Effects of Rexinoids on Thyrotrope Function and the Hypothalamic-Pituitary-Thyroid Axis

0013-7227/06/$15.00/0
Printed in U.S.A.
Endocrinology 147(3):1438 –1451
Copyright © 2006 by The Endocrine Society
doi: 10.1210/en.2005-0706
Effects of Rexinoids on Thyrotrope Function and the
Hypothalamic-Pituitary-Thyroid Axis
Vibha Sharma, William R. Hays, William M. Wood, Umarani Pugazhenthi, Donald L. St. Germain,
Antonio C. Bianco, Wojciech Krezel, Pierre Chambon, and Bryan R. Haugen
Division of Endocrinology, Metabolism, and Diabetes (V.S., W.R.H., W.M.W., U.P., B.R.H.), Department of Medicine,
University of Colorado Cancer Center, University of Colorado Health Sciences Center, Aurora, Colorado 80045; Department
of Physiology (D.L.S.), Dartmouth Medical School, Lebanon, New Hampshire 03756; Thyroid Section (A.C.B.), Division of
Endocrinology, Diabetes, and Hypertension, Brigham and Women’s Hospital, Boston, Massachusetts 02115; and Institut de
Genetique et de Biologie Moleculaire et Cellulaire (W.K., P.C.), Clinique de la Souris and College de France, 67404 Illkirch
Cedex, Communaute Urbaine de Strasbourg, France
Retinoid X receptor (RXR)-selective retinoids (rexinoids) can
cause central hypothyroidism in humans, and this effect has
been confirmed in rodent models. In this report, we characterized the effect of rexinoids on the hypothalamic-pituitarythyroid axis in mice and TSH regulation in a thyrotrope-derived cell line. The synthetic rexinoid (LG 268) suppressed
TSH and T4 levels in mice. Hypothalamic TRH mRNA was
unaffected, but steady-state pituitary TSH␤ mRNA levels
were significantly lowered, suggesting a direct effect of rexinoids on thyrotropes. LG 268 suppressed TSH protein secretion and TSH␤ mRNA in T␣T1 thyrotropes as early as 8 h after
treatment, whereas the retinoic acid receptor-selective retinoid (TTNPB) had no effect. Type 2 iodothyronine deiodinase
(D2) mRNA and activity were suppressed by LG 268 in T␣T1
cells, whereas only D2 mRNA was suppressed in mouse pituitaries. LG 268 suppressed TSH␤ promoter activity by 42% and
T
HE EFFECTS OF vitamin A (retinol) on thyroid hormone
production and action have been known for many
years (1). In the 1940s, Simkins (2) demonstrated that patients
with hyperthyroidism were successfully treated with high
doses of vitamin A. Retinol is converted by alcohol dehydrogenase to retinaldehyde, which is subsequently converted by retinaldehyde dehydrogenase to all-trans retinoic
acid (ATRA). ATRA undergoes isomerization in hepatic microsomes to 13-cis retinoic acid (RA) and 9-cis RA, depending
on the levels of converting enzymes and cellular retinol binding protein. These retinoids can influence expression of many
genes through nuclear receptors [retinoic acid receptor
(RAR) and retinoid X receptor (RXR)]. We have previously
demonstrated that a synthetic RXR-selective retinoid, LG
1069, caused central hypothyroidism in patients with pro-
First Published Online November 23, 2005
Abbreviations: ATRA, All-trans retinoic acid; ChIP, chromatin immunoprecipitation; D1, type 1 iodothyronine deiodinase; D2, type 2
iodothyronine deiodinase; DMSO, dimethylsulfoxide; FBS, fetal bovine
serum; FCS, fetal calf serum; HPT, hypothalamic-pituitary-thyroid; KO,
knockout; LG 268, LG100268; m, mouse; NP40, Nonidet P-40; PVDF,
polyvinyl difluoride; RA, retinoic acid; RAR, retinoic acid receptor; RT,
room temperature; RXR, retinoid X receptor; RXR␥KO, RXR␥-deficient;
SDS, sodium dodecyl sulfate; WT, wild type.
Endocrinology is published monthly by The Endocrine Society (http://
www.endo-society.org), the foremost professional society serving the
endocrine community.
the ⴚ200 to ⴚ149 region accounted for a majority of the LG
268-mediated suppression of promoter activity. The RXR␥ isotype is expressed in thyrotropes. In vitro transfection and in
vivo transgenic studies indicate that any RXR isotype can
mediate TSH suppression by rexinoids, but the RXR␥ isotype
is most efficient at mediating this response. RXR␥-deficient
mice lacked pituitary D2 mRNA suppression by LG 268, but D2
activity remained intact. In summary, RXR-selective retinoids
(rexinoids) have multiple effects on the hypothalamic-pituitary-thyroid axis. Rexinoids directly suppress TSH secretion,
TSH␤ mRNA levels and promoter activity, and D2 mRNA levels but have no direct effect on hypothalamic TRH levels.
Rexinoids also stimulate type 1 iodothyronine deiodinase activity in the liver and pituitary. (Endocrinology 147:
1438 –1451, 2006)
found suppression of serum TSH levels (3). Duvic et al. (4)
confirmed these observations and showed that this effect
appeared to be dose dependent. Liu et al. (5) extended this
observation using a different RXR-selective retinoid
[LG100268 (LG 268)] in rats.
Isotretinoin (13-cis RA) has been used for many years to
treat acne, rosacea, and certain types of cancer. Clinical trials
using isotretinoin have failed to demonstrate any effect on
serum TSH levels (6, 7). Isotretinoin is a weak RAR agonist,
and these data further suggest that the effect of retinoids on
thyrotropes and TSH do not occur through RAR. Isotretinoin
can be interconverted to ATRA and 9-cis RA. In a single case
study, Dabon-Almirante et al. (8) showed that 9-cis RA (RAR
and RXR agonist) suppressed serum TSH in a woman treated
for advanced cervical cancer. Our own data have confirmed
the effect of 9-cis RA on suppression of TSH␤ subunit promoter activity in the TtT-97 mouse thyrotropic tumor model
(9). This is an excellent in vitro thyrotrope model, but the
tumors are difficult to generate and manipulate for extensive
in vitro studies. To explore the hypothesis that RXR-selective
retinoids have direct effects on TSH secretion, TSH␤ mRNA
levels, and TSH␤ promoter activity in thyrotropes, we turned
to the immortalized, thyrotrope-derived pituitary cell line
T␣T1 (10). These cells express TSH␤ and ␣-subunit mRNA,
and treatment with T3 causes a dose- and time-dependent
decrease in TSH␤ mRNA (11). In this report, we examine the
1438
Downloaded from endo.endojournals.org at Harvard Libraries on June 27, 2007
Sharma et al. • Effects of Rexinoids on the HPT Axis
effects of natural and synthetic retinoids on thyrotrope function using the T␣T1 cells as an in vitro model and transgenic
mice as an in vivo model to examine the effects of retinoids
on the hypothalamic-pituitary-thyroid (HPT) axis as well as
deiodinase activity.
Materials and Methods
Experimental animals
Wild-type (WT) and RXR␥-deficient (⫺/⫺) mice (12) were housed in
a pathogen-free transgenic facility at the University of Colorado Health
Sciences Center. Mice were bred on the 129SvJ background and experiments were performed on littermate mice. All animal protocols were
approved by the Animal Care and Use Committee.
Mice were studied at approximately 8 wk of age on a standard ad
libitum chow diet. LG 268 (kindly provided by Ligand Pharmaceuticals,
San Diego, CA) was prepared in wet granulation vehicle and administered by daily oral gavage. Mice were treated for 3 d. On the morning
of the fourth day, mice were again treated by oral gavage and killed 4 h
later after fasting. Blood was collected for plasma and tissues (brain,
liver, pituitary, and hypothalamus) were snap frozen.
Chemicals
9-cis RA, 13-cis RA, and ATRA were purchased from Sigma Chemical
Co. (St. Louis, MO). TTNPB and LG 268 were generously provided by
Ligand Pharmaceuticals. Ethanol was used as a vehicle for 9-cis RA,
13-cis RA, and ATRA in the cell culture experiments. Dimethylsulfoxide
(DMSO) was used as a vehicle for TTNPB and LG 268.
Deiodinase activity
Tissues were homogenized in 0.25 mm sucrose, 20 mm Tris-HCl (pH
7.6), 1.2 mm EDTA using a tissumizer (Tekmar, Co, Cincinnati, OH), and
sufficient buffer to yield approximately a 1:5 homogenate (wt/vol). The
homogenates were centrifuged at 1000 ⫻ g for 5 min. The resulting
supernatants were assayed for 5⬘D activity using 1.0 nm [125I]rT3 as
substrate and 20 mm dithiothreitol as cofactor. Incubations were carried
out for 1 h at 37 C using protein amounts that allowed for a fraction of
deiodination of less than 25%. To distinguish between type 1 iodothyronine deiodinase (D1) and type 2 iodothyronine deiodinase (D2) 5⬘D
activities, 1 mm 6-n-propyl-2-thiouracil and/or 100 nm nonradioactive
T4 were included in the respective incubation medium (13) Activity is
expressed as femtomoles iodide generated per minute per milligram
protein. [125I]rT3 was obtained from PerkinElmer (Norwalk, CT) and
purified by chromatography using Sephadex LH-20 (Sigma) before use.
