Peroxisome Proliferator-activated Receptor α-mediated Pathways Are Altered in Hepatocyte-specific Retinoid X Receptor α-deficient Mice

Abstract Retinoid x receptor α (RXRα) serves as an active partner of peroxisome proliferator-activated receptor (PPARα). In order to dissect the functional role of RXRα and PPARα in PPARα-mediated pathways, the hepatocyte RXRα-deficient mice have been challenged with physiological and pharmacological stresses, fasting and Wy14,643, respectively. The data demonstrate that RXRα and PPARα deficiency are different in several aspects. At the basal untreated level, RXRα deficiency resulted in marked induction of apolipoprotein A-I and C-III (apoA-I and apoC-III) mRNA levels and serum cholesterol and triglyceride levels, which was not found in PPARα-null mice. Fasting-induced PPARα activation was drastically prevented in the absence of hepatocyte RXRα. Wy14,643-mediated pleiotropic effects were also altered due to the absence of hepatocyte RXRα. Hepatocyte RXRα deficiency did not change the basal acyl-CoA oxidase, medium chain acyl-CoA dehydrogenase, and malic enzyme mRNA levels. However, the inducibility of those genes by Wy14,643 was markedly reduced in the mutant mouse livers. In contrast, the basal cytochrome P450 4A1, liver fatty acid-binding protein, and apoA-I and apoC-III mRNA levels were significantly altered in the mutant mouse livers, but the regulatory effect of Wy14,643 on expression of those genes remained the same. Wy14,643-induced hepatomegaly was partially inhibited in hepatocyte RXRα-deficient mice. Wy14,643-induced hepatocyte peroxisome proliferation was preserved in the absence of hepatocyte RXRα. These data suggested that in comparison to PPARα, hepatocyte RXRα has its unique role in lipid homeostasis and that the effect of RXRα, -β, and -γ is redundant in certain aspects.


INTRODUCTION
Peroxisome proliferators including herbicides, plasticizers, hypolipidemic drugs (fibrates), and leukotriene D4 inhibitors play a crucial role in hepatocyte proliferation. The most potent peroxisome proliferator is Wy14,643. These agents cause profound peroxisome proliferation in hepatocytes resulting in hepatomegaly and hepatoma and a rapid transcription of genes encoding the enzymes involved in fatty acid metabolism (for reviews see 1-4). Peroxisome proliferators exert their pleitropic responses via PPARα, a member of the nuclear hormone receptor superfamily (5-10).
Besides peroxisome proliferator, PPARα can also be activated by certain conditions such as starvation, high fat diet, diabetes mellitus under which increased fatty acids are delivered to the liver (11)(12)(13).
RXRs are the required active heterodimeric partners of PPARs (14). Thus, RXR, PPAR, and their ligands are all actively involved in regulating liver gene expression, fatty acid metabolism, lipid transport, and hepatocyte proliferation. Among the three types of RXR, RXRα is the predominant one expressed in the liver.
Absence of PPARα expression in knockout mice prevents the induction of hepatocyte peroxisome proliferation and of fatty acid synthesizing enzymes and β oxidizing enzymes by Wy14,643 (15)(16)(17).
In addition, PPARα deficiency leads to elevated serum cholesterol levels in young adult mice and increased serum triglyceride levels and steatosis in aging mice (18). There is no in vivo model available with which to compare the role of RXRα with PPARα, because of embryonic lethality by guest on July 18, 2018 http://www.jbc.org/ Downloaded from 5 (ChemSyn Science Laboratories, Lenexa, KS). Pelleted mouse chow, which composed of 21.4% protein, 55% carbohydrates, 4% fat, 6.7% ash, 4% fiber, and less than 10% moisture, was commercially prepared containing either 0.0% (control) or 0.1% (wt/wt) Wy14,643 (Bioserv, Frenchtown, NJ). For all the experiments 10-16 weeks old male mice were used. Mice were fed either control or Wy14,643 diet ad libitum for 10 days. For starvation experiment, mouse chow was removed from mice for 48 hours. Animals were housed in groups of two or three in plastic microisolator cages at 25ºC with a 12-h light/12-h dark cycle.
At the end of the treatment, animals were weighed and anaesthetized with pentobarbital (60 mg/kg, ip). Blood samples were obtained by intracardiac puncture. Blood triglycerides and cholesterol levels were determined by automated analysis. The liver was removed immediately, weighed, frozen in liquid nitrogen, and processed for RNA extraction. Part of liver was fixed by formalin and 1.5% glutaraldehyde for light and electron microscopy analysis, respectively.
