Regulatory, Integrative and Comparative Physiology

Induction of overlapping genes by fasting and a peroxisome proliferator in pigs: evidence of functional PPARα in nonproliferating species

Yewon Cheon, Takayuki Y. Nara, Mark R. Band, Jonathan E. Beever, Matthew A. Wallig, Manabu T. Nakamura


Peroxisome proliferator-activated receptor α (PPARα), a key regulator of fatty acid oxidation, is essential for adaptation to fasting in rats and mice. However, physiological functions of PPARα in other species, including humans, are controversial. A group of PPARα ligands called peroxisome proliferators (PPs) causes peroxisome proliferation and hepatocarcinogenesis only in rats and mice. To elucidate the role of PPARα in adaptation to fasting in nonproliferating species, we compared gene expressions in pig liver from fasted and clofibric acid (a PP)-fed groups against a control diet-fed group. As in rats and mice, fasting induced genes involved with mitochondrial fatty acid oxidation and ketogenesis in pigs. Those genes were also induced by clofibric acid feeding, indicating that PPARα mediates the induction of these genes. In contrast to rats and mice, little or no induction of genes for peroxisomal or microsomal fatty acid oxidation was observed in clofibric acid-fed pigs. Histology showed no significant hyperplasia or hepatomegaly in the clofibric acid-fed pigs, whereas it showed a reduction of glycogen by clofibric acid, an effect of PPs also observed in rats. Copy number of PPARα mRNA was higher in pigs than in mice and rats, suggesting that peroxisomal proliferation and hyperresponse of several genes to PPs seen only in rats and mice are unrelated to the abundance of PPARα. In conclusion, PPARα is likely to play a central role in adaptation to fasting in pig liver as in rats and mice.

  • clofibric acid
  • fatty acid oxidation
  • rats
  • mice
  • microarray

regulation of fatty acid oxidation has been an important issue for human health because imbalance of lipid metabolism could contribute to the development of chronic diseases such as obesity, diabetes, and coronary heart disease. Moreover, understanding metabolic adaptation to negative energy balance is a prerequisite for elucidating the physiology of weight loss. Fasting induces dramatic changes in metabolism to maintain energy homeostasis, including the release of fatty acids from adipose tissue, followed by their oxidation. In rats and mice, peroxisome proliferator-activated receptor α (PPARα), a ligand-activated transcription factor, is a key regulator of genes for mitochondrial and peroxisomal fatty acid oxidation, as well as ketogenesis. Targeted disruption of PPARα has demonstrated its essential role in adaptation to fasting in mice (15, 23, 29). PPARα-null mice grow phenotypically normal when they have free access to food. However, when the animals are fasted, they show impairment in the induction of fatty acid oxidation and ketogenic enzymes in liver, resulting in hypoglycemia, hypothermia, elevated serum fatty acids, and abolished serum ketone body increase.

Various long-chain fatty acids, as well as hypolipidemic agents called peroxisome proliferators (PPs) act as ligands of PPARα (10, 24). Because nonesterified fatty acids are increased in liver during fasting (16), fatty acids released from adipose tissue are likely to act as endogenous ligands of PPARα during fasting. On the other hand, PPs not only induce genes for fatty acid oxidation but also cause severe peroxisome proliferation, hepatomegaly, and hepatocarcinogenesis in rats and mice (12). This pathological response is also mediated by PPARα (12).

In contrast to rats and mice, PPs do not induce peroxisome proliferation or tumor in the liver of many other species, such as guinea pigs, swine, monkeys, and humans, although PPs retain a hypotriglyceridemic effect in these species (18, 41). The underlying mechanism of this marked difference in the response to PPs among species is controversial. Some studies reported lower expression of PPARα in the liver of humans (39, 47) and guinea pigs (9, 47) than in rats and mice. Other studies suggest that humans may not have functional peroxisome proliferator response elements in the promoters of genes for peroxisomal enzymes such as acyl-CoA oxidase (ACOX) (27, 49). Because of the paucity of in vivo data, the function of PPARα in nonproliferating species is currently unclear, and the underlying mechanism of species differences in the response to PPs is yet to be investigated.

The objective of this study is to elucidate the physiological functions of PPARα in adaptation to fasting in species that do not undergo peroxisome proliferation by PP administration. To achieve this goal, we investigated the gene expression in pig liver in response to fasting and clofibric acid (a PP), and evaluated the expression pattern against that of rats and mice, the species in which functions of PPARα are well established.


