(Received for publication, January 31, 1997, and in revised form, April 30, 1997)
From the Laboratoire de Biochimie Pharmacologique,
Faculté de Pharmacie, Université de Bourgogne, 7. Bv.
Jeanne d' Arc, Dijon 21033, France and the § Institut de
Biologie Animale, Bâtiment de Biologie, Université de
Lausanne,CH-1015 Lausanne, Switzerland
This study demonstrates that the expression of
the phenol UDP-glucuronosyltransferase 1 gene (UGT1A1) is
regulated at the transcriptional level by thyroid hormone in rat liver.
Following 3,5,3-triiodo-L-thyronine (T3)
stimulation in vivo, there is a gradual increase in the
amount of UGT1A1 mRNA with maximum levels reached 24 h after
treatment. In comparison, induction with the specific inducer,
3-methylcholanthrene (3-MC), results in maximal levels of UGT1A1
mRNA after 8 h of treatment. In primary hepatocyte cultures,
the stimulatory effect of both T3 and 3-MC is also
observed. This induction is suppressed by the RNA synthesis inhibitor
actinomycin D, indicating that neither inducer acts at the level of
mRNA stabilization. Indeed, nuclear run-on assays show a 3-fold
increase in UGT1A1 transcription after T3
treatment and a 6-fold increase after 3-MC stimulation. This
transcriptional induction by T3 is prevented by
cycloheximide in primary hepatocyte cultures, while 3-MC stimulation is
only partially affected after prolonged treatment with the protein
synthesis inhibitor. Together, these data provide evidence for a
transcriptional control of UGT1A1 synthesis and indicate that
T3 and 3-MC use different activation mechanisms.
Stimulation of the UGT1A1 gene by T3 requires
de novo protein synthesis, while 3-MC-dependent
activation is the result of a direct action of the compound, most
likely via the aromatic hydrocarbon receptor complex.
UDP-glucuronosyltransferases (UGTs)1 form a family of microsomal membrane-bound enzymes that catalyze the binding of glucuronic acid on molecules with aromatic groups or aliphatic hydroxy, amino, carboxy, and mercapto groups. Multiple forms of UGTs have been observed in several species, including man (1). The regulation of their synthesis appears to be an important determinant of drug and xenobiotic detoxification and elimination (2). Endogenous compounds, such as retinoic acid, the steroid hormones estrone, and 4-hydroxyestrone, and the thyroid hormones can also be inactivated by glucuronidation (3).
Based on similarities of their predicted amino acid sequences, two families of UGT have been characterized. Family 2, which will not be analyzed in the present work, is composed of multiple steroid UGTs with broad substrate specificity. Family 1 consists of several forms of UGT1 that catalyze glucuronidation of phenol and bilirubin and that are synthesized from the phenol-bilirubin gene complex. The members of the UGT1 gene complex differ in their first exon but share four common exons, i.e. from the second to the fifth. The variable first exon codes for the N terminus of the proteins, which determines substrate specificity. The different UGT1 isozymes result from differential promoter usage and thus from splicing of different first exons to the common exons 2-5. At least 13 and 9 different exons 1 have been identified in the human and rat gene complex, respectively (4-6).
The differential regulation of gene expression by xenobiotics and
endobiotics resulting in the synthesis of different UGT isozymes is not
well characterized. However, one of the UGT1 phenol-glucuronidating isoenzymes is specifically induced by 3-methylcholanthrene (3-MC) in
rat liver (6). It contains an amino-terminal portion encoded by a first
exon called A1, has been identified as a 4-nitrophenol UGT and is
usually named UGT1A1 (7, 8). Recently, we have reported that 3,5,3
triiodo-L-thyronine (T3) also modulates the expression of the UGT1A1 isozyme (9, 10), whereby 4-nitrophenol glucuronidation is selectively increased. In agreement with this observation, this glucuronidation is decreased in thyroidectomized rats. Using quantitative RT-PCR and specific probes of the exon A1 of
UGT1A1, we have found that these variations correlate with concomitant changes in the corresponding mRNA levels (11). However, the molecular basis and the physiological significance of this phenomenon remain to be elucidated. Interestingly, previous studies have implicated the UGT1A1 isoenzyme in the glucuronidation of thyroid
hormones (12-14). Therefore, further investigation of the effect of
thyroid hormone on UGT1A1 expression is required to determine the
mechanism by which thyroid hormones regulate their own metabolism and
stimulate drug glucuronidation and elimination from the liver.
