NFκB Regulates Plasma Apolipoprotein A-I and High Density Lipoprotein Cholesterol through Inhibition of Peroxisome Proliferator-activated Receptor α*

The levels of plasma HDL cholesterol and apoA-I in NFκB p50 subunit-deficient mice were significantly higher than those in wild-type mice under regular and high fat diets, without any significant difference in the level of total cholesterol. To examine the role of NFκBin lipid metabolism, we studied its effect on the regulation of apoA-I secretion from human hepatoma HepG2 cells. Lipopolysaccharide-induced activation of NFκB reduced the expression of apoA-I mRNA and protein, whereas adenovirus-mediated expression of IκBα super-repressor ameliorated the reduction. This IκBα-induced apoA-I increase was blocked by preincubation with MK886, a selective inhibitor of peroxisome proliferator-activated receptor α (PPARα), suggesting that NFκB inactivation induces apoA-I through activation of PPARα. To further support this idea, the expression of IκBα increased apoA-I promoter activity, and this increase was blocked by preincubation with MK886. Mutations in the putative PPARα-binding site in the apoA-I promoter or lack of the site abrogated these changes. Taking these results together, inhibition of NFκB increases apoA-I and HDL cholesterol through activation of PPARα in vivo and in vitro. Our data suggest a new aspect of lipid metabolism and may lead to a new paradigm for prevention and treatment of atherosclerotic disease.

The levels of plasma HDL cholesterol and apoA-I in NFB p50 subunit-deficient mice were significantly higher than those in wild-type mice under regular and high fat diets, without any significant difference in the level of total cholesterol. To examine the role of NFB in lipid metabolism, we studied its effect on the regulation of apoA-I secretion from human hepatoma HepG2 cells. Lipopolysaccharide-induced activation of NFB reduced the expression of apoA-I mRNA and protein, whereas adenovirus-mediated expression of IB␣ superrepressor ameliorated the reduction. This IB␣-induced apoA-I increase was blocked by preincubation with MK886, a selective inhibitor of peroxisome proliferatoractivated receptor ␣ (PPAR␣), suggesting that NFB inactivation induces apoA-I through activation of PPAR␣. To further support this idea, the expression of IB␣ increased apoA-I promoter activity, and this increase was blocked by preincubation with MK886. Mutations in the putative PPAR␣-binding site in the apoA-I promoter or lack of the site abrogated these changes. Taking these results together, inhibition of NFB increases apoA-I and HDL cholesterol through activation of PPAR␣ in vivo and in vitro. Our data suggest a new aspect of lipid metabolism and may lead to a new paradigm for prevention and treatment of atherosclerotic disease.
Nuclear Factor B (NFB) 1 transcription factors are homodimeric and heterodimeric complexes of five family members: p50, p52, c-Rel, RelB, and p65 (RelA). Most of the cells in the body contain only a heterodimeric complex of p50 and p65 (1,2). Activation of NFB is involved in the pathogenesis of many chronic inflammatory diseases, such as asthma and rheumatoid arthritis (3). It may also be involved in some neurodegenerative disorders such as Alzheimer's disease (4), ischemic brain injury (5), a variety of human cancers (6), and atherosclerosis (7).
Mice lacking the p65 subunit (RelA) display embryonic death at 15-16 days of gestation, concomitant with massive degeneration of the liver by apoptosis (8). To the contrary, mice lacking the p50 subunit of NFB show no developmental abnormalities and live for at least 1 year after birth but exhibit multifocal defects in immune responses involving B lymphocytes and nonspecific responses to infection (9). Although many researchers have demonstrated abnormalities in the immune responses of NFB p50 subunit-deficient mice, to our knowledge no one has reported any abnormality in lipid metabolism of them. In this study, we found that the plasma levels of high density lipoprotein (HDL) cholesterol and apolipoprotein A-I (apoA-I) in NFB p50 subunit-deficient mice were significantly higher than those in wild-type littermates under a regular or high fat diet. In addition, the level of apoA-I mRNA in the liver was significantly higher in NFB-deficient mice than in wild-type littermates (data not shown). The plasma levels of apoE and apoB did not appear to differ between these NFB-deficient and wild-type mice.
