Complex Role of the Vitamin D Receptor and Its Ligand in Adipogenesis in 3T3-L1 Cells*

The vitamin D receptor (VDR) and its ligand 1,25-OH2-VD3 (calcitriol) play an essential role in mineral homeostasis in mammals. Interestingly, the VDR is expressed very early in adipogenesis in 3T3-L1 cells, suggesting that the VDR signaling pathway may play a role in adipocyte biology and function. Indeed, it has been known for a number of years that calcitriol is a potent inhibitor of adipogenesis in this model but with no clear mechanism identified. In this study, we have further defined the molecular mechanism by which the unliganded VDR and calcitriol-liganded VDR regulate adipogenesis. In the presence of calcitriol, the VDR blocks adipogenesis by down-regulating both C/EBPβ mRNA expression and C/EBPβ nuclear protein levels at a critical stage of differentiation. In addition, calcitriol allows for the up-regulation of the recently described C/EBPβ corerepressor, ETO, which would further inhibit the action of any remaining C/EBPβ, whose action is required for adipogenesis. In contrast, in the absence of calcitriol, the unliganded VDR appears necessary for lipid accumulation, since knock-down of the VDR using siRNA both delays and prevents this process. Taken together, these data support the notion that the intracellular concentrations of calcitriol can play an important role in either promoting or inhibiting adipogenesis via the VDR and the transcriptional pathways that it targets. Further examination of this hypothesis in vivo may shed new light on the biology of adipogenesis.


Introduction
The regulation of adipogenesis is a key biologic process that is required for both lipid storage and the development of the endocrine adipocyte. Both of these are key processes that determine the morbidity of obesity. Indeed, obesity is the leading risk for the development of Type 2 diabetes as well as an important contributing risk to heart disease and stroke. Thus, further understanding of the biology of adipogenesis might allow for the development of novel targets for new drugs to modify the function of the adipocyte in vivo.
The murine 3T3-L1 cell line has provided an ideal model system to understand adipocyte development. This line differentiates from fibroblast preadipocyte precursors to mature adipocytes when presented with a hormonal cocktail that activates a number of key signaling pathways. These pathways induce a regulatory cascade beginning with the induction of the CCAAT-enhancer binding protein β and δ isoforms (C/EBPβ and δ) which in turn induce C/EBPα and the nuclear receptor peroxisome proliferator-activated receptor γ (PPARγ -present as two isoforms γ1 and γ2) (1). Once expressed, C/EBPα and PPARγ enhance each other's production and are necessary for terminal differentiation (2). In addition, the production of a PPARγ ligand is also necessary to allow PPARγ to activate target genes to allow for differentiation (3).
Using a stably integrated ligand-sensing system we recently demonstrated that endogenous PPARγ ligand activity is induced as early as 24 hours after the induction of differentiation via a cAMP requiring pathway (9).
In order to identify pathways which may lead to PPARγ ligandproduction we used microarrays to identify genes expressed early in adipogenesis that required cAMP production (data not shown). Using this screen, we identified the gene encoding the vitamin D receptor (VDR) as a potential target in this pathway. The VDR, a member of the nuclear receptor superfamily, plays a key role in mineral homeostasis when bound to its ligand, calcitriol (11,12). In addition, the unliganded VDR plays a critical role in hair follicle development indicating that like other nuclear receptors, such as the thyroid and retinoic acid receptor isoforms, the VDR has unique functions in the presence and absence of its ligand (13)(14)(15).
Previous work by a number of groups has established that calcitriol is an inhibitor of adipogenesis in the 3T3-L1 model (16)(17)(18). More recently, work in mice has demonstrated that calcitriol inhibits bone marrow adipogenesis in vivo (19). However, no specific mechanism for the actions of calcitriol in adipocytes has been described. A recent report has suggested that calcitriol-mediated induction of the ER protein Insig-2 in 3T3-L1 adipocytes might lead to the inhibition of adipogenesis by preventing the transcription factor SREBP-1C from reaching the nucleus (20).
While calcitriol inhibits adipogenesis, intriguingly it is the unliganded VDR that is available early in differentiation in 3T3-L1 cells. This suggests that the unliganded VDR may play a unique role in the molecular pathways governing adipogenesis depending upon the availability of intracellular calcitriol. Indeed, the intracellular metabolism of nuclear receptor ligands is a critical determinant of receptor action. A recent striking example is that adipose tissue function can be altered directly by changes in intracellular cortisol levels through the overexpression of 11β-HSD1 which reactivates cortisol in this tissue (21). In a parallel fashion, the intracellular availability of 1 alphahydroxylase, which synthesizes calcitriol from its precursor 25-hydroxy-vitamin D 3 , may be a critical determinant in the action of vitamin D signaling in adipose tissue in vivo (22).
