The Role of Liver X Receptor-α in the Fatty Acid Regulation of Hepatic Gene Expression*

Liver X receptors (LXR) α and β play an important role in regulating the expression of genes involved in hepatic bile and fatty acid synthesis, glucose metabolism, as well as sterol efflux. Studies with human embryonic kidney 293 cells indicate that unsaturated fatty acids interfere with oxysterols binding to LXR and antagonize oxysterol-induced LXRα activity. In this report, we evaluated the effects of unsaturated fatty acids on LXR-regulated hepatic gene expression. The LXR agonist, T1317, induced mRNAs encoding sterol regulatory element-binding protein 1c (SREBP-1c) and two SREBP-1c-regulated lipogenic genes, e.g. fatty-acid synthase and the S14 protein in primary hepatocytes. Treatment of hepatocytes with eicosapentaenoic acid (20:5n-3) suppressed these mRNAs in the absence and presence of T1317. The cis-regulatory elements targeted by T1317 were not required for fatty-acid suppression of FAS or S14 promoter activity. In contrast to SREBP-1-regulated lipogenic genes, 20:5n-3 had no effect on the T1317 induction of ABCG5 or ABCG8 in the rat hepatoma cell line, FTO-2B. These two genes require LXR but not SREBP-1c for their expression. Feeding rats a diet supplemented with fish oil suppressed hepatic SREBP-1c-regulated genes and induced PPARα-regulated genes but had no effect on the LXR-regulated transcripts, CYP7A1, ABCG5, or ABCG8. Transfection studies, using either full-length hLXRα or a chimera containing only the LXRα ligand binding domain, indicate that a wide array of unsaturated fatty acids had little effect on LXRα activity in primary hepatocytes or FTO-2B. These studies suggest that LXRα is not a target for unsaturated fatty acid regulation in primary rat hepatocytes or in liver. Thus, oxysterol/LXR-mediated regulation of transcripts involved in bile acid synthesis or sterol efflux appear insensitive to dietary unsaturated fatty acids. The unsaturated fatty acid suppression of SREBP-1 and its targeted lipogenic genes is independent of LXRα

Liver X receptors (LXR) 1 ␣ and ␤ are ligand-regulated nuclear receptors that play an important role in hepatic bile acid and fatty acid synthesis, glucose metabolism, and sterol efflux (1)(2)(3). Oxysterols, like 22(R)-hydroxycholesterol and 24,25-epoxycholesterol, bind to and activate liver X receptors. Together with RXR, LXR bind DNA regulatory elements, i.e. DR4, and induce the transcription of multiple genes involved in bile acid synthesis (CYP7A), sterol efflux (ABCA1, ABCG5, and ABCG8), glucose metabolism, and de novo lipogenesis. Recent studies with human embryonic kidney (HEK) 293 cells suggest that unsaturated fatty acids bind to LXR␣ (K d ϳ 1-4 M) and antagonize oxysterol activation of the LXR␣ but not LXR␤ (4,5). Unsaturated fatty acids have also been reported to interfere with LXR/RXR binding to DNA regulatory elements (6). Fatty acid interference with oxysterol regulation of LXR has important physiological implications because of the potential to affect the expression of genes involved in bile acid, fatty acid, cholesterol, and glucose metabolism.
