Retinoid X receptor acts as a hormone receptor in vivo to induce a key metabolic enzyme for 1,25-dihydroxyvitamin D3.

We demonstrate here that RNA levels of 25-hydroxyvitamin D3-24-hydroxylase (24-(OH)ase), a key catabolic enzyme for 1,25-dihydroxyvitamin D3, are increased by a highly selective retinoid X receptor (RXR) ligand, LG100268, in mice within hours. Correspondingly, upon LG100268 treatment, kidney 24-(OH)ase enzymatic activity increases 5-10-fold. The endogenous retinoid hormones, all-trans-retinoic acid and 9-cis-retinoic acid, and the synthetic retinoic acid receptor-selective compound, TTNPB, also stimulate 24-(OH)ase. Additionally, we show that LG100268 stimulates transcription of a luciferase reporter plasmid driven by 24-(OH)ase promoter sequences in the presence of RXR in CV-1 cell cotransactivation assays. This first demonstration of a gene that is regulated in the intact animal through an RXR-mediated pathway confirms earlier hypotheses that RXR is a bona fide hormone receptor. Regulation of a key gene in the vitamin D signaling pathway by a retinoid transducer may provide a molecular basis for some of the documented biological effects of vitamin A on bone and vitamin D metabolism.


We demonstrate here that RNA levels of 25-hydroxyvitamin D 3 -24-hydroxylase (24-(OH)ase), a key catabolic enzyme for 1,25-dihydroxyvitamin D 3 , are increased by a highly selective retinoid X receptor (RXR) ligand,
LG100268, in mice within hours. Correspondingly, upon LG100268 treatment, kidney 24-(OH)ase enzymatic activity increases 5-10-fold. The endogenous retinoid hormones, all-trans-retinoic acid and 9-cis-retinoic acid, and the synthetic retinoic acid receptor-selective compound, TTNPB, also stimulate 24-(OH)ase. Additionally, we show that LG100268 stimulates transcription of a luciferase reporter plasmid driven by 24-(OH)ase promoter sequences in the presence of RXR in CV-1 cell cotransactivation assays. This first demonstration of a gene that is regulated in the intact animal through an RXR-mediated pathway confirms earlier hypotheses that RXR is a bona fide hormone receptor. Regulation of a key gene in the vitamin D signaling pathway by a retinoid transducer may provide a molecular basis for some of the documented biological effects of vitamin A on bone and vitamin D metabolism.
Retinoic acid receptors (RARs) 1 have been shown to act as hormone receptors by virtue of the fact that they bind all-transretinoic acid (tRA) with high affinity and activate target genes in the presence of tRA in cell-based cotransactivation assays and in vivo (reviewed in Ref. 1). Retinoid X receptors (RXRs) are also thought to act as hormone receptors (reviewed in Ref. 1) since they bind 9-cis-retinoic acid (9cRA) with high affinity (2,3) and have been shown to stimulate transcription from distinct elements known as RXR-responsive elements in promoter sequences of the apolipoprotein A1 (apoA1) and cellular retinol-binding protein II (CRBPII) genes in cell-based cotransactivation assays in response to 9cRA (2,3), tRA (4,5), or RXR-selective ligands (6,7). However, these genes have not been shown to be regulated by retinoids in vivo. While another gene product (growth hormone) has been shown to increase in cultured cells treated with RXR-and RAR-selective ligands (8), there are currently no known biological target genes of RXR in the intact animal.
RXRs are thought to form homodimers upon interaction with ligand (1) to activate putative RXR-selective genes, such as CRBPII and apoA1, through RXR-responsive elements in the promoters of those genes (2)(3)(4)(5)(6)(7). RXR is also known to participate in heterodimers with a number of other intracellular receptors, including the vitamin D receptor (VDR), RAR, or thyroid hormone receptor to activate a variety of target genes, which are stimulated by 1,25-dihydroxyvitamin D 3 (1,25-(OH) 2 D 3 ), tRA, or thyroid hormone, respectively (1). RXR is thought to act as a silent partner in these interactions, although the effects of liganding of RXR under these circumstances have not been thoroughly explored. Additionally, recent studies describe instances of stimulation of gene transcription by hormone-occupied RXR interacting with a supposedly unoccupied orphan receptor (9,10).