ApoA-I is mainly synthesized in the liver and small intestine (10). Because the levels of plasma apoA-I are positively correlated with hepatic apoA-I mRNA (11)(12)(13), factors influencing the level of apoA-I may be mediated through the level of apoA-I gene expression.
The distinct enhancer region in the human apoA-I gene contains the regulatory element necessary for maximal expression in a human hepatoma cell line, called peroxisome proliferator response element (PPRE) (14). In this cell line activated PPAR␣ binds to this element, thereby enhancing the expression of apoA-I (15)(16)(17). In this report we demonstrated that inactivation of NFB enhances the expression and secretion of apoA-I from HepG2 cells through activation of the transcription factor PPAR␣.

EXPERIMENTAL PROCEDURES
Reagents-Lipopolysaccharide (LPS) and WY-14643 (synthetic PPAR␣ agonist) were purchased from Sigma, and MK886 (selective inhibitor of PPAR␣) was from Wako Pure Chemicals (Osaka, Japan). LPS was dissolved in pure water, whereas WY-14643 and MK886 were dissolved in Me 2 SO. Goat polyclonal anti-apoA-I antibody was purchased from Rockland (Gilbertsville, PA), rabbit anti-apoB antibody was from BioDesign (Saco, ME), and goat anti-apoE antibody and mouse anti-␤-actin monoclonal antibody were from Chemicon International (Temecula, CA).
Animals-Animal experiments were performed in compliance with the Guide for Animal Experimentation and with the approval of the Committee of Animal Experimentation, Ehime University School of Medicine. Animals were maintained in a specific pathogen-free facility under a 12-h dark/light cycle. Homozygous NFB p50 subunit-deficient mice Nfkb1 tm1Bal , NFB Ϫ/Ϫ ) were originally purchased from The Jackson Laboratory (Bar Harbor, ME). The gene-targeting strategy used to generate this mouse line was described previously (9). These mice were back-crossed at least six times to C57BL/6 mice (Clea Japan Inc., Osaka, Japan). Genomic DNA was extracted from the mouse tail and genotyped by PCR following the protocol of The Jackson Laboratory.
Eight-week-old female mice were divided into two groups. Each group, composed of five NFB Ϫ/Ϫ mice and five control mice, was kept on a regular diet containing 5% (wt/wt) fat or a high fat diet containing 1.25% (wt/wt) cholesterol, 17.8% (wt/wt) butter, 1.0% (wt/wt) corn oil, and 0.5% (wt/wt) sodium cholate (Oriental Yeast Co. Ltd., Tokyo, Japan) for 2 weeks. After an overnight fast, mice were anesthetized with ketamine hydrochloride (Sankyo Co. Ltd., Tokyo, Japan) for phlebotomy via the femoral vein. Blood samples were centrifuged at 2500 rpm for 10 min, and about 200 l of plasma was obtained per mouse.
Plasma Lipid Analysis-The concentrations of total cholesterol, triglyceride, and HDL cholesterol in plasma were measured using enzymatic kits (Wako Pure Chemical Industries Ltd.) following the manufacturer's protocols. Plasma lipoprotein profile was determined by polyacrylamide disc-gel electrophoresis (18).
Cell Culture-HepG2 cells were purchased from RIKEN Cell Bank (Ibaraki, Japan) and routinely propagated at 37°C in a humidified atmosphere of 5% CO 2 in Dulbecco's modified Eagle's medium (Sigma) supplemented with 10% fetal calf serum (Invitrogen), 100 g/ml penicillin, and 100 g/ml streptomycin.
Recombinant Adenovirus-Ad5IB (donated by Dr. Iimuro, Kyoto University, Kyoto, Japan) expresses IB␣ super-repressor under the control of the cytomegalovirus promoter, where serine to alanine mutations at amino acids 32 and 36 prevent the phosphorylation that is necessary for it to dissociate from the NFB complex (19). Ad5LacZ, which expresses Escherichia coli ␤-galactosidase under the control of the cytomegalovirus promoter, was used as a control for adenovirusmediated transfection (20). Amplification, purification, and titration of the adenoviruses and adenovirus-mediated gene transfer were carried out as described previously (21). Cells were pretransfected with these adenoviruses 16 h before adding reagents.
