Download Physiological regulation of hypothalamic TRH transcription in vivo is

Physiological regulation of hypothalamic TRH
transcription in vivo is T3 receptor isoform specific
HAJER GUISSOUMA, MOHAMED T. GHORBEL, ISABELLE SEUGNET, TAOUFIK OUATAS,
AND BARBARA A. DEMENEIX1
Laboratoire de Physiologie Générale et Comparée, Muséum National d’Histoire Naturelle, URA
CNRS 90, 75231 Paris, cedex 5, France
Thyroid hormone (tri-iodo-thyronine,
T3) exerts transcriptional effects on target genes in
responsive cells. These effects are determined by
DNA/protein interactions governed by the type of T3
receptors (TRs) in the cell. As TRs show tissue and
developmental variations, regulation is best addressed in an integrated in vivo model. We examined
TR subtype effects on thyrotropin-releasing hormone (TRH) transcription and on the pituitary/thyroid axis end point: thyroid hormone secretion.
Polyethylenimine served to transfect a TRH-luciferase construct containing 554 bp of the rat TRH promoter into the hypothalami of newborn mice. Transcription from the TRH promoter was regulated in a
physiologically faithful manner, being significantly increased in hypothyroidism and decreased in T3treated animals. Moreover, when various ligand binding forms of mouse or chicken TRb and TRa were
expressed with TRH-luciferase, all forms of TRb
gave T3-dependent regulation of TRH transcription,
whereas transcription was T3 insensitive with each
TRa tested. Moreover, chicken TRa increased TRH
transcription sixfold, whereas mouse TRa decreased
transcription. These transcriptional effects had correlated physiological consequences: expression of
the chicken TRa in the hypothalamus of newborn
mice raised circulating T4 levels by fourfold, whereas
mouse TRa had opposite effects. Thus, TR subtypes
have distinct, physiologically relevant effects on TRH
transcription.—Guissouma, H., Ghorbel, M. T.,
Seugnet, I., Ouatas, T., Demeneix, B. A. Physiological regulation of hypothalamic TRH transcription in
vivo is T3 receptor isoform specific. FASEB J. 12,
1755–1764 (1998)
ABSTRACT
Key Words: nonviral gene transfer · polyethylenimine · thyroid
hormone · mouse · central nervous system
A CENTRAL PROBLEM in understanding endocrine
control of transcription is how stage-specific and cellspecific responses arise from ubiquitous signals. In
the case of thyroid hormone (tri-iodo-thyronine, T3)2
-induced responses, current thinking identifies a
number of key players. These include three differentially expressed T3 binding receptor (TR) iso-
forms—TRa1, TRb1, and TRb2—produced from
two TR genes, also called c-erbAa and c-erbAb (1).
In combination with their heterodimeric partners,
these TRs may determine recognition of different
core motifs (thyroid response elements, TREs). Further complexity is brought about by interactions between receptors and other transcription factors either directly or through coactivator or corepressor
proteins (for reviews, see refs 2, 3).
As these TR-dependent interactions will have cellspecific and developmentally dependent dimensions,
we chose to analyze T3-dependent, cell-specific responses by following regulation of a key target gene,
TRH (thyrotropin-releasing hormone), within the
hypothalamus. Although this gene is expressed both
within the hypothalamus and in extrahypothalamic
brain areas (4), we chose to examine hypothalamic
transcription because the hypothalamic neurons expressing TRH are responsible for controlling the hypothalamo/hypophysio/thyroid (HHT) axis and are
most sensitive to negative feedback by T3 (5). Indeed,
the model of TRH transcription has the advantage of
showing clear-cut responses that allow the contribution of different players to be dissected. In particular,
the specific role of TRb in mediating T3-dependent
regulation of this gene has been addressed in primary
hypothalamic neuronal cultures (6) and in cell lines
(7–9).
The distinct effects of TR subtypes on neuronal
TRH transcription seen in chick hypothalamic cultures (6) reflect the spatial expression patterns of TRs
in the chick (10, 11) and rat brain (12), with particularly high levels of expression of TRb in the hypothalamus. To take the physiological analysis a step
1
Correspondence: Laboratoire de Physiologie Générale et
Comparée, Muséum National d’Histoire Naturelle, URA
CNRS 90, 7, rue Cuvier, 75231 Paris, cedex 5, France. E-mail:
[email protected]
2
Abbreviations: ANOVA, one-way analysis of variance; c,
chicken; HHT, hypothalamo/hypophysio/thyroid; m, mouse;
NSB, nonspecific binding; P, postnatal day; PCR, polymerase
chain reaction; PEI, polyethylenimine; PTU, 6-n-propyl-2-thiouracil; RLU, relative light units; RSV, Roux sarcoma virus; T3;
tri-iodo-thyronine; TRH, thyrotropin-releasing hormone; TR;
T3 binding receptor; TREs, thyroid response elements.
0892-6638/98/0012-1755/$02.25 Q FASEB
/ 383c de15 Mp 1755 Thursday Oct 22 08:58 AM LP–FASEB de15
1755
further, into an in vivo context, we used the newborn
mouse. In this model, not only could we examine TR
subtype effects on TRH promoter activity in the hypothalamus in situ, but also the end point of the
HHT axis output, namely, circulating thyroid hormones.
