even with all three sites mutated there was still a 6-fold activation by LXR/RXR. This residual activity is due to LXR induction of SREBP-1c, which subsequently activates the FAS promoter through binding to sites in the proximal promoter (6).
To better understand the mechanism through which the third half-site exerts its effect on LXR activation of FAS, we examined binding of in vitro translated RXR␣ and LXR␣ proteins to a 32 P-labeled DNA probe containing all three nuclear receptor half-sites (Fig. 2B). The addition of both RXR␣ and LXR␣ proteins resulted in a single band (lane 1) that migrated in parallel with a known RXR␣/LXR␣ heterodimer on a consensus DR-4. The addition of a 200-fold excess of unlabeled wild type competitor DNA effectively competed for RXR␣/ LXR␣ heterodimer binding (lane 2); however, competitor DNAs containing mutations in the first and/or second halfsites (lanes 3, 4, 6 -9) were unable to compete significantly for heterodimer binding. In contrast, competitor DNA containing a mutation in the third half-site alone effectively competed for RXR␣/LXR␣ heterodimer binding (lane 5). Thus, the third half-site did not play a significant role in RXR␣/LXR␣ binding to this region of the promoter despite being an important determinant of the FAS LXR response.
Two other known LXR target genes, murine CYP7A1 (24) and human CETP (25), require binding of the monomeric nuclear receptor LRH-1 (also known as fetoprotein transcription factor and CYP7A1 promoter binding factor) at a site proximal the LXR element, for efficient LXR activation. To determine whether LRH-1 could bind to the FAS third half-site and potentiate the LXR response of the FAS gene, we first assessed binding of recombinant LRH-1 protein to the FAS third halfsite by electrophoretic mobility shift assay (Fig. 3A). Incubation of 20 ng of GST-LRH-1 with the labeled third half-site resulted in a single shifted band (lane 2) that migrated in parallel with GST-LRH-1 bound to a known LRH-1 binding site from the CYP8B1 promoter (lane 10). The addition of a 50-and 200-fold excess of unlabeled competitor DNA containing the wild type third half-site sequence effectively competed for GST-LRH-1 binding to the FAS half-site in a dose-dependent manner (lanes 3 and 4) as did competitor DNA containing the LRH-1 binding site from the CYP8B1 promoter (lanes 7 and 8). In contrast, the addition of competitor DNA containing a mutation in the third half-site did not compete for LRH-1 binding (lanes 5 and 6).
We next investigated direct binding of LRH-1 to the endogenous FAS promoter by chromatin immunoprecipitation in normal mouse liver using an antibody to LRH-1 and oligonucleotides that amplify the relevant region of the FAS promoter (Fig. 3B). Here, there was significant enrichment of the FAS promoter by the LRH-1 antibody relative to an IgG control, and this effect was specific as there was no enrichment of DNA from the YY1 locus analyzed in parallel.