Northern Blot Hybridization-Molecular aspects of hepatocyte-specific RXRα mutation were evaluated by northern blotting analysis of RNA levels in the liver for the expression of PPARα target genes. The gene probes used were apoAI and CIII (provided by Dr. J. Auwerx), liver fatty acidbinding protein (provided by Dr. J. Gordon), malic enzyme (provided by Dr. G. Brent), acyl-CoA (AOX) and cytochrome P450 4A1 (CYP4A1) mRNA encoding the key enzymes involved in fatty acid βand ω-oxidation pathways (Fig. 1). PPARα deficiency also resulted in a decreased expression of liver fatty acid binding protein (LFABP) mRNA and a weak induction of apoA1 mRNA. In comparison, RXRα deficiency resulted in inhibition of expression of CYP4A1 and LFABP mRNA. The level of AOX and medium chain acyl-CoA dehydrogenase (MCAD) mRNA remained unchanged. The most striking difference between the PPARα and RXRα deficient mice was that the expression of apoAI and apoCIII mRNA was markedly increased in the absence of RXRα, whereas the induction was very weak, if there was any, in the PPARα-null mice (Fig. 1).
In wild type mice, starvation caused significant induction of PPARα target gene except for the apoCIII gene ( In contrast to the AOX, MCAD, and malic enzyme genes, the basal transcription of the CYP4A1 and LFABP genes can be controlled by PPARα/RXRα at the physiological level. CYP4A1 and LFABP mRNA level was reduced about 3-fold in RXRα deficient mouse livers compared with the wild-type livers (Fig. 3). However, the inducibility of these two genes by Wy14,643 remained the same in mutant mouse livers (Fig. 3). After Wy14,643 administration, there was a 50-and 10-fold induction of CYP4A1 and LFABP mRNA level, respectively, in both wild-type and mutant mouse livers. These data suggest that the basal transcription of the CYP4A1 and LFABP genes is constitutively regulated by PPARα/RXRα or RXRα/RXRα through endogenous ligands such as polyunsaturated fatty acids or 9-cis-retinoic acid in vivo. Therefore, in the absence of RXRα, these genes are expressed at a reduced level. However, when pharmacological levels of exogenous ligands are present, the availability of RXRβ and γ is sufficient to mediate the inductive effect of Wy14,643.
To further understand the role of RXRα in regulating cholesterol and lipid homeostasis, the expression of apoA1 and apoCIII mRNA was examined in Wy14,643 treated mice. In normal cells, PPARα agonists suppress the expression of these genes. RXRα is involved in the basal transcription of the apolipoprotein genes because the basal mRNA levels in normally fed mice were increased in the absence of RXRα (Fig. 4). However, the inhibitory effect of Wy14,643 on apolipoprotein gene expression remained in the absence of hepatocyte RXRα.
Taken together, the expression pattern of these PPARα target genes can be divided into two groups.
In the first group exemplified by the AOX, MCAD, and malic enzyme genes, these genes' basal mRNA level remains unchanged in mutant mouse liver, but the inducibility of the gene by Wy14,643 is decreased remarkably. In the second group, which includes the CYP4A1, LFABP, apoAI and apoCIII genes, the basal mRNA level is altered in the absence of RXRα, but the regulatory effect of Wy14,643 on gene expression remains unchanged in mutant mouse liver.
Reduction of serum cholesterol and triglyceride level by Wy14,643 in hepatocyte-specific RXRα deficient mouse-As a hypolipidemic drug (1-4), Wy14,643 reduces serum cholesterol and triglyceride level. These effects were tested in the hepatocyte RXRα deficient mice. As shown in figure 5, basal serum triglyceride and cholesterol levels were elevated in the RXRα deficient mice, which is consistent with the northern data ( Fig. 1 and 4) demonstrating the induction of apoAI and CIII mRNA in the mutant mouse livers. Administration of Wy14,643 reduced serum triglyceride and cholesterol level not only in wild-type but also in mutant mice. Therefore, Wy14,643 still can exert its hypolipidemic effect even when RXRα is not expressed in the hepatocyte.
Reduced hepatomegaly in the hepatocyte RXRα deficient mice fed Wy14,643 diet-It is well characterized that Wy14,643 causes liver enlargement due to hypertrophy and hyperplasia (hepatomegaly) of hepatocytes (1-4). Furthermore, clofibrate and Wy14,643-induced hepatomegaly is not found in PPARα-null mice (15). In our system, the data was reproducible where Wy14,643 also produced a marked increase in liver weight in the wild-type mouse. The liver/body weight ratio of the wild-type mice increased 2.4-fold after 10 days of Wy14,643 feeding compared with mice fed a standard control diet (Table 1). In contrast, the liver/body ratio of hepatocyte RXRα deficient mouse only increased by 1.6-fold after Wy14,643 treatment. Therefore, the hepatomegaly caused by treatment with the peroxisome proliferator was partially prevented when RXRα was absent.