Only in vivo models can be used for this type of
study. Modulation of TR status can be achieved in
vivo by germinal transgenosis or somatic gene transfer. Transgenic mice have been generated that are
homozygous for targeted inactivation of either the
TRa (13) or the TRb genes (14). These powerful approaches have shown differential effects of TR subtypes, producing distinct phenotypes. However, eliminating gene expression from the begining of
development will affect processes downstream of
early TR action. So to analyze TR effects at a defined,
physiologically relevant moment, somatic gene transfer can have its advantages. Moreover, such methods
allow one to study physiological responses in both
genetically intact animals and modified genetic backgrounds.
Synthetic gene transfer methods, and polyethylenimine (PEI) in particular, can be adapted to these
ends (15). By using PEI-based gene transfer in the
hypothalamus, we show that all TRb isoforms tested
mediated T3-dependent inhibition of transcription
from the TRH promoter, whereas ligand binding
TRa isoforms did not provide T3-dependent inhibition. Moreover, introducing chicken TRa or mouse
TRa into the hypothalamus respectively increased or
decreased basal TRH-luc transcription, and these
transcriptional effects were correlated with changes
in circulating T4. In contrast, expressing TRb, the TR
subtype most represented in the hypothalamus, did
not disturb HHT equilibrium. These results bolster
the hypothesis that TR subtypes have distinct transactivating properties that hold physiological relevance.
Plasmids
Plasmids were prepared by using Jetstar columns (Bioprobe
Systems, Paris, France), suspended in Tris-HCl 10 mM, EDTA
1 mM pH 8, and stocked as aliquots at 027C.
TheTRH-LUC construct contains a rat TRH gene 5* fragment (16) extending from 0554 to /84 base pairs cloned
upstream of the firefly luciferase-coding region (17).
The TREpal-LUC construct contains a synthetic oligodeoxynucleotide sequence encoding a palindromic TRE in SV-luciferase (18). The CMV-LUC expression vector was from Promega, Madison, Wis. The pCMVb-gal plasmid was from
Clontech (Palo Alto, Calif.). Plasmids expressing mouse (m)
TRa1, mTRb1, and mTRb2 receptors were generously provided by Dr. W. Wood (University of Colorado). Each cDNA
is in pRSVL (19) vector, with the Roux sarcoma virus (RSV)
promoter. A plasmid expressing full-length wild-type chicken
TRa1 (c)TRa1 was generously provided by Dr. H. H. Samuels
(New York University Medical Center). This vector has the
Vol. 12
December 1998
Treatment of animals: production of hypothyroid newborn
mice
Female OF1 mice (Iffa Credo, l’Abresle, France) were mated.
To induce fetal and neonatal hypothyroidism, dams were
given 0.05% 6-n-propyl-2-thiouracil (PTU) in drinking water
at day 14 of pregnancy. PTU administration was continued
throughout the lactation period. For evaluating T3 effects on
TRH mRNA levels or transcription, hypothyroid or normal
pups were injected with 250 mg of T3 /100 g of body weight
(in 9‰ saline). Controls received saline (9‰) injections.
Preparation of PEI/DNA complexes
DNA was diluted in 5% glucose. After vortexing, 25 kDa PEI
(Sigma, France) was added (nine equivalents, PEI amine/
DNA phosphate) and the solution was revortexed. The required amount of PEI, according to DNA concentration and
number of equivalents needed, is calculated by taking into
account that 1 mg DNA is 3 nmol of phosphate and that 1 ml
0.1 M PEI is 100 nmol of amine nitrogen.
In vivo gene transfer
Pups were injected on postnatal day 1 (P1). Pups were anesthetized by hypothermia on ice. The pups heads were held by
hand and a small incision was made with iris scissors through
the skin overlying the sagittal suture to expose the skull. A
small hole was made through the skull approximately 1 mm
lateral to the sagittal suture. A glass micropipette was lowered
2 mm through the incision and 2 ml of a 5% glucose solution
containing plasmid/PEI complexes was slowly injected bilaterally into the hypothalamus. Pups were kept under an infrared heat lamp until active.
After 18 h, mice were anesthetized and decapitated. Hypothalami were dissected out for luciferase analyzes or nuclear
receptor extraction. Animals treated with T3 received subcutaneous injections immediately after gene transfer. Controls
received saline at the same time point.
Luciferase assay
Hypothalami were homogenized in 150 ml assay buffer, then
luciferase assay was performed according to the standard protocol (luciferase assay system, Promega) in a single-well luminometer (ILA911, MGM Instruments, Hamden, Conn.). Luciferase was estimated during 10 s with 20 ml brain extract.
Light was expressed as relative light units (RLU).
MATERIALS AND METHODS
1756
RSV promoter in pEXPRESS vector (20). The cTRb2 used was
pSG5-TRb2 (11).
LacZ histochemistry
Pups were decapitated 24 h after transfection and brains were
dissected out. In toto LacZ histochemistry was performed as
described (21) and vibratome sections (100 mm) were prepared.
