 |
INTRODUCTION |
A variety of compounds designated as peroxisome proliferators
elicit pathological changes in the livers of sensitive species. The
effects of exposure include the proliferation of peroxisomes, hepatomegaly, and the development of carcinomas (1-4). Peroxisome proliferation follows the extensive induction of peroxisomal and microsomal enzymes that are involved in fatty acid oxidation, and it is
thought that increased production of reactive oxygen species generated
by these enzymes could contribute to the carcinogenicity of peroxisome
proliferators seen in rats and mice. In contrast to sensitive species,
guinea pigs and primates appear to be resistant to the adverse
consequences of peroxisome proliferator exposure, whereas other species
exhibit intermediate effects (5, 6). Unlike isolated hepatocytes and
tumor-derived liver cell lines from rats and mice, human hepatocytes do
not exhibit peroxisome proliferator-dependent peroxisome
proliferation or induction of the peroxisomal and microsomal enzymes
that contribute to fatty acid oxidation (7-10).
In mice, the targeted disruption of the gene for the peroxisome
proliferator-activated receptor 
(PPAR)1 prevents the
pathological changes and carcinogenicity resulting from peroxisome
proliferator exposure (11, 12). PPAR is a ligand-activated
transcription factor of the nuclear hormone receptor superfamily that
binds as a heterodimer with the retinoid X receptor to specific
enhancer elements in target genes. Response elements for PPAR have
been identified in the genes encoding the microsomal P450
-hydroxylases and components of the peroxisomal long chain fatty
acid 
-oxidation pathway including the long chain fatty acyl-CoA
oxidase (ACO1), the enoyl-CoA hydratase (ECH1), and the 3-oxoacyl-CoA
thiolase (ACAA1) (reviewed in Ref. 13). These enzymes are induced
>10-fold by peroxisome proliferators in rats and mice (reviewed in
Ref. 14). However, it is not known if these genes are regulated by
PPAR in humans.
Fatty acids and peroxisome proliferators are ligand agonists for
PPAR (15, 16), and there is increasing evidence that PPAR plays a
central role in fatty acid homeostasis by regulating the degradation of
fatty acids by mitochondrial as well as by peroxisomal and microsomal
fatty acid oxidation (reviewed in Ref. 13). In addition, PPAR
contributes to the maintenance of energy balance by regulating the
expression of enzymes that participate in mitochondrial fatty acid
oxidation and the formation of ketone bodies from fatty acids in the
fasted state (17-19). The potential role of PPAR in the regulation
of mitochondrial capacity for fatty acid oxidation and ketogenesis in
human liver has not been adequately addressed.
Several studies indicate that the expression level of PPAR in human
liver (20, 21) is much lower than in mouse liver, and it has been
suggested that this difference could contribute to the lack of
peroxisome proliferation and subsequent pathologic effects. PPAR
agonists are used therapeutically for lowering serum triglycerides, and
these hypolipidemic effects observed in humans are known to be mediated
by PPAR in other species (22), suggesting a role for PPAR in the
regulation of at least some aspects of lipid metabolism in humans (3).
The differences between rodents and humans in response to peroxisome
proliferator exposure could reflect differences between species in the
spectrum of target genes regulated by PPAR. However, the
identification of human target genes has been hampered by the lack of
detectable PPAR-mediated responses in isolated human hepatocytes or
in liver-derived cell lines.
The present study was designed to test whether elevated expression of
PPAR in the human hepatoma-derived cell line, HepG2, would unmask
PPAR-responsive human genes. Several stable cell lines were
generated by transfection of HepG2 cells with human PPAR expression
vectors. HepG2 cells were selected because they have been studied
extensively in relation to lipoprotein metabolism and fatty acid
oxidation. Although peroxisomes are present in HepG2 cells and can
oxidize fatty acids, treatment with peroxisome proliferators does not
induce peroxisomal fatty acid oxidation in this cell line (8, 23). The
level of PPAR mRNA expression in HepG2 cells is roughly 3-fold
lower than the values seen for human liver (20), and transient
transfection studies in HepG2 cells using reporter genes under the
control of PPAR response elements (PPREs) indicate that significant
peroxisome proliferator-dependent reporter expression is
evident only when a PPAR expression vector is co-transfected with
the reporter construct (20, 24).
Stable transformants expressing the mouse PPAR-E282G mutant were
also generated. This mutant contains a glycine substitution for
glutamate at position 282, which results in a significantly reduced
ligand-independent transactivation relative to the wild type receptor
(25) and a reduced affinity for a number of ligands (15, 26). However,
PPAR-E282G exhibits similar efficacy as the wild type when activated
by PPAR agonists such as Wy14643 (25). Therefore, this mutant
provides a clearer indication of peroxisome
proliferator-dependent gene regulation than may be evident
using wild-type human PPAR.
The results presented here demonstrate that elevated expression of
PPAR in HepG2 cells increases the expression of genes encoding
several enzymes that catalyze branch points or rate-limiting steps in
the utilization of fatty acids for the formation of ketone bodies. The
PPAR-responsive human genes include the long chain fatty acyl-CoA
synthetase (ACS), carnitine palmitoyl acyl-CoA transferase 1A (CPT1A),
and the mitochondrial HMG-CoA synthase (HMGCS2). In contrast, the
expression of the rate-limiting enzyme of the peroxisomal fatty acid
-oxidation pathway, ACO1, and other components of this
pathway do not appear to be affected by increased PPAR levels.
| |
MATERIALS AND METHODS |
Generation of Stable Cell Lines Expressing PPAR--
HepG2
cells were obtained from the American Type Culture Collection
(Manassas, VA) and maintained in Dulbecco's modified Eagle's medium
(Life Technologies, Inc.) containing 10% fetal bovine serum (Summit,
Ft. Collins, CO). The human PPAR or mouse PPAR-E282G cDNAs
were inserted into the pcDNA3.1 vector (InVitrogen, Carlsbad, CA).
