Differential Modulation of Thyroid Hormone Responsiveness by ...

0013.7227/96/$03.00/O Endocrinology Copyright 0 1996 by The Endocrine

Vol Prmted

137, No. 8 rn U.S.A.

So&t)

Differential Modulation of Thyroid Hormone Responsiveness by Retinoids in a Human Cell Line* T. C. CROWE’F, J. W. BARLOW

N. M. LOIDL,

K. L. PAYNE,

Ewen Downie Metabolic Unit and Department Melbourne, Victoria 3181, Australia

D. J. TOPLISS,

of Medicine,

Monash

ABSTRACT Previous studies have suggested that there is an interrelationship between responses mediated by retinoic acid (RA) and those to thyroid hormone CT,). These experiments have used transfected gene constructs, often in receptor-negative cells. To study the relationship between RA- and T,-mediated responses in intact human cells, we incubated HepGP cells for 4 days in serum-free medium with T, and/or RA or 9-cis-RA. Measured responses were stimulation of secreted sex hormone-binding globulin (SHBG) or inhibition of secreted T,-binding globulin (TBG). T, induced a dose-responsive increase in SHBG secretion that was maximal at 10 nM (206 2 24% of untreated value) and half-maximal at 0.36 2 0.16 nM T,. RA and 9-cis-RA, up to 100 nM, induced a slight fall in SHBG secretion to 79 t 9% and 88 -+ 9%, respectively. T, induction of SHBG secretion was significantly attenuated in cells coincubated with T, (O-10 KIM) and RA. With T, (10 nM) together with RA (3, 10, or 100 nM), the maximal SHBG responses were reduced to 193 ? 24%, 151 ? 5%, and 132 + 30%, respectively. With T, and 9-cis-RA (100 nM), maximal stimulation was 169 ? 20%. Importantly, the effective half-maximal stimulatory concentration of

V

ITAMIN A (retinol) and its biologically active derivatives, such as all-trans-retinoic acid (RA) and 9-cis-RA, collectively called retinoids, have profound effects on vision, reproduction, growth and differentiation of organisms, as well as embryonal morphogenesis (1). Inside cells, vitamin A serves as the precursor for the biologically active RA that is produced by intracellular oxidation. Intracellular isomerases further convert some of RA to 9-cis-RA (2). The diverse biological effects of the retinoids are mediated by structurally related receptors (RAR and RXR) that bind to hormone-responsive elements in target genes and activate transcription in the presence of their respective ligands (3). The RXRs bind 9-cis-RA with high affinity, but do not bind RA, whereas the RARs bind both RA and 9-cis-RA. High levels of RARa messenger RNA (mRNA) are normally present in many tissues, whereas the expression of RARP and RARy is more tissue specific (4). The three RXR genes are all Received November 30, 1995. Address all correspondence and requests for reprints to: Dr. J. W. Barlow, Ewen Downie Metabolic Unit, Alfred Hospital, Commercial Road, Prahran, Victoria 3181, Australia. * This work was supported by a grant from the Alfred Foundation and by the Alfred Hospital Whole-Time Medical Specialists’ Trust. Presented in part at the 11th International Thyroid Congress, Toronto, Ontario, Canada, September 1995, and the Annual Scientific Meeting of the Endocrine Society of Australia, Melbourne, Australia, September 1995. The Alfred Hospital is part of the Eastern Health Care Network. t Present address: University of Leipzig, Clinic for Internal Medicine, Johannesalle 32, D-7010 Leipzig, Germany.