Nuclear receptor extraction
Pools of hypothalami from 12 P1 mice were used in each
experiment. Hypothalami were removed, weighed, homogenized at 27C (0.25 weight/volume) in 250 mM sucrose, 2
M MgCl2, 20 M Tris-HCl (pH 7.6) using a glass-teflon homogenizer, and filtered through nylon Bultex 48 mm. Filtrates were centrifuged (15001g, 10 min). Pellets were
The FASEB Journal
/ 383c de15 Mp 1756 Thursday Oct 22 08:58 AM LP–FASEB de15
GUISSOUMA ET AL.
washed twice in the homogenizing buffer supplemented
with 0.5% triton X-100 and centrifuged (15001g, 10 min).
Pellets were resuspended in binding buffer, rehomogenized (50 mM NaCl, 10% glycerol, 2 mM EDTA, 5 M 2mercaptoethanol, 20 mM tris-HCl, pH 7.6), and centrifuged (15001g, 10 min). This operation was repeated three
times. The final pellets were resuspended in 3 ml of 0.4 M
KCl, 1 mM MgCl2, 20 mM tris-HCl (pH 7.9), stored for 30
min on ice, and centrifuged at 35,0001g for 30 min. Supernatants were dialyzed three times in binding buffer and
centrifuged (50001g, 20 min). Fifty microliters of the supernatants were used to dose protein content with a commercial kit (Bio-Rad, Richmond, Calif.).
T3 binding assay
Samples (350 ml) of nuclear receptor extracts were incubated with [125I]T3 (2.6 nM, specific activity ú1200 mCi/
mg; Amersham, Buckinghamshire England) for 45 min at
257C. After incubation, 3.5 ml binding buffer were added
to each sample, then solutions were filtered at 27C under
vacuum through 0.45 mm Millipore nitrocellulose (HAWP)
filters in order to separate bound and free [125I]T3. Filters
were washed with 3 1 5 ml binding buffer. Radioactivity was
measured in a gamma counter. Nonspecific binding (NSB)
was determined in samples containing the tracer and an
excess of (2.6 mM) stable T3. NSB was substracted from total
binding to calculate the specific binding.
Radioimmunoassay of T4 and T3 hormones
Radioimmunoassay of plasmatic T4 and T3 hormones was carried out as previously described (22).
PCR analysis
Hypothalami of mice transfected with a given TR vector were
homogenized in 150 ml luciferase buffer and centrifuged (10
min, 12,0001g). Phenol/chloroform extraction was carried
out on 120 ml supernatant. After sodium acetate (3 M) precipitation, DNA was resuspended in TE buffer, pH8 (TE is 10
mM Tris-HCl, 1 mM EDTA).
Polymerase chain reaction (PCR) amplifications were performed in a final volume of 50 ml containing 11 reaction buffer, 5 ml template DNA, 2 mM of primers, 200 mM deoxynucleotides (dATP, dCTP, dGTP, dTTP), and 1 unit of Taq
polymerase. The following primers were used to detect ampicillin cDNA sequence present in all the plasmids used in our
experiments: sense, 5*-CCAATGCTTAATCAGTGAGGCACC3*; antisense, 5*-ATTCAACATTTCCGTGTCGCCC-3*. Internal controls were run on the same samples using mouse bactin primers: sense, 5*-GCTGAGAGGGAAATCGTGCG-3*;
antisense, 5*-CCAGGGAGGAAGAGGATGCG-3*. Amplification occurred during 40 cycles (1 min, 947C; 1 min, 607C; and
1 min, 727C). PCR products were separated on 2.8% agarose
gels containing 0.1 mg/ml ethidium bromide.
Statistical analysis of results
In vivo gene transfer results are expressed as means { SEM
per group. Student’s t test was used to analyze differences
between groups. Where appropriate, one-way analysis of
variance (ANOVA) with post hoc multiple comparison analysis by the Bonferroni test for independent variables was
applied. Differences were considered significant at P õ
0.05. In all cases typical experiments are shown, each ex-
periment having been repeated at least three times and providing the same result.
RESULTS
PEI/DNA complex injection into the hypothalamus
We first ensured that PEI/DNA injection resulted in
transgene expression within the hypothalamus. Figure 1A shows a section of a newborn mouse brain
transfected with CMVb-gal. Transgene expression was
seen around the third ventricule in numerous neuron-like cells. Expression was also found in some
other sites close to ventricular spaces, such as regions
of the hippocampus closely bordering the lateral ventricles (Fig. 1B). This is to be expected as complexes
reaching the third ventricule will diffuse through the
ventricules to other brain areas (21). We quantified
transgene distribution by comparing levels ofCMV-luc
expression in the hypothalamus and extrahypothalamic tissue. Luciferase was detected within the hypothalamus, but also in the cortex, striatum, and
hippocampus (Fig. 1C). Generally, levels were lower
in the extrahypothalamic areas, but differences were
not significant. This was in sharp contrast to the result obtained with TRH-luc. As shown in Fig. 1D, expression was fivefold higher in the hypothalamus
than in any other region (Põ0.001) and, most important, only in the hypothalamus was transcription
from this physiological promoter sensitive to T3 treatment.
TRH-luciferase transcription in the hypothalamus is
T3 dependent and regulated in a physiologically
meaningful manner
We next pursued the analysis of physiological regulation of transgenes in the hypothalamus. As seen in
Fig. 2A, transcription from the TRH promoter was
regulated in accordance with the thyroid status of the
animals. In hypothyroid animals, the mean level of
transcription from the TRH-luc construct (3.7{0.37
105 RLU/ hypothalamus, nÅ10) was twice that of normal animals (2.2{0.36 105 RLU/ hypothalamus,
nÅ10,Põ0.05). Treating normal pups with 250 mg of
T3 /100 g body weight further decreased TRH promoter activity by half (1.1{0.27 105 RLU/ hypothalamus, nÅ10, Põ0.05).
