Positive and negative regulation of retinoid X receptor gene expression by thyroid hormone in the rat. Transcriptional and post-transcriptional controls by thyroid hormone.

The 9-cis-retinoic acid receptors (RXRs), belonging to the members of the steroid/thyroid hormone receptor superfamily, act as auxiliary proteins, heterodimerizing with other nuclear receptors such as retinoic acid receptors (RARs), vitamin D receptor, thyroid hormone receptors, and peroxisome-proliferator activated receptor, thereby transactivating target genes in a ligand-dependent manner. We have previously reported that in the rat, thyroid hormone (TH) positively and negatively regulates the hepatic mRNA levels of RXR beta and RXR gamma, respectively. In the present study, we have tried to elucidate the level at which TH regulates the gene expression of RXR beta and RXR gamma in the rat. A RNA synthesis inhibitor (actinomycin D), but not a protein synthesis inhibitor (cycloheximide), blocked the induction of RXR beta mRNA by TH. On the other hand, none of these drugs inhibited the decrease of RXR gamma mRNA levels caused by TH. Nuclear run-on assays showed that the transcription rate of the RXR beta gene was positively regulated by TH, whereas the transcription of RXR gamma gene was not controlled by TH. Taken together, these results indicate that the gene expression of RXR beta is positively regulated by TH at transcriptional level, while the negative regulation of the RXR gamma gene expression by TH may occur at a post-transcriptional level in intact rat. Thus, the RXR-mediated signal transductions may be modulated in part through TH control of the levels of RXR beta and RXR gamma.

lated in part through TH control of the levels of RXRP and R X R y .
The nuclear retinoid receptors belong to a superfamily of nuclear receptors, which are ligand-dependent transcription factors (14). The nuclear retinoid receptors consist of two classes. One class consists of the three retinoic acid receptors Foundation of Tokyo University of Agriculture and by grants-in-aid for * This work was supported in part by a grant from the Hayashida scientific research (to S. M. and S. K.) and priority areas (to S. K.) from the Ministry of Education, Science and Culture of Japan. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
(RARa, -p, and -y),l which bind both all-trans-retinoic acid and 9-cis-retinoic acid with high affinity (5,6). The other class consists of the three 9-cis-retinoic acid receptors ( m a , +, a n d -y ) (6)(7)(8)(9)(10). The RAR a n d RXR homodimers and RAR-RXR heterodimers recognize their cognate response elements, i.e. retinoic acid and retinoid X response elements, in the promoters of the target genes, resulting in ligand-dependent transcriptional activation (11). Moreover, it has recently been shown that vitamin D receptor, thyroid hormone receptor (TR), and peroxisome-proliferator activated receptor require RXR, as an auxiliary protein, to bind to their respective target DNA sequences efficiently, thereby activating transcription in a ligand-dependent manner (12,13). Therefore, RXR is supposed to affect retinoid, vitamin D, thyroid hormone, and peroxisome-proliferator signalings.
Since changing levels of nuclear receptor protein can drastically affect the signaling of its cognate ligand, the study of the regulation of nuclear receptor genes is very important for understanding the signal transduction of that ligand. In our previous report, we had shown that TH (but not retinoid or vitamin D) regulates positively and negatively the mRNA levels of two RXRs ( p and y ) out of the three RXRs in the rat liver (14).
In the present study, we examined whether such regulation of RxR genes by TH might be due to the altered rates of transcription or to the control of the half-life of mRNA. Nuclear run-on assay with isolated hepatic nuclei and drug study using protein synthesis and RNA synthesis inhibitors indicated that TH may regulate the transcription of RXRp gene, whereas the negative effect of TH on RXRy gene expression may occur at a post-transcriptional level. These data imply that TH may control its own receptor signaling, as well as those of retinoid, vitamin D, a n d peroxisome-proliferator in part through the regulation of RxRp a n d RxRy gene expression.
MATERIALS AND METHODS Animals-Wistar rats (3 weeks old) were fed on a 6-propyl-2-thiouracil (PTU)-containing diet for 30 days to induce goiters indicating hypothyroidism, and control rats were fed on the AIN-76 diet. Hypothyroidism was judged by the thyroid grand weight (more than 5-fold increase). The average of body weights for hypothyroid rats was not significantly different from that for non-treated rats (15). For replenishment, these rats were intraperitoneally administrated 100 or 500 pg of 3,5,3'-triiodothyronine (T3), and they were decapitated at the indicated times.