To determine whether LRH-1 might activate the FAS promoter and augment the LXR response, we turned to HEK293T cells because HepG2 cells contain a significant level of endogenous LRH-1. To eliminate any indirect effects mediated through LXR activation of SREBP-1c (an LXR target gene and direct activator of FAS) (Fig. 2 and Refs. 5 and 6), a truncated wild type FAS promoter construct was constructed that contained the three nuclear receptor half-sites but lacked the downstream SREBP recognition motifs (FAS-700/-574) fused to a minimal TATA box containing promoter. This was transfected into HEK293T human embryonic kidney cells with an expression vector for LRH-1 along with the FAS promoter construct containing the three nuclear receptor half-sites (Fig. 4A). Because the RXR/LXR heterodimer is permissive and, thus, can be activated simultaneously by an RXR ligand and an LXR ligand (26), we compared activation of the FAS promoter by a synthetic RXR agonist (LG268), an oxysterol known to activate LXR (22-R-hydroxycholesterol) or both in combination. In the absence of transfected LRH-1, the addition of each ligand or the combination of both resulted in 2-5-fold increases in FAS activation, as is consistent with the FAS being an RXR/LXR target gene. In the presence of transfected LRH-1, however, this modest response increased to 20.5-fold, whereas LXR ligands or LRH-1 alone activated FAS 4.4-and 2.9-fold, respectively. Furthermore, the effect of LRH-1 on the LXR response was specific FIGURE 3. A, recombinant LRH-1 binds to half-site 3 in FAS promoter. Recombinant human LRH-1 was incubated with 32 P-labeled oligonucleotide probes containing the rat FAS half-site 3 (FAS 3) or a known LRH-1 binding site from the CYP8B1 promoter and processed for an electrophoretic mobility shift assay as described under "Materials and Methods." Unlabeled competitor (Com.) DNA was added at either a 50-or 100-fold molar excess as indicated above the corresponding lanes. WT corresponds to the same DNA as used for the 32 P labeling. M3 corresponds to the same DNA with a mutation in half-site 3, and CYP corresponds to the DNA corresponding to the LRH-1 site from the CYP8B1 promoter. B, LRH-1 binding to FAS promoter in liver chromatin by chromatin immunoprecipitation. Fresh mouse liver tissue was subjected to the chromatin immunoprecipitation protocol described under "Materials and Methods," and chromatin specifically precipitated by either an LRH-1 antibody or a control IgG fraction was analyzed for the presence of the FAS promoter or exon 4 from the YY1 gene as a negative control. Q-PCR reactions were performed in triplicate, and the data are plotted as -fold induced with respect to the signal from the IgG negative control. Error bar represents the mean Ϯ S.E. of triplicate reactions. JULY 13, 2007 • VOLUME 282 • NUMBER 28 to FAS, as no such effect was noted on a control promoter construct analyzed in parallel.

LRH-1 Augments LXR Activation of FAS
To better understand the contribution of each half-site to the LRH-1 potentiation effect, we used mutant versions of this FAS reporter construct containing either a double mutation of both halves of the DR-4 or a mutation in the putative LRH-1 (third) half-site (Fig. 4B). As in the previous experiment, LXR activation of the corresponding wild type FAS promoter was dramatically increased in the presence of LRH-1. Furthermore, the addition of an expression vector for SHP, a novel non-DNA binding nuclear receptor known to be a potent inhibitor of LRH-1 activity in other promoters, significantly blunted this response.
The FAS promoter construct containing a mutation in the DR-4 (MUT1/2) was defective for RXR/LXR signaling irrespective of the presence of LRH-1. Importantly, the magnitude of activation by LRH-1 alone on this mutant construct was similar to that seen on the wild type promoter, demonstrating an intact LRH-1 response.
Activation of the promoter construct containing a mutation in the third half-site (MUT 3) by RXR/LXR alone was comparable with that seen on the wild type construct; however, activation by LRH-1 alone was lost as was the concerted activation by both LXR and LRH-1. The residual inhibitory effect of SHP in the absence of the LRH-1 site was likely due to its known ability to inhibit LXR activation (27). Thus, SHP likely inhibits FAS gene expression through interactions with both LRH-1 and LXR.
We next used an siRNA approach to evaluate the effects of specifically reducing endogenous LRH-1 expression on the LXR-mediated induction of FAS in HepG2 cells (Fig. 5A). We used HepG2 cells because an siRNA approach has been used previously to silence LRH-1 gene expression (28). The siRNA specifically reduced endogenous LRH-1 mRNA by 30 -50%, which resulted in a significant blunting of the response of FAS to LXR activation. To rule out off-target effects of the siRNA,  1 g), CMV-␤-galactosidase, and luciferase reporters containing either the rat FAS promoter DNA from Ϫ700 to Ϫ574 fused to the minimal promoter in the pSynTATALuc vector or the minimal promoter vector pSynTATALuc itself as indicated and detailed under "Materials and Methods." Cells were treated with either Me 2 SO (DMSO, as a vehicle control), LG268 (100 nM) (a synthetic RXR agonist), 22-R-hydroxycholesterol (2.5 g/ml), or both (L ϩ 22R) in combination for 16 -18 h and then harvested and assayed for luciferase and ␤-galactosidase activity as described under "Materials and Methods." B, cells were transfected as in A with the following modifications. The rat FAS promoter or mutant versions containing base pair changes in the nuclear receptor half-sites were used as indicated. Additionally, TO901317, a synthetic agonist for LXR, was substituted for the 22-R-hydroxycholesterol as indicated (T), and the LG268 was also added as indicated (L). As noted, the expression vector for SHP was also included (50 ng). Wt, wild type. Error bars represent the mean Ϯ S.E. of triplicate reactions.