In contrast, expression of luciferase from a constitutive promoter (CMV) was unaffected by T3 in the
hypothalamus (Fig. 2B).
Inter- and intraassay transfection variability
Intraassay variability in the transfection experiments
was particularly low; SEMs were routinely °10% of the
mean with n ¢10. This underlines the consistency of
this gene transfer method in the hypothalamus. This
TR ISOFORM EFFECTS ON TRH TRANSCRIPTION IN VIVO
/ 383c de15 Mp 1757 Thursday Oct 22 08:58 AM LP–FASEB de15
1757
Figure 1. Spatial distribution of transgene expression after hypothalamic injection of PEI/DNA complexes in newborn mice. A)
Lac-Z reaction reveals b-galactosidase expression in neuron-like cells throughout the hypothalamus, 18 h postinjection (2 mgCMVbgal) into the hypothalamus. B) b-Galactosidase expression in the hippocampus (h) near the lateral ventricle (same brain as in
panel A). Arrows indicate third ventricle and lateral ventricle in panels A and B, respectively. A) 175; B) 1100. C) Expression
from the CMV promoter is equivalent in different areas of the brain. Luciferase expression in the hypothalamus, striatum/
thalamus (S/T), and frontal cortex (Cr) 18 h postinjection (1 mg CMV-luc) into the hypothalamus. D) Expression from the TRHpromoter is region specific and is only T3-sensitive in the hypothalamus. TRH-luc (1 mg) was injected into the hypothalamus and
luciferase expression assayed 18 h later in the hypothalamus, striatum/thalamus (S/T), and frontal cortex (Cr) in animals
treated with T3 or saline. Means { SEM are given in panels C, D; n ¢ 4 per point.
low intraassay variability obviates the need to add a
constitutively expressed gene to normalize for transfection efficiency as is often the case for most in vitro
work and some in vivo work (23). Interassay variability
could be higher, with different plasmid preparations
providing higher luciferase readings. However, the
same physiological regulations are always observed
whatever the basal level of expression (compare Fig.
2A and Fig. 3A).
1758
Vol. 12
December 1998
Only TRbs, and not TRa1, mediate T3-dependent
regulation of TRH transcription
The effects of different T3 binding TR isoforms on
TRH promoter activity were examined next. We
tested both chicken (c) and mouse (m) forms of TRa
and TRb. As seen in Fig. 3A, cotransfecting a control
plasmid containing the RSV promoter and no coding
sequence had no effect on in vivo transcription from
The FASEB Journal
/ 383c de15 Mp 1758 Thursday Oct 22 08:58 AM LP–FASEB de15
GUISSOUMA ET AL.
the TRH promoter, significant T3-dependent inhibition of transcription being found. When TRH-luc was
expressed with ligand binding forms of mammalian
or avian TR subtypes, there were clear-cut distinctions in TR activities.
First, both mTRa1 and cTRa1 abrogated the negative transcriptional effect of T3. Indeed, no statistical
difference in TRH transcription was found in controls or T3-injected animals (Fig. 3B) in which either
TRa1 was expressed. Moreover, the effects of mTRa1
and cTRa1 on basal transcription were different, with
mTRa1 halving basal TRH-luc transcription and
cTRa increasing transcription by sixfold (Põ0.001 in
each case).
Second, all forms of TRb isoforms mediated negative transcriptional effects of T3 on TRH transcription equivalent to that seen in controls. For each TRb
isoform tested (cTRb2, mTRb1, and mTRb2), statistically significant decreases in transcription were seen
in T3-injected animals as compared to controls (Fig.
3C). Again, as for the a subtypes, the TRb isoforms
differentially affected basalTRH-luc transcription:
cTRb2 decreased basal transcription by fivefold
(Põ0.001), mTRb1 did not modify basal transcription, whereas mTRb2 decreased it by threefold
(Põ0.001).
Both mTRa1 and mTRb2 activate transcription
from a positive TRE-containing promoter in the
hypothalamus
To determine whether these isoform effects were specific to the TRH promoter, we carried out transfection experiments with the same TRs and a positively
regulated TRE-pal construct injected into the hypothalamus. No T3-dependent regulation could be revealed in normal animals (data not shown), probably
because normal levels of T3 maximally activated the
promoter. However, in hypothyroid animals significant T3-dependent activation of TRE-transcription
was seen, and this was equivalent for each TR isoform
tested, whether from mice (Fig. 4) or chicken (data
not shown).
Each TR isoform increases specific T3 binding in
the hypothalamus to an equivalent extent
To verify that differences in TR isoform effects on
TRH transcription were not due to variations in the
efficiency of expression and protein production, we
carried out T3 binding studies on hypothalamic nuclear extracts from control and transfected mice. As
shown in Fig. 5, we found that the [125I]T3 binding
capacity of nuclear extracts prepared from animals
transfected with any one of the TR constructs used
was approximately 2.5 1 the capacity of control, nontransfected mice or mice transfected with an empty
plasmid. Using ANOVA analysis, we found the values
Figure 2. TRH-luc, but not CMV-luc, expression is correlated
with thyroid status. A) TRH-luc transcription was measured in
hypothyroid (Tx), normal (N), and T3-treated animals
(N/T3) 18 h after hypothalamic injection of TRH-luc (1 mg).