The PPAR expression vectors and empty pcDNA3.1 plasmid were each
transfected into HepG2 cells using a modified calcium phosphate co-precipitation method (27). Geneticin (G418) (Life Technologies) was
used at a final concentration of 500 µg/ml to select and maintain positive transformants. The cells were given fresh medium
containing G418 every 2-3 days to ensure that an adequate active
concentration of G418 was present in the culture medium. A microplate
dilution technique was performed to purify transformants. In order to
simplify the clone designations throughout, the following nomenclature has been used: hPPAR-expressing cell lines 251 and 252 are
designated as hPPAR1 and hPPAR2, respectively. The mouse
PPAR-E282G-expressing transformants 15 and 16-13 are
designated as mPPARG1 and mPPARG2. The wild type HepG2 cells (not
transfected) are designated as HepG2, and cells transfected with empty
pcDNA3.1 vector that are resistant to G418 are designated as
pcDNA throughout.
Transient Transfections--
Upper and lower oligonucleotides
that had overlapping, complementary 3'-ends and that together
encompassed nucleotides 1302-1440 of the reported promoter sequence
for the human HMGCS2 gene (GenBankTM accession number
U81851) were annealed, extended, and amplified using Pfu DNA
polymerase (Stratagene, La Jolla, CA). The resulting PCR fragment was
gel-purified and inserted into the SmaI site of p19dLuc to
generate the pLuc-HMGP reporter construct. For the reporter constructs
containing mutations, pLuc-HMGM1 and pLuc-HMGM2, oligonucleotides
containing specific mutations that disrupt either the putative PPAR
A/T-rich binding site or the RXR binding site (Fig. 8A) were
used as described above. The reporter construct containing the
heterologous Herpes simplex thymidine kinase promoter, pLuc-TK-HMG, was
generated by inserting annealed, complementary oligonucleotides
corresponding to the putative human HMGCS2 PPRE (upper strand,
agctCTAAAACTGGGTCAAAGGGCTCAC; lower strand,
agctGTGAGCCCTTTGACCCAGTTTTAG) into the HindIII site of
pLuc-TK. Lowercase letters indicate single strand overhangs used for
cloning. The mutant reporter construct pLuc-TK-HMGM was generated as
above using the human HMGCS2 mutant element (Fig. 8B). The
authenticity of the constructs was verified by sequence analysis. The
pLuc-TK-ACO-AB, pLuc-TK-Z, or pLuc-TK reporter constructs have been
described previously (27). Each luciferase reporter construct was
transfected (2 µg/well for 24-well plates; 5 µg/well for six-well
plates) into HepG2 cells, pcDNA control, or PPAR-transformed cell
lines using a modified calcium phosphate co-precipitation method (27).
A -galactosidase reporter construct (pCMV-;
CLONTECH, Palo Alto, CA) was co-transfected (0.2 µg/well for 24-well plates or 0.5 µg/well for six-well plates) and
used to monitor transfection efficiency. After 16-24 h of incubation
with DNA, the cells were washed with Dulbecco's modified Eagle's
medium and incubated with fresh medium containing 500 µg/ml G418 and
either 50 µM Wy14643 or an equivalent volume of vehicle
(EtOH) for another 20-24 h. Luciferase and -galactosidase activities were then analyzed as described (27). The luciferase activity for each well was normalized to the -galactosidase activity obtained from the same well. Triplicate transfections were carried out
for each treatment in each independent transfection experiment, and at
least three separate transfection experiments were performed for each
cell line.
Preparation of Cell Lysates for Western Blots and Enzyme
Assays--
HepG2 cells and pcDNA or PPAR transformants were each
seeded into T75 flasks in Dulbecco's modified Eagle's medium at a
cell density that would reach ~60% confluence after overnight
culture. After 24 h, the cells were treated with 50 µM Wy14643 or EtOH. The medium was replaced with fresh
medium containing Wy14643 or solvent alone after another 24 h.
After exposure to drug for 48 h, the cells were washed with
Dulbecco's phosphate-buffered saline (Life Technologies) and
trypsinized. For SDS-polyacrylamide gel electrophoresis and Western
blots, the trypsinized cells were suspended in phosphate-buffered
saline containing 0.1 mM EDTA and 10% glycerol. Complete
protease inhibitor mixture (Roche Molecular Biochemicals) was included
prior to sonication twice for 15 s each at 4 °C. The sonicated
lysates were treated with 0.1% Nonidet P-40, and following a 30-min
incubation, they were first centrifuged at 15,000 × g
for 10 min and then centrifuged at 235,000 × g for 30 min at 4 °C. The final supernatant was used for immunoblot analysis.
Protein concentrations were determined with the Pierce BCA reagent
using bovine serum albumin as the standard. Cell lysates (100 µg of
protein/sample) were loaded onto 8% SDS-polyacrylamide gels,
electrophoresed, and transferred electrophoretically to nitrocellulose
membranes. The level of PPAR expression was quantified by developing
the transfer blot sequentially with rabbit anti-mouse PPAR (25) and
125I-labeled Protein A (ICN, Costa Mesa, CA) prior to
analysis using a PhosphorImager SI (Molecular Dynamics, Inc.,
Sunnyvale, CA).