FIGURE 5. FAS mRNA expression is regulated by LRH-1 and SHP.
A, HepG2 cells were transfected with a control (N) or LRH-1 gene specific siRNA construct (L) and treated with or without the synthetic LXR agonist TO901317 as indicated. Cells were harvested for RNA extraction and analyzed for endogenous expression of either FAS or LRH-1 mRNA by Q-PCR as described under "Materials and Methods" (the significant difference for the effect of LXR ligand addition, noted by an asterisk, is associated with a p value of Ͻ0.03). B, hepatic FAS RNA from wild type (WT) or SHP knock-out mice fed a chow diet was measured by Q-PCR, and the values for three individual mice per group are presented (#, p ϭ 0.01). Error bars represent the mean Ϯ S.E. of triplicate reactions.
we also measured mRNA expression of LXR␣ in these cells. As expected, the addition of the siRNA did not decrease LXR␣ levels.
Although LRH-1 knock-out mice die early in embryogenesis and are, thus, unavailable for study (29), SHP knock-out mice are viable (20). Given that SHP is a potent inhibitor of LRH-1 (24,30,31), it follows that if FAS is indeed an LRH-1 target gene, FAS mRNA expression would be elevated in SHP knock-out mice. Consistent with this hypothesis, our animal studies showed that FAS mRNA expression in SHP knockout mice was 6.8-fold higher than in matched wild type controls (Fig. 5B).
In a previous report LRH-1 was shown to augment the LXR response of SREBP-1c (43). However, despite localization of the effect to the proximal 300 bases of the SREBP-1c promoter, an LRH-1 site could not be identified, and no binding studies were performed. To determine the effect of LRH-1 on LXR activation of SREBP-1c in our model system, we transfected HEK293T cells with an SREBP-1c promoter fragment encompassing the LXR element and the putative LRH-1 response region in the presence and absence of expression vector for LRH-1 and compared the LXR response with that of the FAS promoter (Fig. 6A). Both SREBP-1c and FAS showed a 3-fold induction by LXR ligand T0901317. However, the addition of increasing amounts of LRH-1 expression vector had no effect on the SREBP-1c LXR response, whereas FAS expression increased 10-fold. Also of note, no change in SREBP-1c activation was observed by LRH-1 alone, whereas FAS activation increased to 3-fold in the presence of transfected LRH-1 alone. Thus, in our system, LRH-1 does not play a direct role in the SREBP-1c LXR response.
The region of the FAS promoter containing the DR-4 has been previously shown to mediate the FAS response to TR (13). Because RXR/LXR and RXR/TR both preferentially bind DR-4 type nuclear receptor elements, we investigated whether LRH-1 binding to the third half-site also played a role in FAS activation by TR and T 3 . Using HEK293T cells transfected with expression vectors for TR␤, RXR␣, and LRH-1, we discovered that, although TR and T 3 clearly activate the FAS promoter, the FAS T 3 response was not enhanced by the addition of increasing amounts of an expression vector for LRH-1 (Fig. 6B). Thus, LRH-1 potentiation of FAS is specific to the RXR/LXR response.