B). CMV-luciferase transcription was measured in euthyroidand T3-treated animals 18 h after hypothalamic injection of
CMV-luc (1 mg). Means { SEM are given n ¢ 10 per point.
in TR transfected animals to be significantly higher
(Põ0.05) from those in the two control groups.
There were no differences in T3 binding between TR
groups.
Expression of mouse or chicken TRa in the
hypothalamus modifies thyroid status
The end point of the HHT regulatory system is the
secretion of thyroid hormone, principally T4 in mammals (24). To determine the effects of TR subtype
expression within the hypothalamus on thyroid hormone secretion, we expressed various TRs and measured circulating T4 and T3 over the following week.
As seen in Fig. 6, expressing cTRa1 in the hypothal-
TR ISOFORM EFFECTS ON TRH TRANSCRIPTION IN VIVO
/ 383c de15 Mp 1759 Thursday Oct 22 08:58 AM LP–FASEB de15
1759
Figure 3. Regulation of TRH transcription is TR isoform specific. A) Transcription from the TRH-luc reporter (1 mg) in
the newborn mouse hypothalamus, alone or with a control
plasmid (150 ng) with the RSV promoter, but no coding sequence. TRH expression was significantly down-regulated by
T3 in both cases. B) TRH-luc expression is unaffected by T3
when a TRa is coexpressed. C) TRH-luc expression is significantly down-regulated in the presence of physiological concentrations of T3 when a TRb is coexpressed. In all cases, 150
ng expression vector for TRs were injected with 1 mg reporter
construct. Means { SEM are given, n ¢ 10 per point. Each
comparison (T3-dependent regulation in the presence of a
given receptor) was repeated at least three times, with similar
results. The control shown in panels A–C is the same control
for the whole series of experiments displayed. For ease of comparison, data are split into three groups: controls (A) TRa
isoforms (B), and TRb isoforms (C).
amus resulted in a significant, nearly fourfold, increase in circulating T4 at P4, whereas mTRa1 decreased serum T4 at the same time point. Both these
modifications were highly significant (Põ0.001 in
each case). No changes in circulating T4 were seen
in controls (injected with RSV control plasmid) nor
in animals transfected with equivalent amounts of
mTRb1. In cTRa1-injected animals, circulating levels
of T4 returned to control levels by P6. This lack of
cTRa1 effect at this time point was correlated with
loss of plasmid from the hypothalamus. As seen in
Fig. 7, PCR showed plasmid to be present in the hypothalamus of transfected pups for up to P4, but not
P6. Levels of T3, produced principally at the periph1760
Vol. 12
December 1998
ery by desiodation (24), were not significantly different in any group at any time point examined (data
not shown).
DISCUSSION
Nonviral gene transfer into a defined brain area can
be used to study gene regulation and transcription
factor function
The first finding to arise from these studies is that
PEI-based gene transfer into the brain provides physiologically coherent regulation from a promoter that
The FASEB Journal
/ 383c de15 Mp 1760 Thursday Oct 22 08:58 AM LP–FASEB de15
GUISSOUMA ET AL.
Figure 4. Positive transcriptional responses from a palindromic TRE are TR isoform independent. Transcription from a
TRE-luc construct [1 mg/hypothalamus of hypothyroid (Tx)
newborn mouse] is similarly stimulated by T3 when coexpressed with a noncoding RSV construct (control) or with mTRa1
or mTRb2 (150 ng). Means { SEM are given, n Å 4 (controls)
or ¢ 10 (mTRs).
has region-specific expression and hormone-dependent regulation. Two key observations bolster this
idea. First, there is a sharper differential in the expression of a physiological promoter from the site of
Figure 5. Each TR construct increases T3 binding to an equivalent extent. [125I]T3 binding capacity of nuclear extracts of
nontransfected (NT) mice hypothalami or hypothalami transfected with either control plasmid, cTRa1, mTRa1, cTRb2,
mTRb1, or mTRb2. Mice were transfected at P1 into the hypothalamus with 150 ng of each plasmid. Animals were killed
18 h later. The hypothalami were dissected out, nuclear receptors were extracted, and binding studies were carried out
as described in Material and Methods. Means { SEM are given,
n Å 12 per point.
Figure 6. Expression of TRa, but not TRb, in the hypothalamus modulates secretion of T4 by the thyroid. Each TR vector
tested was injected (150 ng) into the hypothalamus of newborn mice (P1) and the animals were killed at the ages shown.
Serum was collected and pooled (four individual samples per
pool) and T4 dosed by radioimmunoassay. Means { SEM of
pooled samples are given, n ¢ 4 for each point.
injection to more distant sites into which the transgenes diffuse than for a constitutive, viral promoter
(compare Fig. 1C, D). Second, only the TRH-luc construct is regulated by T3 and regulation is seen only
in the hypothalamus (Fig. 1C).