3187

J. R. STOCKIGT,

University,

Alfred

AND

Hospital,

T, in the presence of either retinoid (3-100 nM) was unchanged at 0.3 nM T,. In addition, the inhibitory effect of 9-cis RA could not be overcome even with 300 nM T,. The threshold for the RA effect was between 0.3-l nM, with half-maximal inhibition at 30 nM. 9-cis-RA was -lo-fold less potent than RA. Preliminary studies suggested that changes in SHBG messenger RNA levels were similar to those in secreted SHBG. No effect was observed with vitamin D or clofibrate, either alone or combined with T,. Conversely, T, reduced TBG secretion, with maximal suppression to 74 2 5% of the control value at a T, concentration of 10 nM. RA alone reduced TBG secretion to 76% of the control value. RA did not attenuate the effect of T,, and the two agents combined showed no synergism. Neither T, nor RA, alone or in combination, influenced secreted total protein or albumin. RA did not alter the concentration of nuclear T,-binding sites. These data suggest that retinoids act via a gene-dependent mechanism to modulate maximal, but not half-maximal, responses to T, in HepG2 cells with the specificity of RA greater than that of 9-cis-RA. (Endocrinology 137: 3187-3192, 1996)

widely expressed in adult tissues, and at least one isoform is present in every tissue examined to date, an observation that is consistent with the role of RXRs as ubiquitous coregulators of nuclear receptor function (3, 5). Like other members of the steroid-thyroid superfamily of nuclear receptors, the receptors for RA are able to influence gene expression by direct interactions with specific DNA sequences, termed RA response elements. RARs form RAR: RAR homodimers and RAR:RXR heterodimers on these elements as well as form heterodimers with other members of the receptor family, including those for T, (3). Receptors for T, (TRs) and RARs contain specific domains associated with heterodimerization (6). Interactions between these receptors may occur at the response element, as illustrated by the control of the GH gene in rat pituitary cells. GH production in these cells is activated by RA as a single agent that also significantly potentiates the effects of low doses of T, and glucocorticoids (7,8). Other genes that have also been shown to be influenced by RA or T, include those for human alcohol dehydrogenase (9), phosphoenol pyruvate carboxykinase (lo), corticosterone-binding globulin (ll), and placental lactogen (12). RA also regulates its own receptor, RAR (13-15). As well as its direct effects on gene transcription, RA may attenuate other ligand-induced responses, including those of T,. For example, Davis and Lazar (8) showed that RA had little effect on the expression of TRP2, which was markedly decreased by T,, although in the presence of T,, RA was able to completely inhibit the down-regulation of TRP2 mRNA.

3188

MODULATION

OF T, RESPONSIVENESS

They proposed that heterodimerization of endogenous TRs with RARs and/or RXRs could result in a complex that is incapable of mediating transcriptional repression of TR82 mRNA. Similar studies with the aF-crystallin gene were not compatible with this model, but, rather, suggested that an unidentified accessory factor promoted RAR/ RXR binding over TR/ RXR binding (16). The apolipoprotein AI regulatory protein 1 has recently been shown to fulfill this role on a synthetic thyroid hormone response element (TRE) (17). Taken together, these results suggest that modulation of T,-induced gene expression by RA depends on the cell type, the orientation and spacing of the TRE and/or RA response element, and the TR heterodimerization partners. Although there is abundant evidence to suggest an interrelationship between responses mediated by T, and RA, most experimental models designed to study this relationship have used artificial gene constructs in which multiple TREs are attached to reporter genes and transfected into a receptor-negative host cell. In contrast, the system reported here uses unmanipulated cells and is well suited to a more physiological appraisal of the influence of RA on the action of T,. Cells of the human hepatoblastoma line, HepG2, secrete a range of liver proteins that are influenced by T, and are easily detected in cell supernatants. We measured sex hormone-binding (SHBG), which is up-regulated by T, (18, 19), and T,-binding globulin (TBG), which is down-regulated by T, (20), and have investigated the effects of RA on these well defined responses. Materials