To further characterize the key role of the VDR-signaling pathway in adipogenesis we have, in this study, identified the molecular mechanism by which the liganded VDR inhibits adipogenesis and also defined a novel role for the unliganded VDR in promoting adipogenesis. Further work on defining the molecular targets of the VDR in adipose tissue will yield greater insight into its potential in vivo role.
In experiments where calcitriol was added for short periods of time the cells were washed with PBS and replace with the medium present in the control 3T3-L1 cells.
3T3-L1 preadipocytes with the stably integrated ligand-sensing system (5B-2 cells (9)) were cultured as described above or treated with the indicated concentrations of CALCITRIOL and/or rosiglitazone.
Twenty-four hours after induction of differentiation with DIM and the indicated ligands the cells were lysed and assayed for β-galactosidase activity. All experiments were performed in triplicate.
The extent of differentiation, was determined by amount of lipid accumulation, at 5 and 7 days by oil red O (ORO) staining. Briefly, cells were fixed in 10% formaldehyde in phosphate buffered saline (PBS) for 1h, washed with distilled water, and completely dried. Cell were stained with a 0.5% oil red O solution in 60:40 (v/v) isopropanol:H 2 O for 30min at room temperature and washed four times with water, and dried. Differentiation examined by visual inspection and quantified by elution with isopropanol and an optical density (OD) measurement at 590nm.
Cell proliferation was determined by incorporation of crystal violet. Briefly, at indicated time points, cells were aspirated of media and washed with PBS, fixed with 10% formaldehyde for 10min, washed with water, and completely dried. Cells were then stained with 0.1% crystal violet in 30% ethanol for 30min at room temperature, washed with water, and airdried. Crystal violet was then eluted with 10% acetic acid (v/v), diluted 1:40, and an OD measurement was taken at 590nm.
HRP linked anti-mouse, rabbit, and rat were from Amersham and HRP linked anti-goat was from Santa Cruz Biotechnology. Membranes were stripped with 70 mM Tris pH 6.8, 2% SDS and 0.1% beta-mercaptoethanol for multiple probes of the same membrane.

Quantitative PCR
Total RNA was extracted from 3T3-L1 preadipocytes using RNA STAT-60 (TelTest) according to manufacturer's protocol and concentration determined by OD at 260nm. To demonstrate quality, integrity, and accuracy of OD concentration 1.5 µg of each sample was run on a 1.5% agarose/ethidium bromide gel and visualized with UV. Taqman Assay on Demand primer/probe combinations (Applied Biosystems) were used for analysis of VDR, C/EBPbeta, and Insig2 expression (Mm00437297_m1, Mm00843434_s1, and Mm00460121_m1 respectively). Assay on Demand derived QPCR reactions (25 µl) consisted of 5ng RNA (samples diluted to 1ng/µl), 12.5 µl TaqMan Universal PCR Master Mix (Applied Biosystems), 0.25 µl primer/probe, 0.125 µl MulV reverse transcriptase, and 7.125 µl RNase-free water. To normalize sample values, 18S was used as an internal control. Reaction setup for 18S was the same as above except with 0.375 µl of forward and reverse primers, 0.125 µl probe, and 6.5 µl water was used. Sequence for 18S primer and probe: F -AGTCCCTGCCCTTTGTACACA, R -GATCCGAGGGCCTCACTAAAC, Probe-CGCCCGTCGCTACTACCGATTGG. All samples were run in duplicate. QPCR reactions were run on a MX3000p (Stratagene) with the following conditions: 30min at 48˚C, 10min at 95˚C, and 40 cycles of 95˚C at 15 sec and 1min at 60˚C.
Template quantity was determined from gene specific standard curves (50ng -0.1mg total RNA). Relative mRNA concentrations were determined by dividing the initial template quantity of the target gene by the 18S initial template quantity. Results shown are pooled from two separate experiments. The data is displayed as mean +/-SEM. Significance was determined by the student T test.