Reverse Transcriptase-PCR Analysis-HepG2 cells were seeded on 6-well tissue culture plates (3 ϫ 10 5 cells/well) and incubated for 24 h. The cells were transfected with multiplicity of infection 10 Ad5IkB or Ad5LacZ as described previously (21) and incubated for an additional 16 h, exposed to 10 g/ml LPS or vehicle (5% v/v pure water), and cultured for 12 h. Total RNA was extracted from the cells using Isogen (Nippon Gene, Tokyo, Japan) according to the manufacturer's protocol. Oligo-dT primers (Takara Bio Inc., Shiga, Japan) together with 3 g of DNase-treated total RNA and Moloney murine leukemia virus reverse transcriptase (Invitrogen) were used to obtain first strand cDNA. PCR was performed using ExTaq polymerase (Takara Bio Inc.) and the oligonucleotide primers listed in Table I.
Immunoblot Analysis-HepG2 cells were seeded on 6-well tissue culture plates. If adenovirus-mediated transfection had been necessary, the cells were transfected with adenovirus as described above. The cells were exposed to 10 g/ml LPS or vehicle (5% v/v pure water) or 10 M MK886 or vehicle (0.1% v/v Me 2 SO) and cultured for 24 h. Culture media were subjected to SDS-polyacrylamide gel electrophoresis using 4 -20% gradient polyacrylamide gel (Daiichi Pure Chemicals Co., Ltd., Tokyo, Japan) and Western blot analysis with the specific anti-apoA-I antibody, anti-apoB antibody, or anti-apoE antibody, as previously described (25). Mouse plasma was also subjected to SDS-polyacrylamide gel electrophoresis and Western blot analysis by the same method.
Construction of ApoA-I Promoter Plasmids-The human apoA-I promoter fragment between positions Ϫ330 to ϩ69 relative to the transcription start site was prepared by PCR with primers 5Ј-TATAGCTA-GCAACACAATGGACAATGGCAACTG-3Ј and 5Ј-TATAAAGCTTGAA-CCTTGAGCTGGGGAGC-3Ј. The amplified fragment, which contains the PPRE element (26,27), was doubly digested with NheI and HindIII and inserted between the NheI and HindIII sites of pGL2-basic vector (Promega, Madison, WI) to make an "ApoAI-wt" plasmid. Mutation in the PPRE element was generated by PCR using a sense primer (5Ј-A-CTGATCCCTTGTCCCCTGCCCTGCAGCCCCCGCA-3Ј) and an antisense primer (5Ј-AGGGGACAAGGGATCAGTGGGGGCGGGAGGG-GAGT-3Ј) carrying mutations (underlined) to make an "ApoAI-mut" plasmid. The shorter promoter fragment between positions Ϫ142 to ϩ69, which does not contain the PPRE element, was also prepared by PCR with primers 5Ј-TATAGCTAGCAGGGACAGAGCTGATCCTTGA-AC-3Ј and 5Ј-TATAAAGCTTGAACCTTGAGCTGGGGAGC-3Ј and inserted between the NheI and HindIII sites of pGL2-basic vector to make an "ApoAI-short" plasmid.
Transient Transfection and Reporter Gene Assays-HepG2 cells were seeded on 12-well tissue culture plates (1 ϫ 10 5 cells/well). The cells were co-transfected with 0.4 g of the reporter plasmid and pRL-TK (Promega) internal control plasmid using LipofectAMINE plus reagent (Invitrogen), according to the manufacturer's protocol. The cells were subsequently transfected with multiplicity of infection 10 recombinant adenovirus as described above. After the cells were incubated at 37°C for 16 h, they were treated with 10 g/ml LPS, 10 M MK886 or vehicle (0.1% v/v Me 2 SO) and incubated for an additional 24 h. The activities of firefly luciferase from apoA-I promoter-luciferase plasmid and Renilla luciferase from pRL-TK plasmid in the cell extracts were evaluated with a dual luciferase assay kit (Promega) using a luminometer (TD-20/20, Promega) according to the manufacturer's protocol.
Statistical Analysis-Experimental data were evaluated by twotailed Student's t test. All unspecified data were presented as the mean Ϯ S.D.