and

Methods

HepG2 cells (American Type Culture Collection, Rockville, MD) were cultured in DMEM as previously described (18-20). Cells were maintained in serum-free DMEM for experimental manipulations. Cells were incubated with T, (Henning, Berlin, Germany), RA (Sigma Chemical Co., St. Louis, MI), 9-cis-RA (Hoffman LaRoche, Nutley, NJ), vitamin D, (Aldrich Chemical Co., Milwaukee, WI), or clofibrate (Aldrich), either alone or in combination, for 4 days before protein measurement or RNA extraction. SHBG and TBG were measured by immunoradiometric assays from Orion Diagnostica (Espoo, Finland) and Henning, respectively. In the absence of any stimulating factors, the supernatant concentrations of SHBG and TBG were 43 2 18 nM and 13 + 6 mg/liter, respectively. Northern analysis of SHBG mRNA using a SHBG-specific probe was performed as previously described (18). Albumin was measured in cell supernatants by RIA (Techno Genetics, Cassina de’ Pecchi, Italy). Total protein was measured by dye binding (Bio-Rad Laboratories, Richmond, CA). To determine the effect of RA on nuclear T, binding, cells (-1 X lo7 cells/ flask) were incubated in serum-free medium either with or without RA (l-100 nM) for 4 days. Cells were then trypsinized, washed, and incubated in serum-free medium (0.5 ml) containing 1’2511T, (70 DM) and unlabeled T, (10 nM). After 2 h at‘37 C, incubationgwere terminated by placing the tubes on ice. The cells were washed once with 500 ~1 serumfree medium and lysed at 4 C with SMTT buffer (20 rnM Tris, 1.1 mM MgCl,, and 250 mM sucrose, pH 7.85, containing 0.5% Triton X-100), then washed once in the same buffer. The final nuclear pellet, obtained by centrifugation (1000 X g, 10 min at 4 C), was assessed for residual radioactivity. Preliminary studies using this method revealed this binding to be displaceable by unlabeled T, with a K, of about 0.3 nM.

Influence

of retinoids

Results on SHBG and TBG secretion

Retinoids had only a minimal direct effect on SHBG and TBG secretion by HepG2 cells. After 4 days of incubation in

Endo. 1996 Vol 137 . No 8

BY RETINOIDS

serum-free DMEM containing either RA or 9-cis-RA (O-1000 nM), a slight, but significant, decrease in secreted concentrations of both proteins was observed (Fig. 1). Maximal inhibition values at 100 nM RA and 9-cis-RA (*SD; n = 4-5), respectively, expressed as a percentage of the untreated control value, were 79 2 10% (P < 0.001) and 88 Ifr 9% (P < 0.05) for SHBG and 76 + 18% (P < 0.05) and 75 + 11% (P < 0.001) for TBG. Effect of RA on Ts stimulation

of SHBG

secretion

T, positively stimulated the secretion of SHBG to 206 ? 25% (n = 11) of the control value (Fig. 2), with a half-maximal effect at 0.3 nM, corresponding to approximately 40 PM bioavailable T, (18). When cells were incubated with RA in addition to T,, maximal stimulation was reduced. At 10 nM T, together with 3, 10, or 100 nM RA, the maximal response (CSD; n 2 3) was reduced to 193 2 24% (P < 0.05), 151 -+ 5% (P < O.OOl), and 132 t 30% (P < 0.001) of the control value, respectively. Significantly, the half-maximal concentration of T, in the presence of all concentrations of RA remained unchanged at about 0.3 nM. The converse experiment, in which cells were incubated with various concentrations of RA in the presence of a single dose of T,, is shown in Fig. 3. RA overcame T,-induced SHBG secretion with a threshold between 0.3-l nM RA and halfmaximal inhibition at approximately 3 nM. Effect of 9-cis-RA

on Ts stimulation

of SHBG

secretion

Similar experiments were performed with 9-cis-RA, a biologically active isomer of RA. The maximal response to 10 nM T, was attenuated in the presence of 100 nM 9-cis-RA from 206 2 26% to 169 2 38% (*SD; n = 6; P < 0.005), with no change in half-maximal stimulation (Fig. 4). This inhibitory effect of 9-cis-RA could not be overcome even with T, concentrations up to 300 nM. Like RA, the 9-cis isomer (O-300 nM) also overcame maximal T, stimulation. The threshold for this effect was 3 nM SHBG