1591
All animal experiments using more than three rats per group were repeated at least two times. To obtain quantitative data on RXR transcripts, we used densitometric analysis of autoradiograms. The relative abundance of RXR mRNAs is shown as the mean 2 S.D. for at least three samples from each different rat after normalizing with the p-actin mRNA levels Nuclei Isolation and Nuclear Run-on Assay--Rat liver nuclei were prepared from the rats as previously described (19). Briefly, 0.3 g of liver was minced and homogenized in 10 volumes (v/w) of ice-cold bufferA(15 mM Tris-HC1, pH 7.5, 15 mM KCI, 15 mM 2-mercaptoethanol, 2 mM EDTA, 0.5 mM EGTA, 0.15 m~ spermine, 0.5 mM spermidine, 1.9 M sucrose, 0.1% Triton X-100,O.l mM phenylmethylsulfonyl fluoride) by a glasdkflon homogenizer. The homogenates were filtered through nylon gauze to remove debris and then diluted by the addition of an equal volume of buffer B (buffer A without Triton X-100). The sample was centrifuged for 90 min at 90,000 x g. The supernatant was poured off, and the pellet was resuspended, homogenized in 200 pl of buffer C (15 m~ Tris-HC1, pH 7.5, 15 mM NaCl, 60 mM KCI, 15 mM 2-mercaptoethanol, 0.15 m~ spermine, 0.5 m~ spermidine, 0.34 M sucrose, 0.1 mM phenylmethylsulfonyl fluoride), and centrifuged for 5 min at 3,000 x g. The pellet was resuspended at concentration of 1 mg/ml DNA in buffer C and was used as liver nuclei for run-on assay (20). The 20 pl of liver nuclei was added in 40 p1 of run-on buffer (80 mM Tris-HC1, pH 7. Normal rats were injected with T, (500 pg/rat). Control rats were injected with saline. The control and T,injected rats were killed 3, 6, 10, and 24 h after injection. Northern blotting and quantitative analysis were camed out as described for Fig.  1. Relative signal intensity of the transcript levels of the T,-injected rats is expressed by dividing those of the control rats.
0.1% SDS, then washed at 60 "C for 30 min in 0.2 x SSC, 0.1% SDS and treated at 37 "C for 30 min in 2 x SSC containing 5 pdml RNase A. The filters were rinsed in 2 x SSC and washed at 60 "C for 15 min in 0.2 x SSC, 0.1% SDS. The filters were exposed to x-ray film at -80 "C for 5 or 10 days with intensifymg screens. The relative rate of RXR transcription was measured by densitometric analysis of autoradiograms, and shown as the mean f S.D. for at least three samples from each different rat after normalizing with the amount of p-actin transcript.

RESULTS
Thyroid Hormone Regulates the Expression of R X R p and R X R y Genes-In our previous study, we have shown that TH affects the transcript levels of RXRp and RXRy in the adult rat liver (14). To confirm these observations, we first examined the effect of TH on the expression of RXR genes in other tissues. As shown in Fig. 1 ( A and B ) , the levels of RXRp mRNAs clearly decreased in heart and brain of rats fed a PTU-containing diet, and the intraperitoneal administration of T3 (100 pg/rat) to the PTU-treated rats rapidly (within 6 h) restored the levels of RXRp mRNAs, as observed in liver (14). Moreover such positive regulation was confirmed by the induction of the RXRp gene with excess doses of T3 (500 pghat) to normal rats ( Fig. 1, C and Dl. In contrast, the transcript levels of RXRy increased in the tissues examined of rats fed on a PTU-containing diet, and remarkably, decreased upon replenishment of T3 in the PTUtreated rats and the normal rats given T3 (Fig. 1, A-D). These actions of TH on RXR gene expression were observed also in all tissues examined (data not shown). These results support our previous hypothesis that the gene expression of RXRp and RXRy are positively and negatively regulated by TH. Throughout these experiments, no alteration of the transcript levels of RXRa by TH was observed in any tissues examined. We next examined by a time-course study to speculate whether the regulation of RXRp and RXRy genes by TH occurs in a direct or indirect manner. Normal rats were administrated 500 pg of T3 and killed at indicated times. Fig. 2 shows that the brain mRNA levels of RXRp were clearly increased (2-3-fold) 3 h after TH administration, but this induction was reduced quickly. On the other hand, the transcript levels of RXRy were gradually decreased 6-24 h after TH administration. Similar patterns were also observed in the other tissues of the PTUtreated rats given T3 (100 pg/rat) (data not shown). These results suggest that TH may directly modulate the expression of RXRp at level of transcription, while TH may repress the tran- Normal rats were injected with Act D, CHX, or control saline ( C ) . After 1 h, T3 (500 pg/rat) or saline was administrated to the rats, and 4 h later the rats were subjected to analyses. Northern blotting and quantitative analysis were camed out as described for Fig. 1.
scription of the RXRy gene and/or destabilize its mRNA. Effect of Inhibitors on Thyroid Hormone-modulated RXR Gene Expression-% elucidate the level at which TH modulates the expression of RXRp and RXRy genes, the inhibitors of transcription (actinomycin D) and translation (cycloheximide) were used in rats (Fig. 3). Either actinomycin D (Act D; 0.15 mghat) or cycloheximide (CHX, 2 mg/rat) was injected, and after 1 h T3 (500 pg/rat) was administrated. After 4 h of T3 administration, the mRNA levels of RXRs were estimated. Act D abolished the TH-induced increase in the RXRp mRNA levels, whereas the protein synthesis inhibitor CHX did not, indicating that the induction of the RXRp gene by TH may result from an enhanced rate of transcription. In contrast, Act D and CHX did not affect the decrease in RXRy mRNA levels by TH, implying that TH may control RXRymRNA stability (21)(22)(23).