The effect of LXR on lipogenesis involves both direct and indirect mechanisms. LXR/RXR heterodimers bind lipogenic gene promoters, e.g. fatty acid synthase (FAS), or regulate lipogenic gene expression by controlling levels of SREBP-1c (2,3). SREBP-1c is a basic helix-loop-helix-leucine zipper transcription factor that is translated as an ϳ125-kDa precursor (pSREBP-1c) attached to the endoplasmic reticulum (8,9). After proteolytic processing in the Golgi, the active form, nSREBP-1c (ϳ65 kDa), accumulates in the nucleus where it binds sterol regulatory elements (SRE) in promoters of many genes involved in fatty acid and triglyceride synthesis. Transcription of the SREBP-1c gene is induced by insulin (10,11) and oxysterols through LXR (12). Insulin induction of LXR␣ gene transcription might also account for some of this control (13). Much of insulin action on lipogenic gene transcription has been ascribed to the insulin-mediated induction of SREBP-1c (9,14). Unsaturated fatty acid suppression of nuclear SREBP-1c levels is complex. In established cell lines, unsaturated fatty acids inhibit transcription of the SREBP-1 gene (15), enhance mRNA SREBP-1 turnover (16), and interfere with proteolytic processing of SREBPs (17). In primary hepatocytes, unsaturated fatty acids have little impact on SREBP-1c gene transcription (18,19) but enhance mRNA SREBP-1c turnover (16). Over-expression of nSREBP-1c in primary hepatocytes eliminates the polyunsaturated fatty acid (PUFA) effects on several lipogenic genes, indicating that SREBP-1c is a key target for PUFA action on de novo lipogenesis (19,20).
Because LXR␣ is a target for fatty acid inhibition in HEK293 cells, we were interested in evaluating the role LXR␣ played in the fatty acid regulation of hepatic gene expression. Our studies show that under conditions sufficient to suppress SREBP-1c and lipogenic gene expression, certain LXR-regulated transcripts are resistant to PUFA regulation.

MATERIALS AND METHODS
Animals-Male Sprague-Dawley rats (Charles River, Kalamazoo, MI) were maintained ad libitum on a Teklad chow diet. For feeding studies, rats were meal-fed a high carbohydrate diet supplemented with olive oil at 10% w/w for 10 days. The high carbohydrate fat-free diet (glucose replaces sucrose) was obtained from ICN Biomedicals, Inc. (Aurora, OH). Meal feeding involved allowing the rats to eat between 8 a.m. and noon daily. Animals were either maintained on this diet or switched to a high carbohydrate diet supplemented with fish oil (10%, w/w) (Dyets, Bethlehem, PA). After 5 days on the olive oil or fish oil diets, animals were euthanized ϳ2 h after completing the meal, and livers were removed for RNA analysis (19).
Primary Hepatocytes and Transfections-Male Sprague-Dawley rats maintained ad libitum on a Teklad chow diet were used for primary hepatocyte preparation (19). For RNA analysis, cells were plated onto 100-mm type I collagen-coated plates or Primaria plastic dishes (BD Biosciences) at 10 7 cells/plate in Williams E with 10 mM lactate, 10 nM dexamethasone, 100 nM insulin, and 10% fetal calf serum. For transfection studies, cells were plated in the same media onto 6-well type I collagen-coated plates or Primaria plastic dishes at 10 6 cells/well. Transfection conditions have been described previously (19). The ratio of culture medium to cell number was maintained as a constant for the different plating conditions. After a 4-to 6-h attachment period, medium was changed to a serum-free medium, Williams E with 10 mM lactate, 10 nM dexamethasone, and 100 nM insulin. Cells were transfected in this serum-free medium using Lipofectin or LipofectAMINE 2000 (Invitrogen), as described (5). The medium was changed the next morning to Williams E with 25 mM glucose, 10 nM dexamethasone, 100 nM insulin, 250 M fatty acid (NuChek Prep, Elysian, MN), and 50 M bovine serum albumin, very low endotoxin and fatty-acid free (Serological Proteins, Inc., Kankakee, IL), or drug treatment (TO-901317 (T1317), Cayman Chemicals, Ann Arbor, MI).
FTO-2B cells were obtained from L. Reid (University of North Carolina, Chapel Hill, NC) (21) and maintained in Dulbecco's modified Eagle's medium/F12 (Invitrogen) plus 7.5% fetal bovine serum on plastic culture dishes. HEK293 cells were described previously (5). Cells were grown in 6-well plastic culture plates for transfection and 100-mm plastic culture plates for RNA studies. At confluence, cells were transfected in serum-free medium as described above. For RNA studies, cells were incubated overnight in serum-free medium. At the time of treatment, FTO-2B cells received 100 nM insulin and 10 nM dexamethasone along with fatty acids and/or the LXR-agonist, T1317. Cells were transfected using LipofectAMINE 2000 or used for RNA extraction as described above.