That low levels of transcription from the TRH promoter were found in each of the different brain
regions dissected out (cortex and thalamus/stria-
Figure 7. Duration of cTRa1 effects on thyroid status is related
to persistance of plasmid in the hypothalamus. CTRa1 vectors
(150 ng) were injected into the hypothalamus of newborn
mice (P1) and the animals killed at ages shown. The hypothalami were dissected out, DNA extracted and PCR for TR plasmid and b-actin performed as described in Material and
Methods. ct: H2O, d: days.
TR ISOFORM EFFECTS ON TRH TRANSCRIPTION IN VIVO
/ 383c de15 Mp 1761 Thursday Oct 22 08:58 AM LP–FASEB de15
1761
tum) fits with published data. First, pro-TRH mRNA
is found in numerous extrahypothalamic sites (4);
second, in transgenic mice expressing luciferase
downstream from a sequence containing several hundred base pairs of the 5* region and exon 1 of the rat
TRH gene, luciferase expression is found both in the
forebrain and in the hindbrain (25).
The finding that it is only in the hypothalamus that
T3 down-regulates transcription from TRH promoter
corroborates data obtained by in situ hybridization in
the adult rat brain by Koller et al. (5). These authors
showed that modulating thyroid status changed TRH
mRNA production in the hypothalamus, but not in
the thalamus. Thus, although injection into the hypothalamus results in diffusion of PEI/DNA into the
third ventricle and from thence into other brain
regions (21), expression from the TRH promoter reflects endogenous gene expression and, most important, T3-dependent negative feedback occurs only in
the hypothalamus. Moreover, feeding PTU to the
dams rendered the pups hypothyroid (T3 levels were
undetectable in the pups, data not shown), and this
treatment significantly raised TRH transcription.
Thus, modulation of thyroid status resulted in transcription from the TRH-luc construct expressed in the
hypothalamus that faithfully reflected that recorded
for the endogenous gene followed by in situ hybridization (5).
The negative feedback effect of T3 on TRH
transcription is mediated by the b receptor
We next analyzed the differential effects of TR isoforms on transcription. Only TRbs, but not cTRa or
mTRa, mediated T3-dependent inhibition of transcription in this in vivo context. These differential effects were not due to variations in TR expression,
since each TR used produced equivalent increases in
functional T3 binding. This finding of differential effects of TRb vs. TRa in regulating TRH transcription
agrees with previous data from our laboratory (6) obtained using the same TRH-luc in primary cultures of
chick hypothalamic neurons. In this in vitro work, we
tested cTRb and cTRa1, and observed that only cTRb
down-regulated TRH transcription whereas both isoforms were equipotent on transcription from a TRE
construct (6). Similarly, it has been shown that human TRb1, but not TRa1, down-regulates transcription from the human TRH promoter in transient
transfection assays in the CV-1 cell line (7).
The specific action of TRb in regulating the TRH
gene correlates well with the high expression of this
receptor in the hypothalamus (12). However, an additional level of complexity must be taken into account, the production of two isoforms from the cerbAb locus: TRb1 and TRb2 in rodents and humans
(26, 27), and TRb0 and TRb2 in chicken (11). In
mammals, it is the TRb2 isoform that has a particu1762
Vol. 12
December 1998
larly restricted expression, with the mRNA found almost exclusively in hypothalamus in the rat (28). Recently, Langlois et al. (9) ascribed a unique role to
TRb2 in negative regulation by T3 when they showed
hTRb2 to be much more efficient than hTRb1 in ligand-dependent repression of transcription from the
hTRH promoter introduced into CV1 cells. This finding differs from our present data showing the two
mouse isoforms to be equipotent in repressing ligand-dependent TRH transcription in the mouse hypothalamus in vivo. Clearly, the relative contributions
of these isoforms need more detailed analysis.
Expression of TRa in the hypothalamus abolishes
the negative feedback effect of T3 on TRH
transcription
There is consensus across the literature regarding the
distinct actions of TRa1 vs. TRb on TRH transcription, specifically, the lack of T3-dependent regulation
of TRH in the presence of the TRa isoform (6, 8).
These observations are particularly intriguing given
that in each experiment a ligand binding form of
TRa was used, and yet no change in transcriptional
activation of the TRH promoter was seen in the presence or absence of ligand (Fig. 3B; refs 6, 8). Yet we
know from our experiments that ligand-dependent
transcriptional modification can be elicited, as T3-dependent activation results from TRa coexpressed
with the TRE construct (Fig. 4). Many of the differential effects of TRa vs. TRb can be attributed to their
distinct amino termini (7). Although the amino termini of mouse and chicken TRa1 are different (65%
isology), we found that both mTRa1 and cTRa1 had
similar effects in this respect: expression of either one
blocked T3-dependent repression of TRH transcription. However, the effects of these two TRas on basal
levels of TRH-luc transcription were opposite, mTRa1
and cTRa1 respectively lowering and raising TRH
transcription. This result was somewhat unexpected
given the high conservation of sequence between the
two isoforms (90% isology overall). Aligment of the
two sequences shows a few conservative changes in
the first 30 amino acids of the amino terminal and in
the hinge region behind the DNA binding domain.
