| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Departments of 1 Medicine and 2 Pathology, University of Chicago, Chicago, Illinois 60637; 3 Laboratoire de Biologie Moléculaire et Cellulaire de l'Ecole Normale Supérieure de Lyon, 69364 Lyon, France; and 4 Department of Molecular and Cellular Biology, Baylor College of Medicine, Houston, Texas 77030
 |
ABSTRACT |
|---|
 TOP
 ABSTRACT
 INTRODUCTION
MATERIALS AND METHODS
RESULTS
DISCUSSION
REFERENCES
|
|---|
Isoforms of the thyroid hormone
receptor (TR)
and TR
genes
mediate thyroid hormone action. How TR isoforms modulate
tissue-specific thyroid hormone (TH) action remains largely unknown.
The steroid receptor coactivator-1 (SRC-1) is among a group of
transcriptional coactivator proteins that bind to TRs, along with
other members of the nuclear receptor superfamily, and modulate the
activity of genes regulated by TH. Mice deficient in SRC-1 possess
decreased tissue responsiveness to TH and many steroid hormones;
however, it is not known whether or not SRC-1-mediated activation of
TH-regulated gene transcription in peripheral tissues, such as heart
and liver, is TR isoform specific. We have generated mice deficient in
TR and SRC-1, as well as in TR and SRC-1, and investigated
thyroid function tests and effects of TH deprivation and TH treatment compared with wild-type (WT) mice or those deficient in either TR or
SRC-1 alone. The data show that 1) in the absence of TR or TR, SRC-1 is important for normal growth; 2) SRC-1
modulates TR and TR effects on heart rate; 3) two new
TR-dependent markers of TH action in the liver have been identified,
osteopontin (upregulated) and glutathione S-transferase
(downregulated); and 4) SRC-1 may mediate the
hypersensitivity to TH seen in liver of TR-deficient mice.
knockout; resistance to thyroid hormone; thyrotropin
| |
INTRODUCTION |
|---|
|
TOP
ABSTRACT
INTRODUCTION
MATERIALS AND METHODS
RESULTS
DISCUSSION
REFERENCES
|
|---|
THYROID HORMONE
RECEPTOR (TR) and TR function as nuclear transcription
factors that mediate thyroid hormone (TH) action. TRs bind to TH
response elements on thyroid-responsive genes in association with
transcriptional coregulators. Corepressors form part of a
transcriptional complex and recruit histone deacetylases, reducing
transcription (18, 40). The conformational change of TR
produced by TH binding releases the corepressor and permits the
recruitment and binding of a coactivator. The coactivator has a
number of functions that may include intrinsic histone
acetyltransferase activity, recruitment of histone acetyltransferases,
and recruitment of additional transcription factors and RNA polymerase
(21, 23). Several classes of nuclear coactivators have
been described that are important in mediating the response of
mammalian cells to thyroid as well as steroid and retinoid hormones
(2, 7, 14, 17, 41, 43), among which is the steroid
receptor coactivator (SRC)-1, a member of the p160 family of
coactivators (25).
Because some genes are upregulated and others are downregulated by TH
in the same cell, various theories have been proposed to explain the
mechanism(s) of TR-modulated gene expression (see Ref. 44
for review). We propose that specificity of interaction among TR
subtypes with particular cofactors may influence whether there is
stimulation or inhibition of mRNA expression. Determination of the
nature of interaction of TR subtypes with specific cofactors, and
how it affects gene transcription, can be evaluated in vivo by using
animals that are lacking each of the two TR genes with or without
SRC-1. It has been shown that TR knockout mice
(TR
/) have resistance to TH (10-12, 22,
34, 37), whereas mice with disruption of the TR1 and -2
isoforms (TR0/0) are hypersensitive to TH in several of
the tissues examined (22) or less prone to the effects of
TH deprivation (24). On the other hand, mice completely
deficient in both TR and TR exhibit more severe resistance to TH
than those lacking TR only (16). Taken together, these
data suggest that both isoforms play selective and overlapping roles,
both centrally and peripherally. Furthermore, coactivators are
important in TR-mediated TH action in vivo, as demonstrated by a mouse
model with disruption of the SRC-1 gene
(SRC-1/), which has also been shown to produce a
phenotype of reduced hormone sensitivity (38, 42).
Therefore, we ask whether SRC-1 differentially modulates the functions
of TR and TR, and if so, how this effect influences TH action in
the liver and heart.
For this purpose, we generated mice deficient in TR and SRC-1
(TR0/0SRC-1/), as well as mice deficient
in TR and SRC-1 (TR/SRC-1/), and
compared these with wild-type (WT) mice or mice with deficiency in
TR, TR, and SRC-1 alone. Our data provide evidence that, in the
absence of TR or TR, SRC-1 is important for normal growth. We
also show that SRC-1 mediates TR and TR action in the heart. This
study identifies two novel TR-dependent markers of TH action in the
liver, osteopontin (upregulated) and glutathione
S-transferase (downregulated). We have also demonstrated
that SRC-1 may mediate the hypersensitivity to TH seen in
TR0/0 mice.
| |
MATERIALS AND METHODS |
|---|
|
TOP
ABSTRACT
INTRODUCTION
MATERIALS AND METHODS
RESULTS
DISCUSSION
REFERENCES
|
|---|
Generation and Handling of Animals
SRC-1/ mice were generated as reported
previously (42). A targeting vector disrupted the
SRC-1 gene in 129sv ES cells by inserting an in-frame stop
codon at the Met381 position and deleting ~9 kb of
downstream genomic sequence that contains 446 amino acids from
Met381 to Thr826. This eliminated all
functional domains for transcriptional activation, histone
acetyltransferase activity, and interactions with nuclear receptors
CBP, p300, and p/CAF (42). The SRC-1/
construct was maintained in a C57BL/6J mouse strain. The genotype of
mice was confirmed by analysis of tail DNA, as previously described (42).