Animal Studies

Nine pigs (21 days old, male, Yorkshire × Landrace) were purchased from Oak Hill Genetics (Ewing, IL) and adapted for 1 wk. The animals were caged in a room under controlled temperature, humidity, and lighting (22 ± 2°C, 55 ± 5%, and a 12:12-h light-dark cycle, respectively) with free access to water. Food was freely accessible from 8 AM to 8 PM. During the adaptation period, all nine pigs were fed a control diet (phase-one powdered diet, University Feed Mill, University of Illinois at Urbana-Champaign). After adaptation, three pigs were fed the control diet supplemented with clofibric acid (0.5% wt/wt) for 1 wk, while the other six animals remained in the control diet. Food was removed for 24 h before euthanasia from three of the control diet-fed animals (FAST group). The remaining control diet-fed (FED) and clofibric acid-fed (CLFED) groups were fed their respective diets from 8 AM to 11 AM before euthanasia. All animals were anesthetized with 86 mg/kg body wt of pentobarbital sodium (Fatal-Plus, Vortech Pharmaceuticals, Dearborn, MI) via intravenous injection. Blood samples were collected from pigs via jugular venipuncture into heparinized vacutainer tubes (Becton-Dickinson, Rutherford, NJ). Blood was placed on ice until centrifugation at 5,000 g (4°C), and plasma was stored at −80°C for subsequent analysis of 3-hydroxybutyrate and acetoacetate. Three small pieces of liver from individual lobes were excised for histological analysis and immersion-fixed in 10% neutral buffered formalin. The remaining liver samples were immediately immersed in liquid nitrogen. All procedures were completed between 11 AM and 1 PM.

Rats and mice were used for quantification of the copy number of PPARα mRNA, as well as mRNA analysis of selected PPARα-responsive genes. Nine male Sprague-Dawley rats (∼200 g) were acclimated to a meal-fed procedure by feeding a control diet (AIN93G) (40) for 4 h/day for 1 wk. After the adaptation period, three animals were killed before meal as the fasting group. The remaining rats were divided into two groups that received either the control diet or a control diet containing 0.1% Wy14643 (a PP) for 4 h every day. On the fifth day, three rats from each group were killed immediately after the meal. Livers were isolated and frozen at −80°C. Sv/129 mice (8 wk, male, n = 15), after 1 wk of adaptation, received either a control (AIN93G) diet (n = 10) or a control diet containing 0.1% Wy14643 (n = 5) for 1 wk with free access to water. Food was freely accessible from 8 PM to 8 AM. The mice were killed immediately after the meal, except for five of the mice fed the control diet, which were fasted for 24 h before death. Livers were isolated and frozen at −80°C. All of the animal protocols were reviewed and approved by the Institutional Animal Care and Use Committee of the University of Illinois at Urbana-Champaign.

Total RNA Preparation

Total RNA from liver tissue from all animals was extracted by TRIzol reagent (Invitrogen, Carlsbad, CA). The RNA quality and concentration was assessed by spectrophotometry and agarose gel electrophoresis.

Microarray Production

Porcine cDNA library construction.

Porcine skeletal muscle tissue was harvested at gestational days 29, 35, 43, 49, 56, 64, 70, 78, 84, 93, 99, and 106. To construct cDNA libraries, total RNA was extracted by TRIzol reagent (Invitrogen) from tissue samples. Messenger RNA was isolated using an oligo (dT) molecule complementary to poly (A)+ mRNA (Qiagen, Valencia, CA). mRNAs from different gestational days were grouped together: early gestation (days 29, 35, 43, and 49), middle gestation (days 56, 64, 70, and 78) and late gestation (days 84, 93, 99, and 106) stages of development. From these mRNAs, three cDNA libraries were constructed using ZAP Express cDNA Synthesis Kit and ZAP Express cDNA Gigapack III Gold Cloning Kit (Stratagene, Cedar Creek, TX) according to the manufacturer's protocols.

Probe DNA selection, PCR amplification, DNA purification, and amplification.

From the three cDNA libraries, total of 9,781 cDNA clones were obtained and sequenced (Incyte Genomics, Palo Alto, CA). Unique genes were identified by sequence similarity searches (BLAST) against the NCBI Unigene and EST databases. This analysis resulted in 1,618 unique clones of which 1,248 clones were selected for the assembly of the microarray. In addition to this set of clones, 24 porcine EST sequences relevant to this study were obtained from the BACPAC Resource Center at Children's Hospital Oakland Research Institute (Oakland, CA) and added to the microarray collection for a total of 1,272 sequences. After assembly of the unique set, cDNA inserts were amplified by PCR using T7 and T3 primers. PCR products were cleaned on sephadex columns and run on 1% agarose gels to assess quality. PCR products were transferred to 384-well plates, dried, and suspended in 8-μl of 3× saline sodium citrate (SSC) with 1.5 M betaine.

Microarray fabrication.