RNA Isolation and Northern Analysis-RNA was extracted from rat liver, primary hepatocytes, or FTO-2B cells using Triazol (Invitrogen) (5). RNA was separated electrophoretically in denaturating agarose gels, transferred to nitrocellulose, and probed with [ 32 P]cDNAs. Levels of hybridization were quantified by using a Storm 820 PhosphorImager (Amersham Biosciences).
Proteins (50 -100 g) were separated electrophoretically by SDSpolyacrylamide gel electrophoresis (NuPage 4 -10% polyacrylamide-bistris, Invitrogen) and transferred to nitrocellulose membranes. Membranes were incubated with antibodies for SREBP-1c (IgG-2A4) obtained from the supernatant of the hybridoma cell line CRL 2121 (American Type Culture Collection, Manassas, VA). The anti-mouse secondary antibody was obtained from Bio-Rad. The detection system employed the SuperSignal West Pico chemiluminescence kit (Pierce).

Effects of Eicosapentaenoic Acid (20:5n-3) and T1317 on SREBP-1c Expression in Primary
Hepatocytes-The LXR agonist, TO901317 (T1317) was used in all studies with primary hepatocytes and FTO-2B cells. The native oxysterols, 22(R)hydroxycholesterol and 24,25-epoxycholesterol, have no effect on LXR activity or LXR-regulated transcripts in these cells, probably because of their rapid metabolism. T1317 at 5 M was selected as the minimal dose to achieve reliable induction of Results are expressed as relative mRNA abundance, and the mRNA levels in the treated cells are normalized to the level of the corresponding RNA in a rat liver standard from a chow-fed male rat. These results are the mean Ϯ S.D.; n ϭ 9 of three separate studies. Lower panel, cells were treated as above and extracted for microsomal and nuclear proteins (see "Materials and Methods") for immunoblotting. The precursor form of SREBP-1 (pSREBP-1) is found in the microsomal fraction, whereas the nuclear form (nSREBP-1) is recovered in the nuclear fraction. The figure is representative of at least three separate studies. Analysis of the nuclear extracts for HNF-4␣ revealed robust levels of HNF-4 with no consistent treatment effect (not shown).
LXR activity in primary hepatocytes and FTO-2B cells. In HEK293 cells, T1317-induced LXR␣ activity is attenuated by both 20:4n-6 and 20:5n-3, whereas PPAR␣ activity was induced by these fatty acids. These effects on LXR␣ and PPAR␣ activity correlated with significant changes in intracellular non-esterified unsaturated fatty acids (5). In the studies reported here, 20:5n-3 was used as the fatty acid for most studies with primary hepatocytes and FTO-2B cells. 20:5n-3 is a very low abundance unsaturated fatty acid in primary hepatocytes and FTO-2B cells. Its addition at 250 M leads to a Ͼ10-fold increase in the mass of 20:5n-3 in the intracellular non-esterified fatty acid pool of these cells. Changes in 20:5n-3 in the NEFA pool correlate with the suppression of SREBP-1c mRNA, as well as the activation of PPAR␣ and PPAR␣-regulated genes (31).
Treating primary hepatocytes with the LXR agonist, T1317 induced mRNA SREBP-1c as well as the precursor (pSREBP-1) and nuclear (nSREBP-1) forms by 3-to 4-fold (Fig. 1). Treating cells with 20:5n-3 in the presence or absence of T1317 suppressed mRNA SREBP-1c , pSREBP-1, and nSREBP-1 by 50 -80%. This same pattern of control is seen for the 20:5n-3 and T1317 regulation of mRNAs encoding FAS and the S14 protein as well as reporter genes driven by the FAS and S14 promoters (Fig. 2, A and B). Whereas L-pyruvate kinase (LPK) mRNA and LPKCAT reporter gene were suppressed by 20:5n-3 treatment, T1317 did not induce this gene.