TR/ mice were produced, as previously described
(12), by insertion of the LacZ-NeoR cassette downstream to
the splice site in exon 4, eliminating the expression of the DNA and
ligand-binding domains of TR1 and TR2. The TR0/0
mice were produced by insertion of the LacZ-NeoR cassette downstream from exon 3 and replacing exons 5 through 7. This effectively abolished
not only the generation of TR1 and TR2 transcripts but also that
of TR
1 or TR2 by removing a transcription start site in
intron 7 (13). The gene sequence for rev-erbA- protein encoded by the opposite strands for the TR (31) remains
intact. In both sets of mice, the recombinant ES cells were derived
from 129sv mice and were implanted into C57BL/6 recipient blastocysts. C57BL/6 mice were mated to each chimeric mouse and then backcrossed 3-9 times into the same strain, diluting the 129sv background.
The individual knockout mice were backcrossed more than nine times on a
C57BL/6 background to produce a uniform genetic background for WT and
knockout animals. We then crossed the SRC-1/ mice with
TR0/0 and TR/ mice to produce
double-heterozygous animals, 8 times. The double-heterozygous mice were
mated to produced double-homozygous mice, which were backcrossed
3-5 times to each other.
Mice were weaned on the 4th wk after birth and were fed Purina Rodent
Chow (0.8 ppm iodine) ad libitum and tap water. They were housed, 
5
mice per cage, in an environment of controlled 19°C temperature and
12-h alternating darkness and artificial light cycles. All animal
experiments were performed according to protocols approved by the
Institutional Animal Care and Use Committee at the University of Chicago.
Mice were 40-70 days old at the time they were killed. At various
intervals, ~300 µl of blood were obtained by tail vein under light
methoxyflurane (Pitman Moore, Mundelein, IL) anesthesia. Experiments
were terminated by exsanguination via retroorbital vein. Whole blood
was allowed to clot overnight at 4°C, and serum was collected after
centrifugation and stored at 
20°C until analyzed.
Induction of Hypothyroidism and Treatment with TH
TH deficiency was induced in male mice by feeding them a low-iodine (LoI) diet supplemented with 0.15% propylthiouracil (PTU; Harlan Teklad, Madison, WI). On the 10th day, one group of mice maintained on the LoI/PTU diet (>5 mice/genotype) was injected daily for 4 days with vehicle only (1× PBS, controls), and another group received 0.2 µg of 3,3',5-triiodo-L-thyronine (L-T3) · 25 g body wt1 · day1. Experiments were
terminated by exsanguination 14-16 h after the final injection.
L-T3 dissolved in PBS or 0.002% human serum albumin as a vehicle was given by intraperitoneal injection in a total
volume of 0.1-0.3 ml. A stock of L-T3
(Sigma, St. Louis, MO) at a concentration of 1 mg/ml was prepared in a
solution of 50% ethanol-50% 1× PBS containing 5 mM NaOH and kept at
20°C, protected from light. Concentration of
L-T3 was confirmed by RIA (Diagnostic Products,
Los Angeles, CA). Blood samples were obtained at baseline, on the 10th
day after the initiation of the LoI/PTU diet, and at the termination of
the experiment on day 15.
The dose of L-T3 given to TH-deficient animals
was derived from previous experiments. It was optimized to achieve a
partial suppression of serum TSH to make evident the differences
between WT and SRC-1/ mice (22, 34, 38).
This allowed us to examine the effect of deficiency of receptor and
coactivator under identical conditions of TH supply. Metabolism of
T3 was determined in each genotype by measuring serum
T3 levels at 2, 4, 8, and 16 h after injection of
L-T3 (22).
TH and TSH Concentrations in Serum
Serum TSH was measured in 50 µl of serum by use of a sensitive, heterologous, disequilibrium double-antibody precipitation radioimmunoassay, as previously described (28). Samples containing >200 mU of TSH/l were 5- and 50-fold diluted with TSH-deficient mouse serum.Serum thyroxine (T4) and total T3 concentrations were measured by a double-antibody precipitation RIA (Diagnostic Products) with 25 and 50 µl of serum, respectively. The sensitivities of these assays were 0.2 µg T4/dl and 5 ng T3/dl. The interassay coefficients of variation were 5.4, 4.2, and 3.6% at 3.8, 9.4, and 13.7 µg/dl for T4 and 7.7, 7.1, and 6.2% at 32, 53, and 110 ng/dl for T3.
Serum Leptin
Samples were taken from frozen sera (obtained by retroorbital bleed and stored at80°C), and leptin levels were determined by RIA
(Linco Research, St. Charles, MO) with 10 µl of serum diluted into
100 µl in duplicate. Results are reported as serum leptin in
nanograms per milliliter.