Protein concentrations were determined by the method of Bradford (14)
with reagents obtained from Bio-Rad Laboratories (Hercules, CA).
Thyroid function tests
Plasma and media mouse TSH values were measured by RIA (performed by Dr. Samuel Refetoff, University of Chicago, Chicago, IL).
Standards were diluted in plasma from mice treated with thyroid hormone for the plasma measurement and standards were diluted in media
[10% fetal bovine serum (FBS)-DMEM] for measurement of mouse TSH
secreted into the media. Plasma total T4 and total T3 values were measured by standard RIA (Diagnostic Products Corp., Los Angeles, CA).
Cell culture
T␣T1 cells were grown in DMEM (Invitrogen, Life Technologies,
Carlsbad, CA) containing 10% fetal calf serum (FCS, Hyclone, Logan,
UT), 10 mm HEPES buffer solution, 20 U penicillin-streptomycin (Invitrogen). The T␣T1 cells were seeded on Matrigel-coated plates (BD
Biosciences, Bedford, MA), which facilitated adhesion. Matrigel was
diluted 30-fold with DMEM before coating the plates, which were allowed to dry before plating cells. The cells were maintained at 37 C in
an environment of 5% CO2. Monolayer cultures of ␣TSH cells were
maintained in DMEM supplemented with 10% FCS. Replacement with
the same medium containing specified amount of retinoids, was done
Endocrinology, March 2006, 147(3):1438 –1451 1439
48 h before harvesting the cells. TSH levels in the media of cultured cells
were measured after 2 d of treatment with retinoid or vehicle.
RNA measurement by quantitative RT-PCR
Total RNA was isolated from cells using TriReagent (Sigma) as recommended by the manufacturer. The mRNA for mouse (m) TSH␤ and
mouse prepro-TRH was measured by real-time quantitative RT-PCR
using ABI Prism 7700 sequence detection system (PerkinElmer/Applied
Biosystems, Foster City, CA). The sequences of forward and reverse
primers as designed by Primer Express (PE/Applied Biosystems) were
5⬘-CCTGACCATCAACACCACCA-3⬘ and 5⬘-TGGGAAGAAACAGTTTGCCAT-3⬘ [mTSH␤], and 5⬘-CTCCAGCGTGTGCGAGG-3⬘ and
5⬘-TCCCTTTTGCCCGGATG-3⬘ (mTRH), respectively. The TaqMan
fluorogenic probe used was 5⬘-6FAM-GATATCCCGTCATACAATACCCAGCACAG-TAMRA-3⬘ for mTSH␤ and 5⬘-6FAM-CTTGGTGCTGCCTTAGATTCCTGGA-TAMRA-3⬘ for mTRH. Amplification reactions
were performed in MicroAmp optical tubes (PE/Applied Biosystems) in
a 25-␮l mix containing 8% glycerol, 1⫻ TaqMan buffer A [500 mm KCl,
100 mm Tris-HCl, 0.1 m EDTA, 600 nm passive reference dye ROX (pH
8.3) at room temperature], 300 ␮m each of dATP, dGTP, dCTP, and 600
␮m deoxyuridine 5-triphosphate, 5.5 mm MgCl2, 900 nm forward primer,
900 nm reverse primer, 200 nm probe, 0.625 U AmpliTaq Gold DNA
polymerase (PerkinElmer, Foster City CA), 6.25 U Moloney murine
leukemia virus reverse transcriptase (Life Technologies, Inc., Gaithersburg, MD), 10 U RNAsin ribonuclease inhibitor (Promega Corp., Madison, WI), and the template RNA. Thermal cycling conditions were as
follows: reverse transcription was performed at 48 C for 30 min followed
by activation of TaqGold at 95 C for 10 min. Subsequently 40 cycles of
amplification were performed at 95 C for 15 sec and 60 C for 1 min. A
standard curve was generated using the fluorescent data from the 10fold serial dilutions of mTSH␤-cRNA that was synthesized as described
(15). TRH standard curve was generated using control plasmid. Quantities of TSH␤ and TRH in samples were normalized to the corresponding 18s rRNA (PerkinElmer/Applied Biosystems, P/N 4308310).
RAR␤ mRNA was measured in T␣T1 cells as previously described
(16). Quantitative RT-PCR was carried out using SYBR green based on
manufacturer’s recommendations.
Transient transfection studies
5⬘ TSH␤ promoter deletions in pA3 luciferase were generated from
a ⫺1240 to ⫹40 bp TSH␤ promoter construct as previously described
(17–19). T␣T1 cells were cultured to 80 –90% confluency (1.8 ⫻ 106 cells)
for transfection studies. For each 10-cm2 well, 4 ␮g TSH␤-luciferase
plasmid and 8 ␮l Lipofectamine 2000 reagent (Invitrogen) were used as
per the manufacture’s instructions. Each transfection also contained 25
ng Renilla luciferase plasmid (Promega) as an internal transfection control. A Rous sarcoma virus promoter luciferase plasmid and a promoterless pA3 luciferase plasmid were transfected in parallel as positive and
negative controls, respectively. DNA and the lipofectamine reagent were
diluted separately in 200 ␮l of serum-free medium, mixed together, and
incubated at room temperature for 30 min. The culture plates were
washed with PBS and 1.6 ml of media (DMEM, 10% FCS) was added.
The 400 ␮l of plasmid lipofectamine mixture was then added to each
well, and the plates were incubated at 37 C in the presence of retinoid
or vehicle. Cells were harvested after 48 h, subjected to freeze-thaw cell
lysis, and assayed for dual firefly and Renilla luciferase activity in a
Monolight 3010 luminometer using a dual-luciferase reporter assay system (Promega). Firefly luciferase light units were normalized to Renilla
luciferase activity. ␣TSH cells were cultured to 80 –90% confluency (0.8 ⫻
106 cells) on Matrigel-coated plates (described under Cell culture) for
transfection studies. Mouse RXR␥1, RXR␥2, RXR␣, and RXR␤ cDNA
were generated by PCR from plasmids (provided by R. Evans, University of California, San Diego) to generate fragments with NotI overhangs
and were cloned in frame into pCGN2, which contains an amino terminal hemagglutinin epitope (20). Plasmids were fully sequenced to
verify fidelity with original sequences. For each 10-cm2 well, 3 ␮g TSH␤luciferase plasmid (⫺1240 to ⫹40) plus varying amounts of pCGN2 with
pCGN2-RXR isotypes (totaling 1 ␮g) and 8 ␮l of Lipofectamine 2000
reagent were used and transfections were carried out as described above.
Downloaded from endo.endojournals.org at Harvard Libraries on June 27, 2007
1440
Endocrinology, March 2006, 147(3):1438 –1451
Sharma et al. • Effects of Rexinoids on the HPT Axis
FIG. 1. Effect of LG 268 on serum hormone measurements in mice. Mice (four to six in each group) were given vehicle or different doses of LG
268 (x-axis) daily for 3 d by oral gavage. Serum was collected 4 h after the last dose. Data are expressed as TSH milliunits per liter (A), total
T4 (micrograms per deciliter) (B), and total T3 (nanograms per deciliter) ⫹ SEM (C). *, Significant change (based on one-way ANOVA) in hormone
level at a specific dose (1 mg/kg, dark gray; 3 mg/kg, light gray; 10 mg/kg, white) of LG 268, compared with vehicle (black bars). Dashed line
(A) represents the detection limit of the TSH assay (10 mU/liter). Values below 10 mU/liter were assigned a value of 8 mU/liter for statistical
analysis.
Western blot analysis
Nuclear extracts of TtT-97 thyrotropic tumors, T␣T1, and ␣TSH cells
were prepared as described previously (21). Cells incubated under appropriate conditions were washed with ice-cold PBS and cell lysates
were prepared. Protein content of lysates was measured using the DC
protein assay kit (Bio-Rad). Samples containing equal amounts of protein were mixed with 2⫻ Laemmli sample buffer. The proteins were
resolved on a 10% SDS-polyacrylamide gel and transferred to polyvinyl
difluoride (PVDF) membranes. The membranes were blocked with 20
mm Tris-HCl (pH 7.6), 8.5% NaCl, and 0.1% Tween 20 containing 5%
(wt/vol) nonfat dry milk at room temperature (RT) for 2 h and incubated
with RXR isotype-specific antibodies (RXR␣ and RXR␤, Santa Cruz
Biotechnology, Inc., Santa Cruz, CA; RXR␥, Lab Vision, Fremont, CA) in
20 mm Tris-HCl (pH 7.6), 8.5% NaCl, and 0.1% Tween 20 containing 5.0%
milk at 4 C overnight (16). After washing, membranes were incubated
with antimouse IgG (for RXR␤ and RXR␥) and antirabbit IgG (for RXR␣)
conjugated to horseradish peroxidase for 1 h at RT. ECL (Amersham
Pharmacia, Uppsala, Sweden) detection reagents were used for
immunodetection.