All probes were printed in duplicate on Arrayit Superamine slides (TeleChem International, Sunnyvale, CA) using a Cartesian Technologies PixSys 5500 spotter. In addition, housekeeping genes (porcine glyceraldehyde-3-phosphate dehydrogenase and beta-actin) and exogenous control cDNAs from soybean (rubisco small chain 1, AI495218; chlorophyll ab binding protein, BE190670; and monosodium glutamate gene, AJ239127) were spotted on the array. Print quality was assessed visually under a dissecting scope after printing. Slides exhibiting no defects (no spots missing, no spots joined, and all spots uniform in size) were selected. The slides were baked at 80°C for 2 h, washed in 0.2% SDS, denatured in boiling water for 2 min, transferred to 95% ethanol for 15 s, spun dry, and stored.

Microarray Hybridization and Data Analysis

Aminoallyl labeling of RNA.

Total RNA was purified with the RNeasy kit (Qiagen) after treatment with 1 unit of RQ1 DNase (Promega, Madison, WI). RNA samples from the same treatment group were pooled. Twenty micrograms of pooled RNA were combined with 2 μg of dT18 primer (Invitrogen), and spiking controls (5 ng of rubisco small chain 1, 0.5 ng of chlorophyll ab binding protein, and 0.05 ng of monosodium glutamate gene). The mixture was incubated at 70°C for 10 min and chilled on ice for 1 min. The reverse transcription reaction contained 6 μl of 5× first strand buffer, 3 μl of 0.1 M DTT, 0.6 μl of 50× aminoallyl-dNTP mixture (25 mM each of dATP, dCTP, dGTP, 15 mM dTTP, and 10 mM aa-dUTP) and 400 U Superscript II (Invitrogen) in a 30-μl reaction volume. The RNA was degraded by addition of 10 μl of 1 M NaOH followed by 15-min incubation at 65°C and then neutralized by an addition of 10 μl of 1 M HCl. Labeled cDNA was purified using the Qiagen PCR purification kit. Amino allyl-labeled cDNA was coupled with the appropriate Cy dye ester (Amersham Biosciences, Piscataway, NJ; Cy3, #Q13104; Cy5, #15104) and cleaned with the Qiagen PCR purification kit.

Microarray hybridization.

Microarray slides were prehybridized in preheated prehybridization solution (20% formamide, 5× Denhardt's reagent, 6× SSC, 0.1% SDS, 0.05% inorganic pyrophosphate, 25 μg/ml tRNA) at 42°C for 45 min with occasional shaking, then washed by dipping in 200 ml of MilliQ water and followed by an isopropanol wash, spun dry, and hybridized immediately. The Cy3- and Cy5-labeled samples were combined in total 20 μl of sterile water containing 20 μg of porcine C0t-1 DNA, which was prepared using a method by Zwick (51), and 10 μg of poly A RNA. After heating at 95°C for 2 min, the samples were mixed with an equal volume of 2× hybridization buffer (50% formamide, 10× SSC, and 0.2% SDS), and added to the array under a cover slip. The array was assembled in a Corning hybridization cassette and hybridized submerged in a water bath at 42°C overnight. Excess probe was removed by a series of 2-min washes: once in 1× SSC, 0.2% SDS at 42°C; once in 0.1× SSC, 0.2% SDS at room temperature; and twice in 0.1× SSC at room temperature. Hybridized slides were scanned using an Axon GenePix 4000B dual-laser scanner (Axon Instrument, Union City, CA), and images were analyzed with Axon's GenePix software. All three pairwise comparisons between treatments (CLFED-FED, FAST-FED, CLFED-FAST) were replicated on individual slides with dye swapping for a total of six arrays.

Microarray data analysis.

Data analysis was performed using GeneSpring 6.0 (Silicon Genetics, Redwood City, CA) software. Each array was globally normalized to balance intensity differences between the two channels. Data were then further analyzed in Microsoft Excel. Spots with low-fluorescence signals were excluded. The corrected log ratios (Cy5/Cy3) of duplicate spots on each array were averaged. Mean ratios of the two replicate hybridizations (dye swapping) in each comparison were calculated for each gene. Criteria to determine reliability of data were set as P value less than 0.1 for four replicates spots (two per array with dye swap) and average signal above a fluorescent intensity of 1,000 (typically, between 2 and 3 SDs above the mean background across the slide). In addition, a ratio of CLFED-FED to FAST-FED was compared with a direct comparison of CLFED-FAST to further confirm the validity of data. A twofold difference or greater between two treatments was considered significantly different.

Database submission of microarray data.

The microarray data was prepared according to Minimum Information About a Microarray Experiment (MIAME) recommendations (5), deposited in the Gene Expression Omnibus database, and can be accessed at The accession number for the platform is GPL518. Six data sets can be retrieved with accession numbers GSM11011, GSM11012, GSM11013, GSM11014, GSM11015, and GSM11016, which constitute a series accession number GSE714.