For enzyme activity assays, trypsinized cells were suspended in 10 mM Tris, pH 7.4, containing 0.25 M sucrose, 1 mM EDTA, 2 µM phenylmethylsulfonyl fluoride,
and 2 µM 4-(2-aminoethyl)-benzenesulfonyl fluoride
hydrochloride. Whole cell homogenate was obtained using a Radnoti glass
wall tissue grinder (Radnoti, Monrovia, CA). After 20 up-down strokes,
the homogenate was centrifuged at 550 × g for 5 min at
4 °C to remove unbroken cells and nuclei. The resulting supernatant
represents a "crude lysate" that was aliquoted and stored at
80 °C until analysis. The peroxisomal palmitoyl-CoA oxidase
activity in the crude lysate was determined by using palmitoyl coenzyme
A ([1-14C]palmitoyl-labeled) (55 mCi/mmol)
(American Radiolabeled Chemical, St. Louis, MO) as the substrate in the
presence of 2 mM KCN as described by Lazarow (28). The
mitochondrial HMG-CoA synthase activity in crude lysates was analyzed
using the method described by Quant et al. (29) with minor
modifications. The buffer system used was modified to 50 mM
Tris, pH 8.2, and 1 mM EDTA rather than 50 mM
Tris, pH 8.0, with 10 mM MgCl2 in order to
minimize the contribution of cytosolic HMG-CoA synthase activity (30). The absorption coefficient of acetoacetyl-CoA is 3.6 × 103 M
1
cm1 under these conditions.
Reverse Transcription and PCR Amplification--
HepG2 cells and
the stable transformants were cultured and treated with EtOH or 50 µM Wy14643 for 48 h as described above. Total RNA
was isolated from cells using the RNAgents RNA purification kit
(Promega, Madison, WI), and 10 µg of total RNA were utilized for
reverse transcription using a combination of random and oligo(dT) primers (200 ng each) with Moloney murine leukemia virus reverse transcriptase (Stratagene, La Jolla, CA). For PCR, specific primer sets
were synthesized (Table I) and used for assessment of PPAR and
PPAR target gene mRNA levels. As the predicted size of the amplified cDNAs (300-500 nucleotides) and the
Tm values for the primers were similar, uniform
conditions were applied for the amplification of all cDNAs. The
amplification conditions were as follows: one cycle of denaturing at
94 °C for 2 min, annealing at 56 °C for 1 min 30 s, and
extension at 72 °C for 1 min. This was followed by 25 cycles of
denaturing at 94 °C for 30 s, annealing at 56 °C for 1 min
30 s, and extension at 72 °C for 1 min. The reaction products
were analyzed on 1% agarose gels and visualized by ethidium bromide staining.
RNase Protection Assays--
PCR-amplified cDNA fragments
corresponding to PPAR and target genes were subcloned into
pBS-KS(+), sequenced to confirm their authenticity, and used to
generate antisense riboprobes. The human -actin probe used as an
internal control has been described previously (20). The region that is
spanned by each riboprobe corresponds to the primer set location as
listed in Table I. Transcription was
carried out in the presence of [32P]CTP (20 mCi/ml, 800 Ci/mmol) using either T3 or T7 RNA polymerase (Stratagene, La Jolla,
CA). Following in vitro transcription, the reaction mixture
was electrophoresed through 5% acrylamide, 6 M urea gels,
and the probes were retrieved from gel slices with elution buffer
(RPAII kit, Ambion, Austin, TX). For hybridization, 10 µg (target
genes) or 20 µg (PPAR) of total RNA were combined with 4 × 104 cpm of gel-purified specific probe and 2 × 104 cpm of -actin probe that was generated at a lower
specific content. The samples were incubated at 42 °C for 18 h
and then subjected to RNase digestion for 45 min at 37 °C using a
33-fold dilution of RNase A/RNase T1 mix (RPAII kit; Ambion). Protected
fragments were separated on 5% acrylamide, 6 M urea gels
and analyzed with a Molecular Dynamics PhosphorImager.
Electrophoretic Mobility Shift Assay
(EMSA)--
Escherichia coli lysates containing expressed
mPPAR and RXR-MBP (31) were analyzed for binding to
32P-labeled double-stranded oligonucleotides by EMSA as
previously described (27). When supershift assays were performed, 1 µl of anti-mPPAR serum or preimmune serum was added to the
reaction. After a 30-min incubation at room temperature, the reactions
were loaded onto a 4% polyacrylamide (37.5:1) gel containing 0.5× TBE and 1.25% glycerol, and the samples were electrophoresed at 16 mA for
100 min at room temperature. The dried gel was analyzed using a
Molecular Dynamics PhosphorImager, model SI.
| |
RESULTS |
Expression of PPAR in Stable Transformants of HepG2
Cells--
HepG2 cells were transfected with either the pcDNA3.1
expression vector or with the vector harboring the cDNA for either
hPPAR or mPPAR-E282G. The PPAR transformants were enriched by
G418 selection. Several cell lines were cloned from pools of hPPAR transformants by dilution in microplates. Individual wells of the
microplate were examined microscopically to identify those containing
single cells that were then expanded to produce cloned cell lines. The
level of transactivation of a luciferase reporter construct containing
the PPRE from the rat acyl-CoA oxidase gene and the herpes simplex
thymidine kinase promoter, pLuc-TK-ACO-AB, provided an initial
assessment of increased PPAR expression levels in each cell line.