Chromatin immunoprecipitation (ChIP) of RXR
isotypes with the TSH␤ promoter in TtT-97 and
T␣T1 thyrotrope cells
The ChIP method was a modification of the method by Boyd and
Farnham (22). TtT-97 or T␣T1 cells were exposed to vehicle or 1 ␮m LG
268 for 4 h, and then the solution was adjusted to 1% formaldehyde to
cross-link proteins/DNA. The tube was placed on ice and centrifuged
at 4 C at 1500 rpm for 5 min to pellet cells; supernatant was aspirated,
and cells were washed with 10 ml PBS, recentrifuged, and washed with
an additional 50 ml ice-cold PBS. Cells were transferred to a 15-ml
dounce homogenizer in 10 ml cell lysis buffer [5 mm 1,4-piperazine
diethane sulfonic acid (pH 8.0), 85 mm KCl, 0.5% Nonidet P-40 (NP40),
0.5 mm phenylmethylsulfonyl fluoride, and one Complete minitablet
(protease inhibitor cocktail, Roche, Stockholm, Sweden)] and dounced
15 times on ice with the B (tight) pestle to release the nuclei, followed
by incubation on ice for 15 min and douncing five more times. Contents
were transferred to a 15-ml conical polyproplene tube and centrifuged
for 3500 rpm for 5 min at 4 C. The nuclear pellet was resuspended in 10
FIG. 2. Effect of LG 268 on hypothalamic prepro-TRH
and pituitary TSH␤ mRNA in mice. Mice (four in each
group) were given vehicle or 10 mg/kg LG 268 daily for
3 d by oral gavage. The mice were killed 4 h after the last
dose, and hypothalami and pituitaries were collected for
total RNA extraction. Then 100 ng total RNA were used
for quantitative RT-PCR (PRISM 7700, Applied Biosystems) using specific cDNA for standard curves. Data are
expressed as femtograms target mRNA corrected for an
internal standard (nanograms rRNA). *, Significant difference (P ⬍ 0.01) between vehicle and LG 268.
Downloaded from endo.endojournals.org at Harvard Libraries on June 27, 2007
Sharma et al. • Effects of Rexinoids on the HPT Axis
ml nuclear lysis buffer [50 mm Tris-Cl (pH 8), 10 mm EDTA, and 1%
sodium dodecyl sulfate (SDS), 0.5 mm phenylmethylsulfonyl fluoride,
and one Complete minitablet (protease inhibitor cocktail, Roche)], gently
resuspended with a pasteur pipet, and 1.5-ml aliquots were frozen at
⫺80 C in 2 ml microfuge tubes. Tubes were thawed on ice, 200 mg acid
washed glass beads added (Sigma; G1277), and chromatin sonicated
25–30 times for 15-sec pulses with a microtip using a Fisher sonicator at
4 C to an average size of 300-1000 bp. Tubes were centrifuged to pellet
debris and supernatants Approximately 70 ␮g of chromatin (1–2 ␮g/␮l)
was diluted 10-fold with immunoprecipitation dilution buffer [0.01%
SDS, 1% NP40, 1.2 mm EDTA, 16.7 mm Tris-Cl (pH 8.0), and 167 mm
NaCl] and precleared twice with 50 ␮l protein A⫹G beads (Santa Cruz)
containing 3.3 ␮g salmon sperm DNA for 1 h at 4 C.
Supernatants were removed to a new tube, and 2.5 ␮g of antibody for
RXR isotypes (same as used for Western blot) or rabbit IgG were added and
tubes rotated overnight at 4 C. To the tube was added 25 ␮l protein G
microbeads (MACS separation, Miltenyi Biotec, Gladbach, Germany) for
1 h with rotation and complexes captured on microcolumns with a magnetic separator. Immunoprecipitated chromatin was washed with five 1-ml
aliquots of ChIP wash buffer [100 mm Tris-Cl (pH 8), 500 mm LiCl, 1% NP40,
1% deoxycholic acid] and eluted with three 100-␮l aliquots of 50 mm
NaHCO3 and 1% SDS. Cross-links were reversed by adding 10 ␮g RNase
A and adjusting to 0.3 m NaCl before a 4-h incubation at 65 C, followed by
addition of 6 ␮l 0.5 m EDTA, 6 ␮l Tris-Cl (pH 6.5), and 6 ␮l 20 mm proteinase
K (Sigma) and incubation at 42 C for 90 min. DNA was purified on minipurification columns (QIAGEN, Valencia, CA) and eluted in 60 ␮l 10 mm
Endocrinology, March 2006, 147(3):1438 –1451 1441
Tris-Cl (pH 8) and 1 mm EDTA as recommended by the commercial supplier. PCR was performed on 4 – 8 ␮l of each sample for 25–29 cycles using
Taq Gold polymerase (Applied Biosystems) using oligonucleotide primers
for TSH␤ (⫺219/⫺135) sense 5⬘-AGAAGAGAGGAAGATGCATGCTATAAT-3⬘, antisense 5⬘-TCATACTGAACCCCAAATAAAACTTG-3⬘
with an annealing temperature of 55 C or specific for the coding region of
glyceraldehyde 3-phosphate dehydrogenase sense 5⬘-ATGGTGAAGGTCGGTGTGAACG-3⬘, antisense 5⬘-CCTTCTCCATGGTGGTGAAGAC-3⬘
with an annealing temperature of 53 C.
Statistics
Statistical analyses between individual measurements were performed using the Student’s t test, and measurements over multiple time
or dosing points or comparison between transgenic animals and drug
or RXR isotype transfections were performed using one-way ANOVA
followed by Fisher’s least significant differences (protected t tests) using
the program GB-Stat or pairwise multiple comparison (Tukey test) using
the program SigmaStat 2.03 (Point Richmond, CA). P ⬍ 0.05 was considered to be statistically significant.
Results
Effect of an RXR-selective retinoid on the HPT axis
Mice were treated with increasing amounts of LG 268 daily
for 3 d by oral gavage. Figure 1 shows the results of 0, 1, 3,
FIG. 3. Effect of LG 268 on tissue deiodinase
mRNA and enzyme activity in mice. Mice (four
in each group) were given vehicle or 10 mg/kg
LG 268 daily for 3 d by oral gavage. The mice
were killed 4 h after the last dose and liver,
brain, and pituitaries were collected for total
RNA extraction and deiodinase enzyme activity. A, 100 ng total RNA were used for quantitative RT-PCR (PRISM 7700, Applied Biosystems) using specific cDNA for standard
curves. Data are expressed as picograms target mRNA corrected for an internal standard
(nanograms rRNA). B, 25–100 ␮g protein extract were used in each 5⬘D activity assay.
Enzyme activity is expressed as femtomoles
per minute per milligram (D1) and femtomoles
per hour per milligram (D2). *, Significant difference (P ⬍ 0.05) between vehicle and LG 268.
Downloaded from endo.endojournals.org at Harvard Libraries on June 27, 2007
1442
Endocrinology, March 2006, 147(3):1438 –1451
and 10 mg/kg䡠d LG 268 on plasma TSH, T4, and T3 levels.
There was a significant dose effect of LG 268 on TSH (P ⫽
0.03) and T4 (P ⬍ 0.001) levels. T3 levels were not significantly
suppressed by LG 268 (P ⫽ 0.064). To determine the direct
effects of LG 268 on the hypothalamus and pituitary, TRH
and TSH␤ mRNA levels were measured after treatment
(Fig 2). Hypothalamic TRH mRNA was not suppressed by
LG 268, whereas pituitary TSH␤ mRNA was decreased by
77% (P ⬍ 0.01), indicating a direct effect of LG 268 on
thyrotropes and not an indirect effect through TRH
suppression.
Sharma et al. • Effects of Rexinoids on the HPT Axis
To further explore the difference between T4 and T3 levels
after treatment with LG 268, tissue deiodinase mRNA and
activity was measured (Fig 3). D1 mRNA in the liver was
significantly increased by LG 268, and D2 in the pituitary was
significantly decreased, whereas brain D2 mRNA was unaffected. Both D1 activity in the liver and D2 activity in the
brain were significantly increased by treatment with LG 268
(P ⫽ 0.03 and 0.02, respectively), whereas D2 activity in the
pituitary was not significantly changed. The increase in liver
D1 mRNA and activity may explain why T3 levels are not as
low as T4 levels after treatment with LG 268.