Measurements of Growth, Heart Rate, and Energy Expenditure
Mice from 2 to 9 wk of age were weighed on a 200-g balance on the 1st day of each week. In addition, the length of each mouse was determined from the tip of the nose to the base of the tail under light gas anesthesia at the time of weighing. Heart rates of mice were determined at baseline, after 14 days of a LoI/PTU diet, and on a PTU diet following 14-16 h after the fourth daily intraperitoneal injection of T3 (0.2 µg/25 g). Animals were anesthetized with chloral hydrate (4 mg/10 g body wt ip), and heart rate was determined using a Hewlett-Packard model 78534AA monitor/terminal with a chart speed of 25 mm/s. Body temperature was maintained by keeping mice on a heating pad during measurement. Energy expenditure (EE) was determined at baseline by measurement of change in body weight and food consumption over 4 days as previously described (5, 37). EE was calculated according to the formula
|
![= [food consumption (g/day) × 4.058* × 0.8**]](/content/vol284/issue1/fulltext/E36/img002.gif)
|
![± [weight change (g/day) × 7***]](/content/vol284/issue1/fulltext/E36/img003.gif)
|
Isolation of Liver mRNA
Livers from animals were immediately frozen on dry ice and stored at80°C. For RNA extraction, ~80 mg of liver tissue from individual mice were homogenized in 1 ml of TRIzol (Life Technologies, Rockville, MD) with a Polytron tissue homogenizer. Total RNA was extracted according to the protocol provided with the TRIzol reagent. Concentration (A260) of the total RNA was determined, and
RNA was stored in 1/10 volumes of 3 M NaOAc and 3 volumes of 100% ethanol at 80°C.
Microarray Analysis of Mouse Liver
Six mice (male, 70 days old) of three genotypes (WT, TR/, or SRC-1/) were treated with
PTU and intraperitoneal L-T3, as described in
Induction of Hypothyroidism and Treatment with TH.
Mice were killed on the 15th day, and livers were removed, immediately
frozen on dry ice, and stored at 80°C. Livers from three mice were
pooled into one sample used to extract RNA and make cDNA. The cDNA was hybridized to Clontech's mouse Atlas Microarray V1.2 (1,172 genes) and
resolved on a phosphorimager (Clontech, Palo Alto, CA). Each microarray
compared expression in TH-deprived liver tissue with that in
L-T3-treated mouse liver for each genotype.
With use of threshold ratios for significance of 1.67 for each
comparison (WT with and without L-T3,
TR/ with and without T3, and
SRC-1/ with and without L-T3),
genes were identified in each group that were either up- or
downregulated in response to TH. In all three groups, 7.8-8.9% of
the total 1,172 genes displayed a potential response to TH. In WT and
SRC-1/ mice, a majority of responsive genes were
upregulated by TH, whereas in the TR/ mice, a
majority of responsive genes were downregulated by TH (Table
1).
|
TaqMan RT-PCR of Genes Expressed in Liver
To quantitate mRNA expression of various genes in livers of different genotypes, 2 µg of total RNA were reverse transcribed using the First-Strand Synthesis Superscript Kit (Life Technologies) according to the provided protocol. Reverse transcription (RT) was performed using random hexamers. cDNAs obtained from the RT reaction were diluted with RNAse-free water to a concentration of 1 ng/µl. TaqMan fluorescent probe/primer sets were designed using Primer Express 1.5 (Applied Biosystems, Foster City, CA) and mRNA sequences taken from GenBank. Specificity was confirmed by the Basic Logical Alignment Search Tool search. Primer/probe sets were then obtained for osteopontin, glutathione S-transferase (GST), split hand-split foot (SHSF), Ets-related transactivation factor (ERF), and 5'-deiodinase (MegaBases, Evanston, IL) genes (Table 2). Equal loading of wells was controlled using a commercially available probe/primer set for 18S ribosomal RNA (Applied Biosystems). Detection of mRNA was performed with sequence detector software and the ABI 7000 Sequence Detection System (Applied Biosystems), which is capable of reading two fluorophores (sample probe and 18S ribosomal control probe) simultaneously in each well.
|
Ten nanograms of each reverse-transcribed cDNA sample were run in
duplicate, and the reaction was performed with TaqMan Universal Mix and
96-well optical plates (Applied Biosystems). Each duplicate sample
represents reverse-transcribed total RNA from individual mouse liver
samples. The threshold cycle (Ct) is the first cycle, in a
40-cycle reaction, at which fluorescence is detected. For each sample,
there are two recorded threshold cycles, the first corresponding to
amplification of 18S rRNA (VIC fluorophore), and the second to a
specific gene of interest (FAM fluorophore). Normalization of data
involved subtraction of rRNA Ct from that of the specific
gene being amplified per well, because amplification is logarithmic.
For each mouse genotype analyzed, at least five individual liver sample
RNAs were run in duplicate. To calculate results, the average
Ct for WT mice was determined. Individual mouse liver data
were reported as degrees of increase or decrease from this WT average.
Assays were repeated 
3 times, and the data were normalized and merged.
TaqMan quantitative RT-PCR expression data for the four genes identified by microarray analysis (osteopontin, GST, SHSF, and ERF) are reported as percent change of PTU-treated mice from littermates on a PTU diet treated with T3 (as described in Induction of Hypothyroidism and Treatment with TH). The mean degree of change of PTU-treated animals of each genotype was determined relative to that of WT mice on PTU. Percent change was calculated by dividing each genotype's T3-treated mean (calculated against WT PTU) by its own PTU mean (±SE). Each group of animals (PTU and PTU+T3) had at least five mice.