Several stable lines that exhibited significant transactivation of the
transiently transfected reporter relative to both HepG2 and pcDNA
controls were selected for further analysis. The results of the
transient transfection assay for two examples of human PPAR clones
are shown in Fig. 1. Both of the hPPAR transformants exhibit relatively strong ligand-independent
transactivation of the transiently transfected reporter construct,
which is consistent with the phenotype exhibited by the human receptor
in transient co-transfection assays with the reporter (20). This
transactivation is enhanced by roughly 2-fold or less when Wy14643, a
PPAR agonist, is included in the culture medium. No effect is seen
for the reporter construct lacking the PPRE, pLuc-TK. A significantly
lower ligand-independent transactivation of the reporter construct is
evident for mPPARG1 cells; however, these cells exhibit similar
levels of ligand-dependent activation of reporter
transcription to those seen for the hPPAR transformants (Fig. 1). As
a result, Wy14643 treatment of the mPPARG1 cells elicits a 30-fold
response in the expression of the reporter compared with the roughly
2-fold or lower ligand-dependent responses seen for the
hPPAR transformants. These differences are consistent with the
results of previous comparisons between the mutant and wild-type PPAR
phenotypes (25).
View larger version (17K):
[in this window]
[in a new window]
|
Fig. 1.
Characterization of PPAR
stable transformants of HepG2 cells by transient transfection of
luciferase reporter constructs. A luciferase reporter construct
containing the rat acyl-CoA oxidase PPRE (AB) or the same
vector lacking the PPRE (TK) was co-transfected with the
pCMV--gal vector into either wild type HepG2 cells
(HepG2), pcDNA control cells (pcDNA),
human PPAR-overexpressing stable clones
(hPPAR1 and hPPAR2),
or mouse PPARG1 cells. Cells were then treated with either the PPAR
agonist Wy14643 (filled bars) or ethanol vehicle
(open bars). The expression of -galactosidase
activity was used to monitor the transfection efficiency as described
under "Materials and Methods." The results represent triplicate
determinations from three independent transfections. The
error bars indicate S.D. from the mean. The
average normalized luciferase data was expressed relative to the data
obtained with each cell line transfected with the pLuc-TK control
reporter and treated with ethanol.
|
|
PPAR protein is readily detected by immunoblotting in those cell
lines stably expressing PPAR (Fig. 2).
In contrast, PPAR was not detected under these conditions using
equivalent amounts of protein lysate from the control cells or from
human liver. The amount of protein detected by immunoblotting in the
hPPAR and mPPARG transformants is similar to that seen for
equivalent amounts of protein lysate prepared from mouse liver.
View larger version (26K):
[in this window]
[in a new window]
|
Fig. 2.
PPAR protein
expression in HepG2 PPAR stable
transformants. Cell lysates were prepared as described under
"Materials and Methods" and subjected to immunoblot analysis.
A, proteins were separated on an 8% SDS-polyacrylamide gel
(100 µg of protein/lane) and transferred to nitrocellulose. Bound
antibody against mouse PPAR was visualized using
125I-labeled Protein A, and the immunoblots were analyzed
using a PhosphorImager. The apparent molecular weight of the
immunoreactive protein corresponded with either in vitro
translated or bacterially expressed mPPAR (not shown) included in
the same blot. B, the histogram displays the amount of
PPAR protein detected. The results are expressed relative to the
value obtained from the sample designated mouse liver 1.
|
|
Elevated PPAR Expression Elicits Increased Transcription of
Endogenous Genes--
Peroxisome proliferators have been shown to
regulate the expression of HMGCS2, CPT1A, and ACS in other
species (32-34). In the PPAR stable transformants, the most
dramatic effect was seen on the expression of HMGCS2. In HepG2 and
pcDNA control cells, mRNA corresponding to HMGCS2 was not
observed by RT-PCR or RNase protection assays and remained undetectable
following Wy14643 treatment (Fig. 3). In
contrast, HMGCS2 mRNA was readily detectable in the
hPPAR-overexpressing transformants, and the mRNA levels were
further increased 2-fold by Wy14643 treatment. Increased amounts of
HMGCS2 mRNA were also observed in the mPPARG1 transformant following treatment with Wy14643, whereas little or no mRNA was detected in the absence of the agonist (Fig. 3), which is consistent with the phenotypic differences between wild-type PPAR and
mPPAR-E282G. Increased HMGCS2 enzyme activity was also observed in
hPPAR-overexpressing transformants relative to HepG2 and pcDNA
control cells (Fig. 4). In the absence of
Wy14643, the enzyme activity obtained from mPPARG1 cell lysates was
similar to the basal levels observed in HepG2 and pcDNA transformed
cells. However, a significant induction of HMGCS2 enzyme activity was
observed when the mPPARG1 cells were exposed to Wy14643. These
results indicate that elevations of HMGCS2 mRNA levels lead to
increases in HMGCS2 enzyme activity.
View larger version (42K):
[in this window]
[in a new window]
|
Fig. 3.
Expression of mitochondrial HMG-CoA synthase
mRNA in HepG2 PPAR stable
transformants. A, RNase protection assay. Total RNA
samples (10 µg) isolated from cells treated with Wy14643 (+) or the
solvent alone, ethanol (), were hybridized to antisense riboprobes
specific for human HMGCS2 and -actin prior to digestion with RNase.
The protected fragments were separated on 5% acrylamide, 6 M urea gels and analyzed using a PhosphorImager.
B, the histogram shows the average HMGCS2 to -actin
ratios from RNase protection assays carried out on three independent
preparations of RNA. The error bars indicate the
S.D. from the mean. C, RT-PCR amplification of HMGCS2
cDNA from total RNA. The products correspond to the predicted size
of 391 base pairs.
|
|
View larger version (12K):
[in this window]
[in a new window]
|
Fig. 4.