In the present study, we examined whether regulation of UGT1A1 synthesis by thyroid hormone is due to increased rates of gene transcription or to increased mRNA stability. Nuclear run-on assays with isolated hepatic nuclei from whole animals and studies using inhibitors of RNA and protein synthesis in primary hepatocyte cultures revealed that thyroid hormones regulate transcription of the UGT1A1 gene without affecting the half-life of the mRNA. Furthermore, the hormone-dependent increase in the transcription rate is dependent upon de novo protein synthesis. We also observed that 3-MC-stimulated UGT1A1 gene transcription was affected by cycloheximide only after prolonged treatment (16 h).
UGT1A1 DNA and -actin DNA
probes were synthesized by PCR as described previously (11). Amplified
cDNAs corresponding to the expected 507-bp fragment (nucleotides
6-513) of UGT1A1 exon 1 and 222-bp fragment (nucleotide 2818-3153) of
-actin were inserted into the SmaI site of the
pBluescript vector (Stratagene), creating pBS-UGT1A1 and pBS-
act,
respectively.
Total RNA was extracted from Wistar
rat liver as described by the supplier of the extraction kit (Eurobio,
Les Ulis, France), and RNA integrity was assessed by agarose gel
electrophoresis. Total RNA (50 µg) was denatured, electrophoresed on
a formaldehyde, 1% agarose gel and transferred onto a Gene Screen
membrane (NEN Life Science Products) using 3 M sodium
chloride and 0.15 M sodium citrate (20 × SSC).
Membranes were baked at 80 °C for 2 h and prehybridized
overnight at 65 °C in a rotary hybridization incubator with the
following solution: 0.2% polyvinylpyrrolidone
(Mr 40,000), 0.2% bovine serum albumin, 0.2%
Ficoll (Mr 400,000), 0.05 M Tris-HCl (pH 7.5), 1 M NaCl, 0.1% sodium pyrophosphate, 1% SDS,
10% dextran sulfate (Mr 500,000) and denatured
salmon sperm DNA (100 µg/ml). The membranes were then hybridized for
48 h at 65 °C with the same buffer plus the appropriate
cDNA probe labeled with [-32P]dCTP by random
priming as described by the supplier of the labeling kit (Appligene).
After the hybridization reaction, membranes were washed twice in 2 × SSC at room temperature for 5 min with constant agitation, twice in
2 × SSC with 1% SDS at 65 °C for 30 min, and twice in
0.1 × SSC at room temperature for 30 min and subjected to
autoradiography at
70 °C for 3 days. The relative amounts of
mRNA were estimated by densitometric scanning of the Northern blot
autoradiograms. The blots were rehybridized with a
-actin probe, and
the amount of actin mRNA in each lane was used as a standard to
quantify UGT1A1 mRNA.
All steps were performed
at 4 °C. Nuclei from Wistar rat liver were freshly prepared by
sucrose gradient centrifugation essentially as described by
Corthésy and Wahli (15) with some modifications as described
below. The pelleted nuclei were resuspended at a concentration of 2 × 106 in 200 µl of transcription mix (92 mM
Tris-HCl (pH 8.0), 20 mM NaCl, 4 mM
MnCl2, 10 mM phosphocreatine, 0.35 M (NH4)2SO4, 1.44 units
of human placental ribonuclease inhibitor (Promega, Zurich, Switzerland), 5 µg of creatine phosphokinase, plus, for each
reaction, 1 mM each of ATP, GTP, and CTP, and 100 µCi
[-32P]UTP (400 Ci/mM, 20 µCi/µl)).