SREBP-1 binds the S14 promoter at the SRE located at Ϫ139 to Ϫ129 bp; it is near a Y-box (Ϫ104 to Ϫ99 bp) that binds nuclear factor-y (19,28,32). Together, these elements play an obligatory role in S14 gene transcription. Mutation of either element essentially abrogates S14 gene transcription (28,32). The SRE/nuclear factor-y elements are in the S14 PUFA-RR, a region previously identified as the principal target for PUFA suppression of S14 gene transcription (33). The FAS promoter has distinct regulatory elements for both SREBP and LXR (3). Our goal was to determine whether the T1317 effect on S14 promoter activity is simply caused by T1317/LXR-mediated induction of SREBP-1c. Accordingly, a detail promoter analysis was performed to locate the cis-regulatory target for T1317.
Deletion and mutation analyses show that the T1317-RR is located in a region previously identified as a glucose regulatory region (GlRR, Ϫ1.6 to Ϫ1.4 kb) (Fig. 3A) (34). The E-box in the GlRR binds glucose-regulated binding proteins as well as SREBP-1c (35,36). Deletion of the GlRR (Ϫ1.6 to Ϫ1.4 kb) or mutation of the E-box within (at Ϫ1440 bp) has no effect on PUFA suppression of S14CAT activity (Fig. 3B). In the context of the thymidine kinase promoter, neither the GlRR or the thyroid hormone regulatory region (TRR) are sensitive to 20: 5n-3 suppression. Preliminary studies suggest that at least two T1317 elements are located within the GlRR. Studies are underway to define these elements. A similar analysis using FAS-CAT reporter genes indicted that the FASCAT activity is suppressed 50% by 20:5n-3 treatment. Like the S14 promoter, the presence and absence of the LXRE at Ϫ669 bp (3) did not influence promoter sensitivity to 20:5n-3. 2 These studies show that although S14 and FAS have distinct cis-regulatory elements for T1317/LXR and SREBP-1c, the T1317/LXR-cis regulatory elements are not required for PUFA control of these transcripts in rat primary hepatocytes. Blots were probed with [ 32 P]cDNAs for FAS, S14, or LPK. Levels of hybridization were quantified by PhosphorImager analysis. Results are expressed as relative mRNA abundance; the mRNA levels in the treated cells are normalized to the level of the corresponding RNA in a rat liver standard from a chow-fed male rat. These results are the mean Ϯ S.D.; n ϭ 9 of three separate studies. B, primary hepatocytes were transfected with fatty acid synthase, S14, and L-pyruvate kinase reporter genes containing promoter elements fused to the CAT reporter gene. FASCAT contains Ϫ2363 to ϩ16 bp of the FAS promoter, S14CAT contains Ϫ2800 to ϩ19 bp of the S14 promoter, and LPKCAT contains Ϫ4300 to ϩ14 bp of the L-pyruvate kinase promoter. After transfection, hepatocytes were treated with 20:5n-3 and T1317 as described above. After a 24-h treatment, cells were harvested for CAT and protein assays. Results are reported as fold change in CAT activity. The results are the mean of three independent studies involving triplicate samples.
various lipogenic genes are easily examined in primary rat hepatocytes, certain LXR-regulated transcripts (e.g. CYP7A, ABCG5, and ABCG8) decrease to nearly undetectable levels when compared with their expression in liver. The reason for this decline is likely because of the loss of key transcription factors controlling expression of these transcripts. Fortunately, ABCG5 and ABCG8 are well expressed and regulated by T1317 in FTO-2B cells (25). Because the T1317 regulation of these transcripts is independent of SREBP-1c, we examined the effect of 20:5n-3 on the T1317-mediated induction of SREBP-1c, ABCG5, and ABCG8 in FTO-2B cells (Fig. 4).