Rates of HMGCS2 enzyme activity in
PPAR transformants. HMGCS2 enzyme
activity was determined from three or more independent cell lysate
preparations obtained from cultures treated with Wy14643
(filled bars) or with ethanol vehicle
(open bars). *, a statistically significant
difference (p < 0.01) relative to pcDNA control
cells treated with ethanol using a one-tailed Student's t
test.
|
|
CPT1A and ACS mRNA levels are also elevated in the PPAR
transformants. The expression of CPT1A mRNA was elevated 2-4-fold in the hPPAR transformants and in the mPPARG1 transformant
following treatment with Wy14643 as
judged by RNase protection assays (Table II and Fig.
5). These results suggest that the human
CPT1A gene is also PPAR-responsive. Human ACS mRNAs were
elevated 3-8-fold (Table II) in the hPPAR transformants and in the
mPPARG1 transformant following exposure to Wy14643 (Fig. 5). These
results demonstrate that human genes encoding enzymes that catalyze
branch points or rate-limiting steps in the production of ketone bodies
from fatty acid oxidation during starvation have retained PPAR
responsiveness.
View larger version (13K):
[in this window]
[in a new window]
|
Fig. 5.
The effect of Wy14643 on the transcription of
selected target genes in the mPPARG1
transformant. RNase protection assays were performed on 10 µg of
total RNA isolated from the mPPARG1 transformant using probes
specific for ACAA1, ACO1, ACS, CPT1A, ECH1, or FABP1. RNA was obtained
from cultures grown in the presence (filled bars)
or absence (open bars) of Wy14643. The values
shown are the average of results obtained for three independent RNA
preparations and are expressed relative to the values obtained for the
pcDNA control treated with solvent (EtOH). The error
bars indicate the S.D. from the mean. *, a statistical
significance of p < 0.01 based on comparison of the
value obtained in the presence of Wy14643 with that of the solvent
control using a one-tailed Student's t test.
|
|
The effect of elevated PPAR levels on the expression of human genes
encoding the liver fatty acid-binding protein (FABP1) and components of
the peroxisomal and microsomal fatty acid oxidation pathways was also
examined, since these pathways have been shown to be regulated by
PPAR in other species. The rate-limiting enzyme in the peroxisomal
-oxidation of long chain, saturated fatty acids is ACO1, one of two
fatty acyl-CoA oxidases in peroxisomes. The other, ACO2, oxidizes
branched chain fatty acids. The levels of the ACO1 mRNA in the
human PPAR transformants were elevated roughly 2-fold over the
control cells (Table II), and further increases were not apparent after
treatment with Wy14643. In contrast, the ACO1 mRNA levels were not
elevated in the mPPARG1 transformant even after Wy14643 treatment
(Fig. 5). These results suggest that the modest increase in ACO1
mRNA levels that are not enhanced by agonist may reflect an
indirect effect of higher PPAR levels. When ACO2 mRNA levels
were assessed, no significant differences were observed between control
cells and the PPAR transformants (results not shown). The second
enzyme in the peroxisomal long chain fatty acid oxidation pathway,
ECH1, was only marginally elevated in the hPPAR transformants (about
2-fold; Table II), and no effect was seen in the mPPARG1
transformant in the presence or absence of Wy14643 (Fig. 5).
Significant elevations of mRNAs encoding the third enzyme of the
peroxisomal fatty acid pathway, ACAA1, were not seen. The level of
human FABP1 mRNA was elevated ~2-3-fold in the hPPAR
transformants relative to the levels seen in HepG2 cells and the
pcDNA control transformant (Table II). However, FABP1 mRNA
levels were not significantly elevated in the mPPARG1 transformant
or increased by treatment with Wy14643 (Fig. 5).
PPAR has also been shown to regulate the expression of genes
encoding the microsomal cytochrome P450 fatty acid -hydroxylases of
the 4A subfamily in other species (35, 36). To date, only one human
P450 4A, CYP4A11, has been characterized. RT-PCR analysis did not
detect the mRNA for this enzyme in HepG2 cells or in the PPAR
transformants in the presence or absence of Wy14643. Recently, a
variant CYP4A11 gene was identified in the human genome that is thought
to be a polymorphic form (37). A specific primer set for this new
sequence was also tested. Messenger RNA corresponding to this putative
variant was not detected. Thus, we could not assess whether CYP4A11
is regulated by PPAR.
The peroxisomal palmitoyl CoA oxidase activity in cell lysates prepared
from the cell lines was also characterized (Fig.
6). The rate of cyanide-insensitive fatty
acid oxidation was 2-3-fold higher in both hPPAR cell lines that
exhibited elevated levels of the ACO1 mRNAs. However, no increases
in enzyme activity were seen when the cells were treated with Wy14643.
In addition, no increases over control cells were seen for the
mPPARG1 cells in the presence or absence of Wy14643.
View larger version (16K):
[in this window]
[in a new window]
|
Fig. 6.
Rates of peroxisomal fatty acid
-oxidation in the PPAR
transformants. The catalytic rate of palmitoyl-CoA oxidase
was determined for three independent preparations of cell lysates
obtained from the indicated cell lines. *, rates that differ
significantly (p < 0.01) from the rates obtained for
the pcDNA control treated with solvent (EtOH) using a one-tailed
Student's t test.
|
|
Although the hypolipidemic effect of PPAR agonists is largely
thought to reflect the induction of lipoprotein lipase in extrahepatic tissues, studies using isolated, perfused rat livers indicate that the
secretion of triglycerides from liver decreases when ketogenesis
increases in response to clofibrate. The reciprocal relationship
between ketogenesis and triglyceride secretion was attributed to a
reduction in the availability of fatty acids for triglyceride synthesis
resulting from increased oxidation of fatty acids (38). Thus, changes
in the secretion of triglycerides from the liver could also contribute
to the lipid-lowering effects of fibrates. The observation that PPAR
regulates the expression of mitochondrial fatty acid oxidation and
ketogenesis suggests that this pathway could contribute to the
hypolipidemic effects of PPAR agonists in humans.