In vitro elongation of nascent mRNA was carried out at
26 °C for 30 min. The reaction was stopped by addition of 4 µl of
MgCl2 at 0.5 M and 3 µl of DNase I (2 mg/ml, Promega) and followed by an additional incubation (for 5 min at 26 °C) to remove template DNA. The mixture was deproteinized by treatment with 20 µg/ml proteinase K in the presence of 100 µg/ml Escherichia coli tRNA (Boehringer, Rotkreuz, Switzerland)
for 30 min at 37 °C. Nuclear RNA, including the RNA labeled during the elongation reaction, was purified as described by Bender et al. (16). 5 µg of pBS-UGT1A1 and pBS-
act were linearized with EcoRI and XbaI, respectively, and denatured
before immobilization on a gene screen membrane using a dot-blot
apparatus. Then, the filters were baked at 80 °C for 2 h. Prior
to hybridization with radiolabeled nuclear transcripts, the filters
were treated for 6 h at 65 °C in 50 mM Tris-HCl (pH
8.0), 0.3 M NaCl, 10 mM EDTA, 0.2% SDS, 0.2%
Ficoll, 0.2% polyvinylpyrrolidone, 1% sodium pyrophosphate, and 1%
E. coli tRNA. Hybridization reactions were performed in 5-ml
propylene tubes (Falcon) in a final volume of 1 ml containing 50 mM Tris-HCl (pH 8.0), 0.3 M NaCl, 10 mM EDTA, 0.2% SDS, 0.02% Ficoll, 0.02%
polyvinylpyrrolidone, 0.1% sodium pyrophosphate, and 0.1% E. coli tRNA in the presence of 8 × 106 cpm of
labeled mRNA, at 65 °C for 3 days using a rotary hybridization incubator. After hybridization, the filters were first washed twice 30 min at 65 °C with 2 × SSC buffer. The filters were then washed
in the same buffer containing 0.4 µg/ml of RNase A and 10 units/ml of
RNase T1 at 37 °C for 15 min to hydrolyze the nonhybridized mRNA. Finally, the filters were washed at 37 °C for 1 h in
1 mM Tris-HCl, pH 7.5, 150 mM NaCl, 1 mM EDTA, 1% SDS, and 50 µg/ml of proteinase K and
subjected to autoradiography. The relative amounts of
32P-labeled mRNAs bound to the filters were estimated
by laser densitometric scanning of the autoradiograms.
Rat hepatocytes were isolated by collagenase perfusion of livers from 200-250-g Wistar rats (17), and cell viability was higher than 85% by trypan blue exclusion test. The hepatocytes were cultured in monolayer (1.5 × 105 cells/cm2) in Williams E medium (Life Technologies, Inc.) supplemented with 5% fetal calf serum (10%, v/v), fatty acid-free bovine serum albumin (0.2%, w/v), NaHCO3 (26 mM), glutamine (2 mM), glucose (3 g/liter), and antibiotics, at 37 °C in a humidified atmosphere of 5% C02, 95% air. After a 4-h incubation period, the medium was replaced by serum-free medium, and treatments with T3 (0.1 µM in 0.9% NaCl), 3-MC (10 µM in corn oil), and vehicles alone as control were started.
RT-PCR AnalysisComplementary DNA was synthesized from RNA
samples by mixing 1 µg of total RNA, 100 pmol of random hexamers as
primers in the presence of 50 mM Tris-HCl buffer (pH 8.3),
75 mM KCl, 3 mM MgCl2, 10 mM dithiothreitol, 200 units of Moloney murine leukemia virus reverse transcriptase, 40 units of RNase inhibitor, and 1 mM of each dNTP in a total volume of 20 µl. Samples were
incubated at 37 °C for 60 min and then diluted to 100 µl with
sterile diethylpyrocarbonate-treated H2O. The reverse
transcriptase was inactivated by heating at 95 °C for 5 min. A
10-µl aliquot of the reverse transcription reaction was then used for
subsequent PCR co-amplification and for quantitative analysis with
specific primers for UGT1A1 isoform and for the -actin internal
standard according to the sequences determined by Iyanagi et
al. (7) and Nudel et al. (19), as described previously
(11). Absence of genomic DNA contamination was verified by using a
control tube supplemented with 10 µg of RNase.