diet-Using light and electron microscopy, the liver morphology of the wild type and RXRα deficient mice was evaluated (Fig. 6). Compared with wild-type mouse livers, RXRα deficient mouse livers had normal morphology under light and electron microscope (Fig. 6a-d). Treatment of wild-type mice with Wy14,643 resulted in pale pink staining enlarged cells which had increased homogeneous cytoplasm. The cytoplasmic rough endoplasmic reticulum was strikingly reduced (Fig. 6e). Further, the number and the size of peroxisome were significantly increased after the administration of Wy14,643 as demonstrated by electron microscopy (Fig. 6f). In contrast, under light microscopy, the mutant mouse liver contain both normal and enlarged cells after administration of Wy14,643 (Fig.   6g). Electron microscopy revealed that Wy14,643 still induced hepatocyte perxisome proliferation in RXRα deficient mice (Fig. 6h).  Table 2. In PPARα knockout mice, basal serum cholesterol level is elevated to the same extent (1.6-fold induction) as in the hepatocyte specific RXRα knockout mice. However, young adult male PPARα -null mice have normal serum triglyceride and apoCIII level (16,18). Serum triglyceride level only elevates in aged animals (6-12-month-old), and the level is higher in females (2-fold induction) than males (1.5-fold induction) (18). In contrast, in hepatocyte RXRα deficient mice, a 1.7-fold induction of serum triglyceride level and a remarkable induction of apoCIII gene expression were observed in 2-month-old male mice. The early induction in serum triglyceride level defines the unique and important role of hepatocyte RXRα in controlling lipid homeostasis. It is possible that the effect of RXRα in regulating apoCIII gene expression and serum triglyceride level is mediated through dimerization with PPARγ rather than PPARα.
In PPARα -null mice, peroxisome proliferators such as clofibrate and Wy14,643 are completely unable to induce hepatomegaly and hepatocyte peroxisome proliferation, and have no effect in regulating the expression of PPARα target genes including AOX, bifunctional enzymes, CYP4A1, CYP4A3, LFABP, apoAI, and apoCIII (15)(16)(17). These data suggest that the effect of PPARα is unique in peroxisome proliferator-mediated pathways, and that PPARβ and γ cannot replace PPARα . In contrast, in vivo, the roles of RXR α, β, and γ appear to be at least partially redundant.
Based on our results, the PPARα/RXRα target genes can be categorized into several groups. The first group of genes includes AOX and malic enzyme. The basal transcriptional rate of these genes is controlled by PPARα , but not by RXRα. The second group of genes is CYP4A1, LFABP, and apoAI. Within this group, the basal transcriptional rate of the genes is constitutively maintained by PPARα as well as by RXRα through endogenous ligands. The third group of gene includes apoCIII.
The basal transcriptional rate of the apoCIII gene is controlled by RXRα, but not by PPARα. Since RXRα controls the basal transcription of the CYP4A1, LFABP, and apoAI genes, but has no effect on the AOX, MCAD, and malic enzyme genes, these data suggest that in vivo at the physiological level RXRα is crucial for microsomal ω-hydroxylation of fatty acids, fatty acid transport, and cholesterol and fatty acid homeostasis, whereas RXRα may only become important for AOX and MCAD-mediated fatty acid β-oxidation and malic enzyme-mediated lipogenesis when pharmacological dose of PPARα ligand is employed.
Even though RXRβ and γ are able to substitute RXRα, the total amount of RXRs is critical in mediating the action of RXRs because in the absence of RXRα, fatty acid is not utilized efficiently in response to starvation and Wy14,643 can not fully exert its effects. RXR dimerizes with more than ten different kinds of receptor. Activation one of these RXR-mediated pathways might alter other pathways in opposite directions. When the pool of RXRs is decreased, many RXR-mediated regulatory pathways may be impaired. Based on our data, it seems that the level of RXR, rather than the type of RXR, has a major impact in mediating the effect of peroxisome proliferator. It is crucial to understand the regulation of the RXR genes.
RXR can be freely activated in permissive heterodimers with PPAR (37) although it also can be silent in non-permissive heterodimers with the thyroid hormone receptor or the vitamin D receptor (38). It would be interesting to test if 9-cis-retinoic acid has the same effect as Wy14,643 on RXRα deficient mice. 9-Cis retinoic acid can activate RXR/RAR and RXR/RXR, and that would further deprive the availability of RXR to PPARα. Therefore, challenge the mutant mice with 9-cis retinoic acid may produce more phenotypes.
Taken together, nuclear factors might have unique, redundant, synergistic, or antagonistic effects. These effects depend on the relative level of the receptors, presence of hormones, or the pathological condition. Comprehension of the regulation of liver gene transcription provides insight into the understanding of the molecular mechanisms leading to liver physiology, function, development, differentiation as well as proliferation.  Total RNA (20 µg) was extracted from representative livers of wild-type and hepatocyte RXRα deficient mice fed control or Wy14,643 (0.1%) rodent diet for 10 days. Northern blot hybridization was performed using the indicated cDNA probes. The relative fold changes of the message levels after normalization to 18S rRNA level are indicated below each panel. Fig. 4 The expression of the apoAI and CIII genes in wild type and hepatocyte RXRα deficient mouse livers.
Representative northern blots demonstrate the expression of apoAI and CIII mRNA in the livers of wild-type and hepatocyte RXRα deficient mice fed control or Wy14,643 (0.1%) rodent diet for 10 days. Total RNA (20 µg) was electrophoresed and hybridized with the indicated cDNA probes. The relative fold changes of the message levels after normalization to 18S rRNA level are indicated below each panel. Each value represents the mean + S.D. of 4 mice. When existence, statistically significant differences between treated and untreated animals of the same genotype (*), as well as between wild type and deficient mice (**) are indicated by asterisks (P < 0.05).