FIG. 4. Effect of natural and synthetic retinoids on TSH
secretion and TSH␤ mRNA levels in T␣T1 thyrotrope cells.
T␣T1 cells (5 ⫻ 105) were grown to 70% confluence in
DMEM and 10% FBS. Fresh media were added with vehicle
(DMSO) or 1 ␮M retinoid. After 48 h, media were collected
for TSH assay and cells collected for total RNA extraction.
A, TSH assay on the media was performed as previously
described (15), using DMEM and 10% FBS as the diluent.
B, Quantitative RT-PCR for TSH␤ mRNA was performed as
previously described (15). mRNA levels are expressed as
attograms of TSH␤ per nanograms of rRNA. C, TSH␤
mRNA was measured after 48 h treatment of T␣T1 cells
with vehicle or increasing doses (0.001–10 ␮M) LG 268.
Results are an average (⫹SEM) of four separate experiments. *, Significant suppression by retinoid (P ⬍ 0.05); **,
significantly higher TSH, compared with vehicle control.
Downloaded from endo.endojournals.org at Harvard Libraries on June 27, 2007
Sharma et al. • Effects of Rexinoids on the HPT Axis
Endocrinology, March 2006, 147(3):1438 –1451 1443
Effect of retinoids on TSH protein and mRNA in a
thyrotrope cell line
TSH protein levels were measured in media from T␣T1
cells after 48 h in the presence of 1 ␮m retinoid or vehicle (Fig.
4A). 9-cis RA suppressed TSH levels by 35%, whereas 13-cis
RA and ATRA had no effect. The synthetic RXR-selective
ligand, LG 268, suppressed TSH levels by 60%, but the synthetic RAR-selective ligand, TTNPB, did not suppress TSH
levels (levels were significantly higher, compared with vehicle), suggesting that the suppression of TSH in the thyrotrope by retinoids is RXR mediated in T␣T1 cells. The suppressive effect of LG 268 on TSH levels was also seen after
6 d of treatment (data not shown).
T␣T1 cells were collected after 48 h of treatment with 1 ␮m
retinoid, and total RNA was prepared for analysis. Quantitative RT-PCR was performed using mouse TSH␤ sense RNA
to generate a standard curve (15). TSH␤ mRNA levels were
suppressed 64% by 9-cis RA (Fig. 4B). 13-cis RA had no effect
and ATRA had only a modest effect (30% suppression) on
TSH␤ mRNA levels. These natural retinoids can be interconverted by isomerase enzymes, and it is difficult to determine whether the effects are primarily through RAR or
RXR. The RXR-selective LG 268 suppressed TSH␤ mRNA
levels 74%, whereas the RAR-selective TTNPB had no effect,
indicating that retinoid-induced suppression of TSH␤
mRNA levels in thyrotropes is mediated through an RXRmediated mechanism. LG 268 concentrations as low as 0.01
␮m suppressed TSH␤ mRNA levels, and the maximal effect
appeared to be with 0.1 ␮m LG 268 (Fig 4C).
To determine how rapidly LG 268 suppresses TSH secretion and mRNA levels in T␣T1 thyrotropes, cells were exposed (in fresh media) to 1 ␮m LG 268 for 0 – 48 h, and at each
time point, media were collected and cells were harvested for
RNA collection. TSH was measurable in the media as early
as 1 h (Fig 5A), and levels continued to increase throughout
the 48 h. There was a significant effect of LG 268 on TSH
secretion in the media (P ⬍ 0.001). Specific significant inhibition of secretion was seen at 24 and 48 h (P ⬍ 0.05). There
was a significant effect of LG 268 on TSH␤ mRNA levels (Fig
5B, P ⬍ 0.001). Specific significant suppression was seen at
8, 24, and 48 h (P ⬍ 0.05), indicating that the decrease in TSH
secretion is preceded by a decrease in TSH␤ mRNA.
Effect of RAR-selective and RXR-selective retinoids on
mRNA levels of the ␣- and ␤-subunits of TSH, D2, and
RAR␤ in the T␣T1 thyrotrope cells
To determine the broader effect of retinoids on gene
regulation in the thyrotrope, mRNA levels of three genes
(␣- and ␤-subunits of TSH and D2) were measured in T␣T1
cells after 48 h of treatment with vehicle (DMSO) or 1 ␮m
LG 268 (RXR selective) or TTNPB (RAR selective) retinoid.
RAR␤ mRNA (RAR-responsive gene) was used as a positive control for TTNPB. Figure 6 shows that mRNA levels
for both subunits of TSH and D2 were significantly decreased by treatment with LG 268, whereas TTNPB had no
effect. As expected, TTNPB increased RAR␤ mRNA levels,
indicating that the RAR-signaling pathway is intact in
these cells. D2 activity was also significantly decreased
(42%) by LG 268 but not TTNPB (data not shown). This
FIG. 5. Temporal effect of LG 268 on TSH secretion and TSH␤ mRNA
levels in T␣T1 thyrotrope cells. T␣T1 cells (5 ⫻ 105) were grown to 70%
confluence in DMEM and 10% FBS. Fresh media were added with
vehicle (DMSO) or 1 ␮M LG 268. At different time intervals (0 – 48 h),
media were collected for TSH assay and cells collected for total RNA
extraction. A, TSH assay on the media was performed using DMEM
and 10% FBS as the diluent. B, Quantitative RT-PCR for TSH␤
mRNA was performed as previously described (15). Results are an
average (⫹SEM) of two separate experiments performed in triplicate.
*, Significant suppression by retinoid (P ⬍ 0.05).
would also suggest that TSH suppression by retinoids is
not mediated through an increased D2 activity and intracellular T3 levels in thyrotropes.
Effect of an RXR-selective retinoid on TSH␤ promoter
activity in T␣T1 cells
To determine whether the RXR-selective retinoid LG 268
can suppress TSH␤ promoter activity and identify important
regions in T␣T1 thyrotropes, cells were transfected with the
progressive TSH␤ 5⬘ flanking deletions in the presence or
absence of 1 ␮m LG 268 (Fig. 7); ⫺1240 to ⫹40 bp of the
mTSH␤ 5⬘-flanking region was used as the largest promoter
construct, and this promoter activity was inhibited 42% by
Downloaded from endo.endojournals.org at Harvard Libraries on June 27, 2007
1444
Endocrinology, March 2006, 147(3):1438 –1451
Sharma et al. • Effects of Rexinoids on the HPT Axis
FIG. 6. Effect of RXR-selective and RAR-selective retinoids
on ␣- and ␤-TSH subunit, D2, and RAR␤ mRNA levels in
T␣T1 thyrotrope cells. T␣T1 cells (5 ⫻ 105) were grown to
70% confluence in DMEM and 10% FBS. Fresh media were
added with vehicle (DMSO) or 1 ␮M RXR-selective (LG 268)
or RAR-selective (TTNPB) retinoid. After 48 h cells were
collected for total RNA extraction. Quantitative RT-PCR
(SYBR green) was performed using specific primers and
serially diluted TtT-97 mRNA as a standard control. Results are expressed as percent mRNA levels, compared with
vehicle for each mRNA, and are an average (⫹SEM) of four
separate experiments. *, Significant suppression by retinoid (P ⬍ 0.05).
LG 268. 5⬘ deletions between ⫺550 and ⫺200 resulted in a
small loss of inhibition of promoter activity by LG 268. However, deletion from ⫺200 to ⫺149 resulted in almost complete
loss of the inhibition of promoter activity by LG 268, suggesting that this region is important for mediating the effects
of LG 268 on TSH␤ promoter activity in these T␣T1
thyrotropes.
RXR isotype protein levels in thyrotrope-derived cell types
The RXR-selective retinoid LG 268 suppresses TSH protein levels, TSH␤ mRNA levels, and TSH␤ promoter activity in mice and T␣T1 thyrotrope-derived cells. This
effect is believed to be mediated by RXR, of which there
are three isotypes (RXR␣, RXR␤, and RXR␥). To determine
which isotype(s) are present, Western blot analysis was
performed on nuclear protein extracts from three thyrotrope-derived cells. TtT-97 and T␣T1 thyrotropes respond
to retinoid treatment with decreased TSH␤ promoter activity (3, 9), whereas ␣TSH cells lacks this response (9).
Figure 8 shows that RXR␣ is expressed at similar levels in
all three cell types. RXR␤ protein is detectable only in the
FIG. 7. Effect of LG 268 on TSH␤ promoter activity in T␣T1 cells.
mTSH␤ promoter-luciferase reporter constructs were generated as
previously described (15). Transient transfection was performed using Lipofectamine 2000 (Invitrogen). Cells were harvested 48 h after
transfection and promoter activity was measured as luciferase activity. A, 5⬘ deletion constructs of the ⫺1240 to ⫹40 mTSH␤ promoter
are displayed on the y-axis. Percent inhibition of promoter activity by
1 ␮M LG 268, compared with vehicle control, is shown on the x-axis.