To study the consequences of liganding of RXR in vivo, we began by testing the effects that an RXR-selective compound might have on the activity of a vitamin D-inducible gene. 25-Hydroxyvitamin D 3 -24-hydroxylase (24-(OH)ase) is one such gene product that is induced by 1,25-(OH) 2 D 3 (11), acting through VDR-RXR heterodimers that bind to a vitamin D response element within the promoter of the gene (12,13). The effects of in vivo administration of various retinoids on the levels of the 24-(OH)ase gene product were examined. We found that RXR-selective ligands are able to stimulate 24-(OH)ase RNA levels and enzymatic activity within hours in normal or vitamin D-deficient mice. These results describe a novel RXR-target gene, which is also the first demonstration of such a gene in the intact animal.
Dosing of Mice-BALB/c female mice (Harlan-Teklad, 8 -12 weeks old, three to four per dosing group) were placed on a control diet (low phosphate, low calcium; Harlan Teklad 94326) or a diet that renders them 1,25-(OH) 2 D 3 -deficient (minus vitamin D, low phosphate, low calcium, 0.8% strontium (15); Harlan Teklad 94325) for 14 days prior to dosing with compounds ( Fig. 2) in sesame oil vehicle by oral gavage. Dosing for the 8-day experiment (Fig. 2a) was once daily in the a.m., and sacrifice was ϳ3-4 h after the dose on the 8th day. The 32-h study (Fig. 2b) included drug doses that were at 8 a.m., noon, and 4 p.m. on the 1st day and 8 a.m. and noon on the 2nd day with sacrifice ϳ3 h later. For 8-h studies (Fig. 3, Tables I and II) mice were fed a normal diet (Harlan-Teklad 7012), and dosing was performed twice 4 h apart; sacrifice was 3-4 h after the last dose.
Northern Analysis-RNA was extracted (16), and Northern analyses were performed as by standard techniques. Probes were as follows: 900-bp EcoRI/XbaI DNA fragment of the rat 24-(OH)ase cDNA in pUC-r24-(OH)ase (11); 1.4-kb human glyceraldehyde-3-phosphate dehydrogenase (GAPDH) fragment (Clonetech); RAR␤ probe, a 1.4-kb NcoI/SacI fragment from YEpTRPhRAR␤, a yeast expression vector constructed using hRAR␤ cDNA (17); apoA1 probe, a polymerase chain reaction-* 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.
Circulating Levels of Compounds-Plasma concentrations of LG100268 and 9cRA following dosing were determined using reversed phase high pressure liquid chromatography (HPLC). Equal volumes of plasma from three mice were combined to produce each 0.5-ml sample. 9cRA determination was as described previously (20).
LG100268 determination was as follows. Samples were precipitated with 5 volumes of methanol and then subjected to solid phase extraction. Solvent was evaporated, and the residue was solubilized in mobile phase (acetonitrile, 10 mM ammonium acetate, acetic acid (75:25:1)). Analyses were conducted using a Rainin Microsorb-MV C-18 column (250 ϫ 4.6 mm) and an Alltech direct connect guard column maintained at 40°C. Detection was at 281 nm, and the flow rate was 1.2 ml/min. The response of the instrument was linear between 0.25 and 25 M.
Enzyme Activity Studies-Mice were dosed twice daily (4 h between doses) with 3 mg/kg LG100268 or vehicle for 8 h or 3.5 days (eight mice per group). Sacrifice was 3 h after the last dose. Kidneys were perfused with cold 0.25 M sucrose in phosphate-buffered saline, pH 7.4, pooled from four mice, homogenized, and precipitates were prepared as described previously (21); incubations (0.5 ml; 12 mg/ml protein) with 50 M 25-(OH)D 3 were for 1 h at 37°C (21). Purification and quantitation of generated 24,25-(OH) 2 D 3 were by solid phase extraction and HPLC (21).