Rats were treated intraperitoneally with T3
dissolved in 0.9% NaCl at a dose of 50 µg/kg/day for 8 days and the
relative amounts of liver UGT1A1 mRNA were assayed by Northern blot
analysis. In parallel, we monitored UGT1A1 mRNA levels in rats
treated with 3-MC, a strong UGT1A1 inducer, and with vehicles alone as
positive and negative controls, respectively. Hybridization of mRNA
to the UGT1A1 cDNA probe reveals a prominent transcript of about 2.5 kilobase pairs (Fig. 1A). Following
treatment with T3 and 3-MC, the levels of the UGT1A1
transcript increase two and four times, respectively, when normalized
to the actin signal (Fig. 1B). These results are in
agreement with the quantitative RT-PCR analysis reported previously
(11) and correlate well with the increase in both the UGT activity
toward 4-nitrophenol and the UGT1A1 protein levels as measured by
immunoblot analysis (11).
To study further T3 induction of UGT1A1 expression, the
time course of UGT1A1 mRNA accumulation in response to
T3 (500 µg/kg in one injection) and to 3-MC (80 mg/kg in
one injection) as a positive control, was followed in rat liver by
semiquantitative RT-PCR analysis. Injection of the vehicles alone
served as negative controls. Fig. 2 shows the levels of
UGT1A1 mRNA at different time points after treatment using as
reference the -actin mRNA levels, which remain unchanged in the
rat liver after T3 and 3-MC treatments. There is a gradual
T3-induced increase in UGT1A1 mRNA, with a maximum
level found at 24 h (about 3-fold), which then declines to reach
nearly control levels at 48 h after treatment (Fig.
2B). In contrast, 3-MC leads to a quicker and stronger
increase of the UGT1A1 mRNA level that is already close to maximum
8 h after treatment (5-fold increase). The induced level remains
high up to 36 h and decreases thereafter (Fig. 2B).
NaCl or corn oil alone has no effect. These results provide evidence
that both T3 and 3-MC are able to increase UGT1A1 mRNA
levels in an inducer-specific and time-dependent manner and
that thyroid hormone is not as strong an inducer as 3-MC.
Transcriptional Induction of UGT1A1 Expression by T3
To determine the molecular mechanism by which
accumulation of UGT1A1 mRNA is regulated in rat liver, nuclear
run-on assays were performed first using hepatocyte nuclei isolated
16 h after a single injection of T3 (500 µg/kg of
body mass) or 3-MC (80 mg/kg of body mass). In this assay, only
transcripts that are already initiated in vivo are
faithfully elongated in vitro, giving an accurate estimation
of the relative transcriptional activity at the time point of nuclei
isolation. The radiolabeled nascent RNA transcripts were isolated and
hybridized to immobilized UGT1A1 and -actin cDNA. Densitometric
analysis of the resulting hybridization signals reveal that both
T3 and 3-MC treatments augment the levels of nascent UGT1A1
transcripts relative to
actin transcripts, reflecting increased
transcription rates by a factor of 3 and 6 for T3 and 3-MC,
respectively (Fig. 3). These results indicate that
T3 regulates UGT1A1 expression at the transcriptional level and confirm results obtained previously with 3-MC (8). The difference
between the transcription induction factors of T3 and 3-MC
correlate well with the difference in induced levels of UGT1A1 mRNA
described above (compare results of Fig. 2 with those of Fig. 3).