Data Presentation and Statistics
Values are reported as means ± SE. Initially, a two-way or one-way ANOVA calculation was done to determine whether there was a significant interaction between treatment and genotype on each of the parameters measured. If significant interaction was detected, the effect of treatment was examined separately for each genotype, and vice versa. The Tukey-Kramer method, at 5% significance (Statview v. 5.0, SAS Institute), was used to control for multiple comparisons. To stabilize the variance of data for serum TSH and leptin, these data were analyzed on a logarithmic scale.| |
RESULTS |
|---|
|
TOP
ABSTRACT
INTRODUCTION
MATERIALS AND METHODS
RESULTS
DISCUSSION
REFERENCES
|
|---|
Thyroid Function Tests
Thyroid function tests of different mouse genotypes at baseline and after treatment with PTU or PTU and intraperitoneal T3 are shown in Table 3. Experiments were performed in adult male mice because of previously reported sex and age differences in TH levels in WT mice (5, 28). Serum total T4 and T3 levels, as well as TSH concentrations, were significantly higher in TR/ and
SRC-1/ compared with WT mice, as previously reported
(12, 34, 38). TR0/0 mice demonstrated
decreased serum T4 with normal TSH compared with WT mice
(P < 0.005). In the combined
TR/SRC-1/ mice, serum levels of
T4, T3, and TSH were 1.7, 2.7, and 2.8 times
greater, respectively, than in TR/ mice or 2.7, 2.8, and 10 times greater, respectively, than in the SRC-1/
mice. Mice deficient in both SRC-1 and TR had serum T4
and T3 values that were 2.0 and 1.7 times greater than
TR0/0, respectively, and TSH values were also increased
by 2.3 times. Levels of free T3 and T4 showed
similar differences (35).
|
TH deprivation resulted in marked increases in serum TSH in all
genotypes, although serum TSH levels in TR0/0 and
TR0/0SRC-1/ mice did not increase as
much as in WT mice, and TSH levels in TR/SRC-1/ mice increased threefold
more than in WT (P < 0.0001). T4 levels in all
mice decreased to <0.25 µg/dl after the PTU/LoI diet. Treatment with
TH resulted in variable suppression of absolute serum TSH values
depending on the genotype of the mouse. The
TR/SRC-1/ mice had the
least suppression, indicative of the highest degree of resistance to
TH, and TR0/0 mice showed the highest suppression of
serum TSH; the latter did not reach statistical significance because of
the large standard deviation of WT mice.
Growth in Mice of Different Genotypes
Length and body weight from age 2 to 9 wk were measured in mice of different genotypes (Fig. 1). SRC-1/ mice grew at the same rate and achieved similar
lengths at week 9 compared with WT mice. At week
5, lengths of TR/ and TR0/0 mice
were 14% less than those of WT mice (significance at the 5%
confidence level) and remained from 5 to 10% less compared with WT
mice at week 9. Therefore, both TR and TR are required for normal linear growth. TR is the major determinant of overall body length, because in the absence of TR (with or without SRC-1), these animals have the most stunted length (Fig. 1A).
TR0/0SRC-1/ mice show more retarded
linear growth than TR0/0 mice (Fig. 1B).
|
Weight gain remained similar for TR/,
SRC-1/, and WT mice through week 9. For
animals to achieve normal body weight, SRC-1 appears to be an important
modifier in the absence of either TR or TR. TR0/0
mice have a 28% reduction in body weight at 3 wk and by 9 wk still
maintain a 17% reduction in body weight (Fig. 1D). In the absence of both SRC-1 and TR or TR, mice have greater reduction in growth. Specifically, TR/SRC-1/
mice begin to have a decline in body weight at 8 wk (P < 0.0001), a change not seen in the SRC-1/ or
TR/ mice.
EE and Serum Leptin
EE experiments were performed in mice of 6-10 wk of age (Table 4). TR/SRC-1/ mice have the highest EE,
1.4 times that of WT mice (significance at the 5% confidence level).
This, along with the high TH levels in these mice, implies that the
TH-mediated EE occurs through TR. This corresponds to the decrease
in body weight seen in the TR/SRC-1/ mice
(Fig. 1C). TR0/0 mice tend to have lower EE,
but it does not reach significance. In the absence of SRC-1, TR
and/or TR mice also maintain normal EE.
|
Serum leptin concentrations in nonfasting mice were 2.1-fold higher in
SRC-1/ mice (P = 0.0472) and 2.5-fold higher
in TR/SRC-1/ mice (P = 0.020) compared with WT mice (Table 4, Fig.
2). The higher leptin levels in these
animals suggest that TH-mediated increase in leptin occurs in the
absence of SRC-1 and TR.
|
The contrasting effect of TH and genotype in EE and leptin
concentration is demonstrated in Fig. 2. Note that EE did not increase very much despite significant increases in TH levels in the different genotypes, except for that in the combined
TR/SRC-1/ mice. On
the other hand, regulation of serum leptin levels is less dependent on
genotype, but more dependent on TH levels, where the increase in TH is
relatively independent of the genotype.