Results are the average (⫹SEM) of three separate experiments performed in triplicate.
TtT-97 cells, whereas RXR␥ is expressed in the rexinoidresponsive TtT-97 and T␣T1 cells but not the rexinoidunresponsive ␣TSH cells. To determine which receptors
are associated with the TSH␤ promoter in vivo, we performed ChIP for each receptor in TtT-97 and T␣T1 cells.
Figure 9 shows that RXR␣ and RXR␥ are associated with
the TSH␤ promoter, but RXR␤ is not (RXR␤ is expressed
in TtT-97 cells). The amplified band for RXR␤ is no different from the nonspecific IgG control in any of the experiments, suggesting no direct interaction of RXR␤ and
FIG. 8. Western blot analysis of RXR isotypes in thyrotrope-derived
cells. Sixty micrograms of nuclear protein extract from each thyrotrope-derived cell line was size separated on a 10% SDS-polyacrylamide gel and transferred to a PVDF membrane. Membranes were
incubated overnight with RXR isotype-specific antibodies (RXR␣ and
RXR␤, Santa Cruz Biotechnology; RXR␥, Lab Vision). After washing,
membranes were incubated with antimouse IgG conjugated to horseradish peroxidase for 1 h at RT. Enhanced chemiluminescence (Amersham Pharmacia) detection reagents were used for immunodetection. Lamin B was used as a protein loading control. Results are
representative of three separate experiments.
Downloaded from endo.endojournals.org at Harvard Libraries on June 27, 2007
Sharma et al. • Effects of Rexinoids on the HPT Axis
Endocrinology, March 2006, 147(3):1438 –1451 1445
FIG. 9. In vivo occupancy of RXR isotypes with the
TSH␤ promoter in TtT-97 and T␣T1 thyrotrope cells.
The ChIP method is outlined in Materials and Methods.
TtT-97 or T␣T1 cells were exposed to vehicle or 1 ␮M LG
268 for 4 h, and then the solution was adjusted to 1%
formaldehyde to cross-link proteins/DNA. After washing, lysis and sonication (25–30 times for 15-sec pulses
with a microtip), 70 ␮g of chromatin were diluted in
immunoprecipitation buffer, precleared over A/G Sepharose beads, and immunoprecipitated using 2.5 ␮g of specific antibody. After washing, chromatin was eluted in
60 ␮l, cross-link was reversed, DNA was purified, and
4 – 8 ␮l were subjected to PCR with specific primers for
TSH␤ (⫺395 to ⫺31; 29 cycles) or glyceraldehyde-3phosphate dehydrogenase (GAPDH; 25 cycles) as a nonspecific control. IgG was used as a nonspecific antibody
control. Results are representative of four separate experiments.
the TSH␤ promoter in vivo. RXR␣ and RXR␥ interact with
the TSH␤ promoter in the absence and presence of
rexinoid.
93.7 ⫾ 5.9 ng/dl) were similar between the two groups.
Overall, plasma TSH, T4, and T3 levels were not different in
the WT and RXR␥-deficient mice, but TSH suppression was
less in the RXR␥KO mice at the lowest dose of LG 268 (1
Effects of RXR isotypes on rexinoid-mediated suppression of
TSH␤ promoter activity in ␣TSH cells
We have previously shown that transient transfection of
RXR␥ into ␣TSH cells reconstituted the effects of 9-cis RA on
TSH␤ promoter activity, whereas RXR␤ did not have this
effect (9). Rexinoid-responsive (TtT-97 and T␣T1) and nonresponsive (␣TSH) cells all express RXR␣ (Fig. 8), but only
the rexinoid-response cells express RXR␥. To determine
whether there is an RXR isotype effect on TSH␤ promoter
activity suppression by LG 268, ␣TSH cells were transiently
transfected with plasmids (pCGN2) containing RXR␥1,
RXR␥2, RXR␣, or RXR␤ cDNA. Different amounts of plasmid
were transfected to achieve equivalent amounts of each RXR
isotype protein (data not shown). Figure 10 shows that 1 ␮g
of pCGN2-RXR␥1 and 400 ng of each of the other RXR isotype plasmids generate similar amounts of protein by Western blot analysis. RXR␥1 mediates a 64% suppression of
TSH␤ promoter activity by 1 ␮m LG 268 in ␣TSH cells,
compared with empty vector (Fig. 11, P ⬍ 0.001). Similar
amounts of other RXR isotypes also mediated suppression of
TSH␤ promoter activity by LG 268, compared with empty
vector (P ⬍ 0.001), but to a lesser degree than RXR␥1 (RXR␥2,
34%; RXR␣, 33%; RXR␤, 24%). Higher amounts (1 ␮g) of each
transfected RXR isotype mediated a greater suppression of
TSH␤ promoter activity by LG 268 (RXR␥2, 45%; RXR␣, 81%;
RXR␤, 61%), suggesting that any isotype can mediate this
response, but the RXR␥1 isotype is the most efficient receptor
to mediate TSH␤ promoter activity suppression by rexinoids.
Effects of an RXR-selective retinoid on the HPT axis of
RXR␥-deficient mice
RXR␥-deficient (RXR␥KO) and littermate WT mice were
treated with increasing amounts of LG 268 for 3 d and thyroid
function tests were measured (Fig. 12). Baseline levels of TSH
[WT 28 ⫾ 7.5, knockout (KO) 30 ⫾ 4.3 mU/liter], T4 (WT
2.35 ⫾ 0.17, KO 2.08 ⫾.14 ␮g/dl), and T3 (WT 97.2 ⫾ 7.0, KO
FIG. 10. Western blot analysis of RXR isotypes in transiently transfected ␣TSH cells. Whole-cell protein extracts were prepared as previously described (16). Forty micrograms of protein (in duplicate from
each condition) was size separated on a 10% SDS-polyacrylamide gel
and transferred to a PVDF membrane. ␤-Actin was used as a protein
loading control. A, One thousand nanograms pCGN2 or 1000 ng of
each isotype cDNA in pCGN2 vector were transiently transfected into
␣TSH cells, and cells were incubated for 48 h before harvesting. B,
One thousand nanograms pCGN2-RXR␥1 or 400 ng pCGN2-RXR␥2,
RXR␣, or RXR␤ plus 600 ng pCGN2 were transiently transfected into
␣TSH cells, and cells were incubated for 48 h before harvesting.
Downloaded from endo.endojournals.org at Harvard Libraries on June 27, 2007
1446
Endocrinology, March 2006, 147(3):1438 –1451
Sharma et al. • Effects of Rexinoids on the HPT Axis
FIG. 11. Effects of LG 268 and RXR␥ isoforms on TSH␤
promoter activity in ␣TSH cells. Transient transfections
using 3 ␮g mTSH␤ (⫺1240 to ⫹40) promoter-luciferase
plasmid with 1 ␮g pCGN2, 1 ␮g pCGN2-RXR␥1, or 400
ng pCGN2-RXR␥2, RXR␣, or RXR␤ plus 600 ng pCGN2
were carried out as described in Materials and Methods.
Cells were incubated in the presence of vehicle (DMSO)
or 1 ␮M LG 268 for 48 h and then harvested for luciferase
assays. Results are expressed as percent TSH␤ promoter
activity, compared with DMSO control. Results are the
average (⫾SEM) of six separate experiments performed
in duplicate. *, Significant difference in TSH␤ promoter
activity suppression by RXR␥1, compared with each of
the other receptor isotypes (P ⬍ 0.001).
mg/kg䡠d). Pituitary TSH␤ mRNA levels were measured after
treatment (Fig. 13), and mice lacking RXR␥ had blunted
suppression of TSH␤ mRNA levels at each treatment dose of
LG 268, although TSH␤ mRNA was suppressed by LG 268
in RXR␥KO mice. These data suggest that RXR␥ is required
for the in vivo effect of rexinoids at low dose of rexinoids, but
other receptor isotypes can mediate this effect at higher doses
of rexinoid. D2 mRNA levels were measured from pituitaries
of mice (WT and RXR␥KO) after treatment with 3 d vehicle
or 10 mg/kg䡠d LG 268. Mice lacking RXR␥ had complete loss
of LG 268-mediated suppression of D2 mRNA levels (Fig.
14A), indicating that the RXR␥ receptor isotype is required
for this in vivo effect. D2 activity (Fig. 14B) was significantly
higher in the RXR␥KO mice (P ⬍ 0.05), and D2 activity was
slightly but not significantly higher in these mice after treatment with LG 268. Pituitary D1 activity did not differ between WT and RXR␥KO mice, and retinoid treatment significantly increased activity in both groups of animals (Fig.