It is possible that these conservative changes are sufficient to produce distinct interactions with the transcriptional machinery. Indeed, a 10 amino sequence
(amino acids 21 to 30) in the amino terminal of the
cTRa has been shown to be essential for an interaction between the TR and TFIIB (29). Given the complexity of the interactions of TR with the transcriptional machinery and accessory proteins such as
coactivators (for review, see ref 2) or corepressors (3)
and that TRs show developmental and spatial regulations, it will be interesting to see whether exploring
TR action in vivo will produce a better definition of
The FASEB Journal
/ 383c de15 Mp 1762 Thursday Oct 22 08:58 AM LP–FASEB de15
GUISSOUMA ET AL.
the respective roles of receptor subtypes in modulating target gene transcription.
5.
TR effects on TRH-luc transcription have correlated
repercussions on thyroidal status
6.
A final, striking observation resulting from these studies is that modulating TR expression in the hypothalamus results in correlated changes in thyroid
hormone production. Indeed, expressing cTR increased basal transcription from the TRH promoter
by sixfold and the effect on the HHT axis was a fourfold increase in circulating T4 at P4. In contrast, expression of mTRa1 halved both TRH-luc transcription and T4 secretion. Thus, there would appear to
be a direct link between the effects of the different
receptor subtypes on basal levels of transcription
from the exogenous TRH promoter and their effects
on endogenous TRH gene (as deduced from modulation of the HHT axis end point: T4 production).
No effect on circulating T3 was seen. This is probably
because most T3 is not produced in the thyroid, but
from peripheral deiodination (24). No doubt, the
changes in T4 production produced by modulation
of TR expression can be compensated for by changes
in deiodination activity in the target tissues, thereby
maintaining an euthyroid T3 status overall.
In conclusion, our results confirm that the restricted expression of TRb2 correlates with its action
in mediating negative feedback effects of thyroid hormone on TRH transcription. Moreover, misexpression of TRa1 in the hypothalamus results in up or
down-regulation of TRH transcription according to
the species origin of the TRa1 used, with alterations
in TRH transcription played out down the HHT
axis. Finally, this versatile methodology opens up
new possibilites for dissecting the physiological relevance of transcription factor function in integrated contexts.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
We thank Drs. C. Glass (San Diego), H. Samuel (New York),
W. Wood (Colorado), B. Vennström (Stockholm), and S. Lee
(Boston) for providing plasmids. We gratefully acknowledge
the support of ARC and AFM. M.T.G. and H.G. were supported by fellowships from the Ligue contre le Cancer. We
thank Drs. Sylvie Dufour and Nathalie Becker for helpful discussions and support.
18.
19.
20.
REFERENCES
21.
1.
2.
3.
4.
Brent, G. A., Moore, D. D., and Larsen, P. R. (1991) Thyroid
hormone regulation of gene expression. Annu. Rev. Physiol. 53,
17–35
Glass, C. K., Rose, D. W., and Rosenfeld, M. G. (1997) Nuclear
receptor coactivators. Curr. Opin. Cell Biol. 9, 222–232
Wolffe, A. P. (1997) Transcriptional control. Sinful repression.
Nature 387, 16–17
Segerson, T. P., Hoefler, H., Childers, H., Wolfe, H. J., Wu, P.,
Jackson, I. M. D., and Lechan, R. M. (1987) Localization of thyrotropin-releasing hormone prohormone messenger ribonu-
22.
23.
TR ISOFORM EFFECTS ON TRH TRANSCRIPTION IN VIVO
/ 383c de15 Mp 1763 Thursday Oct 22 08:58 AM LP–FASEB de15
cleic acid in rat brain by in situ hybridization. Endocrinology 121,
98–107
Koller, K. J., Wolff, R. S., Warden, M. K., and Zoeller, R. T. (1987)
Thyroid hormones regulate levels of thyrotropin-releasing-hormone mRNA in the paraventricular nucleus. Proc. Natl. Acad.
Sci. USA 84, 7329–7333
Lezoualc’h, F., Hassan, A. H., Giraud, P., Loeffler, J., Lee, S. L.,
and Demeneix, B. A. (1992) Assignment of the beta-thyroid hormone receptor to 3,5,3*-triiodothyronine-dependent inhibition
of transcription from the thyrotropin-releasing hormone promoter in chick hypothalamic neurons. Mol. Endocrinol. 6, 1797–
1804
Hollenberg, A. N., Monden, T., and Wondisford, F. E. (1995)
Ligand-independent and -dependent functions of thyroid hormone receptor isoforms depend upon their distinct amino termini. J. Biol. Chem. 270, 14274–14280
Hollenberg, A. N., Monden, T., Flynn, T. R., Boers, M., Cohen,
O., and Wondisford, F. E. (1995) The human thyrotropin-releasing hormone gene is regulated by thyroid hormone through
two distinct classes of negative thyroid hormone response elements. Mol. Endocrinol. 9, 540–550
Langlois, M., Zanger, K., Monden, T., Safer, J. D., Hollenberg,
A. N., and Wondisford, F. E. (1997) A unique role of the b-2
thyroid hormone receptor isoform in negative regulation by thyroid hormone. J. Biol. Chem. 272, 24927–24933
Forrest, D., Sjöberg, M., and Vennström, B. (1990) Contrasting
developmental and tissue specific expression of a and b thyroid
hormone receptor genes. EMBO J. 9, 1516–1528
Sjöberg, M., Vennström, B., and Forrest, D. (1992) Thyroid hormone receptors in chick retinal development: differential expression of mRNAs for a and N-terminal variant b receptors.