Previous studies indicated that the predominant SREBP-1 subtype in FTO-2B cells is SREBP-1c (4). RNA protection studies indicate the ratio of SREBP-1c to -1a is Ն4-fold. 2 T1317 induced mRNA SREBP-1c as well as the precursor and nuclear forms of SREBP-1c 2-to 3-fold in FTO-2B cells (Fig. 4A). 20:5n-3 suppressed mRNA SREBP-1c and the nuclear form of SREBP-1c by 50 -70% in the absence and presence of T1317. The 20:5n-3 and T1317 regulation of SREBP-1c in FTO-2B cells is similar to that seen in primary hepatocytes (Fig. 1). Although T1317 induced mRNA ABCG5 and mRNA ABCG8 ϳ2-fold, 20:5n-3 had no effect on the level of these transcripts (Fig. 4B). This finding suggests that LXR, per se, is not a target for unsaturated fatty acid antagonism in FTO-2B cells.

20:5n-3 Regulation of LXR␣ Activity in Primary Hepatocytes and FTO-2B Hepatoma Cells-
In an effort to examine the effect of 20:5n-3 on LXR␣ activity directly, we used a transfection approach. Because our previous studies established that only LXR␣, and not LXR␤, was affected by unsaturated fatty acid treatment (5), our studies focused on LXR␣. Accordingly, primary hepatocytes were transfected with an expression vector containing the LXR␣ ligand-binding domain fused to the Gal-4 DNA-binding domain and the MH-TK-luciferase (LUC) reporter containing four Gal 4 regulatory elements (Fig. 5). The use of this chimeric receptor allows for an evaluation of fatty acid effects on LXR activity without the requirement for RXR heterodimerization. After treatment of primary hepatocytes with T1317, LXR activity was induced 4-fold. No significant fatty acid-mediated antagonism of LXR␣ activity was detected until 20:5n-3 levels reached 1 mM. At lower levels (0.25 mM), which are levels typically used to examine PUFA regulation of lipogenic gene transcription, 20:5n-3 augments LXR␣ activity FIG. 3. Role of the T1317 regulatory region in the PUFA control of S14 promoter activity. A, mapping the T1317 regulatory element in the S14 promoter. Primary hepatocytes were transfected with CAT reporter genes containing various elements from the S14 promoter. A schematic of the S14 promoter is illustrated at the top of the figure, and the composition of each promoter-reporter gene construct is illustrated in the figure. The TRR and the GlRR were excised from the S14 promoter and fused to the thymidine kinase (TK) promoter to yield TRR-TK-CAT and GlRR-TK-CAT, respectively. The reporter gene at the bottom of the figure has the GlRR (Ϫ1506 to Ϫ1408 bp) removed and replaced by an NsiI restriction site. After transfection, hepatocytes were treated with 5 M T1317, as described above. After a 24-h treatment, cells were harvested for CAT and protein assays. Results are reported as fold change in CAT activity. The results are the mean of three independent studies involving triplicate samples. T1317-RR, T1317-regulatory region; PUFA-RR, PUFA-regulatory region. B, the S14 T1317-regulatory region is not involved in PUFA suppression of S14CAT activity. Primary hepatocytes were transfected with CAT reporter genes containing various elements from the S14 promoter or S14 promoter elements fused to the thymidine kinase promoter. The composition of the promoter elements is illustrated in the figure. The S14 reporter constructs lacking the entire GlRR or the E-box in the GlRR are replaced by the NsiI restriction site. After transfection, hepatocytes were treated with 250 M 20: 5n-3, as described above. After a 24-h treatment, cells were harvested for CAT and protein assays. Results are reported as % inhibition of CAT activity by 20:5n-3. The results are the mean Ϯ S.D. of two independent studies involving triplicate samples.
by 30%. A similar study with FTO-2B cells revealed no evidence of 20:5n-3 interference with T1317-mediated induction of LXR␣ activity. In each case, fatty acid treatment had no effect on basal LUC activity in the absence of T1317. For comparison, a recent dose-response analysis of 20:4n-6 effects on LXR activity in HEK293 cells showed that the IC 50 for the antagonism of 20:4n-6 was 22 M 2 . Clearly, LXR␣ is considerably more sensitive to PUFA inhibition in HEK293 cells than in primary hepatocytes. More importantly, the inhibitory effect of 20:5n-3 on LXR␣ activity is only seen when very high non-physiological levels of the fatty acid are added. Such levels are likely never reached in vivo.