The changes in liver secretion of triglyceride elicited by clofibrate
could also reflect direct regulation of triglyceride and lipoprotein
secretion by PPAR in the liver. PPAR agonists are reported to
modestly induce apolipoprotein A-I and A-II mRNAs and suppress
apolipoprotein C-III mRNAs in human hepatocytes (39-41). Resulting
changes in the composition of lipoprotein particles are also likely to
increase the catabolism of triglycerides in extrahepatic tissues (42,
43). RT-PCR analysis using a range of conditions did not detect
significant differences in the mRNA levels for these
apolipoproteins between the HepG2 or pcDNA controls and the
hPPAR or mPPARG transformants. Also, the addition of Wy14643 to
the cultures did not have a significant impact on mRNA levels (data
not shown). RNase protection assays confirmed the results of the RT-PCR
study for apolipoprotein CIII (Table II). The suppression of
apolipoprotein C-III mRNA levels by fibrates in HepG2 cells was
reported to be dependent on the serum used in the culture medium (44),
and it is possible that different culture conditions might confer a
Wy14643 dependence to the suppression of the apolipoprotein CIII RNAs.
Additionally, RT-PCR assays (data not shown) did not demonstrate
noticeable differences for mRNAs encoding the mitochondrial medium
chain acyl-CoA dehydrogenase or the stearoyl-CoA desaturase 1, which
have been reported to be regulated by PPAR in other species (45,
46). The lipogenic malic enzyme gene is regulated by PPAR in rats
(47). Similar to the results obtained for CYP4A11, mRNA encoding
the cytosolic malic enzyme could not be detected in HepG2 cells or the
PPAR transformants by RT-PCR, precluding an assessment of the
responsiveness of this gene to PPAR regulation in humans.
Identification of a PPRE in the Proximal Promoter of the Human
HMGCS2 Gene--
The proximal promoter of the rat HMGCS2 gene contains
a functional PPRE (33). Examination of the corresponding region of the
human HMGCS2 gene indicated that the essential features of this element
were likely to be conserved (Fig. 7).
EMSA experiments performed with this putative human PPRE demonstrated
PPAR/RXR binding. Inclusion of anti-PPAR antiserum in
the EMSA resulted in nearly complete supershifting of the complexes.
Double-stranded oligonucleotides corresponding to the human HMGCS2 PPRE
that contained mutations were also synthesized. Based on previous
studies (27, 31), disruption of the A/T-rich region in the 5'-end of
the extended PPRE abrogates PPAR binding and transcriptional
activation. As expected, these mutations to the human HMGCS2 PPRE
abolished PPAR/RXR heterodimer binding (Fig. 7). Mutations introduced
into the second half site of hHMGCS2 PPRE also interfered with
PPAR/RXR binding (data not shown). These results suggest that a
functional PPRE may be present in the human HMGCS2 gene.
View larger version (56K):
[in this window]
[in a new window]
|
Fig. 7.
Rat and human mitochondrial HMG-CoA synthase
genes exhibit a conserved PPRE in the proximal promoter region.
A, a putative PPAR binding site located between nucleotides
1323 and 1340 based on the numbering given in the GenBankTM
sequence, accession number U81851, of the human HMG-CoA synthase gene
(170 to 153 nucleotides relative to the start site of translation)
was identified by comparison with the sequence of the PPRE in the
promotor of the rat HMGCS2 gene (rHMGCS2) (33). The oligonucleotides
used in the EMSA experiments are aligned for comparison.
Lowercase letters indicate mutations introduced
in the hHMGCS2m PPRE. A PPRE consensus sequence and the CYP4A6 PPRE are
shown for comparison. B, EMSA. Lysates were prepared from
Escherichia coli expressing either mPPAR or an
hRXR-maltose-binding fusion protein. The lysates containing PPAR and
RXR were mixed and incubated with 10 fmol of 32P-labeled,
double-stranded oligonucleotide either with (+) or without ()
preimmune serum (preimmune) or anti-PPAR antibody (immune) at room
temperature for 30 min before loading onto a 4% acrylamide gel as
described under "Materials and Methods." The mobility of
PPAR/RXR-DNA complexes is indicated at the left. PPAR/RXR
heterodimers clearly bind to the hHMGCS2 PPRE and can be supershifted
by antibody against PPAR. Mutations in the 5' extended binding site
immediately 5' of the two half-sites (hHMGCS2m) abolishes
binding.
|
|
A reporter construct containing a portion of the human HMGCS2 proximal
promoter region that encompassed the putative PPRE and included the
transcription start site was constructed based on the deletion analysis
that identified a PPRE in the rat HMGCS2 gene (33). This study
indicated that a portion of the rat promoter from 116 to +28
exhibited extensive transactivation by PPAR. The corresponding
region of the human HMGCS2 promoter, nucleotides 130 to +8, was
inserted into the p19dLuc vector. When this reporter construct was
transiently transfected into hPPAR-overexpressing cells, significant
transactivation was observed (Fig.
8A). This transactivation was
further enhanced when Wy14643 was included in the culture medium. A
clear ligand-dependent response was also observed for the
mPPARG transformant. As expected, the mutations introduced into the
A/T-rich region of HMGCS2 PPRE or the RXR binding site, as described
above in the EMSA binding experiments, abolished PPAR-mediated
transactivation when introduced into the HMGCS2 promoter (Fig.