14C).
Discussion
In this report, we have characterized the effects of RXRselective retinoids (rexinoids) on the HPT axis and deiodinase activity in mice as well as TSH protein levels, TSH␤
mRNA levels, and TSH␤ promoter activity in the thyrotropederived T␣T1 cell line. These results demonstrate that rexinoids can directly affect thyroid function at different levels,
and this is primarily mediated through RXR. It has long been
known that high doses of vitamin A can interfere with the
effects of thyroid hormone on metabolism (2, 23). Shadu and
Brody (23) demonstrated that high doses of vitamin A caused
significant decreases of thyroid weight in rats and postulated
that this effect may be through a central mechanism. Vitamin
A (retinol) can be converted to different natural retinoids,
which exert cellular effects through two classes of nuclear
hormone receptors RAR and RXR (1). A recent study by
Sherman et al. (3) showed that an RXR-selective retinoid
(bexarotene) can cause clinically significant central hypothyroidism (low T4 and low TSH) in patients treated with this
retinoid. An understanding of this mechanism will provide
insights into the role of retinoids and receptors in thyrotrope
function and may be useful as therapy in patients with disorders of TSH regulation including TSH-secreting adenomas
and the syndrome of thyroid hormone resistance.
Two groups have shown that patients treated with isotretinoin (13-cis RA) had no change in serum TSH levels (6, 7). In
contrast, Sherman et al. (3) clearly showed that an RXRselective retinoid suppressed serum TSH levels in patients
treated for cancer. Liu et al. (5) demonstrated that another
RXR-selective retinoid (LG 268) can dramatically decrease
serum TSH levels in rats as early as 30 min after administration, suggesting that the effect of this RXR-selective retinoid on thyrotropes occurs, at least in part, through inhibition of secretion. In the present study, we show that LG 268
also decreases circulating TSH and T4 levels in mice, and 3 d
of treatment decreases TSH␤ mRNA in the pituitary, but has
no effect on hypothalamic TRH levels, indicating that the
effect of retinoids on TSH suppression is directly on the
thyrotropes and not through hypothalamic regulation. Furthermore, we used the T␣T1 thyrotrope model to show that
LG 268 directly suppresses TSH␤ mRNA levels and TSH
secretion in thyrotropes, and this effect is seen as early as 8 h.
These in vitro effects are in contrast with the in vivo observations of Liu et al. (5). The differences may be due, in part,
to effects of rexinoids on clearance of TSH, which we would
not have in the T␣T1 model. TTNPB, an RAR-selective retinoid, did not suppress TSH secretion or TSH subunit mRNA
in T␣T1 cells, indicating that the observed effect is mediated
through RXR. This may also explain the negative clinical
studies with isotretinoin, which is primarily an RAR agonist.
There is limited information on the effects of retinoids on
deiodinase enzymes and no studies exploring the effects of
RXR-selective retinoids. Farwell and Leonard (24) showed
that retinoids (ATRA, 13-cis RA, retinol) had no effect on D2
activity in rat astrocytes, but a 2-fold stimulation was seen
with retinoids in the presence of cAMP. Our own data show
that brain D2 activity increases 2-fold with LG 268, whereas
mRNA levels are unchanged. We also observed that the
RXR-selective retinoid LG 268 decreased D2 activity in the
T␣T1 thyrotrope cells, but RAR-selective TTNPB had no
effect. This would suggest that the effect of retinoids on D2
in thyrotropes is through an RXR-mediated mechanism, and
Downloaded from endo.endojournals.org at Harvard Libraries on June 27, 2007
Sharma et al. • Effects of Rexinoids on the HPT Axis
Endocrinology, March 2006, 147(3):1438 –1451 1447
FIG. 12. Effect of LG 268 on serum hormone measurements
in WT and RXR␥KO mice. Mice (four to five of each genotype
in each group) were given vehicle or different doses of LG 268
(x-axis) daily for 3 d by oral gavage. Serum was collected 4 h
after the last dose. Data are expressed as percent, compared
with vehicle-treated mice (⫾SEM). *, Significant difference
(P ⬍ 0.05) in TSH levels between WT and RXR␥KO mice at
the 1 mg/kg䡠d dose (t test).
Downloaded from endo.endojournals.org at Harvard Libraries on June 27, 2007
1448
Endocrinology, March 2006, 147(3):1438 –1451
Sharma et al. • Effects of Rexinoids on the HPT Axis
FIG. 13. Effect of LG 268 on pituitary TSH␤ mRNA in WT
and RXR␥KO mice. Mice (four to five in each group) were
given vehicle or different doses of LG 268 (x-axis) daily for
3 d by oral gavage. The mice were killed 4 h after the last
dose, and pituitaries were collected for total RNA extraction. One hundred nanograms total RNA were used for
quantitative RT-PCR (ABI PRISM 7700) using specific
cDNA for standard curves. Data are expressed as percentage TSH␤ mRNA level corrected for an internal standard
(nanograms rRNA), compared with vehicle control. *, Significant difference (P ⬍ 0.05) between vehicle and LG 268.
this is different from the effects observed in the brain (upregulated D2 activity only). D2 is regulated by thyroid hormone through two distinct mechanisms: T4 decreases D2
mRNA and increases protein degradation through a ubiquitination pathway (13, 25). Our data with rexinoids can be
explained by the effects of rexinoids on mRNA levels (pretranslational) and the effects of T4 on protein degradation
(posttranslational). In the pituitary, LG 268 directly decreases
D2 mRNA, whereas lower serum T4 levels in the animals
result in decreased D2 protein degradation. These opposing
effects result in no significant change in pituitary D2 activity.
This hypothesis is further strengthened by the observations
in T␣T1 cells (decreased D2 mRNA and activity) because the
T4 levels in the media are unaffected by treatment with LG
268. Interestingly, D2 mRNA in the brain is not altered by LG
268, suggesting a tissue-specific effect of rexinoids on D2
mRNA. D2 activity in the brain, however, is increased presumably through the lower T4 levels and decreased D2 protein degradation.
Taruoura et al. (26) found no effect of high-dose etretinate
(ATRA precursor) on D1 activity in rats treated with this
retinoid, but other retinoids were not tested. Schreck et al.
showed that D1 activity is increased in a HepG2 liver cell line
by ATRA and 9-cis RA (27), which is consistent with what we
found in livers of mice treated with LG 268. Retinoid stimulation of D1 appears to occur in cell lines from different
types of cancer as well (27, 28). This effect may occur through
a direct stimulation of gene transcription based on studies of
the D1 promoter (29, 30). Low T4 levels in our treated mice
may explain the higher D2 activity in the brain, but the effects
are the opposite for liver D1, suggesting a direct mechanism
of LG 268 on liver D1. This stimulatory effect of LG 268 on
D1 activity is not confined to the liver because pituitary D1
activity also increased in WT and RXR␥KO mice.
The direct effects of retinoids on thyrotrope function have
been studied using natural retinoids that can be interconverted by isomerization, which makes dissection of specific
RAR and RXR pathways difficult. Breen et al. (31) examined
the effects of ATRA on TSH␤ mRNA levels and promoter
activity in a rat and in vitro model. Rats were treated for a
total of 60 d with either a vitamin A-deficient diet or a
supplemented diet with retinyl palmitate. The investigators
observed a significant increase in total T4 in the vitamin
A-deficient animals, and TSH␤ mRNA levels were noted to
be 2-fold higher in the vitamin A-deficient rats. TSH␤ promoter activity was inhibited by 56% using 0.5 ␮mol ATRA
in a CV-1 cell line transfection model. This inhibition of
promoter activity required the presence of both RAR and
RXR. This same group went on to localize the retinoid-responsive region of the TSH␤ promoter between ⫺209 and ⫹9
(32). They also demonstrated an additive and possibly synergistic effect between RA and T3 on TSH␤ promoter activity.
Our group examined the effects of 9-cis RA on TSH␤ promoter activity in a mouse thyrotropic tumor model (9). 9-cis
RA significantly decreased TSH␤ promoter activity in the
TtT-97 thyrotropic tumor cells. We identified the ⫺200 to
⫺149 region of the mouse TSH␤ 5⬘ flanking DNA as necessary for mediating the effect of 9-cis RA on TSH␤ promoter
activity in this model. The effects of different natural and
synthetic retinoids on TSH protein secretion or TSH␤ mRNA
levels have not been thoroughly explored in the TtT-97 thyrotropic tumor model. This hyperplastic thyrotrope tumor is
difficult to maintain and produce, and cells do not divide or
survive long in primary culture.
Alarid et al. (10) developed an immortalized, thyrotropederived pituitary cell line called T␣T1. This cell line was
generated by expression of the Simian virus 40 early region
coding sequence for both large and small T antigens under
direction of the ␣-subunit glycoprotein hormone promoter
(⫺5.5 to ⫹49). These cells express TSH␤ and ␣-subunit
mRNA, and treatment with T3 causes a dose- and timedependent decrease in TSH␤ mRNA, suggesting that this
represents an excellent model of a functional thyrotrope (11).