II

0.1

Retinoid

I

Concentration,

TBG

10

100

nM

II

Retinoid

0.1

I

Concentration,

10

100

nM

FIG. 1. Effects of retinoids on the secretion of SHBG (left panel) and TBG (right panel) by HepG2 cells. Cells were incubated in serum-free medium with either RA (0) or 9-cis-RA (0) for 4 days. Each point represents the mean + SD of at least four experiments, each performed in triplicate. Data are normalized to the untreated control value, which is designated 100%. RA and 9-cis-RA (100 nM) induced a slight fall in SHBG secretion to 79 2 10% and 88 2 9%, respectively, and in TBG secretion to 76 2 18% and 75 + ll’%, respectively. ‘@,P < 0.05; **, P < 0.001 (compared to untreated control value).

MODULATION


BY RETINOIDS

3189

+/ 225

e

T

TJ Alone

T+3nMRA s

h

.IB z

200

+z-

I

A

T

+ IOnMRA

175

i z

150

E wl

125

--o-

RA+

10nMT3

+

9ckR4+

10nMT3

75 100

d’ 0

1

.l

TR Concentration,

nM

FIG. 2. Influence of RA on T,-stimulated SHBG secretion in HepG2 cells. Cells were incubated for 4 days in serum-free medium with T, (O-10 nM) alone (0) or together with R.A 13 (‘I), 10 CO), or 100 (0) nM]. Maximal stimulation in the presence of T, alone was 220 2 16% of the untreated control value at 10 nM T,. The maximal response was attenuated in the presence of RA (3,10, or 100 nM) to 193 ? 24%, 151 + 5%, and 132 2 30%, respectively (ASD; n 2 3). *, P < 0.001; **, P < 0.05 (compared to the identical dose of T, in the absence of RA).

9-cis-RA, with half-maximal inhibition at approximately 30 nM (Fig. 3). Effects of vitamin

D and clofibrate

on SHBG

Retinoid Concentration,

10

FIG. 3. Influence of RA and 9-cis-RA on T,-stimulated SHBG secretion in HepG2 cells. Cells were incubated for 4 days in serumfree medium together with retinoids (O-300 nM) and T, (10 nM. Increasing concentrations of both RA (0) and 9-cis-RA (0) were able to overcome the T, stimulation of SHBG secretion. The threshold for the effect occurred at 0.3-l nM RA or l-3 nM 9-cis-RA. Half-maximal inhibition was at -3 nM RA or -30 nM 9-cis-RA. The 200% point shown on the y-axis represents SHBG measured in the presence of T, alone. For 9-cis-RA and RA concentrations of O-10 nM, each point represents the mean (t-SD) of three to five experiments, each performed in triplicate. For RA concentrations of 30 nM and above, the points represents the mean of triplicate observations in a single experiment.

secretion

To investigate the specificity of the retinoid effect, we performed similar experiments with compounds that interact with the vitamin D receptor and peroxisome-proliferator activated receptor (PPAR). The ligand for PPAR is unknown, but is thought to involve the drug clofibrate (21). Clofibrate is either metabolized to the active ligand or promotes the production of a compound that interacts with PPARs. Neither vitamin D nor clofibrate altered T,-induced SHBG secreted by the cells after 4-day incubation in serum-free DMEM (Fig. 5). Influence

of

retinoids

on TBG secretion

nM

in HepG2

cells

The data presented above suggest that RA and 9-cis-RA have marked effects on T, stimulation of SHBG secretion by HepG2 cells. BecauseT, has multiple effects on gene expression in hepatic cells, we also studied the influence of RA on T, down-regulated secretion of TBG. Exposure of cells to either RA or 9-cis-RA (O-100 nM) caused a small decrease in TBG secreted into the cellular medium (Fig. 1). T, (100 nM) decreased TBG secretion to 74 + 5% (*SD; n = 6; Fig. 6). In cells incubated with RA (O-10 nM) together with T, (1, 10, or 100 nM), no apparent amelioration of the T,-induced reduction of TBG secretion was observed (Fig. 6). These data suggest that the combined effects of RA and T, are additive at best.