RESULTS
Plasma Lipids-There was no significant difference in the concentration of total cholesterol in the plasma of female NFB-deficient (NFB Ϫ/Ϫ ) mice and wild-type (NFB ϩ/ϩ ) controls after 6 h of food removal. After the mice were fed a high fat diet for 2 weeks, both types of mice had an about 2.5-fold increase in the fasting cholesterol level, and there was no significant difference between NFB-deficient mice and wildtype littermates. NFB-deficient mice showed a significantly higher HDL cholesterol level than wild-type littermates under both a regular and high fat diet. NFB-deficient mice also showed a significantly lower triglyceride level than wild-type littermates only under a regular diet (Table II). Polyacrylamide disc-gel electrophoresis of whole plasma confirmed that the ratio of HDL to total lipoprotein was significantly higher in NFB-deficient mice than in wild-type control, whereas there was no significant difference in the ratios of low density lipoprotein (LDL) plus intermediate density lipoprotein and very low density lipoprotein (Table III).
Plasma Apolipoproteins-The levels of plasma apoA-I, apoB-100, and apoE were determined by immunoblot with specific antibodies (Fig. 1A). Plasma apoA-I levels showed a 10-fold increase in NFB-deficient mice compared with control under a regular diet and a 2-fold increase under a high fat diet (Fig.  1B). There were no significant differences in the levels of apoE and apoB between these mouse types irrespective of diet.
mRNA of Apolipoproteins in HepG2 Cells-To activate NFB we treated HepG2 cells with LPS, and to inactivate NFB directly we used adenovirus-mediated overexpression of IB␣ super-repressor. When intracellular NFB was activated by LPS treatment, the level of apoA-I mRNA in HepG2 cells was significantly decreased. Adenovirus-mediated overexpression of IB␣ super-repressor, which leads to inactivation of NFB, ameliorated this LPS-induced decrease. There was no significant difference in the levels of apoB and apoE mRNA, although NFB was activated by LPS and/or inhibited by IB␣ superrepressor (Fig. 2, A and B).
Apolipoproteins Secreted from HepG2 Cells-The level of apoA-I protein secreted from HepG2 cells was significantly decreased by LPS treatment. Adenovirus-mediated overexpression of IB␣ super-repressor ameliorated this decrease. There  ApoA-I  5Ј-ATGAAAGCTGCGGTGCTGACC-3Ј  5Ј-CACCTTCTGGCGGTAGAGCTCC-3Ј  22  ApoE  5Ј-ACTGGCACTGGGTCGCTTT-3Ј  5Ј-GTTGTTCCTCCAGTTCCGATT-3Ј  23  ApoB  5Ј-TAGACACCAACTTCTTCCACG-3Ј  5Ј-GGCGACCTCAGTAATTTTCTTG-3Ј  23  ␤-Actin  5Ј-CAAGAGATGGCCACGGCTGCT-3Ј  5Ј-TCCTTCTGCATCCTGTCGGCA-3Ј  24 was no significant difference in the levels of apoB and apoE, although NFB is activated by LPS and/or inhibited by IB␣ super-repressor (Fig. 2, C and D). PPAR␣ Inhibitor Blocked the Effect of NFB on ApoA-I Secretion-A control study in which HepG2 cells were treated with 10 M WY14643, a PPAR␣ agonist, for 24 h showed that apoA-I protein secretion from the cells was increased. This increase was blocked by treatment with 10 M MK886, a selective PPAR␣ inhibitor (28). These results indicate that apoA-I secretion from HepG2 cells is regulated by activation/inactivation of PPAR␣ (Fig. 3, A and B). Likewise, when IB␣ superrepressor was overexpressed in the cells, which leads to inactivation of NFB, the secreted apoA-I protein was increased. This increase was blocked by treatment with 10 M MK886 (Fig. 3, C and D). These results suggest that NFB inactivation induces apoA-I secretion from HepG2 cells through activation of PPAR␣. Unlike apoA-I, there was no difference in the levels of apoB and apoE in the culture medium, although NFB was inhibited by IB␣ super-repressor or PPAR␣ was inhibited by MK886 (Fig. 3C).