RNA Interference
3T3-L1 preadipocytes were transfected with siRNA oligo duplexes one day post confluence (day -1) with Lipofectamine 2000 (Invitrogen). Generally 100 nM (unless otherwise indicated) siRNA was transfected with 4 µl lipofectamine per well of a six well plate with fresh media. Each experiment contained equivalent samples transfected with a non-targeting control siRNA pool (Dharmacon) and samples not treated with lipofectamine. siRNA oligonucleotide duplexes for each gene of interest were purchased from Dharmacon as either a four duplex pool (Insig2) or an optimized single duplex (VDR,Sense:GAAUGUGCCUCGGAUC UGUUU,Antisense:ACAGAUCCGAGG CACAUUCUU)). Transfection efficiency was monitored using fluorescent-tagged oligonucleotides (Blockit, Invitrogen) transfected as described above and visualized with a mercury lamp fluorescent microscope.

The VDR is Induced Early in Adipogenesis Via a cAMP Pathway
To confirm our microarray data suggesting that the VDR is expressed as early as 6 hours after the induction of differentiation with DIM we analyzed VDR expression via Q-PCR in differentiating 3T3-L1 cells. As shown in Figure 1A, VDR mRNA levels reach their maximum 6 hours after the addition of DIM and then decline rapidly. Consistent with its mRNA expression are VDR protein levels in 3T3-L1 nuclear extracts. VDR nuclear protein begins to accumulate at 4 hours and is maximally present at 12 hours. Thereafter it quickly declines and disappears 2 days into differentiation ( Figure 1B). In contrast nuclear protein levels of histone deacetylase 1 (HDAC1) are relatively constant during differentiation. We next looked at nuclear VDR protein in response to each component of the differentiation cocktail. While dexamethasone and insulin had no effect in inducing the VDR, IBMX strongly induced the VDR 6 hours after its addition ( Figure 1C). This data is consistent with the cAMP pathway being responsible for the early induction of the VDR in 3T3-L1 cells, and with similar effects of this pathway in other cell types on VDR expression (24,25).
To determine whether VDRsignaling influences formation of PPARγ ligand activity in 3T3-L1 cells we analyzed ligand activity in 3T3-L1 5B2 cells which are stably transfected with a PPARγ ligand sensing system which activates a β-galactosidase reporter in the presence of a PPARγ ligand (10). As shown in Figure 2, 24 hours after the addition of DIM, β-galactosidase levels increase by close to 8-fold consistent with the production of an endogenous PPARγ ligand. Indeed, 500 mM rosiglitazone induced a 4-fold induction of the same reporter. In the presence of DIM and calcitriol reporter activation is inhibited by approximately 40% consistent with the notion that the VDR-signaling system lies upstream of PPARγ ligand formation. To control for possible squelching we added calcitriol to the rosiglitazone treated cells and saw no effect indicating that liganded-VDR was not blocking the liganded Gal4-PPARγ directly or via competition for limiting cofactors.

Liganded-VDR Blocks Adipogenesis by Lowering C/EBPβ Levels
To confirm the ability of calcitriol to inhibit adipogenesis we examined a range of concentrations of calcitriol. Concentrations of calcitriol from 100nM down to 1 nM are sufficient to inhibit adipogenesis as shown by the lack of oil red O staining ( Figure 3A). This data demonstrates that the serum used during DIM-induced adipogenesis contains very little calcitriol, supporting the notion that the VDR expressed early in adipogenesis is unliganded.
We next examined the levels of VDR in the presence of calcitriol. As expected calcitriol acts to stabilize the VDR in nuclear extracts and prevent its degradation, as has been demonstrated previously (26,27). Because of this effect the VDR is present throughout most of the adipogenic period.
However, a transcriptional component appears to also play a role, as VDR transcription is reinitiated at day 4 and increases further at day 5 ( Figure 3B). Thus, the presence of calcitriol through both transcriptional and post-transcriptional mechanisms induces the production and stability of the VDR explaining in part its ability to potently block adipogenesis.
To ensure that calcitriol was not blocking mitotic clonal expansion, which occurs in the first few days after DIM addition and appears to be necessary for adipogeneis (28) we measured DNA content using crystal violet staining. As shown in Figure 3C, calcitriol has a small effect on expansion but the treated cells still go through a near doubling process in the first 48 hours, indicating that the liganded-VDR is not fully blocking this process.
In order to determine the mechanism by which the liganded-VDR inhibits adipogenesis we examined the expression of key members of the transcriptional cascade that induce the adipogenic program. Not surprisingly both C/EBPα and PPARγ levels were very low in calcitriol treated nuclear extracts ( Figure 4A).