225

X m

125

100

75

o-s-

L.Ol

I

.l

+ 100 nM 9cisR4

I

I

I

1

10

100

T3 Concentration,

nM

FIG. 4. Influence of 9-&s-RA on Ts stimulation of SHBG secretion in HepG2 cells. Cells were incubated for 4 days in serum-free medium with T, alone or together with 9-&s-RA (100 nM). Maximal stimulation in the presence of T, alone was 206 + 26% of the untreated control value at 10 nM T,. The maximal response was attenuated in the presence of 100 nM 9-c&RA to 169 IT 20. Each point represents the mean 2 SD of five or six experiments, each performed in triplicate. *, P < 0.05. ** P < 0.005 (compared to the identical dose of T, in the absence of Rk).

MODULATION


BY RETINOIDS

Endo. 1996 Vol 137 . No 8

600

01

i

I

T3 Concentration,

10

I

nM

/

Ligand

IO

100

Concentration,

nM

FIG. 5. Effect of vitamin D or clofibrate on To-stimulated SHBG secretion in HepG2 cells. Neither vitamin D (100 nM; n ) nor clofibrate (100 nM; 0) altered T, responsiveness over a range of concentrations (leftpanel). Both agents were without effect with a fixed concentration of T, (10 nM; rig&panel). Vitamin D alone (0) or clofibrate alone (0; right nanel) was without effect. Each point represents the mean 2 SD orthree observations, each performed in triplicate.

100

.Ol

+

T, alone

+2d

+ IOnMRA +FtA IOOnM

.I

I

T3 Concentration,

IO

nM

FIG. 6. Effects of RA and 9-c&=RA on To-induced TBG secretion in HepG2 cells. Cells were incubated for 4 days in serum-free medium containing T, (O-10 nM), either alone or together with BA (10 or 100 nM). Maximal inhibition of TBG secretion (%SD; n = 61, expressed as a percentage of the untreated control value for T, (10 nM), was 73.5 2 5.1%. For T, (10 nM) and RA (10 nM) together (n = 31, inhibition was 74.3 + 9.2% (P = NS). For T, (10 nM) and R4 (100 nM) together (n = 2), inhibition was 68.4%.

No detectable than es in secreted total protein, secreted albumin, or nuclear [F?]Ts binding were observed in the presence of RA compared with levels in untreated controls (Fig. 7). Discussion The biological actions of retinoids are mediated by RARs and probably RXRs, whereas the effects of T, are mediated by TRs. Although the interactions between TRs and RARs and the determinants of their specific binding to responsive elementsin target genesare still poorly understood, the data presented here suggestthat the interrelationship between the

1

IO

100

Retinoid concentration, nM 7. Concentrations of total protein (m) or albumin (0) in supernatants of HepG2 cells treated with FU (O-100 nM) for 4 days and (0) in cell nuclei after RA treatment, displaceable 1iz51]T, binding expressed as a percentage of the control value. Points represent the mean 2 SD of three experiments performed in triplicate (W and 01 or the mean t range of two experiments performed in duplicate (01. No change in cell number was observed in the nuclear binding experiments (not shown). Control values were: total protein, 413 + 9 pg/ml; and albumin, 2.2 k 0.3 pg/ml. The mean level of specific nuclear bound [iz51]T3 was 2260 cpm for untreated cells. FIG.