Heart Rate
Heart rates were measured in mice of different genotypes at baseline and after 14 days of a LoI/PTU diet or LoI/PTU with T3 treatment (Table 5). As previously reported at baseline (22), we found that TR0/0 mice had relative bradycardia compared with WT
mice (371 ± 28 vs. 524 ± 10 beats/min, respectively;
P < 0.0001). Unexpectedly, SRC-1/ mice also
had bradycardia (388 ± 12; P < 0.0001). Deletion of both TR and SRC-1 resulted in the lowest mean heart rate. However, this value was not significantly different from values in the TR0/0 and SRC-1/ mice just described. We
also found baseline heart rates in TR/ mice to be
lower than those in WT (P = 0.02). Mice showed decreases in
heart rate in response to TH deprivation (LoI/PTU, Table 5), except for
animals deficient in TR. Although all genotypes increased heart rate
in response to TH treatment, TR0/0SRC-1/
mice had a modest increase of 17.6 ± 3.8% compared with
27-43% observed in the other genotypes. These data suggest that
SRC-1 facilitates the TR-mediated TH action in the heart.
|
Identification of TH-Responsive Genes by Microarray Analysis
TH-responsive genes specifically mediated by TR or SRC-1 were
identified by microarray analysis. Microarrays of mRNA from liver of
WT, SRC-1/, and TR/ mice were
analyzed, comparing each genotype in hypothyroid and TH-treated states.
Analysis of over 1,100 genes revealed 92-105 genes that were
affected by TH in each group (Table 1). Using high stringency (>2-fold
change), we identified two genes that were TR dependent (i.e., in
which there was no effect of TH treatment in TR/
mice): osteopontin (upregulated in WT mice) and GST (downregulated in
WT mice). Also, we identified two genes that were SRC-1 dependent (i.e., in which there was no effect of TH treatment in
SRC-1/ mice): SHSF (upregulated in WT mice) and ERF
(downregulated in WT mice) (Table 6).
|
Gene Expression in Liver with TH Withdrawal and TH Treatment
Confirmation of gene expression observed in the microarray analyses was done by TaqMan quantitative real-time PCR with the same and additional livers.Osteopontin (J04806).
WT and SRC-1/ mice had a 76 and 30% increase,
respectively, in osteopontin expression with T3 treatment.
RNA from livers of TR/ and
TR/SRC-1/ mice had no response to
T3, confirming a TR dependence for the TH-mediated
upregulation of osteopontin demonstrated by microarray. Although not
evaluated on the microarray, TR also appears to be necessary for
regulation of osteopontin expression, as TR0/0 and
TR0/0SRC-1/ mice failed to stimulate
expression with TH treatment (Table 7).
SRC-1 is not absolutely required for induction of osteopontin by TH,
although it may facilitate the response, a 1.36 ± 0.16-fold (30.6 ± 0.16%) increase in SRC-1/ compared with
a 1.86 ± 0.39-fold increase (82.2 ± 50.7%) in WT.
|
GST (J03958).
GST expression appeared to be downregulated by TH on the basis of data
from the microarray in WT mice, and it was also TR dependent.
Results from quantitative analysis are consistent with this data,
showing an 0.21 ± 0.02-fold (81.7 ± 2.1%) decrease in GST
expression in response to TH. In the absence of TR, there was a
small, paradoxical increase in GST expression in response to TH (Table
7). Interestingly, in the absence of both TR and SRC-1, TH-mediated
downregulation was restored (0.23 ± 0.04-fold; 77.8 ± 5.7%). This result indicates that the two receptor subtypes work
together to facilitate downregulation of GST in response to TH.
However, in the presence of only the TR (TR/
mice), the additional presence of SRC-1 inhibits TH-mediated downregulation of GST. By eliminating the SRC-1 along with TR, the
inhibition of TH-mediated GST repression is removed.
SHSF (U41606).
In WT mice, SHSF expression was increased by 1.5 ± 0.08-fold with
TH treatment, and although a blunted response was seen in SRC-1/ mice (1.19 ± 0.09), it was not
significantly different from that in WT mice. The other genotypes also
showed reduced responses (Table 7).
ERF (U58533).
ERF was shown to be SRC-1 dependent by microarray, and we have
confirmed this to be the case by quantitative PCR (Table 7). However,
unlike the TH-mediated downregulation of ERF seen by microarray in WT
and TR/ mice, quantitative RT- PCR results
demonstrated a modest increase (1.59 ± 0.22 and 1.90 ± 0.38, respectively) in ERF expression. Furthermore,
TR0/0 mice did not demonstrate any effect of TH.
5'Deiodinase (5'DI, NM007860).
Although not included as part of the microarray, we chose to study
5'DI, a gene well known to be upregulated by TH. To evaluate the role
of the TR subtypes and SRC-1 in mediating 5'DI expression, we
investigated the response of 5'DI to TH in mice deficient in either TR
subtype, in SRC-1, or in SRC-1 and each of the TR subtypes with TH
deprivation and TH treatment (Fig. 3).