Transactivation of ApoA-I Promoter by Inhibition of NFB Was Mediated by PPAR␣-To confirm whether this induction of apoA-I mRNA by inhibition of NFB really derives from activation of the apoA-I promoter and to determine whether this induction is mediated by the transcription factor PPAR␣, we constructed human apoA-I promoter-luciferase plasmids carrying the wild-type PPRE element (ApoAI-wt) or a mutant PPRE element (ApoAI-mut) and a shorter plasmid that does not carry the PPRE element (ApoAI-short) (Fig. 4A). HepG2 cells were transiently transfected with one of these promoter-   reporter plasmids, transfected with adenovirus expressing IB␣ super-repressor or LacZ, and then treated with LPS, MK886, or vehicle (5% v/v pure water and/or 0.1% v/v Me 2 SO). The relative promoter activity in the cells where IB␣ superrepressor was overexpressed was significantly higher than the activity in LacZ-overexpressing cells (2.7-fold, p Ͻ 0.01, Fig.   4B). When the cells were transfected with ApoAI-wt plasmid and IB␣ super-repressor, treatment with 10 M MK886 significantly decreased relative promoter activity (2.7-1.4-fold, p Ͻ 0.01). When the cells were transfected with ApoAI-mut plasmid or apoAI-short plasmid, overexpression of IB␣ superrepressor appeared to increase relative promoter activity a little, but these increases were much less than the increase when ApoAI-wt plasmid was used and were not significant. DISCUSSION Cholesterol circulates in plasma as two main lipoprotein particles, LDL and HDL. LDL transports cholesterol synthesized in the liver to the peripheral tissues, whereas HDL returns cholesterol from the peripheral tissues to the liver for bile acid excretion. The major structural component of the HDL particle is apoA-I, whereas that of the LDL particle is apoB (29,30). Plasma HDL cholesterol and apoA-I levels are correlated with each other (31)(32)(33)(34), and they are negatively correlated with the prevalence of coronary heart disease (35)(36)(37) and atherosclerosis (31, 38 -42).
Inhibitors of the NFB pathway include a variety of agents, including antioxidants, proteasome inhibitors, decoy oligonucleotides, IB phosphorylation and/or degradation inhibitors, and IB super-repressor (43). Some of them act as general inhibitors, whereas others act as specific inhibitors. Among them, adenovirus-mediated overexpression of IB␣ super-repressor is a good tool for direct and selective inhibition of NFB activity in vitro (44). In most cell types, the NFB p50-p65 dimer is bound to one of the closely related endogenous inhibitory proteins, collectively referred to as IB, and is held inactive in the cytoplasm. A variety of extracellular stimuli, including viral infection, LPS, cytokines, and stress factors have been reported to activate NFB through phosphorylation of IB at serine-32 and serine-36, dissociation of IB from the complex, and translocation of the p50-p65 dimer from the cytoplasm to the nucleus. Phosphorylation of IB also leads to its ubiquitination and proteasomal degradation. Adenovirus-mediated overexpression of IB␣ super-repressor, where serine-32 and serine-36 of the IB␣ are substituted by alanines, suppresses phosphorylation and degradation of IB␣ (19), which leads to selective and constitutive inactivation of NFB.
Transcriptional regulation of apoA-I by activated PPAR␣ may be a species-specific phenomenon. In rats, Staels et al. (45) report that treatment with a PPAR␣ agonist, fenofibrate, markedly reduced hepatic apoA-I mRNA. In contrast, Vu-Dac et al. (46) report that fibrates increase human apoA-I production due to stimulation of apoA-I gene expression in the liver.