To determine whether other unsaturated fatty acids affect LXR␣ activity in primary hepatocytes, transfected cells were treated with various mono-and polyunsaturated fatty acids (at 250 M) in the absence and presence of T1317 (Fig. 6A). In no case did any fatty acid tested affect LXR␣ activity. Of the fatty acids tested, all but 18:1n-9 suppress SREBP-1c mRNA levels in primary hepatocytes (19).
Others reported that, in addition to fatty acid effects on oxysterol binding, unsaturated fatty acids interfered with LXR-RXR heterodimerization and DNA binding (6). To determine whether unsaturated fatty acids interfered with the activity of the full-length LXR␣, primary hepatocytes were transfected with a full-length LXR␣ expression vector and the LXRE-TK-Luc reporter vector (Fig. 6B). For the reporter plasmid to respond to T1317, endogenous RXR must be recruited to the promoter by LXR␣. Accordingly, treatment of primary hepatocytes with T1317 induced LUC activity 15-fold. Treatment of cells with 18:1n-9 or 20:5n-3 in the absence of T1317 had no effect on basal LUC activity (not shown). Cotreatment with T1317 and either 18:1n-9 and 20:5n-3 did not inhibit LUC activity. Thus, the capacity of the full-length LXR␣ to activated promoters containing LXRE is not impeded by the addition of unsaturated fatty acids to primary hepatocytes.
Effects of Fish Oil on LXR-regulated Gene Expression in Rat Liver-In an effort to assess the in vivo effects of fatty acids on LXR-regulated hepatic gene expression, we fed rats fish oil. Fish oil is enriched in 20:5n-3 and 22:6n-3. The addition of fish oil to diets has well established effects on PPAR␣ and SREBP-1c-regulated gene expression in rat liver (19,23). In this study, rats were meal-fed a high carbohydrate diet containing either olive or fish oil for 5 days (Fig. 7). Fish oil feeding induced hepatic mRNA levels for several PPAR␣ target genes, i.e. cy- tochrome P450-4A, mitochondrial HMG-CoA synthase, and acyl-CoA synthetase-1 Ն3-fold (Fig. 7), as well as acyl-CoA oxidase. 2 The mRNAs encoding SREBP-1c, FAS, S14, GPAT, and LPK were suppressed by 50 -80%. SREBP-1c regulates the transcription of FAS, S14, and GPAT but not LPK (9,19). Fish oil treatment had no effect on transcripts encoding PPAR␣ and HNF-4␣. More importantly, fish oil feeding did not affect hepatic levels of the LXR-regulated transcripts CYP7A1, ABCG5, or ABCG8. Reverse transcriptase-PCR analysis for ABC-A1 also shows this transcript is insensitive to fish oil treatment. 2 This in vivo analysis agrees favorably with our previous analysis of these transcripts in primary hepatocytes (Figs. 1 and 2) and FTO-2B hepatoma cells (Fig. 4). In vivo, n3-PUFA induced PPAR␣-regulated transcripts and suppressed SREBP-1c-regulated transcripts but had no effect on those LXR-regulated transcripts that do not require SREBP-1c. The in vivo studies corroborate the cell culture studies with primary hepatocytes and FTO-2B hepatoma cells. DISCUSSION LXR␣ and -␤ play a key role in regulating the transcription of multiple genes involved in bile acid and fatty acid synthesis, glucose metabolism, and sterol efflux (1,3,7,12,25,37,38). As such, these nuclear receptors influence a broad array of hepatic metabolic events that, in turn, influence whole body glucose, lipid, and cholesterol metabolism. Moreover, the liver is a major target for fatty acid-regulated gene expression (39). Finding that unsaturated fatty acids antagonized the activity of LXR␣, but not of LXR␤, suggested that the spheres of influence of unsaturated fatty acids may be extended potentially beyond their well known effects on the expression of genes involved in fatty acid synthesis and oxidation (39). In this report, we used several approaches to evaluate the fatty acid regulation of LXR in liver. The new information reported here includes: (i) the T1317-mediated induction of hepatic mRNA SREBP-1c , as well as the precursor and nuclear forms of SREBP-1, is suppressed by 20:5n-3 treatment (Fig. 