There was a 300% increase in 5'DI expression in WT,
SRC-1/, and TR0/0SRC-1/
mice in response to TH. An even greater response (>2,400%) was observed in the TR0/0 mice, consistent with the
hyperresponsiveness observed with other markers of TH action in these
animals. This increased response was abrogated when SRC-1 was deleted
along with the TR. This result indicates that the hypersensitivity
seen in the TR0/0 mice may be due, in part, to the
presence of SRC-1, because in the absence of both, the increased
response is ablated. TR/ mice had a markedly reduced
response (only ~2.6% of the WT with TH), a reduction further
compounded by deficiency in both TR and SRC-1 (1.3% of WT with TH).
|
| |
DISCUSSION |
|---|
|
TOP
ABSTRACT
INTRODUCTION
MATERIALS AND METHODS
RESULTS
DISCUSSION
REFERENCES
|
|---|
These studies investigate the in vivo effect of TR subtypes in
peripheral tissues, specifically by focusing on animal growth, heart
rates, EE, and markers of TH action in the liver and heart. To do this,
we have bred mice deficient in the TR or TR, in SRC-1, or in each
of the TR subtypes and SRC-1 together. These animals were generated in
a mixed sv129 and C57BL/6J background. Therefore, single- and
double-heterozygous mice were backcrossed with WT littermates to
generate a uniform genetic background closer to C57BL/6J. The role of
the TR subtypes in growth has been studied previously in TR-deficient
animals by our laboratory and others (10-12, 16, 30).
In addition, mice deficient in SRC-1 have no reported changes in linear
growth and weight (42). Here, we show that, at 5 wk, mice
deficient in TR or SRC-1 have body weights similar to those of WT
mice. However, we show somewhat decreased linear growth in
TR/ mice relative to WT mice at 5 wk (86.8%,
significance at the 5% confidence level) and linear growth at 9 wk
(90.9%, significance at the 5% confidence level). Although we and
others have previously reported no differences in body weight and
linear growth in TR/ mice, the trends are there, and
with our larger number of mice in this study, differences reach
significance. These changes are much milder than in mice deficient in
TR at 5 wk (74.6%; 86.8%), a result consistent with work showing
that TR1 is important for intestinal crypt development at weaning, a
phenotype that seems to improve (27). At 9 wk,
TR0/0 mice still show decreased linear and body weight
growth (86.8%; 95.0%), similar to what has been shown previously
(12, 13). However, after full maturity, both TR- and
TR-deficient mice have retarded growth. Although
SRC-1/ mice achieve adult linear and body weights that
are not different from those of WT animals, loss of SRC-1 in
combination with either TR subtype
(TR0/0SRC-1/ or
TR/SRC-1/) results in linear and
weight growth retardation compared with WT at both 5 wk (86.7/92.2%
and 69/76.9%, respectively) and 9 wk (77.7/73.0% and 90.8/90.0%,
respectively). Taken together, these data indicate that both TR
subtypes have important roles in governing linear and body weight
growth. In addition, SRC-1 plays a cooperative role in growth with each
TR subtype in the absence of the other. In the absence of SRC-1, the
presence of other cofactors may compensate for its loss, but
insufficiently when there is additional loss of either TR. Although we
did not investigate expression of growth hormone in TR
subtype/SRC-1-deficient mice, we speculate that these levels may be
normal, as has been shown in TR0/0 (13),
TR1/ and TR/ (13),
and SRC-1/ (32) mice. Additionally,
animals deficient in both TR and SRC-1 show increased EE over all
other genotypes on a normal diet (P < 0.001). This
corresponds with much higher baseline levels of TH (Table 3) and
progressive weight loss in these mice at 9 wk. The use of a TR
isoform-specific ligand, GC-1, failed to maintain core body temperature
and reduced stimulation of uncoupling protein in brown adipose tissue
of hypothyroid mice (29). Taken together, these data
indicate that TR mediates the increased EE observed in response to
TH, supporting previous observations (39), and that SRC-1
is not necessary for this action of TH. The increased EE and weight
loss in these animals do not correspond to illness, as these animals
demonstrate fertility and litter size at homozygosity that are similar
to those of WT, without increased mortality in adult animals, at least
through 60 wk, the longest we have maintained them.
That TR is required for normal heart rate has been shown previously
(15, 19, 20, 22, 30). In addition, it has also been
reported by our laboratory and others that absence of TR does not
cause such a decline in basal heart rates and, in fact, results in a
slight increase in heart rate attributed to increased circulating
levels of TH in these mice (Table 3) (15, 19, 20, 33, 35,
36). However, there appear to be a number of differences in
baseline heart rates of WT mice dependent on strain, even in papers by
the same laboratories. SRC-1 appears to be essential for maintaining
heart rate, as in its absence, alone and in combination with each of
the receptor subtypes, heart rate is decreased. In the absence of
SRC-1, TR and TR are unable to maintain normal heart rates,
presumably by competing for a limited supply of cofactors or an
affinity of the TR2, which does not bind TH, for other
cofactors. In the absence of TR alone, TR and SRC-1 are
unable to maintain WT heart rate. However, in the absence of both TR
and SRC-1, there is a marked increase in serum TH (Table 3) over that
in TR/ mice available to bind to TR1, which does
not happen in the absence of SRC-1 alone. This seems likely, in that
TR/SRC-1/ mice achieve WT heart
rates when treated with exogenous TH. Despite the proposed role for
SRC-1 in regulating heart rate that is inferred by these data, we have
previously shown that other TH-dependent genetic markers in heart,
specifically SERCA2, MHC, and MHC, are unaffected by the absence
of SRC-1 (32).
To discern the mechanism of TH action in the liver, we sought to
identify genetic markers that were TR dependent vs. SRC-1 dependent.