In contrast, C/EBPα accumulation in the nuclear extracts from untreated cells is dramatically enhanced on day 2 after the addition of DIM and stays at high levels throughout the adipogenic program. Similarly, PPARγ expression becomes noticeable at day 1 and progressively increases during the remainder of differentiation.
Thus, calcitriol appears to have its effects upstream of the induction of C/EBPα and PPARγ.
We next assessed the protein levels of C/EBPβ in nuclear extracts from calcitriol treated and control cells. As is shown in Figure 4B, C/EBPβ protein levels begin to increase 6 hours after the induction of the adipogenic program, though some basal expression is present at time 0. C/EBPβ protein levels peak during day 2 and then fall off, presumably after inducing adequate amounts of C/EBPα and PPARγ to maintain effective adipogenesis. In contrast, in calcitriol treated nuclear extracts C/EBPβ protein levels rise in a similar fashion to control extracts but are decreased on day 2 suggesting that the liganded-VDR functions to block C/EBPβ accumulation and thus prevents full activation of C/EBPα and PPARγ levels. While the data presented highlight the 34 KD LAP isoform of C/EBPβ we could also visualize the smaller LIP isoform which was expressed at lower levels and did not change in proportion with calcitriol treatment. In the same experiments we also examined C/EBPδ levels. In the DIM treated extracts there was little C/EBPδ at time 0 but this progressively increased to a peak at 4 hours and then rapidly fell. This early pattern of induction was not changed in the calcitriol treated nuclear extracts consistent with calcitriol affecting only C/EBPβ levels.
To determine if calcitriol signaling regulates C/EBPβ at the level of transcription we next looked at C/EBPβ mRNA levels. In control cells, C/EBPβ mRNA is induced strongly 6 hours after the addition of DIM in the absence of calcitriol and is maintained until day 2 after which it begins to fall towards its basal level (Figure 4B). In the presence of calcitriol, C/EBPβ mRNA is also induced. However, it is also downregulated much more quickly beginning at day 1 and progressing through day 2 such that it has returned to basal levels. These data are consistent with the protein levels seen in nuclear extracts. Thus, calcitriol either directly or indirectly negatively regulates C/EBPβ gene expression in differentiating 3T3-L1 cells.
We also considered whether calcitriol-VDR signaling may also regulate pathways which may modify C/EBPβ action.
Recently, Rochford et al demonstrated that the transcriptional corepressor ETO/MTG8 is expressed in preadipocytes and is down-regulated via insulin signaling in early adipogenesis (29). When overexpressed, ETO/MTG8 is a potent inhibitor of adipogenesis through its ability to directly interact with C/EBPβ and inhibit its action on the C/EBPα promoter. To analyze the role of calcitriol on ETO/MTG8 expression we looked at ETO/MTG8 levels in nuclear extracts from differentiating 3T3-L1 cells. As described, ETO levels decrease dramatically in the first 12 hours after the addition of the differentiating cocktail in control cells. In contrast, in the presence of calcitriol (Figure 5), ETO expression is maintained throughout adipogenesis and increases after Day1.
We next considered whether the actions of calcitriol are reversible such that its removal would allow differentiation to resume. As is seen in Figure 6, the addition of calcitriol to DIM treated 3T3-L1 cells for as brief a period as 6 hours is effective in blocking differentiation. Longer treatments from 12 hours up to 3 days are clearly more effective. To control for the addition and removal of calcitriol, we added the ligand for 1 hour, and saw no effect on adipogenesis (data not shown). Not surprisingly, the effects of short term pulses of calcitriol are mediated by the inhibition of the induction of C/EBPα and PPARγ that would normally occur in DIM treated cells by Day 3 (Figure 6B).