1); (ii) the T1317 regulation of SREBP-1c nuclear content parallels its effects on FAS and S14 gene expression and promoter activity (Fig. 2); (iii) the T1317/ LXR regulated cis-acting elements in the S14 and FAS genes are not required for fatty acid regulation of these promoters (Fig. 3); (iv) the T1317 induction of LXR-regulated transcripts, i.e. CYP7A, ABCG5, ABCG8, and ABCA1, is not affected by unsaturated fatty acids in FTO-2B cells and rat liver ( Fig. 4  and 7); (v) hLXR␣ activity is antagonized by PUFA only at high, non-physiological levels that are Ͼ20-fold of that found for the PUFA antagonism of hLXR␣ in HEK293 cells (Fig. 5). Based on these findings, we conclude that LXR␣ is not a target for fatty acid antagonism in rat liver, rat primary hepatocytes, or rat FTO-2B hepatoma cells. Thus, unsaturated fatty acids activate PPAR␣ regulatory networks and suppress SREBP-1c regulatory networks, but do not impede hepatic LXR regulatory networks in primary hepatocytes or in vivo.
Because 20:5n-3 is a minor unsaturated fatty acid in the total lipid fraction and NEFA pool of hepatocytes, its addition to primary hepatocytes leads within 90 min to a Ͼ50and Ͼ10-fold increase in mass of this fatty acid in the total lipid and intracellular NEFA fractions, respectively (31). The level of 20:5n-3 in the NEFA pool is sustained for ϳ6 h, after which it declines to levels ϳ3-fold above basal values by 24 h. The change in cellular 20:5n-3 in the NEFA fraction correlates well with the activation of PPAR␣, the induction of PPAR␣-regulated transcripts (31), and the suppression of SREBP-1c mRNA. 2 The 250 M dose of 20:5n-3 used here is sufficient to suppress mRNAs for SREBP-1c and lipogenic genes and to induce PPAR␣ and PPAR␣-regulated transcripts but is insufficient to affect LXR␣ activity. Only very high, non-physiologic levels of 20:5n-3 inhibited LXR␣ activity (Fig. 5). Clearly, LXR␣ activity in primary hepatocytes is far less sensitive to PUFA than PPAR␣, SREBP-1c, or the PUFA regulation of LXR␣ in HEK293 cells (ED 50 Յ 50 M) (5). Finding that full-length LXR␣ was insensitive to PUFA treatment (Fig. 6B) argues against effects of PUFA on LXR/RXR heterodimerization and DNA binding. Interestingly, in HEK293 cells, LXR␣ and PPAR␣ are equally sensitive to PUFA action, suggesting that both receptors might serve as sensors to intracellular NEFA levels (5). Although PPAR␣ can be considered as a monitor of intracellular NEFA (31), the studies reported here indicate that LXR␣ is not responsive to major changes in intracellular NEFA in liver or primary hepatocytes.
The fact that we found no evidence for unsaturated fatty acid regulation of LXR activity in liver does not exclude LXR␣ from being sensitive to fatty acid control in another cell context. Clearly, fatty acids affect LXR␣ activity in HEK293 cells (4,5). Although the mechanism for this control has been ascribed to competitive binding, in vivo mechanisms may be more complicated involving receptor-coactivator interaction or the regulation of other signaling mechanism that do not exist in liver or primary hepatocytes, e.g. cyclooxygenase and lipoxygenase (29). Previous efforts to examine fatty acid effects on targets of LXR action, i.e. CYP7A, did not reveal specific effects of unsaturated fatty acids (40 -42). In fact, the effects of fatty acids on CYP7A have been attributed to PPAR␣ (42), not LXR. A recent report on the LXR-regulated gene, ABCA1, in macrophage suggests that fatty acid treatment inhibited cholesterol efflux by increasing ABCA1 protein degradation, without effects on ABCA1 mRNA (43). These observations have clinical significance because if fish oil interfered with bile acid synthesis or sterol efflux, then serum cholesterol levels would rise, significantly increasing the risk for atherosclerosis and coronary disease. If anything, fish oil or eicosa pentaenoic acid/docosahexaenoic acid treatment of rats and humans is cardioprotective and has pronounced hypolipemic effects (44,45).