8A). When the putative HMGCS2 PPRE was inserted upstream of
the herpes simplex thymidine kinase promoter and luciferase reporter
gene (pLuc-TK-HMG) and transiently transfected into cells
overexpressing hPPAR or mPPARG, the PPRE conferred PPAR- and
Wy14643-dependent transactivation to the heterologous promoter (Fig. 8B). As expected, the mutations introduced
into the A/T-rich region of HMGCS2 PPRE (pLuc-TK-HMGM) abolished
PPAR-mediated transactivation (Fig. 8B). These results
indicate that the human HMGCS2 gene contains a functional PPRE in the
proximal promoter region.
View larger version (34K):
[in this window]
[in a new window]
|
Fig. 8.
The putative hHMGCS2 PPRE confers
PPAR responsiveness to the HMGCS2 promoter and
to the heterologous thymidine kinase promoter. A, the
luciferase vector control (pLuc) or reporter constructs containing the
human HMGCS2 promoter with or without mutations in the putative hHMGCS2
PPRE site (pLuc-HMGPM1, pLuc-HMGPM2, pLuc-HMGP) were each
co-transfected with the pCMV--Gal vector into pcDNA, hPPAR1,
and mPPARG2 cells. -Galactosidase activity was used to monitor
transfection efficiencies and to normalize luciferase activities as
described under "Materials and Methods." Lysates were obtained from
cultures treated with Wy14643 (+Wy) or with vehicle only
(Wy). The normalized luciferase activity is expressed
relative to the data obtained from the pcDNA cell line transfected
with pLuc-HMGP and treated with ethanol. B, a single copy of
the putative hHMGCS2 PPRE with or without mutations (pLuc-TK-HMGM1,
pLuc-TK-HMG) was inserted into the luciferase reporter construct
containing the heterologous promoter from the herpes simplex thymidine
kinase gene (pLuc-TK). Each of the reporter constructs was
co-transfected with the pCMV--Gal vector into pcDNA, hPPAR1,
and mPPARG2 cells. The normalized luciferase activity is expressed
relative to the data obtained from the pcDNA cell line transfected
with pLuc-TK and treated with ethanol. The results in A and
B represent triplicate determinations from a representative
experiment and are expressed as the mean ± S.D. Similar results
were obtained in four other independent transfections. The nucleotide
sequence of the region between 111 and 89 that contains the PPRE is
shown. The arrows over the sequences
indicate the DR1 motif. An X under the
sequence indicates the position of mutations that are
delineated by lowercase letters.
|
|
| |
DISCUSSION |
The present study was designed to examine whether human genes
encoding enzymes and proteins that control the rate of fatty acid
oxidation by mitochondrial, peroxisomal, and microsomal pathways exhibit a capacity for regulation by PPAR. Elevated hPPAR levels in HepG2 cells elicited increased expression of three of the genes examined, HMGCS2, CPT1A, and ACS. In each case, the mPPARG1 cells exhibited strong ligand-dependent responses. However,
ligand-specific increases for CPT1A and ACS were not as readily
apparent in the hPPAR cell lines. This is concordant with the
phenotypic differences seen between hPPAR and mPPAR-E282G when
PPAR expression plasmids and reporter constructs were used in
transient co-transfection assays (20).
The largest increase mediated by hPPAR was seen for the
mitochondrial form of HMG-CoA synthase, which is the rate-limiting enzyme in ketone body formation from fatty acids (48). HMGCS2 mRNA
could not be detected in control cells but was abundantly expressed in
stably transformed cells overexpressing hPPAR. An increase in
PPAR-regulated transcriptional activity in response to fatty acid
and fibrate agonists has been demonstrated for the rat HMGCS2 gene
(33). In mice, this enzyme is elevated in response to starvation, and
this effect is blocked in PPAR null mice. Fed PPAR-deficient mice
also exhibit reduced levels of HMGCS2 in liver (49). Taken together,
these results indicate that the expression of HMGCS2 in mice is highly
dependent on the expression of PPAR. Our results suggest that human
HMGCS2 is also responsive to regulation by PPAR. In contrast, the
abundance of mRNAs encoding the cytosolic enzyme, HMGCS1, was not
affected by increased expression of PPAR or by treatment with
Wy14643 (results not shown).
The rat HMGCS2 gene exhibits a PPRE in the proximal promoter that binds
PPAR/RXR heterodimers (33). As shown here, the sequence of the
corresponding response element in the human gene is highly similar to
the rat PPRE, and the human element exhibits strong PPAR/RXR
binding. The putative human PPRE confers transactivation by PPAR to
a luciferase reporter gene that is under the control of the
heterologous TK promoter, indicating that this response element is
functional. In addition, this PPRE is active in the context of the
natural human HMGCS2 proximal promoter, and mutations that disrupt
binding of PPAR/RXR to the PPRE block transactivation of both the TK
and native promoter reporters by PPAR. Taken together, the data
suggest that this element could contribute to the dramatic elevation of
human HMGCS2 mRNA levels in the hPPAR cell lines.