Finally, to determine if the dramatic induction of the Insig-2A isoform by calcitriol could also be playing role in the ability of the liganded VDR to block adipogenesis we asked whether Insig-2 lies downstream of calcitriol (20). Since both Insig 2A and 2B encode the same protein we first asked whether total Insig-2 mRNA is regulated by calcitriol (30). To do this we used Q-PCR with the assay directed at a coding exon that both isoforms share. Using this assay we saw only a moderate two-fold induction in Insig-2 mRNA 24-36 hours after the addition of DIM and calcitriol (not shown) suggesting that Insig 2A makes up only a small portion of total Insig-2 with Insig 2B making up the majority. We next used siRNA against Insig-2 to determine whether its induction plays a role in calcitriol action in inhibiting differentiation. To accomplish this we used a siRNA pool which dramatically reduced Insig-2 mRNA expression ( Figure 7A). As is shown in Figure 7B knockdown of Insig-2 is pro-adipogenic leading to a substantial induction of oil red O staining 5 days after the addition DIM. However, siRNA directed against Insig-2 has no effect on the ability of calcitriol to inhibit adipogenesis. Thus, the actions of calcitriol do not appear to require induction of Insig-2.

The Role of the Unliganded VDR in Adipogenesis
Our data demonstrates that the early expression of the VDR in the adipogenic program occurs in the absence of appreciable amounts of calcitriol present in the media and serum, thus the VDR likely acts early in adipogenesis as an unliganded receptor. To determine whether the unliganded-VDR plays a role in adipogenesis we used siRNA to downregulate its expression in differentiating 3T3-L1 cells. Confluent 3T3-L1 cells were transfected with varying amounts of siRNA duplexes 24 hours prior to the induction of adipogenesis with DIM. As shown in Figure 8A, successful knockdown of the VDR can be accomplished using this technique and levels of the VDR remain at low levels when compared to control cells 24 hours after the induction of adipogenesis. We next examined the role of the unliganded VDR in adipogenesis. We allowed groups of the transfected cells with varying amounts of the VDR siRNA to differentiate and performed oil-red O staining at Day 8. We found that 100 nM siRNA was most effective at decreasing both VDR levels and oil-red O staining at Day 8 (13% when quantified by optical density after elution in isopropanol).
To further examine the role of the unliganded VDR in adipogenesis we also wanted to look at the rate of adipogenesis by examining lipid accumulation earlier than Day 8.
To do this we again transfected 3T3-L1 cells with either VDR siRNA or scrambled duplexes (and also a mock transfection control) and looked at VDR protein levels at Day 1 after the addition of DIM and oil red O staining at Day 5. As demonstrated in Figure 9A, VDR protein levels were significantly reduced at Day 1 in cells transfected with VDR siRNA as compared to control cells and oil red O staining was also decreased by 35% in these cells as compared to control cells. To confirm this data we repeated the dose response experiment and stained again at Day 5. Again, increasing doses of VDR siRNA caused a reduction in oil red O staining consitent with an effect on lipid accumulation ( Figure 9B).
While oil red O staining is partially impaired we also wanted to examine markers of adipocyte differentiation in cells treated with VDR siRNA. As shown in Figure 9C, VDR levels are substantially reduced over the course of differentiation by VDR siRNA. However, induction of the PPARγ isoforms is not blocked by VDR siRNA suggesting that the role of the unliganded VDR is not complete in blocking the adipogenic program both in context of oil red O staining and induction of adipogenic markers.

Discussion
The 3T3-L1 cell line is a key model that has revealed regulatory routes by which transcription factors induce adipogenesis. While it has been long been known that the VDR and its cognate ligand calcitriol inhibit adipogenesis in this model the mechanism has remained elusive. Given the role of the VDR in regulating tumor growth and proliferation in malignancy, the mechanism by which the unliganded and liganded VDR regulate adipogenesis will provide further key insight into the role of the vitamin D signaling pathway in health and disease (12). In addition, the data presented here demonstrate unique roles for the VDR depending upon the availability of its ligand, calcitriol, suggesting that the intracellular concentration of calcitriol might be a target for modulation in the treatment of obesity.