Real-Time Quantitative PCR Analysis

To verify the results of the microarray experiment, as well as to complete the list of gene expressions involved in target metabolic pathways, we measured mRNA expression of key genes using a real-time quantitative PCR method with SYBR Green fluorescence dye (Applied Biosystems, Foster City, CA). Two micrograms of total RNA were reverse-transcribed into complementary DNA by Superscript II reverse Transcriptase (Invitrogen, Carlsbad, CA) using random hexamers (Applied Biosystems). Reactions were carried out at 25°C for 10 min, followed by 50 min at 42°C, and finally 70°C for 15 min. The primer pair for each gene was designed by Primer Express software (Applied Biosystems) with melting temperature of 58–60°C. The primer sequences are listed in Table 1. All primers were purchased from MWG Biotech (High Point, NC). Real-time PCR was performed using an ABI Prism 7700 sequence detection system (Applied Biosystems). The mRNA abundance relative to 18S rRNA was determined using the comparative critical threshold method according to manufacturer's instructions. A validation experiment was performed before sample analysis to verify the equal amplification efficiency of target genes and 18S. Each reaction mixture (25 μl) contained 5 ng of cDNA, 800 nM forward and reverse primers, and 1× SYBR Green PCR Master Mix (Applied Biosystems). A standard SYBR Green Assay program for thermal cycling was used: 5 s at 95°C for one cycle and 40 cycles at 95°C for 15 s followed by 1 min at 60°C.

View this table:
Table 1.

Primer sequences used for real-time PCR

Quantification of the Copy Number of PPARα and 18S

To determine the absolute copy number of the target transcript, a plasmid DNA for PPARα and 18S was used to generate a standard curve. The following plasmids were provided as kind gifts: pig PPARα (AF175309) from Dr. Harry Mersmann; rat PPARα from Dr. Ronald Evans (25); mouse PPARα from Dr. Stephen Green (21); and pig 18S (BI183569) from Dr. Daniel Pomp (6). Rat 18S plasmid (CA339498) was purchased from Open Biosystems (Huntsville, AL), and mouse 18S (BQ442953) plasmid was purchased from Invitrogen. The copy numbers of plasmid DNA template were calculated according to the molecular weight of the plasmid and then converted into the copy numbers based on the Avogadro's number (1 mole = 6.022 × 1023 molecules). A standard curve was constructed by plotting the threshold cycle (Ct) vs. the copy numbers of a standard plasmid added as a template. The copy number of PPARα and 18S mRNA in each sample was calculated based on its Ct value with its plasmid DNA standard curve. The copy number of PPARα mRNA was then normalized to 18S to minimize variability due to differences in the efficiency of reverse transcription.


After fixation in 10% neutral buffered formalin for 24 h, tissue samples from pig liver were trimmed, dehydrated in graded alcohols, imbedded in paraffin and sectioned at 4-μm thickness. The sections were mounted on negatively charged glass slides, deparaffinized, and stained with either hematoxylin and eosin or periodic acid-Schiff. The slides were examined using an Olympus BX-41 microscope with an Olympus 750 camera, and images were processed by a Pax-it Capture and Archiving System.

Determination of Liver Glycogen Content

Liver glycogen content in pig liver was measured by a phenol-sulfuric acid colorimetric assay, as described by Lo et al. (31). Briefly, frozen liver was powdered under liquid nitrogen. After dissolving in 30% (wt/vol) KOH saturated with Na2SO4, samples were boiled for 15 min until a homogenous solution was obtained, then cooled on ice for 30 min. After precipitation of glycogen by 95% ethanol, samples were centrifuged at 840 g, and glycogen pellets were resuspended in distilled water. A 5% phenol solution and 96% sulfuric acid solution were added, samples were incubated at room temperature for 10 min, and the absorbance was read at 490 nm.

Plasma Ketone Body Analysis

3-Hydroxybutyrate and acetoacetate in pig plasma were determined by an enzymatic method (34, 48). Plasma was deproteinized by adding the same volume of perchloric acid (ca. 10% wt/vol), and were then centrifuged at 3,000 g at 4°C for 15 min. Supernatant was neutralized with KOH (ca. 20% wt/vol). The enzyme 3-hydroxybutyrate dehydrogenase catalyzes the conversion of acetoacetate to 3-hydroxybutyrate with concomitant reduction of NADH to NAD+ for acetoacetate assay. The same enzyme was used for the reverse reaction in 3-hydroxybutyrate assay. The concentration of ketone bodies is calculated from the amount of NAD+/NADH converted during the time required to complete the reaction using the absorption coefficient of NADH at 340 nm.


Data were presented as means ± SD. Comparisons were made by using one-way ANOVA coupled with Fisher's protected least significant difference post hoc test using Statview 5.0.1 (SAS Institute, Cary, NC). Statistical significance was set as P < 0.05.