It has been shown in vitro that coactivators are used by TR to
mediate TH action (reviewed in Ref. 23). We took livers from WT, TR/, and SRC-1/ mice that
were made hypothyroid with a LoI/PTU diet or were hypothyroid and
treated with TH, and we subjected them to microarray analysis. From this analysis, we identified four TH-dependent genes; two were
determined to be TR dependent (osteopontin, upregulated, and GST,
downregulated) and two were SRC-1 dependent (SHSF, upregulated, and
ERF, downregulated). Osteopontin is a secreted glycoprotein with an RGD
domain characteristic of integrin-binding proteins. It has been shown
to be an important chemokine in inflammation, a potential oncogene in
renal cancers, a stable component in mineralized tissues (interacting
with the vitamin D receptor and the retinoid X receptor) and smooth
muscle, with an additional presence in the anterior pituitary (6,
26). A direct role for osteopontin regulation by TH has not been
shown. Data here confirmed microarray results showing that osteopontin
expression was indeed TR dependent, as mice deficient in TR
(TR/ and
TR/SRC-1/) showed no response to TH
(Fig. 3A). We also showed that osteopontin may also be
regulated by the TR, as mice deficient in TR
(TR0/0 and TR0/0SRC-1/)
show a decrease in osteopontin in response to TH (16.9 and 26.3%,
respectively), distinguishing the roles for TR subtypes in controlling
this gene. Flores-Morales et al. (9) demonstrated that
40% of TH-responsive genes identified in an expression profile were
TR independent, also suggesting a role for TR in modulating gene
expression. In addition, we investigated GST by RT-PCR. Beckett and
colleagues (3, 4) showed that, by depriving rats of selenium in their diets (inhibiting T3 production in liver
by 5'DI) or by use of the PTU diet, there was in increase in expression of GST. However, there has been no detailed study on the effects of
administration of exogenous L-T3 on GST
expression. Here, we show in TR/ mice that
there is no downregulation of GST as seen in other genotypes, which
suppress GST with TH treatment by ~75% (Table 7). Interestingly, in
the absence of TR and SRC-1, suppression by TH absent in
TR/ mice is restored. From this, we conclude that,
in TR/ mice, the presence of SRC-1 inhibits
TR-mediated suppression of GST, and when SRC-1 is also removed, the
TR, possibly acting with another coregulatory molecule, mediates
TH-induced GST suppression.
Of the two TH-responsive genes identified to be SRC-1 dependent, we
were unable to confirm the microarray data with RT-PCR. In WT mice,
SHSF increased by 40% without any response in
SRC-1/ mice (Table 7). However, there were
also blunted responses to TH in the other genotypes investigated,
indicating that SHSF will not be a good marker for further use. ERF was
only partially SRC-1 dependent by RT-PCR; however, we saw differences
between TaqMan and microarray in the response of WT and
TR/ mice to TH, showing increases in ERF expression
(Table 7) vs. decreases seen by microarray (Table 6). It is possible
that ERF would be a good marker for future use, but it would require
further investigation. Importantly, we confirmed the viability of two new TR-dependent markers (osteopontin and GST) in the presence of
TH. These genes have not been previously identified in two other
studies of TH-responsive gene expression (8, 9). We have
seen other genes reported by microarray that were not reproduced by
other methods (Weiss RE and Sadow PM, unpublished data) and note that the previous studies confirmed only a small number of the
genes with Northern analysis that were identified by microarray (8, 9).
In addition to studying new markers identified by microarray, we also
investigated by RT-PCR liver expression of 5'DI, an enzyme whose
expression has been known to be upregulated by TH via the TR
(1). Interestingly, we saw sharp increases in 5'DI expression in TR0/0 mice, far greater than in
any other genotype (Fig. 3). Macchia et al. (22) recently
reported that mice deficient in all known TR isoforms have
hypersensitivity to TH. It was speculated that a potential mechanism
for hypersensitivity in these animals would be elimination of the
inhibitory TR2. However, as 5'DI data would indicate, the
hypersensitivity conferred upon TR-deficient mice (evidenced by vast
TH-induced increases in 5'DI expression) is ablated in the absence of
both TR and SRC-1, in which
TR0/0SRC-1/ mice increase 5'DI in
response to TH to levels similar to those of WT and
SRC-1/ animals. This result indicates that the
hypersensitivity seen in TR0/0 mice may be due to more
than absence of the inhibitory TR2. In fact, it is likely that the
hypersensitivity is due to an increased availability of SRC-1 to
interact with the TR, an event that may be controlled in
TR-competent mice by squelching of the coactivator by the TR.
A model of TH action in liver is shown in Fig.
4. 5'DI represents a liver gene
upregulated by TH. TR1 appears to be the key isoform to mediate TH
action on 5'DI in the liver, because in its absence, there is a major
reduction in 5'DI induction with TH. In TR0/0 mice,
there is a hyperresponse in 5'DI to TH. We propose that this TH
hypersensitivity is due to relief of inhibition by TR2, as
originally hypothesized by Macchia et al. (22). However, we extend this hypothesis to include that SRC-1 is necessary for this
action in liver and that SRC-1 inhibits TR2 activity, as TR0/0SRC-1/ mice are not hypersensitive
to TH-induced 5'DI expression. This result might best be confirmed by
overexpression of SRC-1 in vivo in liver and observation of the effect
on gene expression. Additionally, we propose that SRC-1 facilitates
TR-induced 5'DI expression but is not necessary. It is
possible that, in the absence of SRC-1, alternative coactivators,
such as TIF-2 or SRC-3, could compensate.