During periods of fasting, the substrates of HMGCS2, acetyl-CoA and
acetoacetyl-CoA, are produced from the oxidation of fatty acids in
liver mitochondria, leading to increased production of ketone bodies
for distribution to other tissues. In mice nullizygous for PPAR,
fasting results in hypoglycemia and liver steatosis, reflecting
diminished fatty acid oxidation and ketogenesis in liver (19). Several
aspects of fatty acid oxidation are disrupted in PPAR null mice that
include a diminished, constitutive expression of several components of
the mitochondrial fatty acid oxidation pathway (17, 50). These results
suggest that, in mice, PPAR is a significant regulator of the
mitochondrial capacity for fatty acid -oxidation. The flux of long
chain fatty acids through the mitochondrial -oxidation pathway is
regulated by CPT1, a component of the long chain fatty acylcarnitine
translocases (51). The activity of CPT1 is modulated by the
concentration of malonyl-CoA, a potent inhibitor of the enzyme that is
formed in the first committed step of fatty acid synthesis (52). In
addition, CPT1 is induced by peroxisome proliferators in rat liver
(32). The predominant liver enzyme, CPT1A, is one of two variants of
CPT1 that have been identified. The other, CPT1B, is found
predominantly in muscle and adipose tissue. Although a PPRE has been
characterized for the human CPT1B gene (53-55), RT-PCR experiments
indicated that CPT1B mRNAs were present at low levels in HepG2
cells and were not affected by elevated expression of PPAR in the
presence or absence of Wy14643 (results not shown). In contrast, the
abundance of CPT1A mRNA was elevated 2-4-fold in the hPPAR
transformants and following the treatment of the mPPARG1 cells with
Wy14643, suggesting that the human CPT1A gene is also
PPAR-responsive.
The fatty acyl-CoA substrates for CPT1 are formed by the esterification
of free fatty acids to CoA by ACS. ACS is induced by peroxisome
proliferators in rat liver, and a PPRE has been identified in the
promoter of the rat gene (34). The expression of human ACS mRNAs
was elevated in the hPPAR cell lines and could be augmented further
by treatment with Wy14643. These changes are commensurate with an
increased capacity for mitochondrial fatty acid oxidation. Given the
importance of adaptive changes in the expression of ACS, CPT1, and
HMGCS2 to the maintenance of energy homeostasis during fasting, it is
not surprising that the corresponding human genes are also capable of
being regulated by PPAR.
In contrast, consistent and significant effects of elevated PPAR
levels on the expression of genes encoding components of the
peroxisomal long chain fatty acid -oxidation pathway, ACO1, ECH1,
and ACAA1, were not clearly evident. Effects of 2-fold or less were
seen for the two hPPAR cell lines shown in Table II. Examination of
a larger panel of cell lines that exhibit elevated expression of human
PPAR indicated that the modest elevation of these mRNAs was not
consistently seen in other transformants. In addition, a
ligand-dependent increase was not seen with the mPPARG1
cells or with either of the two hPPAR cell lines, suggesting that
the small differences observed might not be directly attributable to
PPAR regulation of these genes in the hPPAR cell lines.
We and others have noted that the expression level of PPAR in human
liver is much lower than in mouse liver, and it has been suggested that
this difference could contribute to the lack of peroxisome
proliferation and subsequent pathologic effects (20, 21). Like humans,
guinea pigs are refractory to peroxisome proliferation and also exhibit
lower hepatic expression of PPAR (5). However, transient
transfection of primary cultures of guinea pig hepatocytes with
expression vectors for either mouse, guinea pig, or human PPAR
conferred increases in the expression of rat acyl-CoA oxidase gene
reporter constructs and also in the enzymatic activity of endogenous
acyl-CoA oxidase in response to peroxisome proliferators (56). These
results contrast with our data for the human ACO1 gene in the PPAR
HepG2 transformants. Other evidence for the functional impact of lower
PPAR levels is apparent in data from mice indicating that the
induction of the peroxisomal fatty acid oxidation pathway is lower in
extrahepatic tissues that also express lower levels of PPAR (57).
Moreover, the expression of the murine ACO1 gene is elevated in PPAR
null mice in response to agonists that are relatively specific for
PPAR
and PPAR
. The observed response is lower than that seen for
PPAR agonists in wild-type mice and is likely to reflect the
relatively low levels of PPAR and PPAR when compared with the
level of PPAR in mouse liver (58, 59). These findings suggest that
quantitative differences in PPAR expression alter the extent of
enzyme induction and that the relatively high expression of PPAR in
mouse liver could contribute to the pathogenicity of peroxisome
proliferators in this species.
Our results indicate that increasing the expression of PPAR in HepG2
cells to levels found in mouse liver is not sufficient to confer
peroxisome proliferator-responsiveness to the ACO1 gene or the human
genes encoding other components of peroxisomal fatty acid oxidation in
HepG2 cells. The absence of an effect on this pathway could reflect
either of two likely possibilities. First, functional PPREs may not
have been conserved in the orthologous human genes during evolution of
the species. In this regard, the human ACO1 gene was reported to
contain a functional PPRE (60). However, the functionality of this PPRE
has been questioned (61). Alternatively, differences between HepG2
cells and normal liver in the expression of critical transcription
factors or necessary auxiliary proteins could limit the capacity of
PPAR to regulate the expression of the genes encoding the
peroxisomal fatty acid oxidation pathway. The lack of detectable
expression of mRNAs for CYP4A11 or the malic enzyme is clearly
indicative of differences in gene expression between HepG2 cells and
normal liver. In this respect, the positive effects of PPAR
overexpression on the expression of HMGCS2, CPT1A, and ACS can be more
readily interpreted than the absence of an effect on peroxisomal fatty
acid oxidation.
During fasting or starvation, activation of mitochondrial ketogenic
capacity represents an important physiologic regulatory pathway for
utilization of existing energy stores through mobilization and
catabolism of free fatty acids. The results presented here indicate
that human genes encoding key enzymes for mitochondrial fatty acid
oxidation and ketogenesis can be regulated by human PPAR.
in HepG2 Cells*
,
TOP
ABSTRACT
INTRODUCTION
These authors contributed equally to this work.