General Characteristics of Clofibric Acid-Fed Pigs

Food consumption of a pig in all groups was about 1.2 kg/day. Thus the dose of clofibric acid was about 500 mg·kg body wt−1·day−1. As shown in Table 2, there was no difference in body weight between the fed and clofibric acid groups. Mean values of liver weight and the liver weight-to-body weight ratio were slightly higher (approximately +10%) in the clofibric acid group and slightly lower (approximately −10%) in the fasted group than the fed group, although none reached statistical significance (Table 2). Plasma 3-hydroxybutyrate and acetoacetate concentration was not significantly different among groups.

View this table:
Table 2.

Body weight, liver weight, liver glycogen, and blood ketone body concentrations in pigs

As shown in Fig. 1, hepatocytes from 24-h fasted pigs had dense eosinophilic cytoplasm with minimal or no vacuolation, whereas hepatocytes in fed pigs had clear, vacuolated cytoplasm, indicating accumulation of glycogen. Hepatocytes from clofibric acid-treated pigs had cytoplasmic characteristics midway between the other two groups, with partly vacuolated cytoplasm. Consistent with this observation, liver glycogen content in the fasted and clofibric acid groups was 3 and 31% of the fed group, respectively (Table 2).

Fig. 1.

Light microscopic image of pig liver sections stained with hematoxylin and eosin. Animals were fed either a control diet or a diet containing 0.5% clofibric acid for 1 wk. A and D: control animals fasted for 24 h; B and E: control animals in fed state; C and F: clofibric acid-treated animals in a fed state. A–C, periportal area; D–F, perivenous area. Scale bar, 25 μm. Extensive vacuolation in fed state of control animals (B, E) indicates accumulation of glycogens, whereas fasted liver (A, D) shows dense staining. Clofibric acid-fed group (C, F) shows less vacuolation than the control-fed group. Aside from the decreased vacuolation, little change was observed in clofibric acid-treated groups.

Other than the magnitude of vacuolation, pigs exposed to clofibric acid had few changes in hepatocyte morphology on light microscopic examination. No apparent peroxisome proliferation, hepatocyte hypertrophy, or hyperplasia was observed with hematoxylin-eosin (Fig. 1) or periodic acid-Schiff staining (data not shown).

Both Fasting and Clofibric Acid Induced Genes for Mitochondrial Fatty Acid Oxidation and Ketogenesis in Pig Liver

Tables 3 to 5 summarize changes in mRNA abundance in pig liver by fasting and clofibric acid administration relative to the control group. The results of microarray analysis were generally in good agreement with those of the real-time PCR assay.

View this table:
Table 3.

Expression of genes involved with fatty acid oxidation in pig liver

Table 3 shows genes related to fatty acid oxidation. Among mitochondrial enzymes, carnitine palmitoyltransferase 1 (CPT1A) and 3-hydroxy-3-methylglutaryl-CoA synthase 2 (HMGCS2), the rate-limiting enzymes for mitochondrial beta-oxidation and ketogenesis, respectively, were induced by both fasting and clofibric acid (Table 3). Other genes involved with mitochondrial fatty acid oxidation, such as hydroxyacyl-CoA dehydrogenase/3-ketoacyl-CoA thiolase/enoyl-CoA hydratase α (HADHA) and electron-transferring-flavoprotein dehydrogenase (ETFDH) were also induced by fasting and clofibric acid in a similar degree (Table 3). In contrast, no change was observed in genes involved with the citric acid cycle by either fasting or clofibric acid treatment, indicating a specific induction of genes for fatty acid oxidation in mitochondria by these treatments (Table 4).

View this table:
Table 4.

Expression of genes involved with citric acid cycle and glucose metabolism in pig liver

Little Induction of Genes for Peroxisomal and Microsomal Fatty Acid Oxidation by PPs in Pigs

In contrast to a strong induction by PPs in rats and mice, genes encoding peroxisomal fatty acid oxidation enzymes were induced less than twofold in all genes examined in pig liver, although most genes showed a trend of upregulation in a similar degree by both fasting and clofibric acid (Table 3). However, peroxisomal membrane protein 70 kDa (ABCD3), a long-chain fatty acid transporter in peroxisomal membrane (20), was induced fourfold by both fasting and clofibric acid (Table 3). In addition, a strong induction by both fasting and clofibric acid was observed in the mRNA of lactate dehydrogenase (LDH)-B, which is considered to facilitate shuttling NADH generated by peroxisomal fatty acid oxidation from peroxisome to mitochondria (4) (Table 3). Upregulation of ABCD3 and LDHB may increase flux of peroxisomal fatty acid oxidation without changes in β-oxidation enzymes in the same manner as the β-oxidation in mitochondria by upregulation of CPT-1. Catalase, a peroxisomal enzyme that degrades H2O2 produced by reactive oxygen species, was modestly induced by clofibric acid (2×), whereas it was significantly decreased during fasting (Table 3).