In this report, we have demonstrated that these T␣T1 cells
secrete TSH protein into the media, and these protein levels
are affected by treatment with retinoids. We have further
shown that treatment of T␣T1 cells with an RXR-selective
retinoid (LG 268) resulted in a significant decrease in TSH
levels in the media and TSH subunit mRNA, and treatment
of these cells with an RAR-selective retinoid (TTNPB) did not
decrease TSH levels or subunit mRNA in this thyrotrope
model. These data would suggest that TSH production
Downloaded from endo.endojournals.org at Harvard Libraries on June 27, 2007
Sharma et al. • Effects of Rexinoids on the HPT Axis
Endocrinology, March 2006, 147(3):1438 –1451 1449
FIG. 14. Effect of RXR␥ on rexinoid-mediated suppression of pituitary D2 mRNA. Mice (four to six in each
group) were given vehicle or 10 mg/kg䡠d LG 268 (x-axis)
daily for 3 d by oral gavage. A, Pituitary D2 mRNA
(quantitative RT-PCR, TtT-97 mRNA as standard
curve) is expressed as picograms mRNA per nanogram
rRNA. *, Significant difference between WT and
RXR␥KO mice (P ⬍ 0.05). B, Pituitary D2 activity. C,
Pituitary D1 activity. Protein extract (25–100 ␮g) was
used in each 5⬘D activity assay. Enzyme activity is expressed as femtomoles per hour per milligram.
and/or secretion in thyrotropes can be directly affected by
rexinoids, and this occurs through an RXR-mediated mechanism. This effect is seen as early as 8 –24 h of treatment. Liu
et al. (5) showed that rats treated with a single dose of LG 268
had suppressed TSH levels as early as 30 min, which is quite
different from our observations in the T␣T1 cells. One explanation for this discrepancy could be that clearance of TSH
is rapidly increased by LG 268, which would not be seen in
our cell line. Direct effects on secretion may not occur until
8 h after treatment. RXR-selective retinoids also decrease
TSH subunit mRNA levels in the thyrotrope-derived T␣T1
cells, and RAR-selective retinoids have no effect. The natural
retinoids had a variable effect on TSH␤ mRNA levels.
We have further shown that LG 268 suppresses TSH␤
promoter activity in these cells, and this occurs primarily
through the ⫺200 to ⫺149 region, confirming the observation
Downloaded from endo.endojournals.org at Harvard Libraries on June 27, 2007
1450 Endocrinology, March 2006, 147(3):1438 –1451
in the TtT-97 thyrotropic tumor model (9). These data would
suggest that RXR-selective retinoids can affect TSH␤ gene
transcription and that specific elements in the proximal promoter are responsible for a majority of this effect. Inspection
of this region reveals many potential nuclear receptor halfsites, but no classical retinoid-response elements (9). A report
by Castelein et al. (33) showed that the DR1 element was a
high-affinity binding site for the RXR homodimer, but other
elements had high-affinity as well (DR2, PAL0, DR6, DR0).
The ⫺200 to ⫺149 region of the TSH␤ promoter contains two
putative binding sites for the RXR homodimer (DR0, DR7).
Suppression of TSH␤ promoter activity by rexinoids may
occur through the DR0 or DR7 element, a novel negative
regulatory element or perhaps through a non-DNA binding
mechanism. This does not appear to occur through a classical
RAR/RXR heterodimer because the RAR-selective ligand
TTNPB had no effect. We predict that the effect of rexinoids
on thyrotrope function occurs through RXR homodimers or
heterodimers between RXR and other partners (peroxisomal
proliferator-activated receptor, liver X receptor, farnesoid X
receptor, vitamin D receptor, and others).
We and others have shown that one particular RXR isotype, RXR␥, is uniquely expressed in thyrotropes and thyrotrope-derived cells (9, 34, 35). To further explore this observation in the T␣T1 thyrotrope model, we performed
Western blot analysis on three thyrotrope-derived cell types.
All three cell types expressed RXR␣ protein, whereas only
the TtT-97 cells expressed RXR␤ protein. RXR␥ expression
was limited to the retinoid-responsive TtT-97 and T␣T1 cells,
and this receptor was not detected in the retinoid-nonresponsive ␣TSH cells (9). We have previously shown that
RXR␥-deficient mice have higher serum TSH and T4 levels,
suggesting that this receptor is important in the retinoidmediated suppression of TSH (15). Introduction of each RXR
isotype into the thyrotrope-derived ␣TSH cells, which lack
RXR␤ and RXR␥ but express RXR␣, could mediate suppression of TSH␤ promoter by LG 268, suggesting RXR isotype
redundancy in mediating this effect. RXR␥1 mediated a
greater suppression of TSH␤ promoter activity than any
other RXR isotype when similar amounts of protein were
expressed. RXR␥1 contains a unique N-terminal region that
may be required for optimal suppression of TSH␤ gene transcription by retinoids.
To explore this observation in an in vivo model, we compared the effects of LG 268 on littermate RXR␥KO and WT
mice. After 3 d of treatment, the mice had similar effects of
LG 268 on TSH, T4, and T3 levels, indicating that RXR␥ is not
absolutely required for the effects of retinoids on TSH, T4,
and T3 over this period of time. Pituitary levels of TSH␤
mRNA were differentially affected. The lowest dose of LG
268 (1 mg/kg䡠) suppressed mRNA levels and serum TSH in
WT mice but had less of suppressive effect in the RXR␥deficient mice, whereas higher doses suppressed TSH␤
mRNA and serum TSH levels in all mice. These data are
consistent with our in vitro transfection data in ␣TSH cells,
which shows that any RXR isotype can mediate rexinoid
suppression of TSH␤ promoter activity, but RXR␥ appears to
be the most efficient receptor to mediate this response. A
novel and unexpected finding in our studies was the complete loss of LG 268-mediated suppression of D2 mRNA in
Sharma et al. • Effects of Rexinoids on the HPT Axis
the pituitaries of RXR␥KO mice. These studies indicate that
RXR isotypes are not completely redundant and there are
RXR␥-specific effects of retinoids on thyroid and thyrotrope
function.
In addition to the direct effect of retinoids to suppress
TSH␤ promoter function, an increase in pituitary T3 content,
through changes in either serum thyroid hormone levels
and/or pituitary deiodinase expression, could decrease serum TSH levels. In that regard, pituitary D2 activity is unchanged by treatment of mice with LG 268. This, combined
with the lower serum T4 levels, make the D2 an unlikely
source of additional T3. In contrast, pituitary D1 activity is
increased approximately 60% in retinoid-treated mice. However, the role of the D1 in contributing to the pituitary T3
content and TSH expression remains uncertain. Recent evidence from D1 knockout animals suggests this enzyme is
either not involved or involved to only a limited extent in
TSH regulation; TSH levels in D1-deficient animals are normal (36).
In summary, RXR-selective retinoids (rexinoids) have
multiple effects on the HPT axis. Rexinoids directly suppress
TSH secretion, TSH␤ mRNA levels and promoter activity,
and D2 mRNA levels but have no direct effect on hypothalamic TRH levels. Rexinoids also stimulate D1 activity in the
liver and pituitary.
Acknowledgments
We thank Cynthia Kramer and Andrew Berenz for technical
assistance.
Received June 13, 2005. Accepted November 17, 2005.
Address all correspondence and requests for reprints to: Bryan R.
Haugen, M.D., University of Colorado at Denver and Health Sciences
Center, MS 8106, P.O. Box 6511, Aurora, Colorado 80045. E-mail:
[email protected]
This work was supported by National Institutes of Health Grant
DK54383. We acknowledge use of the Gene Expression Core Facility
DNA Sequencing Core Facility and Animal Care Facility of the University of Colorado Cancer Center. The T␣T1 cells were generously
provided by Dr. Pamela Mellon (University of California, San Diego, San
Diego, CA). TTNPB and LG100268 (LG 268) were generously provided
by Ligand Pharmaceuticals (San Diego, CA).
Results from this work were presented in part at the 75th Annual
Meeting of the American Thyroid Association, Palm Beach, Florida,
September 16 –21, 2003.
The authors have no conflict of interest.
References
1. Haugen BR 2004 The effect of vitamin A, retinoids and retinoid receptors on
the hypothalamic-pituitary-thyroid axis. In: Beck-Peccoz P, ed. Syndromes of
hormone resistance on the hypothalamic-pituitary-thyroid axis. Boston: Kluwer; 149 –163
2. Simkins S 1947 Use of massive doses of vitamin A in the treatment of hyperthyroidism. J Clin Endocrinol Metab 7:574 –585
3. Sherman SI, Gopal J, Haugen BR, Chiu AC, Whaley KNP, Duvic M 1999
Central hypothyroidism associated with retinoid X receptor selective ligands.