Since the repression effect of LG100268 on 1,25-(OH) 2 D 3 induction of 24-(OH)ase occurred only after longer time periods, a direct negative effect of the RXR ligand on the ability of the VDR-RXR heterodimer to stimulate 24-(OH)ase seemed unlikely. An alternate possibility might be that liganded RXR does not heterodimerize with VDR; complete depletion of the pool of unliganded RXR available to VDR may take longer than 32 h. However, since LG100268 induces 24-(OH)ase RNA in the absence of 1,25-(OH) 2 D 3 at both short and long time periods, it may also act via a VDR-independent mechanism to stimulate 24-(OH)ase. Since 24-hydroxylation is the initial enzymatic step in the catabolism of 1,25-(OH) 2 D 3 , increased 24-(OH)ase activity over time might be expected to result in lower cellular concentrations of 1,25-(OH) 2 D 3 and, therefore, less VDR activity on its target genes, including 24-(OH)ase.
To test this hypothesis we set out to determine whether the induction of 24-(OH)ase by LG100268 also occurred in the non-vitamin D-deficient state. Mice were fed a normal diet and treated with LG100268 and various other retinoids. TTNPB, a known RAR-selective ligand of high potency (23,24), was tested to determine if the retinoid effect on 24-(OH)ase was limited to RXR-selective ligands. Mice were also treated with 9cRA and tRA to determine if the increased levels of 24-(OH)ase RNA observed upon treatment with the synthetic RXR-selective compound could also be elicited by the naturally occurring endogenous retinoid hormones. Three-to five-fold increases in kidney 24-(OH)ase RNA levels were detected from mice treated for a period of 8 h with two doses (spaced 4 h apart) of LG100268, 9cRA, or tRA in comparison with vehicletreated mice (Fig. 3a, lanes 1-25). While no increase was detected with the potent RAR-selective ligand TTNPB at 1 g/kg (data not shown), TTNPB at 10 g/kg did elicit a 3-fold increase in 24-(OH)ase RNA levels (Fig. 3a, lanes 26 -28). Potency comparisons between the various retinoids based upon their elicited efficacies (-fold inductions of RNA levels) are not possible without taking into consideration the complicated differences in bioavailability, metabolism, and excretion of the compounds in vivo.
While LG100268 is immaculately specific for RXR as assessed by both ligand binding assays (K i versus [ 3 H]9cRA for RXRs ϭ 3 nM; for RARs Ͼ10 M) and in cell-based cotransac- tivation assays (Ref. 14 and see Fig. 4, lanes 1-4) to ensure that LG100268 was acting as an RXR-selective ligand in vivo in our system, Northern analysis was performed on the RAR␤ gene product. RAR␤ transcripts have been shown to be up-regulated by ligands that bind to RAR, but not by RXR-selective compounds, in F9 cells (8). Fig. 3b shows that while 9cRA (lanes 9 -12) and tRA (lanes 13-16) effect 1.5-2-fold increases in the levels of kidney RAR␤ RNA, LG100268 does not (lanes 5-8). Therefore, LG100268 exhibits the properties of an RXR-selective compound in this system. Since RXR-selective (LG100268) and RAR-selective (TTNPB) compounds, as well as the endogenous retinoids 9cRA and tRA, increase 24-(OH)ase RNA levels, both RXR and RAR are able to participate in the induction of 24-(OH)ase.
Certain genes including apoA1 and CRBPII have been postulated to be activated through RXR-driven pathways by virtue of the fact that reporter plasmids driven by sequences from within their promoters have been shown to be stimulated by 9cRA or RXR-selective compounds in the presence of cotransfected RXR through a defined element in cell-based cotransactivation assays (2)(3)(4)(5)(6)(7). However, since this initial work 4 years ago, there have not been any published reports showing that these genes are regulated in vivo. To test whether these genes are regulated in vivo in our system, Northern analyses were performed using RNA from liver (apoA1) and intestine (CRB-PII) of the mice treated with two or five doses of the various retinoids for 8 or 32 h, respectively (Fig. 3c). ApoA1 (Fig. 3c,  lanes 1-16) and CRBPII (Fig. 3c, lanes 17-24) RNA levels did not change with retinoid treatment on the 8-or 32-h schedule in comparison with untreated animals. Therefore, in BALB/c mice under the conditions of these experiments no significant up-regulation of apoA1 and CRBPII was observed with retinoid treatment. Additionally, apoA1 RNA levels from rat liver also were not observed to change after dosing of rats with three different concentrations of RXR-selective compound for durations ranging from 4 h to 5 days. 2 Therefore, to date, 24-(OH)ase is the only gene that has been shown to be regulated by RXR-selective ligands in the intact animal.