effects of T, and RA on gene transcription are gene dependent. RA alone had a small but significant effect on SHBG secretion. In the presence of T,, SHBG secretion was inhibited by retinoids in a dose-dependent fashion. Both RA and 9-cisRA were able to inhibit the T, effect, with RA showing a potency greater than that of 9-cis-RA by approximately lofold in terms of both threshold concentration of effect and half-maximal inhibitory dose. T,-induced SHBG secretion was virtually abolished by 100 nM RA. It is notable that the maximal responsivenessto T, was reduced without a change in the half-maximal effect. As the RARs are the predominant receptor for RA binding, this suggeststhat at least one of the RARs plays an important role in the modulation of TR-induced responses.Moreover, although 9-cis-RA is the natural ligand for binding to RXRs, it also binds to RARs with a higher affinity than it does to RXRs (22). This implies that both naturally occurring retinoid isomers function to regulate SHBG expression via a RARmediated pathway. An important consideration in relation to our observations is the cell-dependent interconversion between RA and 9-cisRA. Although some mammalian cells are able to convert 9-cis-RA to RA, studies specifically in HepG2 cells were unable to demonstrate the reverse transition of RA to 9-cis-RA (22). This suggeststhat in the cell system used in this study, an enzymatic pathway that converts 9-cis-RA to RA may exist, whereas the reverse reaction doesnot. Accordingly, the effect of 9-cis-RA on the modulation of T, action may be due to the conversion of 9-cis-RA to RA and subsequent action through a RAR-dependent pathway. The situation in vivo may be somewhat different. Urbach and Rando (2) demonstrated isomerization of RA to 9-cis-RA in bovine liver homogenates. Given that isomerization does not occur in

MODULATION


HepG2 cells, it is likely that this conversion pathway is specifically inactivated in z&o. In preliminary studies, retinoids also reduced Ts-stimulated SHBG mRNA production without changing the mRNA stability (data not shown). This suggests that the response elicited by T, and RA combined results from changes in gene transcription. RA did not alter the number of nuclear T,binding sites or the level of secreted total protein or secreted albumin. The most plausible hypothesis to explain these data is that exposure of these cells to RA results in the regulation of a cofactor involved in T, induction of the SHBG gene. Because maximal, rather than half-maximal, secretion was reduced, our results suggest that RA influences the number of sites of the putative factor rather than its affinity for the transcriptional machinery. It is not yet clear which transcriptional cofactors are modified by RA, although the nuclear binding studies suggest that it is not TRs. RA modulates the RAR concentration (13-15), but RXR is regarded as a ubiquitous (3, 5) and perhaps essential coregulator of T, action, which, if modulated by RA, would result in the type of responses observed here. Alternatively, RA may modulate levels of a nuclear corepressor protein. Candidate proteins have recently been described (23, 24) that could fulfill this role, but factors controlling intranuclear concentrations of repressors have not yet been studied. Our results do not allow us to distinguish between these possibilities of RA modulation by activation or repression. The expected failure to demonstrate any change in T, stimulation of SHBG secretion by either vitamin D or clofibrate suggests that at a physiological level, neither of these ligands has been implicated in altering serum SHBG. Although Schrader et al. (25) recently demonstrated that vitamin D receptors can heterodimerize with TR and that these complexes have an independent discriminatory role that is dictated by the target DNA-binding site, no evidence for this effect on SHBG gene trans-activation via a change in secreted SHBG was observed. Although PPARs can heterodimerize with TRs and retinoid receptors (26), there was also no suggestion from this study for a role for PPAR in modulation of thyroid hormone responses. No effect of RA on the T, down-regulation of TBG was observed. T, reduced TBG secretion by approximately 30% of the control level, whereas RA and 9-cis-RA were both able to inhibit the secretion of TBG by approximately 20%. T, also reduces TBG mRNA in HepG2 cells (20). The combined effect of T, and RA on TBG secretion was no more than additive. Accordingly, it appears that T, modulation of TBG synthesis is influenced to only a minor degree by retinoids. If SHBG gene transcription depends on TR heterodimerization, and TBG transcription is unaffected by RA, this observation would imply that the mechanism of heterodimerization on the TBG gene is different from that on the SHBG gene. Distinct pathways for T, induction have recently been observed in rat pituitary cells (27), and factors other than retinoids may play a more important role in mediating T,-induced gene down-regulation. These include the steroid hormone receptors (28), the AP-1 complex, and HNF-1, whose sites of action have previously been identified in the TBG gene (29).