Development 114, 39–47
Bradley, D. J., Young, W. S., III, and Weinberger, C. (1989) Differential expression of and thyroid hormone receptor genes in
rat brain and pituitary. Proc. Natl. Acad. Sci. USA 86, 7250–7254
Fraichard, A., Chassande, O, Plateroti, M., Roux, J. P., Trouillas,
J., Dehay, C., Legrand, C., Gauthier, K., Kedinger, M., Malaval,
L., Rousset, B., and Samarut, J. (1997) The T3Ra gene encoding
a thyroid hormone receptor is essential for post-natal development and thyroid hormone production. EMBO J. 16, 4412–4420
Forrest, D., Hanebuth, E., Smeyne, R. J., Everds, N., Stewart, C.
L., Wehner, J. M., and Curran, T. (1996) Recessive resistance to
thyroid hormone in mice lacking thyroid hormone receptor b:
evidence for tissue-specific modulation of receptor function.
EMBO J. 15, 3006–3015
Abdallah, B., Hassan, A, Benoist, C., Goula, D., Behr, J., and
Demeneix, B. A. (1996) A powerful non-viral vector for in-vivo
gene transfer into the adult mammalian brain: polyethyleneimine. Hum. Gene Therapy 6, 1515–1524
Lee, S. L., Stewart, K., and Goodman, R. H. (1988) Structure of
the gene encoding rat thyrotropin releasing hormone. J. Biol.
Chem. 263, 16604–16609
Wood, W. M., Kao, M. Y., Gordon, D. F., and Ridgway, E. C.
(1989) Thyroid hormone regulates the mouse thyrotropin b-subunit gene promoter in transfected primary thyrotropes. J. Biol.
Chem. 264, 14840–14847
Glass, C. K., Lipkin, S. M., Devary, O. V. and Rosenfeld, M. G.
(1989) Positive and negative regulation of gene transcription by
a retinoic acid thyroid hormone receptor heterodimer. Cell 59,
697–708
de Wet, J. R., Wood, K. V., Deluca, M., Helinski D. R., and Subramani, S. (1987) Firefly luciferase gene: structure and expression in mammalian cells. Mol. Cell. Biol. 7, 725–737
Forman, B. M., and Samuels, H. H. (1991) pEXPRESS: a family
of expression vectors containing a single transcription unit active
in prokaryotes, eukaryotes and in vitro. Gene 105, 9–15
Goula, D., Remy, J. S., Erbacher, P., Wasowicz, M., Levi, G., Abdallah, B., and Demeneix, B. A. (1998) Size, diffusibility and
transfection performance of linear PEI-DNA complexes in the
mouse central nervous system. Gene Therapy 5, 712–71721.
Demeneix, B. A., and Henderson, N. E. (1978) Serum T4 and T3
in active and torpid ground squirrels, Spermophilus richardsoni.
Gen. Comp. Endocr. 35, 77–85
de Luze, A., Sachs, L., and Demeneix, B. (1993) Thyroid hormone-dependent transcriptional regulation of exogenous genes
transferred into Xenopus tadpole muscle in vivo. Proc. Natl. Acad.
Sci. USA 90, 7322–7326
1763
24.
25.
26.
27.
1764
McNabb, F. M. A. (1992) Thyroid Hormones, pp. 283, PrenticeHall, Englewood Cliffs, New Jersey.
Balkan, W., Tavianini, M. A., Gkonos, P. J., and Roos, B. A. (1998)
Expression of rat thyrotropin-releasing-hormone (TRH) gene in
TRH-producing tissues of transgenic mice requires sequences
located in exon 1. Endocrinology 139, 252–259
Hoding, R. A., Lazar, M. A., Wintman, B. I., Darling, D. S., Koening, R. J., Larsen, P. R., Moore, D. D., and Chin, W. W. (1989)
Identification of a thyroid hormone receptor that is pituitaryspecific. Science 244, 76–79
Sakurai, A., Miyamoto, T., and DeGroot, L. J. (1992) Cloning
and characterization of the thyroid hormone receptor b1 gene
promoter. Biohem. Biophys. Res. Commun. 185, 78–84
Vol. 12
December 1998
28.
29.
Lechan, R. M., Qi, Y., Berrodin, T. J., Davis, K. D., Schwartz,
H. L., Strait, K. A., Oppenheimer, J. H., and Lazar, M. A. (1993)
Immunocytochemical delineation of thyroid hormone receptor
2 like immunoreactivity in the rat central nervous system. Endocrinology 132, 2461–2469
Hadzic, E., Desai-Yajnik, V., Helmer, E., Guo, S., Wu, S., Koudinova, N., Casanova, J., Raaka, B. M., and Samuels, H. H. (1995)
A 10-amino-acid sequence in the N-terminal A/B domain of thyroid hormone receptor alpha is essential for transcriptional activation and interaction with the general transcription factor
TFIIB. Mol. Cell. Biol. 15, 4507–4517
The FASEB Journal
/ 383c de15 Mp 1764 Thursday Oct 22 08:58 AM LP–FASEB de15
Received for publication March 13, 1998.
Revised for publicaiton July 8, 1998.
GUISSOUMA ET AL.