The VDR like other NR family members including TRs, RARs, and PPARγ is bound to regulatory elements in target genes as heterodimers with the retinoid x receptor (RXR) in both the absence and presence of their cognate ligands (22). In the absence of ligand these receptors recruit the nuclear corepressors, NCoR and SMRT, which act as platforms for a multiprotein complex that mediates transcriptional repression via histone deacetylation (31)(32)(33)(34). Indeed, the unliganded VDR is known to recruit NCoR and SMRT and appears to recruit a novel complex containing the Williamssyndrome transcription factor (WSTF) (35). Furthermore, the unliganded VDR plays a distinct role in hair follicle development in vivo establishing it as a bona fide mediator independent of the presence of calcitriol. Indeed, the unliganded VDR's actions in the hair follicle may be through its recruitment of the tissue-specific corepressor hairless (36,37). In the presence of ligand, an allosteric shift repositions the helices of the NR ligand-binding domain which displaces the corepressors and allows for the targeted recruitment of a coregulatory complex that allows for transcriptional activation via histone acetylation and methylation (38)(39)(40). While this model explains the action of the VDR on most targets, certain targets genes are paradoxically repressed by the liganded VDR such as the 25(OH)D 3 1α-hydroxylase promoter. In this case the liganded VDR is able to bind to and block the activation function of VDIR which binds to and activates the 1α-hydroxylase promoter (41). Thus, the VDR has important biologic activity in the absence of ligand and has the ability to either inhibit or activate gene expression in the presence of ligand, In this study we demonstrate unique functions of both the unliganded and liganded VDR in adipogenesis in 3T3-L1 cells. Surprisingly it is the unliganded VDR which may promote adipogenesis, while the liganded VDR represses C/EBPβ expression and inhibits adipogenesis. While the unliganded VDR is not required for adipogenesis it appears to enhance the process and allow for more lipid accumulation. The mechanism by which the VDR acts is not clear but the possibility exists that an adipocyte-specific cofactor mediates the effects of the unliganded VDR on lipid accumulation. Certainly VDR knock-out mice appear to have normal amounts of adipose tissue. However, due to the dietary needs of these animals it is unclear how they would respond to a nutritional challenge such as a high fat diet.
Our data demonstrate that the VDR is expressed early in adipogenesis, driven by the cAMP pathway which is a known to activate VDR expression in other cell types. The mechanism by which this pathway increases VDR transcription is not clear though a cAMP response element is present in the VDR promoter (24). Interestingly, the early response gene, krox20, has recently been demonstrated to be essential for adipogenesis and to be expressed in 3T3-L1 cells in the first 4 hours after the addition of DIM (42). Krox20 appears to function by enhancing C/EBPβ expression, though, the region of the C/EBPβ promoter that mediates this effect lacks a krox20 binding site. However, the VDR promoter contains a consensus krox20 site near the transcriptional start-site. It is tempting to speculate that the early induction of krox20 allows for VDR expression which in turn plays a role in full induction of C/EBPβ in the absence of calcitriol (24,42). A precedent for ligandindependent activation of gene expression by NRs exists such that the TR-isoforms activate transcription of target genes such as TSHβ in the pituitary and TRH in the hypothalamus (43)(44)(45).
In contrast to the role of the unliganded VDR, the liganded VDR is a potent inhibitor of adipogenesis in 3T3-L1 cells. While VDR levels decline rapidly in the absence of calcitriol, the presence of calcitriol maintains VDR expression through at least Day 5 of the adipogenic program. This is in keeping with the wellestablished role of calcitriol in stabilizing the VDR protein itself (26,27). Thus, the liganded VDR is able to persistently act to inhibit adipogenesis. The mechanism by which the liganded VDR inhibits adipogenesis is multifactorial but appears focused on the C/EBPβ signaling pathway. Our data demonstrate that the liganded-VDR causes a substantial fall in both the induction of C/EBPβ mRNA and the amount of C/EBPβ protein present beginning between Day 1 and 2 of the differentiation program. This occurs at a critical period when C/EBPβ is required for the induction of C/EBPα and PPARγ. Thus given the required role of this pathway in adipogenesis, it is likely that C/EBPβ levels are too low to induce C/EBPα and PPARγ. In addition calcitriol inhibits PPARγ ligand formation consistent with the previously described role of C/EBPβ in this process (46) (9).
While calcitriol-VDR signaling directly inhibits C/EBPβ production, the remaining C/EBPβ seen would have little transcriptional effect based on the high levels of ETO/MTG8, present in the calcitriol treated cells.
Indeed, ETO directly blocks transcriptional activation mediated by C/EBPβ based on its ability to interact directly with C/EBPβ on target promoters and recruit corepressors such as NCoR and SMRT to target genes (29,47). Further work will be required to understand the mechanism by which calcitriol-VDR signaling reactivates ETO/MTG8 expression.
While disruption of the C/EBPβ pathway is critical for the effects of the liganded-VDR the mechanism by which the liganded-VDR blocks C/EBPβ expression is not clear.