Thyroid Hormone Regulates Danscription of the R X R p
Gene-The drug study depicted in Fig. 3 suggests that TH may regulate the gene expression of at least RXRp at transcriptional level. To test this hypothesis, we directly determined the effect of TH on the transcription rate of the RXR genes using nuclear run-on assay with the isolated liver nuclei from the PTU-treated rats and the rats replenished with TH for 4 h. As shown in Fig. 4 ( A and C ) , the transcription rate of RXRp was clearly decreased in the PTU-treated rats but was quickly restored by T3 replenishment. Moreover, administration of TH to normal rats enhanced 2-fold the transcription rate of RXRp gene (Fig. 4, B and D), strongly suggesting that TH controls the transcription of the RXRp gene. However, the transcription rates of RXRa, RXRy, and p-actin genes were not affected by TH. These data suggest that the negative regulation of RXRy gene expression by TH might be due to destabilization of its mRNA.

DISCUSSION
We had previously shown that, in the liver of rats treated with P T U , administration of TH enhanced the mRNA levels of RXRp but reduced those of RXRy without affecting RXRa gene expression (14). In the present study, those positive and negative effects of TH on the RXR gene expression could be also confirmed in various tissues (Fig. 1). Furthermore, we investigated a possible mechanism of gene regulation of RXRp and RXRy by TH. Nuclear run-on assay using the isolated liver nuclei revealed that TH directly controls the transcription rate @ FIG. 4. Nuclear run-on analysis of T,-regulated FtXR gene expression. A and C , goiters were induced in the rats by feeding on a PTU-containing diet (PTU), and 3,5,3'-triiodothyronine (2'3) (100 pg) was intraperitoneally administrated to the PTU-treated rats for replenishment. After 4 h of administration, the treated rats (PTU and 7'3) and control rats ( C ) were killed and subjected to analyses. B and D, an excess amount (500 pg) of T, was given to normal rats (2'3). After 4 h of T3 administration, the treated rats (2'3) and control rats ( C ) were killed and subjected to analyses. Isolation of liver nuclei and nuclear run-on assay were camed out as described under "Materials and Methods." The pSG5 (parent expression vector) was used for negative control. Representative results on the nuclear run-on assay are shown in panels A and B. The transcriptional rates of RXR genes are expressed as the mean * S.D. of more than D where the induction of RXRP gene by TH was inhibited by actinomycin D, but not cycloheximide. In contrast, it is most likely that the negative regulation of RXRy gene by TH occurs at a post-transcriptional level such as control of the half-life of mRNA, because transcription of the R X R y gene was not affected by TH in the nuclear run-on analysis and drug study. As it is well described that TH regulates the expression of target genes a t transcriptional and post-transcriptional levels (21-231, the molecular mechanism of the hormonal regulation of RXRP and R X R y genes by TH may be classified into the well characterized mechanisms already described in the other target genes. RxRs play critical roles in the signal transductions for retinoid, vitamin D, thyroid hormone, and peroxisome-proliferator, since they serve as auxiliary proteins heterodimerizing with the nuclear receptors of those ligands (12,13). In addition to various combinations of RXRs with those receptors upon sequence-specific DNA binding, the binding of 94s-retinoic acid and the ligand on the receptor dimers is involved in the positive and negative controls of receptor-mediated transactivation. Recent studies have demonstrated that the ligand (94s-retinoic acid) for RXR positively and negatively modulates ligand-induced transactivation mediated with the receptors for TH and vitamin D in a response element-specific manner (24)(25)(26)(27)(28). Although such ligand-induced regulations on receptor-mediated transactivation have been studied extensively in the transient expression assay, the effects of ligands on the gene expression of nuclear receptors themselves have not been fully investigated. The present study clearly shows that the gene expression of RXRP and R X R y is regulated by TH through distinct levels, indicating that TR.RXR-mediated signal transductions may be controlled in a thyroid hormone response element-specific manner not only by TH-induced transactivation through TR.RXR heterodimer but also by the altered levels of RXRs by TH. Moreover, these results imply a novel action of TH that the RXR-mediated signal transductions for retinoid, vitamin D, and peroxisome-proliferator may be modulated through the altered levels of RXRP and R X R y by TH.