|
From these studies, we conclude that 1) in the absence of
TR or TR, SRC-1 is important for normal growth; 2)
SRC-1 partially mediates the TH effect on heart rate by TR and
TR; 3) we have identified two new TH-responsive markers
in the liver mediated by TR: osteopontin (upregulated) and
glutathione S-transferase (downregulated); and 4)
we have shown that SRC-1 may mediate the hypersensitivity to TH seen in
TR0/0 mice, as demonstrated by 5'deoiodinase expression
in liver.
| |
ACKNOWLEDGEMENTS |
|---|
We are indebted to Prof. Samuel Refetoff for advice and guidance during this project and to Dr. T. Karrision for help with statistical analyses. We also gratefully acknowledge the technical assistance of Kevin Cua with preliminary energy expenditure and heart rate experiments.
| |
FOOTNOTES |
|---|
This work was supported in part by grants from the National Institutes of Health: DK-58281 (to R. E. Weiss), DK-58242 (to J. Xu), and HD-078587 (to B. W. O'Malley), and from the Ministry of Research ACI 283 (to J. Samarut), and by the Seymour J. Abrams Thyroid Research Center.
Address for reprint requests and other correspondence: R. E. Weiss, Thyroid Study Unit, MC 3090, Univ. of Chicago, 5841 S. Maryland Ave., Chicago, IL 60637 (E-mail: rweiss{at}medicine.bsd.uchicago.edu).
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.
September 17, 2002;10.1152/ajpendo.00226.2002
Received 23 May 2002; accepted in final form 7 September 2002.
| |
REFERENCES |
|---|
|
TOP
ABSTRACT
INTRODUCTION
MATERIALS AND METHODS
RESULTS
DISCUSSION
REFERENCES
|
|---|
1.
Amma, LL,
Campos Barros A,
Wang Z,
Vennstrom B,
and
Forrest D.
Distinct tissue-specific roles for thyroid hormone receptors beta and alpha1 in regulation of type 1 deiodinase expression.
Mol Endocrinol
15:
467-475,
2001
2.
Anzick, SL,
Kononen J,
Walker RL,
Azorsa DO,
Tanner MM,
Guan XY,
Sauter G,
Kallioniemi OP,
Trent JM,
and
Meltzer PS.
AIB1, a steroid receptor coactivator amplified in breast and ovarian cancer.
Science
277:
965-968,
1997
3.
Beckett, GJ,
Nicol F,
Proudfoot D,
Dyson K,
Loucaides G,
and
Arthur JR.
The changes in heptaic enzyme expression by selenium deficiency and hypothyroidism in rats are produced by independent mechanisms.
Biochem J
266:
743-747,
1990[ISI][Medline].
4.
Beckett, GJ,
Russell A,
Nicol F,
Sahu P,
Wolf CR,
and
Arthur JR.
Effect of selenium deficiency on hepatic type I 5-iodothyronine deiodinase activity and hepatic thyroid hormone levels in the rat.
Biochem J
282:
483-486,
1992[Medline].
5.
Bergmann, BM,
Kushida CA,
Everson CA,
Gilliland MA,
Obermeyer W,
and
Rechtschaffen A.
Sleep deprivation in the rat: II. Methodology.
Sleep
12:
5-12,
1989[ISI][Medline].
6.
Ehrchen, J,
Heuer H,
Sigmund R,
Schafer M-H,
and
Bauer K.
Expression and regulation of osteopontin and connective tissue growth factor transcripts in rat anterior pituitary.
J Endocrinol
169:
87-96,
2001[Abstract].
7.
Eng, FC,
Barsalou A,
Akutsu N,
Mercier I,
Zechel C,
Mader S,
and
White JH.
Different classes of coactivators recognize distinct but overlapping binding sites on the estrogen receptor ligand binding domain.
J Biol Chem
273:
28371-28377,
1998
8.
Feng, X,
Jiang Y,
Meltzer P,
and
Yen PM.
Thyroid hormone regulation of hepatic genes in vivo detected by complementary DNA microarray.
Mol Endocrinol
14:
947-955,
2000
9.
Flores-Morales, A,
Gullberg H,
Fernandez L,
Stahlberg N,
Lee NH,
Vennström B,
and
Norstedt G.
Patterns of liver gene expression governed by TR.
Mol Endocrinol
16:
1257-1268,
2002
10.
Forrest, D,
Erway LC,
Ng L,
Altschuler R,
and
Curran T.
Thyroid hormone receptor is essential for development of auditory function.
Nature Genet
13:
354-357,
1996[ISI][Medline].
11.
Forrest, D,
Hanebuth E,
Smeyne RJ,
Evereds N,
Stewart CL,
Wehner JM,
and
Curran T.
Recessive resistance to thyroid hormone in mice lacking thyroid hormone receptor : evidence for tissue-specific modulation of receptor function.
EMBO J
15:
3006-3015,
1996[ISI][Medline].
12.
Gauthier, K,
Chassande O,
Plateroti M,
Roux JP,
Legrand C,
Pain B,
Rousset B,
Weiss R,
Trouillas J,
and
Samarut J.
Different functions for the thyroid hormone receptors TR and TR in the control of thyroid hormone production and post-natal development.
EMBO J
18:
623-631,
1999