N Engl J Med 340:1075–1079
4. Duvic M, Martin AG, Kim Y, Olsen E, Wood GS, Crowley CA, Yocum RC
2001 Phase 2 and 3 clinical trial of oral bexarotene (Targretin capsules) for the
treatment of refractory or persistent early-stage cutaneous T-cell lymphoma.
Arch Dermatol 137:581–593
5. Liu S, Ogilvie KM, Klausing K, Lawson MA, Jolley D, Li D, Bilakovics J,
Pascual B, Hein N, Urcan M, Leibowitz MD 2002 Mechanism of selective
retinoid X receptor agonist-induced hypothyroidism in the rat. Endocrinology
143:2880 –2885
6. Marsden JR, Trinick TR, Laker MF, Shuster S 1984 Effects of isotretinoin on serum
lipids and lipoproteins, liver and thyroid function. Clin Chim Acta 143:243–251
Downloaded from endo.endojournals.org at Harvard Libraries on June 27, 2007
Sharma et al. • Effects of Rexinoids on the HPT Axis
7. O’Leary TJ, Simo IE, Kanigsberg ND, Ooi TC 1986 Lack of effect of isotretinoin on thyroid-function tests. Clin Chem 32:913–914
8. Dabon-Almirante CL, Damle S, Wdaler S, Hupart K 1999 Related case report:
in vivo suppression of thyrotropin by 9-cis retinoic acid. Cancer J Sci Am
5:171–173
9. Haugen BR, Brown NS, Wood WM, Gordon DF, Ridgway EC 1997 The
thyrotrope-restricted isoform of the retinoid X receptor (␥1) mediates 9-cis
retinoic acid suppression of thyrotropin beta promoter activity. Mol Endocrinol 11:481– 489
10. Alarid ET, Windle JJ, Whyte DB, Mellon PL 1996 Immortalization of pituitary
cells at discrete stages of development by directed oncogenesis in transgenic
mice. Development 122:3319 –3329
11. Yusta B, Alarid ET, Gordon DF, Ridgway EC, Mellon PL 1998 The TSH␤
subunit gene is repressed by thyroid hormone in a novel thyrotrope cell line,
mouse T␣T1 cells. Endocrinology 139:4476 – 4482
12. Krezel W, Dupe V, Mark M, Dierich A, Kastner P, Chambon P 1996 RXR ␥
null mice are apparently normal and compound RXR␣⫹/⫺, RXR␤⫺/⫺,
RXR␥⫺/⫺ mutant mice are viable. Proc Natl Acad Sci USA 93:9010 –9014
13. Croteau W, Davey JC, Galton VA, St. Germain DL 1996 Cloning of the
mammalian type II iodothyronine deiodinase. A selenoprotein differentially
expressed and regulated in human and rat brain and other tissues. J Clin Invest
98:405– 417
14. Bradford MM 1976 A rapid and sensitive method for the quantification of
microgram quantities of protein utilizing the principle of protein dye binding.
Anal Biochem 72:248 –254
15. Brown NS, Smart A, Sharma V, Brinkmeier ML, Greenlee L, Camper SA,
Jensen DR, Eckel RH, Krezel W, Chambon P, Haugen BR 2000 Thyroid
hormone resistance and increased metabolic rate in the RXR-␥-deficient
mouse. J Clin Invest 106:73–79
16. Haugen BR, Larson LL, Pugazhenthi U, Hays WR, Klopper JP, Kramer CA,
Sharma V 2004 Retinoic acid and retinoid X receptors are differentially expressed in thyroid cancer and thyroid carcinoma cell lines and predict response
to treatment with retinoids. J Clin Endocrinol Metab 89:272–280
17. Wood WM, Kao MY, Gordon DF, Ridgway EC 1989 Thyroid hormone regulates the mouse thyrotropin ␤ subunit gene promoter in transfected primary
thyrotropes. J Biol Chem 264:14840 –14847
18. Wood WM, Ocran KW, Kao MY, Gordon DF, Alexander LM, GutierrezHartmann A, Ridgway EC 1990 Protein factors in thyrotropic tumor nuclear
extracts bind to a region of the mouse thyrotropin ␤-subunit promoter essential
for expression in thyrotropes. Mol Endocrinol 4:1897–1904
19. Haugen BR, Gordon DF, Nelson AR, Wood WM, Ridgway EC 1994 The
combination of Pit-1 and Pit-1T have a synergistic stimulatory effect on the
thyrotropin ␤ subunit promoter but not the growth hormone or prolactin
promoters. Mol Endocrinol 8:1574 –1582
20. Gordon DF, Lewis SR, Haugen BR, James A, McDermott MT, Wood WM,
Ridgway EC 1997 Pit-1 and GATA-2 interact and functionally cooperate to
activate the thyrotropin ␤ subunit promoter. J Biol Chem 272:24339 –24347
21. Haugen BR, Wood WM, Gordon DF, Ridgway EC 1993 A thyrotrope-specific
Endocrinology, March 2006, 147(3):1438 –1451 1451
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
variant of Pit-1 transactivates the thyrotropin ␤ promoter. J Biol Chem 268:
20818 –20824
Boyd KE, Farnham PJ 1999 Coexamination of site-specific transcription factor
binding and promoter activity in living cells. Mol Cell Biol 19:8393– 8399
Shadu DP, Brody S 1947 Excess vitamin A ingestion, thyroid size and energy
metabolism. Am J Physiol 149:400 – 404
Farwell AP, Leonard JL 1992 Dissociation of actin polymerization and enzyme
inactivation in the hormonal regulation of type II iodothyronine 5⬘-deiodinase
activity in astrocytes. Endocrinology 131:721–728
Bianco AC, Salvatore D, Gereben B, Berry MJ, Larsen PR 2002 Biochemistry,
cellular and molecular biology, and physiological roles of the iodothyronine
selenodeiodinases. Endocr Rev 23:38 – 89
Taruoura D, Higueret P, Garcin H, Higueret D 1991 Effects of a synthetic
retinoid, etretinate (RO 10 –9359), on the peripheral metabolism of thyroid
hormones in rats. Food Chem Toxicol 29:595–599
Schreck R, Schnieders F, Schmutzler C, Kohrle J 1994 Retinoids stimulate
type I iodothyronine 5⬘-deiodinase activity in human follicular thyroid carcinoma cell lines. J Clin Endocrinol Metab 79:791–798
Garcia-Solis P, Aceves C 2003 5⬘Deiodinase in two breast cancer cell lines:
effect of triiodothyronine, isoproterenol and retinoids. Mol Cell Endocrinol
201:25–31
Jakobs TC, Schmutzler C, Meissner J, Kohrle J 1997 The promoter of the
human type I 5⬘-deiodinase gene—mapping of the transcription start site and
identification of a DR⫹4 thyroid-hormone-responsive element. Eur J Biochem
247:288 –297
Zhang CY, Kim S, Harney JW, Larsen PR 1998 Further characterization of
thyroid hormone response elements in the human type 1 iodothyronine deiodinase gene. Endocrinology 139:1156 –1163
Breen JJ, Matsuura T, Ross AC, Gurr JA 1995 Regulation of thyroid-stimulating hormone ␤-subunit and growth hormone messenger ribonucleic acid
levels in the rat: effect of vitamin A status. Endocrinology 136:543–549
Breen JJ, Hickok NJ, Gurr JA 1997 The rat TSH␤ gene contains distinct
response elements for regulation by retinoids and thyroid hormone. Mol Cell
Endocrinol 131:137–146
Castelein H, Janssen A, Declercq PE, Baes M 1996 Sequence requirements for
high affinity retinoid X receptor ␣ homodimer binding. Mol Cell Endocrinol
119:11–20
Sugawara A, Yen PM, Qi Y, Lechan RM, Chin WW 1995 Isoform-specific
retinoid-X receptor (RXR) antibodies detect differential expression of RXR
proteins in the pituitary gland. Endocrinology 136:1766 –1774
Sanno N, Sugawara A, Teramoto A, Abe Y, Yen PM, Chin WW, Osamura RY
1997 Immunohistochemical expression of retinoid X receptor isoforms in human pituitaries and pituitary adenomas. Neuroendocrinology 65:299 –306
Galton VA, Schneider MJ, St. Germain DL, Wu SY, Changes in thyroid
hormone (TH) economy in type 1 deiodinase (D1) knock-out mice. Proc 76th
Annual Meeting of the American Thyroid Association, Vancouver, Canada,
2004, p 715 (Abstract 105)
Endocrinology is published monthly by The Endocrine Society (http://www.endo-society.org), the foremost professional society serving the
endocrine community.
Downloaded from endo.endojournals.org at Harvard Libraries on June 27, 2007