DISCUSSION
Liver receptor homolog-1 was described as CYP7A1 promoter binding factor, a liver-enriched transcription factor essential for expression of the CYP7A1 gene (32). Interestingly, although a required component of CYP7A1 gene expression, LRH-1 had very little activation potential of its own, and it was subsequently shown to significantly augment the response of the murine CYP7A1 promoter to LXR/oxysterol transactivation (24,31). These studies also identified the LRH-1 central role in bile acid homeostasis by demonstrating that LRH-1 is a key target of SHP inhibition. LRH-1 is now recognized as an important component in negative feedback regulation of bile acid biosynthesis.
LXR␣ was also originally identified as a liver-enriched nuclear receptor (33,34). It is activated through the binding of oxysterol agonists (35) and is an important regulator of cholesterol homeostasis (36). When cellular cholesterol levels are elevated, oxysterol levels rise and serve as physiologic agonists for LXR (37). In response, the expression of genes involved in cholesterol storage as well as cholesterol metabolism and excretion are activated. Key LXR target genes involved in cholesterol storage include SREBP-1c (35) and FAS (6). LXR target genes involved in cholesterol metabolism and excretion include murine CYP7A1 (38) and ABCA1 (39), ABCG5 and ABCG8 (40), apoE (41), and human CETP (42).
Because the identification of LRH-1 as a key factor for the LXR/oxysterol response of murine CYP7A1, the LXR responses of two other genes have been shown to be augmented by LRH-1: human CETP (25) and mouse SREBP-1c (43). CETP plays an important role in balancing human lipoprotein metabolism, and SREBP-1c is a key transcription factor responsible for orchestrating the lipogenic gene program (44).
The LXR and LRH-1 sites are conserved in the human, mouse, and chicken FAS promoters (Fig. 1). In contrast, the LXR site is not present in the human CYP7A1 gene (45), and the mouse genome does not encode a functional CETP gene. Thus, FAS is unique in that activation by both LXR and LRH-1 is conserved across species.
In our studies the LXR response of the SREBP-1c promoter was not altered by LRH-1 (Fig. 6). The LRH-1-responsive elements in murine CYP7A1, human CETP, and human and murine FAS are within 50 bases of the LXR response element. The SREBP-1c promoter used here contains several hundred base pairs surrounding the LXR site. Thus, it is unlikely that the lack of response is due to not using a large enough promoter fragment in the assay. It is possible, however, that the lack of

LRH-1 Augments LXR Activation of FAS
LRH-1 responsiveness of the SREBP-1c promoter in our studies is due to the absence of some transcription factor in HEK293T cells that is present in the primary hepatocytes and cultured hepatoma cells used in the previous study. Alternatively, the effect on the SREBP-1c promoter might be indirect and/or require higher levels of LRH-1 than is required for the direct effect on the FAS promoter.
Why a distinct subset of LXR target genes depends upon LRH-1 for efficient activation in the presence of oxysterols is a compelling question. Because LRH-1 is selectively expressed in a handful of tissues, it has been postulated that the requirement for LRH-1 may allow for tissue specificity of the LXR/oxysterol response (24). In support of this hypothesis, CYP7A1, CETP, SREBP-1c, and FAS are all expressed at high levels in the liver but in a fairly restricted pattern in most other tissues. Interestingly, this group of LXR target genes is also sensitive to bile acid repression through SHP inhibition. Thus, regulation of a subset of LXR target genes involved in cholesterol metabolism by LRH-1 also provides an effective mechanism for selectively coupling their expression to bile acid homeostasis.
A recent report implicated FXR-mediated bile acid regulation of lipogenesis as the mechanism behind the triglyceridelowering effect of bile acids (43). In this study supplementing the chow diet of mice with cholic acid for 1 day resulted in a significant increase in SHP mRNA and a corresponding decrease in SREBP-1c mRNA levels (43). Expression of FAS (as well as several other lipogenic genes) also declined and was attributed to the decrease in SREBP-1c. They further analyzed the SREBP-1c promoter and noted a SHP-dependent decrease in promoter activation, thus providing a straightforward mechanism for bile acid repression of triglyceride synthesis.