BY RETINOIDS

3191

Finally, it is important to note that our data are entirely consistent with the idea of the involvement of rapidly turned over proteins in T, responsiveness (30). The mechanism by which such proteins influence the effects of T, on transcription of one gene, but not another within the same cell, is currently the subject of investigation in our laboratory. Acknowledgment The 9-cis-RA used in this study Hoffman LaRoche (Nutley, NJ).

was a generous

gift from

Dr. A. Levin,

References 1 Ross AC, Gardner EM 1994 The function of vitamin A in cellular growth and differentiation, and its roles during pregnancy and lactation. Adv Exp Med Biol 352:187-200 Urbach J, Rando R 1994 Isomerisation of all-trams-retinoic acid to 9.cls-retinoic acid. Biochem J 299:459-465 Giguere V 1994 Retinoic acid receptors and cellular retinoid binding proteins: complex interplay in retinoid signahng. Endocr Rev 15:61-79 Smith M, Parkinson D, Cheson B, Friedman M 1992 Retinoids in cancer therapy. J Clin 0~01 10:839-864 Mangelsdorf D, Borgmeyer U, Heyman R, Zhan J, Ong E, Kakizuka A, Evans R 1992 Characterisation of three RXR genes that mediate the action of 9-cis RA retinoic acid. Genes Dev 6:329-344 6 Spanjaard RA, Darling DS, Chin WW 1991 L&and-binding and heterodimerization activities of a conserved region in the ligand-binding domain of the thyroid hormone receptor. l’roc Nat1 Acad Sci 88:8587-8590 Bedo G, Santisteban P, Aranda A 1989 Retinoic acid regulates growth hormone gene expression. Nature 339:231-234 Davis K, Lazar M 1992 Selective antagonism of thyroid hormone action by retinoic acid. J Biol Chem 2673185-3189 Harding P, Duester G 1992 Retinoic acid activation and thyroid hormone repression of the human alcohol dehydrogenase gene ADH3. J Biol Chem 267:14145-14150 10. Lucas PC, Forman BM, Samuels HH, Granner DK 1991 Specificity of a retinoic acid response element in the phosphoenolpyruvate carboxykinase gene promoter: consequences of both retinoic acid and thyroid hormone receptor binding. Mol Cell Biol 11:5164-5170 11. Elfahime E, Felix J, Kock B 1994 Antagonistic effects of retinoic acid and triiodothyronine in the expression of corticoid-binding globulin (CBG) by cultured fetal hepatocytes. J Steroid Biochem Mol Biol 48:467-474 12. Stephanou A, Handwerger S 1995 Retinoic acid and thyroid hormone regulate placental lactogen expression in human trophoblast cells. Endocrinology 136:933-938 13. Swisshelm K, Ryan K, Lee X, Tsou H, Peacocke M, Sager R 1994 Downregulation of retinoic acid receptor p in mammary carcinoma cell lines and its up-regulation in senescing normal mammary epithelial cells. Cell Growth Differ 5:133-141 14. Tsou H, Lee X, Si S-P, Peacocke M 1994 Regulation of retinoic acid receptor expression in dermal fibroblasts. Exp Cell Res 211:74-81 15. Davis K, Lazar M 1993 Induction of retinoic acid receptor-p by retinoic acid is cell specific. Endocrinology 132:1469-1474 16. Tini M, Tsui L-C, Giguere V 1994 Heterodimeric interaction of the retinoic acid and thyroid hormone receptors m transcriptional regulation on the yF-crystallin everted retinoic acid response element. Mel Endocrinol 8:1494-1506 17. Rhee M, Ikeda M, Chin WW 1995 Retinoid X receptor-a and apolipoprotein AI regulatory protein 1 differentially modulated 3,5,3’-triiodothyronone-induced transcription. Endocrinology 136:2697-2704 L, Blok R, Hamblin S, Barlow J 1992 Effects of thyroid hormone on 18. Raggatt sex hormone-binding globulin gene expression in human cells. J Chn Endocrinol Metab 75:116-120 19. Barlow JW, Crowe TC, Cowan NL, Raggatt LE, Topliss DJ, Stockigt JR 1994 Stimulation of sex hormone-binding globulin mRNA and attenuation of corticosteroid-binding globulin mRNA by triiodothyronine in human hepatoma cells. Eur J Endocrinol 130:166-170 20. Crowe TC, Cowen NL, Loidl NM, Topliss DJ, Stockigt JR, Barlow JW 1995 Down-regulation of thyroxine-binding globulin messenger ribonucleic acid by 3,5,3’-tniodothyronine in human hepatoblasoma cells. J Clin Endocrinol Metab 80:2233-2237 21. Bocos C, Gottlicher M, Gearing K, Banner C, Enmark E, Teboul M, Crickmore A, Gustafsson JA 1995 Fatty acid activation of peroxisome prolifer&or-activated receptor (WAR). J Steroid Biochem Mol Biol 53:467-473 22. Allegretto E, McClurg M, Lazarchik S, Clemm D, Kerner S, Elgort M, Boehm M, White S, Pike J, Heymann R 1993 Transactivation properties of retmoic acid and retinoid X receptors in mammalian cells and yeast. J Biol Chem 268~26625-26633