Effect of T3 on UGT1A1 mRNA Stability
Next,
we tested whether T3 stimulation is confined to induction
of transcription only or also involves an increase in mRNA stability, a mechanism used in the control of some thyroid
hormone-responsive genes. To this end, hepatocytes in primary culture
were treated for 6 and 16 h with T3 in the presence or
absence of the RNA synthesis inhibitor actinomycin D (Fig.
4). In parallel, the same experiment was performed in
cultures treated with 3-MC (Fig. 5), as well as with the
vehicles alone as controls. The decrease in UGT1A1 mRNA levels
after inhibition of RNA synthesis was measured by quantitative RT-PCR
as a function of time. The results obtained with T3 and
3-MC are very similar (Figs. 4A and 5A): UGT1A1
mRNA decays slowly and independently of the presence or absence of the inducers. In Figs. 4B and 5B, the amounts of
UGT1A1 mRNA (log values) are plotted as function of time, allowing
the determination of the RNA half-life (t1/2)
assuming first-order kinetics. The values obtained with the
T3 experiments are 9.5 ± 1.5 h
(r = 0.994) in T3-treated hepatocytes and
7 ± 1 h (r = 0.993) in control hepatocytes and, thus, are not significantly different from each other (Fig. 4B). This result indicates that T3 does not
exert its control on UGT1A1 mRNA levels by RNA stabilization.
Similarly, the 3-MC treatment has no effect on mRNA stability,
since we determined half-lives of 7 ± 1 h and 8 ± 1 h in treated and nontreated cells, respectively (Fig.
5B).
T3-induced UGT1A1 Transcription Is Dependent on de Novo Protein Synthesis
The slow increase in UGT1A1 mRNA levels
after T3 injection suggested that the transcriptional
activation of the UGT1A1 gene might require de
novo protein synthesis. This possibility was tested using
hepatocytes in primary culture incubated for 6 and 16 h with 0.1 mM T3 alone or in the presence of 10 mg/ml
cycloheximide, an inhibitor of protein synthesis. In the presence of
cycloheximide, the induction of the UGT1A1 gene by T3 is
blocked (Fig. 4, A and C), providing evidence
that de novo protein synthesis is required for UGT1A1
mRNA induction. A different result is observed with 3-MC, since
cycloheximide treatment does not suppress UGT1A1 mRNA stimulation
after 6 h of treatment and has only a weak effect after a 16-h
treatment (Fig. 5, A and C). These results
suggest that de novo synthesis is not required for
stimulation by 3-MC but that after prolonged treatment one factor(s)
becomes rate-limiting and has to be produced de novo
to maintain full induction. Importantly, we observe that cycloheximide
treatment, even for 16 h, does not affect -actin gene
expression.
In this work we provide evidence that T3 regulates UGT1A1 gene expression at the transcriptional level and that de novo protein synthesis is required for this stimulation. These results offer a molecular explanation for our recent observation that administration of T3 to adult rats enhances the activity of phenol UGT in the liver (11). First, these conclusions are based on a nuclear run on assay using isolated liver nuclei which showed that T3 appears to act on UGT1A1 gene expression mainly at the transcriptional level by increasing the rate of synthesis of UGT1A1 mRNA. Furthermore, the slow progressive induction of UGT1A1 by thyroid hormone is consistent with de novo protein synthesis, which is required for this transcriptional regulation. Second, we observed no change in the stability of UGT1A1 mRNA after T3 treatment, which suggest that post-transcriptional regulation does not contribute to the T3-dependent increase in UGT1A1 activity.