Cytochrome P450 IVA (CYP4A) catalyzes microsomal omega-oxidation of fatty acids (43). The induction of CYP4A mRNA by clofibric acid was less than twofold, whereas it was significantly decreased in the fasting state (Table 3).

Other Gene Groups Regulated by Fasting or Clofibric Acid

In general, fasting showed stronger effects than clofibric acid on genes involved with glucose metabolism in pig liver. Genes for gluconeogenesis, glucose-6-phosphatase, and fructose-1,6-bisphosphatase 2, were induced by fasting but not by clofibric acid (Table 4). Also, suppression of the glucokinase mRNA and induction of glycogen synthase by fasting was much stronger than by clofibric acid (Table 4). On the other hand, clofibric acid increased mRNA of pyruvate dehydrogenase kinase 4 (PDK4) and glycogen phosphorylase (PYGL), whereas neither gene was induced by fasting (Table 4).

Among other genes investigated in pig liver, clofibric acid and fasting showed an opposite effect on the mRNA expression of three acyl-CoA desaturases: stearoyl CoA desaturase (SCD), Δ6-desaturase and Δ5-desaturase (Table 5), enzymes that catalyze unsaturated fatty acid synthesis (37). Other genes involved with fatty acid metabolism were unchanged by either fasting or clofibric acid (Table 5). No change was observed in genes involved with the cholesterol metabolism by either fasting or clofibric acid treatment (Table 5). Ornithine aminotransferase, which degrades ornithine, was induced only by fasting (Table 5).

View this table:
Table 5.

Expression of genes involved with other metabolism in pig liver

Table 6 shows mRNA expression of selected genes in rats. ETFDH, catalase, and glucokinase showed similar response to pigs, whereas ABCD3 and PDK4 were strongly induced by a PP (>10×). PDK4 is an inducible isozyme that reduces glucose oxidation by inactivating pyruvate dehydrogenase. Strong induction of PDK4 by PPs is consistent with other studies (7–33×) (7, 26).

View this table:
Table 6.

Gene expression in rat liver

Copy Number of PPARα mRNA Was Higher in Pigs Than in Rats and Mice

Copy number of PPARα mRNA was higher in pigs than in rats and mice. Because much lower PPARα expression was reported in nonproliferating species, we compared the PPARα expression in livers from pigs, rats, and mice. As shown in Fig. 2, all three species show similar expression patterns, a higher expression in fasting than in fed and PP administration. Whereas the copy number of PPARα mRNA was nearly identical between rats and mice, it was two- to threefold higher in pigs than in rats and mice in all fed, fasted, and PP-treated conditions (Fig. 2).

Fig. 2.

Quantification of PPARα mRNA expression in livers from pigs, rats, and mice. Animal studies for pigs, rats, and mice are described in the experimental procedures. PPARα mRNA was measured using real-time quantitative PCR with SYBR Green fluorescence dye. Copy number of PPARα mRNA was determined by generating a standard curve for each species with a plasmid containing pig, rat, or mouse PPARα. 18S plasmids from each species were used to quantitate total RNA. *Significant difference compared with rat and mouse (P < 0.05) by Fisher's protected least significant difference post hoc test. PPs, peroxisome proliferators.


The present study revealed that the major species difference exists in the response of several peroxisomal and nonperoxisomal genes to PPs. These genes were hyperresponsive in rats and mice, whereas they are weakly responsive or unresponsive in pigs and likely in humans as well. Magnitudes of induction of these genes by PPs would explain the species differences in the development of peroxisome proliferation and hepatocarcinogenesis by PP administration.

Clofibric acid feeding for 1 wk did not increase liver weight of pigs (Table 2) as previously reported (41), and histological examination showed no sign of hypertrophy or hyperplasia (Fig. 1). In contrast, feeding PPs to rats and mice reproducibly increases liver weight by 50% or more in 1–2 wk (16, 22, 28) with marked peroxisome proliferation, hepatocyte hypertrophy and hyperplasia, eventually leading to hepatocarcinogenesis (12, 28). Gene expression patterns summarized in Fig. 3 include differential effects of PPs on peroxisomes between species. In mice and rats, PPs strongly induce peroxisomal genes, including ACOX1 (>10×) (7, 16, 28), enoyl-CoA hydratase/3-hydroxyacyl-CoA dehydrogenase (EHHADH) (40–150×) (1, 16, 28) and ABCD3 (13×) (Table 6). These genes that participate in peroxisomal β-oxidation were induced only modestly by clofibric acid in pig liver (Table 3). In addition to peroxisomal genes, our study also identified other genes that differ between species in the response to PPs and are involved in fatty acid metabolism. When peroxisome proliferators were administered to rats and mice, there is a drastic increase (>10×) in the expression of CYP4A (7, 16, 28), CD36 (7) and PDK4 (7, 26) (Table 6), whereas modest (PDK4) or no induction (CYP4A, CD36) was observed in pig liver by clofibric acid (Tables 3 and 4, Fig. 3). It is noteworthy that incubation of primary culture of human hepatocytes with PPs gave no induction of ACOX1, EHHADH, ABCD3, or CYP4A11 (27). A human hepatoma cell line, HepG2 cells express very low PPARα. When PPARα was expressed at higher levels by a stable transfection of the PPARα gene, HepG2 cells showed similar response to PPs to human hepatocytes (19, 27). The unresponsiveness of these genes to PPs in human hepatocytes and HepG2 cells is very similar to the gene expression pattern of pigs.