In this report they clearly demonstrated that this speciesspecific regulation of apoA-I expression was because of sequence differences in two distinct enhancer regions in the rodent and human apoA-I promoter. Similarly, using transgenic mice containing 21 copies of an 11-kilobase human genomic DNA fragment, Berthou et al. (47) also clearly demonstrate that 7 days of treatment with fenofibrate (5% wt/wt) increased plasma human apoA-I up to 750% and HDL cholesterol up to 200% in the transgenic mice, whereas it decreased plasma mouse apoA-I to 59% in non-transgenic mice. Fibrates are widely used hypolipidemic drugs, and they activate PPAR␣ that binds to the PPRE element on the human apoA-I promoter and positively regulate the expression of apoA-I. Fibrates also induce the expression of Rev-erb␣, which binds to a RebRE site in the rat apoA-I promoter and then repress the expression of apoA-I. They showed that the transcription from human apoA-I gene was promoted via PPAR␣ binding to a positive PPRE. Because of three single nucleotide differences, this site is not conserved in rats and mice. In contrast, rodent apoA-I transcription is repressed by Rev-erb␣, whose binding site RebRE is adjacent to the TATA box in the rodent apoA-I promoter but not in the human apoA-I promoter (48).
However, the question arises as to whether the three single nucleotide differences completely inactivated rodent PPRE or not. In our study, the levels of plasma apoA-I and HDL cholesterol were significantly higher in NFB p50 subunit-deficient mice than wild-type littermates. To support our data, Bisgaier et al. (49) report that 7 days of treatment of rats with a novel PPAR␣ activator PD72935 at a daily dose of 100 mg/kg significantly increased serum apoA-I and HDL cholesterol to 148 and 185% compared with control, respectively. Formation of the p65-p50 dimer and its translocation to the nucleus are completely lost in NFB p50 subunit-deficient mice. Therefore, unlike the fenofibrate-treatment, nuclear PPAR␣ could not interact with the p65 subunit, and the transcriptional activity of PPAR␣ might be selectively on a high level (50). Similarly, unlike with fenofibrate treatment, lack of the NFB-p50 gene did not seem to induce the expression of Rev-erb␣. To further support our concept, specific inhibition of NFB by IB␣ superrepressor significantly induced apoA-I expression in murine hepatoma Hepa1-6 cells, and this induction was blocked by treatment with a PPAR␣ inhibitor, MK886 (data not shown). In such a way our result that the lack of NFB p50 subunit increased plasma apoA-I and HDL cholesterol in mice is mostly compatible with the results of Berthou et al. (47) and Bisgaier et al. (49).
To assess our hypothesis that inactivation of NFB facilitates expression of apoA-I through activation of the transcription factor PPAR␣, we performed reverse transcriptase-PCR analysis to semi-quantify apoA-I mRNA in HepG2 cells and immunoblot analysis to quantify its secretion into the medium from HepG2 cells. Activation of NFB by LPS treatment significantly decreased apoA-I at both the mRNA and protein levels. Specific inactivation of NFB by overexpression of IB␣ super-repressor ameliorated the decreases in both levels (Fig.  2). This IB␣ super-repressor-induced apoA-I elevation was completely blocked by pretreatment of the cells with MK886, a selective PPAR␣ inhibitor (Fig. 3, C and D). When the NFB p65-p50 dimer translocates into the nucleus, the p65 subunit interacts with PPAR␣ and represses PPAR␣ transactivation of a PPRE-driven promoter (50). In the cells where IB␣ superrepressor is overexpressed, the dimer tightly binds to the super-repressor, and translocation of the dimer into the nucleus and interaction of the p65 subunit with PPAR␣ are suppressed (50 -52).
To further support our hypothesis, activity of the human apoA-I promoter was also increased 2.7-fold by overexpression of IB␣ super-repressor. The 2.7-fold increase was reduced to 1.5-fold by pretreatment with MK886, a selective PPAR␣ inhibitor. When the apoA-I promoter, carrying mutations in the PPRE element, or the shorter promoter, which did not contain the PPRE element, was used instead of the wild-type promoter, the increase in promoter activity induced by IB␣ super-repressor was blocked. These results indicate that NFB activation/ inactivation controls expression of apoA-I at the transcriptional level through PPAR␣.
We have demonstrated here that lack of the NFB p50 subunit in mice increased plasma apoA-I and HDL cholesterol without any difference in total cholesterol in vivo. We have also demonstrated that adenovirus-mediated inactivation of NFB increased the expression of apoA-I at the mRNA and protein level in vitro. The reciprocal function between PPAR␣ and NFB to the PPRE in the apoA-I promoter was responsible for this transactivation. In summary, this report demonstrated that NFB activity indirectly controls apoA-I expression both in vivo and in vitro.