3192

MODULATION


23. HGrlein AJ, NBlr AM, Heinzel T, Torchia J, Gloss B, Kurokawa R, Ryan A, Kamei Y, SSderstrSm M, Glass CK, Rosenfeld MG 1995 Ligand-independent repression by the thyroid hormone receptor mediated by a nuclear receptor co-repressor. Nature 3771397-404 24. Chen JD, Evans RM 1995 A transcriptional co-repressor that interacts with nuclear hormone receptors. Nature 377~454-457 25. Schr;ider M, Miiller K, Nayeri S, Kahlen J-P, Carlberg C 1994~ Vitamin D,-thyroid hormone receptor heterodimer polarity directs l&and sensitivity of transactivation. Nature 370:382-386 26. Juge-Aubry CE, Gorla-Bajszczak A, Pemin A, Lemberger T, Wahli W, Burger AG, Meier CA 1995 l’eroxisome proliferator-activated receptor mediates cross-talk with thyroid hormone receptor by competition for retinoid X receptor. J Biol Chem 270:18117-18122

BY RETINOIDS

Endo. Vol 137.

1996

No 8

27. Maia AL, Harney JW, Larsen PR 1995 Pituitary cells respond to thyroid hormone by discrete, gene-specific pathways. Endocrinology 136:1488-1494 28. Yen PM, Wilcox EC, Chin WW 1995 Steroid hormone receptors selectively affect transcriptional activation but not basal repression by thyroid hormone receptors. Endocrinology 136:440-445 29. Hayashi Y, Mori Y, Hanssen 0, Sunthomthepvarakul T, Weiss R, Takeda K, Weinburg M, Seo H, Bell G, Refetoff S 1993 Human thyroxine-binding globulin gene: complete sequence and transcriptional regulation. Mel Endocrinol 7:1049-1060 30. Hamblin P, Santos A, Wong N, Schwartz H, Oppenheimer J 1987 Triiodothyronine regulation of multiple rat hepatic genes: requirement for ongoing protein synthesis. Mel Endocrinol 1:397-402