Fig. 3.

Summary of target genes induced by PPs and fasting in pigs, rats, and mice. Pigs largely share a common pattern of gene induction with rats and mice when animals are fasted or treated with PPs: 1) induction of genes for fatty acid oxidation by both fasting and PPs, 2) effects on glucose metabolism largely by fasting, and 3) induction of three desaturases by PPs. The major difference is a response of several genes (ABCD3, PDK4, ACOX, EHHADH, CYP4A, and CD36) to PPs: hyperresponsive in rats and mice, and unresponsive in pigs. Genes that share common response between pigs and rats/mice were shown in bold letters. Genes in shaded area markedly differ between species in the response to PPs. *Suppressed. Data on rats and mice come from Refs. 14, 7, 8, 16, 26, 28, 29, 32, 42, and 50.

With the exception of the peroxisomal and nonperoxisomal genes discussed above, pigs, mice, and rats largely share common gene expression patterns in response to fasting and PPs (Fig. 3). In the present study, both clofibric acid treatment and fasting induced the genes involved with mitochondrial fatty acid oxidation and ketogenesis in pig liver, including genes for key regulatory enzymes CPT1A and HMGCS2 (Table 3). This overlapping induction of mitochondrial genes by both PPs and fasting is also observed in rats and mice (Fig. 3). PPs increased mRNA of CPT1A (∼5×), HMGCS2 (∼5×), HADHA (∼3×), medium-chain acyl-CoA dehydrogenase (∼2×) (7, 16, 26), and ETFDH (Table 6) in rats and/or mice. Fasting also increased CPT1A (5–10×), HMGCS2 (∼5×), medium-chain acyl-CoA dehydrogenase (∼4×) (2, 16, 29), and ETFDH (Table 6). Importantly, studies with PPARα-null mice have demonstrated that PPARα plays an essential role in metabolic adaptation to fasting by inducing mRNA of these mitochondrial enzymes for fatty acid oxidation and ketogenesis during fasting (2, 17, 29). Therefore, the common response of these genes to fasting and PPs observed across species suggests that PPARα mediates metabolic adaptation to fasting in pigs, as well as rats and mice by inducing genes for mitochondrial β-oxidation and ketogenesis (Fig. 3). Moreover, in primary culture of human hepatocytes and in PPARα-expressed HepG2 cells, CPT1A and HMGCS2 were induced by PP treatment (19, 27), suggesting that PPARα mediates the induction of mitochondrial β-oxidation in human liver as well.

Very low expression of PPARα mRNA was reported in liver of other nonproliferating species (humans and guinea pigs) compared with rats and mice, and was suggested as a possible mechanism of differential response to PPs among species (9, 39, 47). On the contrary, our quantification of mRNA copy numbers revealed that the expression of PPARα mRNA was higher in pigs than in rats and mice (Fig. 2), suggesting that a degree of PPARα expression is not an underlying mechanism for the absence or low magnitude of induction of certain genes by clofibric acid in pigs. Species variations could simply account for the differences between the present study and previous reports in the degrees of PPARα expression among nonproliferative species. It should be noted, however, that all previous results (9, 39, 47) could be inconclusive because a cDNA clone was not used as a standard for the determination of copy numbers. The correction of probe/amplification efficiency with cDNA standard would be critical for a cross-species comparison that uses different probes/primers between species. Regarding the PPARα protein in human liver, both presence (39) and absence of the protein were reported (19). The absence of PPARα protein in human liver (19) was based on the determination in which human and mouse liver proteins were probed with an antibody against mouse PPARα. Purified human PPARα protein was not used as a reference. Induction of CPT1A and HMGCS2 in human hepatocytes in response to PPs suggests the presence of PPARα protein in human liver (27).

The mechanism that confers differential responses to peroxisome proliferators between proliferating and nonproliferating species is currently unknown. One likely mechanism is a differential regulation of individual genes. Woodyatt et al. reported that a putative peroxisome proliferator response element in the promoter of human ACOX was inactive (49). Other possible candidates that may be responsible for the species differences include truncated PPARα protein and differences in amino acid sequence in PPARα ligand-binding domain. An alternatively spliced PPARα mRNA that lacks ligand-binding domain was identified in tissues from humans (11) and pigs (44) but not from rats and mice. A cell culture study showed that the truncated human PPARα protein acted as a competitive inhibitor (11). A few critical amino acids of ligand-binding domain of PPARα differ between proliferating and nonproliferating species, although the amino acid sequence of the domain is, in general, well conserved across species (44). Mutating these amino acids in human PPARα to the mouse counterpart increases transactivation potency (36).