Finally, we have identified a prospective cis-regulatory region for LXR action in the S14 promoter. This region, tentatively identified as a T1317-RR, is located in the glucose-regulatory region of the S14 promoter, located between Ϫ1.4 and Ϫ1.6 kb upstream from the transcription start site. This region contains an E-box that is known to bind glucose-regulated binding proteins as well as SREBP-1c (7,35). Additional studies will be required to identify the specific minimal cis-regulatory targets for T1317 action and whether LXR/RXR binds these elements. Despite this limitation, deletion of the entire GlRR eliminates T1317 regulation of S14 promoter activity but fails to impact PUFA control of S14 promoter activity (Fig. 3). Previous studies established that the PUFA-RR, binding both SREBP-1c and NF-Y, is indispensable for S14 gene transcription. Moreover, this region is the principal target for PUFA control of the S14 promoter activity (28,32,33). These findings indicate that the key elements involved in PUFA control of S14 are distinct from the GlRR containing T1317-regulated factors. The fatty acid synthase promoter contains an LXRE at Ϫ669 bp that is distinct from two SREs (Ϫ150 and Ϫ65 bp) (3). Analysis of FASCAT reporter genes indicates that the LXRE is not FIG. 6. Effect of unsaturated fatty acids on LXR␣ activity in primary hepatocytes. Primary hepatocytes were transfected with CMX-Gal4-hLXR␣ and the TK-MH100X4-Luc reporter (A) or CMX-hLXR␣ and the LXREx3-Luc reporter plasmid (B). After an overnight transfection, cells were treated without or with the LXR agonist 5 M T1317 and without and with various unsaturated fatty acids at 250 M for 24 h. BSA was included at a 5:1 ratio of fatty acid to BSA. After the 24-h treatment period, cells were lysed and assayed for luciferase activity. Results are expressed as relative luciferase activity and are representative of Ͼ2 separate studies. Mean ϩ S.D.; n ϭ 6. Male Sprague-Dawley rats were meal-fed high carbohydrate diets supplemented with 10% olive or 10% fish oil (w/w) for 5 days. RNA was extracted and assayed for various transcripts (see "Materials and Methods"). Results are expressed as fold change in mRNA. The expression of the various transcripts was normalized to olive oil-fed rats. Mean ϩ S.D., n ϭ 3. The results are representative of several studies. required for PUFA suppression of FASCAT activity. 2 Based on these studies, T1317/LXR-regulatory elements in the FAS and S14 promoters are dispensable for PUFA control of these promoters.
In summary, we have examined the role LXR␣ plays in PUFA regulation of hepatic gene expression. Although LXR␣ is clearly a target for PUFA antagonism in HEK293 cells (4,5), it is not a target for fatty acid regulation in rat liver, primary hepatocytes, or FTO-2B hepatoma cells. This conclusion is based on the lack of 20:5n-3 antagonism on LXR␣ activity or the LXR-regulated transcripts, ABCG5 and ABCG8, at doses sufficient to repress SREBP-1c mRNA or to induce PPAR␣ activity. The fact remains that LXR␣ activity is well antagonized by PUFA in HEK293 cells. This antagonism might be important for LXR action in some tissues. However, the absence of any effect on LXR in liver suggests that fatty acid effects on bile acid and fatty acid synthesis, sterol efflux, or glucose metabolism cannot be explained by abrogated LXR action.