In another study, Moschetta et al. (46) demonstrated that FAS expression is increased in FXR knock-out animals where expression of SHP is very low compared with wild type controls, suggesting as our data also does (Fig. 5) that FAS is under tonic SHP inhibition. Because SREBP-1c levels were increased in the FXR knock-out mice, the elevation of FAS was attributed to the indirect effect of elevated SREBP-1c. It is important to note that unlike FXR knock-out animals, chow-fed SHP knock-out animals have normal SREBP-1c levels (20) and yet have elevated FAS expression (Fig. 5). Because SREBP-1c levels were altered in both this study and by Moschetta et al. (46), it was not possible to determine whether there was also a direct effect of SHP on FAS transcription.
In our study we have identified an LRH-1 site in the FAS promoter that is involved both in potentiation of the LXR response as well as in transcriptional repression by SHP. Our data taken all together establishes FAS as the first direct lipogenic gene target for LRH-1 and reveals that SHP repression of lipogenesis occurs both directly at the FAS promoter and indirectly by modulating SREBP-1c levels (43).
When the synthetic LXR agonist T0901317 was first evaluated, it was shown to potently induce hepatic FAS mRNA and increase serum triglyceride levels when administered to mice over 3 days. This was recognized to be due to activation of FAS by direct effects of LXR. However, by 7 days of treatment with T0901317 in this study, both serum triglycerides and hepatic FAS (but not SREBP-1c) mRNA levels returned to pretreat-ment levels (6). Although a mechanism for this biphasic response was not understood at the time, our data identifying FAS as a gene target of LRH-1/SHP regulation now predicts this result. The addition of T0901317 would initially increase expression of LXR target genes including FAS, SREBP-1c, and CYP7A1. As a result, both triglyceride synthesis and bile acid synthesis (in rodents) are stimulated. When signaling through the LXR pathway is allowed to continue over 7 days, elevated flux of cholesterol into bile acids eventually triggers expression of SHP and other FXR target genes. In this case FAS would be subject to SHP repression, which would limit hepatic fatty acid biosynthesis and triglyceride accumulation. In this way, uncontrolled LXR-induced lipogenesis is effectively kept in check.
Interestingly, in a companion study, we showed that the nuclear receptor half-site in the FAS promoter that binds LRH-1 is the 5Ј half of a nuclear receptor IR-1 element that directly binds the FXR/RXR heterodimer and stimulates FAS in response to bile acids (23). Although the precise rationale for the overlapping nuclear receptor sites is impossible to determine definitively, it is interesting to note that, whereas LRH-1 binding to this site would in effect make FAS sensitive to bile acid-induced repression by SHP, FXR/RXR binding to this site would induce FAS expression under conditions of elevated bile acids. Ironically, the opposing effects of bile acids on FAS promoter activation may actually be adaptive. When cholesterol levels are elevated, oxysterol levels increase and result in elevated signaling through the LXR pathway. FAS expression increases under these conditions to facilitate esterification of cholesterol. In this way, the initiating stimulus (e.g. excess cholesterol) turns on the biochemical pathway that mitigates its toxic effects. When bile acid flux is acutely elevated, bile acidmediated up-regulation of SHP results in a drop in FAS (and CYP7A1) due to inhibition of LRH-1 even in the presence of activating oxysterol ligands for LXR. SHP inhibition of CYP7A1 blocks de novo biosynthesis of bile acids, thus preventing further bile acid accumulation. However, if bile acid flux remains elevated, chronic repression of bile acid biosynthesis eventually leads to accumulation of cholesterol. This prediction is supported by Watanabe et al. (43), who reported an increase in hepatic cholesterol levels in mice fed 0.5% cholic acid for 3 weeks. Although inhibition of bile acid biosynthesis is favorable for minimizing the toxic effects of bile acids, increased fatty acid biosynthesis is necessary under these conditions to esterify the excess cholesterol no longer permitted to enter the bile acid biosynthetic pathway. Thus, bile acid activation of FAS (via direct FXR/RXR binding to the IR-1) provides an SHP-independent mechanism for re-initiation of cholesterol esterification in the context of chronically elevated bile acids.