The present study also suggests an involvement of PPARα in the expression of various other genes in pigs as in rats and mice. Fatty acid desaturases are a unique group of genes that are induced by PPs but not by fasting in rats and mice (3, 16, 26, 32) (Fig. 3). We showed that mRNA expression of all three desaturases was significantly higher in clofibric acid-treated group than in fasting (Table 5, Fig. 3). The induction of desaturases by clofibric acid in pigs provides another support for the presence of functional PPARα in pigs. Although all three mammalian desaturases are dependent on sterol regulatory element binding protein-1 (32, 38, 45), studies with PPARα-null mouse have shown that the presence of PPARα is required for the induction of desaturases by PPs (13, 30). Functional peroxisome proliferator response element has been identified in promoters of mouse SCD-1 (35) and human Δ6-desaturase (46).

Liver glycogen content was low in both fasted and clofibric acid-treated pigs (Fig. 1, Table 2), suggesting involvement of PPARα in hepatic glycogen metabolism. Clofibrate treatment also lowered glycogen content in rat liver (14). This reduction in liver glycogen by fasting and PP administration in both pigs and rats is another common feature that may be mediated by PPARα, although the gene expression pattern does not indicate an underlying mechanism of this effect. Another common gene expression across species was observed in the genes for glucose metabolism. Fasting but not clofibric acid-induced genes for gluconeogenesis, such as glucose-6-phosphatase and fructose-1,6-bisphosphatase 2, and suppressed mRNA of glucokinase, the primary enzyme for glucose utilization in liver (Table 4, Fig. 3). In rats and mice, glucose-6-phosphatase was also induced only by fasting and not by PPs (2, 8, 26), and glucokinase was suppressed by fasting (2) (Table 6, Fig. 3). Although PPs did not induce/suppress these genes for glucose metabolism as fasting did, PPARα-null mice showed a blunted response of these genes to fasting (2), suggesting a permissive effect of PPARα on these genes during fasting.

ABCD3 and LDHB are among the genes with the highest induction by both fasting and clofibric acid in pig liver (Tables 3 and 4, Fig. 3). Overexpression of ABCD3 in the peroxisomes of Chinese hamster ovary cells increases oxidation of long-chain fatty acid (16:0) but not very long-chain fatty acid (24:0), suggesting that ABCD3 is a long-chain specific fatty acid transporter and regulates oxidation rate of fatty acids in peroxisomes (20). LDH isozymes form tetramers LDH-A4 and LDH-A3B in rat liver, and the latter form is primarily present in peroxisomes (4). Treating rats with PPs increases the protein level and activity of LDH in rat liver homogenate and peroxisomes (4, 42). Because the citric acid cycle is absent in peroxisomes, lactate shuttles the reducing equivalent generated by fatty acid oxidation from peroxisomes to mitochondria (33). Therefore, increased intraperoxisomal LDH is likely to facilitate peroxisomal β-oxidation by regenerating NAD. Induction of ABCD3 and LDHB by both PPs and fasting further supports the role of PPARα in the adaptation to fasting in pigs (Fig. 3).

In conclusion, we have presented several lines of evidence that support the presence of functional PPARα and its active role in metabolic adaptation to fasting in pigs. Pigs share the following characteristics with rats and mice, in which physiological roles of PPARα are well established: 1) genes for mitochondrial fatty acid oxidation and ketogenesis were induced by both fasting and clofibric acid, 2) PPARα mRNA is abundantly expressed in liver, 3) all three fatty acid desaturases were induced by clofibric acid but not by fasting, and 4) liver glycogen was decreased by both clofibric acid and fasting. Moreover, genes that potentially increase the flux of long-chain fatty acid oxidation in peroxisomes were also induced by both fasting and clofibric acid in pigs. Although pigs, rats, and mice largely share responses to fasting and PPs, several peroxisomal and nonperoxisomal genes that are markedly induced by PPs in rats and mice were weakly induced or nonresponsive in pigs. This hyperresponse of a set of genes to PPs, not abundance of PPARα, is the likely cause of peroxisome proliferation and hepatocarcinogenesis seen only in rats and mice. A similarity in the gene response to PPs between pig liver and human hepatocytes/hepatoma cells suggests that PPARα also has a functional role in adaptation to fasting in humans.


This study is supported in part by Scientist Development Grant from American Heart Association and USDA Hatch Funds to M. T. Nakamura.


  • The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.


View Abstract