T3 regulates the expression of several other target genes at the transcriptional level. These include the genes encoding rat growth hormone (21), rat malic enzyme (22), phosphoenolpyruvate carboxykinase (23), chicken fatty acid synthase (24), and rat apolipoprotein AI (25, 26). T3 mediates its effects through nuclear T3 receptors (TRs) that are ligand-dependent transcription factors and that bind to thyroid hormone response elements (TREs) in the promoter of these genes (27, 28). In addition, the TRs bind to TREs as homodimers (29, 30) or heterodimers with accessory proteins such as the retinoid X receptor (31, 32). Interestingly, our results suggest that the mode of action of T3 on UGT1A1 gene expression is different, since de novo protein synthesis is required for stimulation. Although our observations do not formally exclude a direct action of T3 on UGT1A1 expression via the thyroid hormone receptor, they clearly show that if it were to occur it would not be sufficient, since the hormone is controlling the de novo production of a factor that is required for the stimulation of UGT1A1 gene transcription.
The biological significance of the T3 enhancement of the transcriptional rate of the UGT1A1 gene, and the subsequent increase in the expression of the corresponding isozyme, remain speculative. Indeed, little is known about the effects of UGT1A1 activity on thyroid hormone homeostasis in rats. Nevertheless, glucuronidation has been found to be a major pathway for biotransformation of the iodothyronines (33, 34), conjugates being subsequently excreted in the bile. Recent studies reported that multiple UGT isoenzymes are involved in this pathway. Thyroxine (T4), the main secretory product of the thyroid gland is glucuronidated by 4-nitrophenol (UGT1A1) and bilirubin UGTs, while T3, which is produced by outer ring deiodination of T4 in peripheral tissues, is glucuronidated by androsterone UGT (14, 15). In addition, Visser et al. (13) reported that the MC-type inducers also increase T3 glucuronidation. These findings suggest that the T3-stimulated increase in UGT1A1 activity following T3 treatment could lead to a lowering of the T4 and T3 levels in blood and, thus, participate in the regulation of thyroid hormone action. The physiological or pathological relevance of thyroid stimulation of UGT1A1 gene expression could now also be explored further in Gunn rats in which this gene is non-functional.
Comparison of the response to T3 and 3-MC stimulation is interesting, since it has been reported that 3-MC is an effective inducer of UGT1A1 gene expression also acting at the transcriptional level (8). Herein, nuclear run on assays on isolated liver nuclei confirm this finding and show that 3-MC enhances the transcriptional rate of the UGT1A1 gene. In addition, we present original data providing important information on the mechanism of 3-MC action: 3-MC has no effect at the post-transcriptional level, and although de novo protein synthesis is not required for transcriptional induction, a factor(s) that is most likely unstable, becomes rate-limiting, and its continuous synthesis is necessary for full induction. The nature of this factor(s) remains to be elucidated. 3-MC is known to act via a cytosolic protein, called aromatic hydrocarbon (Ah) receptor. 3-MC binds to the Ah receptor, and the Ah receptor-ligand complex translocates into the nucleus after interaction with the Ah receptor nuclear translocator. Once in the nucleus, the activated complex binds to the AhRE (also termed xenobiotic or dioxin response elements (XREs or DREs, respectively) and turns on transcription of the Ah-responsive genes (reviewed in Refs. 35-38). Interestingly, with respect to our results, it has been demonstrated that induction of mRNA encoding mouse Ugt1.6, which corresponds to the rat UGT1A1, was dependent upon the Ah receptor complex (18), which recognizes one or more XREs in the UGT1A1 promoter (8).
In conclusion, the data presented herein further our understanding of the regulation of UGT1A1 gene expression and demonstrate a major difference in the mode of action of T3 and 3-MC. The effect of 3-MC appears to be direct, in the sense that it is independent of de novo protein synthesis during the initial phase of the stimulatory process. However, a rate-limiting factor is required to obtain maximal induction, and it needs to be produced by de novo synthesis. In comparison, the regulating action of T3 depends on a factor being synthesized after hormonal treatment. The identification of this factor opens a new research avenue on UGT1A1 gene expression control.
We are grateful to Dr. R. Planells (INSERM U 38, Marseille, France) for helpful technical advice and to Dr. E. Beale (Institut de Biologie Animale, Lausanne, Switzerland) for reading the manuscript.