Circulating levels of LG100268 and 9cRA were determined by performing pharmacokinetic studies on mice treated with LG100268 or 9cRA. Two doses of 30 mg/kg LG100268, 3 mg/kg LG100268, or 10 mg/kg 9cRA were given 4 h apart, mice were sacrificed, and blood was obtained at various time points. Table  I shows that doses of 30 mg/kg LG100268 resulted in plasma concentrations of ϳ1-13 M LG100268 throughout the duration of the study. Ten-fold lower doses of LG100268 (3 mg/kg) effected the same -fold increase in 24-(OH)ase RNA levels as did higher LG100268 doses of 12 mg/kg (Fig. 3a) and 30 mg/kg (data not shown) and resulted in circulating LG100268 levels below the limits of quantitation of the assay (Ͻ250 nM). Doses of 10 mg/kg of 9cRA resulted in plasma concentrations of ϳ100 -600 nM 9cRA at the indicated time points. Ten-fold lower doses of 9cRA (1 mg/kg) also effected induction of 24-(OH)ase RNA (Fig. 3a); circulating levels were not measured at that dose. Therefore, the determined circulating levels of LG100268 and 9cRA after administration of effective doses for exertion of a biological effect (i.e. increase of 24-(OH)ase RNA) are in the low nanomolar range.
To ascertain if the observed increases in 24-(OH)ase RNA corresponded with an increase in kidney 24-hydroxylase enzyme activity, metabolic conversion assays were performed. Table II shows that enzymatic conversion of 25-(OH)D 3 to 24,25-(OH) 2 D 3 in kidney extracts from mice treated with LG100268 (3 mg/kg twice daily for 8 h or 3.5 days) was 5-10fold greater than the conversion observed in kidney extracts from normal untreated mice. The retinoids utilized in this 2 P. Syka, unpublished data.
To test whether the retinoids are able to stimulate transcription from 24-(OH)ase promoter sequences and to confirm that RXR is required for that activation, cotransfection experiments were performed in CV-1 cells with a promoterless luciferase vector containing human 24-(OH)ase promoter sequences (13). Saturating concentrations of 9cRA or LG100268 elicited 7-10fold inductions of luciferase activity in CV-1 cells that had been cotransfected with an expression vector producing RXR␥ and a reporter plasmid containing 6 kb of 24-(OH)ase promoter sequence (13) (Fig. 4, lanes 5 and 6) while 1.5-2-fold induction was observed without transfected receptor, representative of endogenous RXR present in CV-1 cells (data not shown). 9cRA was also able to stimulate promoter activity in the presence of RAR␥ (ϳ9-fold (Fig. 4, lane 7)) or RXR␥ and RAR␥ (ϳ14-fold (Fig. 4, lane 8)). 1,25-(OH) 2 D 3 induced transcription (ϳ10-fold) from a 400-bp 24-(OH)ase promoter sequence in the presence of VDR, as previously demonstrated (lane 9) (13). This reporter was also activated by 9cRA in the presence of RXR (lane 10). A reporter vector with the 400-bp 24-(OH)ase promoter sequence containing three mutations in the upstream half-site of the vitamin D responsive element no longer binds to VDR, and the 1,25-(OH) 2 D 3 -induced transactivation through the promoter sequence is abolished (13) (Fig. 4, lane 11). 9cRA was able to stimulate luciferase activity in the presence of RXR via the mutant promoter (Fig. 4, lane 12) as well as it did through the wild type promoter sequence (Fig. 4, lane 10). Therefore, retinoid ligands acting through their respective receptors are able to stimulate transcription from the 24-(OH)ase promoter in a heterologous system utilizing cis element(s) that appear to be distinct from or overlapping with that utilized by VDR. Hence, effectors of vitamin A signal transduction, including the RXRselective ligand, LG100268, acting through the nuclear retinoid receptors, induce a key gene involved in vitamin D signaling, thereby adding a new level of complexity to the cross-talk between the two systems.