Given that C/EBPβ mRNA levels accumulate during the first 12-24 hours of differentiation in the presence or absence of calcitriol it is likely that the liganded-VDR induces the expression of an adipocyte-specific negative regulator of C/EBPβ transcription rather than directly inhibiting C/EBPβ transcription. This is supported by a recent report demonstrating that calcitriol induces C/EBPβ mRNA levels in osteosarcoma and kidney cells, suggesting the cell-specific expression of regulatory cofactors may influence liganded VDR action on the C/EBPβ gene (48). Further work will focus on genes induced by calcitriol in 3T3-L1 cells in order to clarify this mechanism. In addition, identification of pathways induced by calcitriol will allow for insight into the mechanism by which calcitriol-VDR signaling also inhibits PPARγ ligand formation. While not primary in the ability of calcitriol to inhibit adipogenesis, identification of further elements this pathway may facilitate efforts to identify the endogenous PPARγ ligand.
In addition to the results presented here a recent report has also described the ability of calcitriol to inhibit C/EBPα and PPARγ expression in 3T3-L1 cells (49). Furthermore, these investigators also demonstrated the ability of the VDR to block PPARγ transcriptional activity suggesting that calcitriol may function on multiple levels to regulate adipogenesis. This report, did not demonstrate regulation of C/EBPβ by calcitriol, but this was not quantitatively assessed.
In summary, we have demonstrated that the liganded-VDR represses both C/EBPα and PPARγ expression via inhibition of C/EBPβ expression and action and is a potent inhibitor of adipogenesis. In contrast, the unliganded receptor is not required for adipogenesis but may play a role in some aspect of the process. These data suggest that the intracellular levels of calcitriol play a key role in adipocyte formation. Thus, altering the production of intracellular calcitriol may allow for the modulation of adipogenesis in vivo.

Abbreviations
VDR, vitamin D receptor; PPAR, peroxisome proliferator-activated receptor; C/EBP, CCAAT/enhancer binding protein; NR, nuclear receptor; TR, thyroid receptor; RAR, retinoic acid receptor; HDAC, histone deacetylase  Two days after reaching confluence dexamethasone, insulin and IBMX was added to 3T3-L1 cells and total RNA was isolated at the indicated time points. Q-PCR for the VDR and 18S RNAwas performed and the data normalized (VDR/18S). B. Using an identical paradigm nuclear extracts were isolated from 3T3-L1 cells and equal amounts were run on SDS-PAGE and probed with either a VDR or HDAC1 antibody. C. 3T3-L1 cells were exposed to either insulin, dexamethasone or IBMX and nuclear extracts isolated 6 and 12 hours later and VDR levels were analyzed by Western.  B. Groups of 3T3-L1 cells were allowed to differentiate in the presence of 10 nM calcitriol and either mRNA or nuclear extracts were isolated at the indicated time points. Q-PCR was performed for VDR and 18S and the data was quantified as described above. VDR nuclear protein levels were determined by Western analysis. C. 3T3-L1 cells were allowed to differentiate for the indicated amount of time after which the cells were stained with crystal violet to determine DNA amounts and cell proliferation. Each point was performed in duplicate.

Figure 4:
The calcitriol-VDR pathway inhibits adipogenesis by negatively regulating C/EBPβ protein and mRNA levels. A. Nuclear extracts were obtained from differentiating 3T3-L1 cells at the indicated time points and Western analysis was performed for C/EBPα and PPARγ either in the presence or absence of 10 nM calcitriol. B. mRNA and nuclear extracts were isolated from differentiating 3T3-L1 cells at the indicated time points in the presence or absence of 10 nM calcitriol. Q-PCR was performed for C/EBPβ and 18s mRNA and the data normalized (C/EBPβ/18s). * P< 0.05   Confluent 3T3-L1 pre-adipocytes were transfected with siRNA directed against insig-2, a scrambled control or were not transfected. 24 hours later DIM was added either alone or with 10nM calcitriol. Parallel groups of cells were lysed 24 hours after the addition of DIM to collect mRNA for Q-PCR to demonstrate efficient knock-down of insig-2 (A.) or were allowed to differentiate for 5 days and then fixed and stained with oil red O (B.) Figure 8: Knockdown of VDR levels in 3T3-L1 cells using siRNA. A. One-day postconfluent 3T3-L1 cells were transfected with a variety of concentrations of VDR siRNA duplex. Twenty-four hours later (two days post-confluence, time 0) differentiation cocktail was added and nuclear extracts were isolated 24 hours later. Equal amounts of nuclear extracts were analyzed by Western for VDR and HDAC1. B. Parallel groups of 3T3-L1 cells were transfected with the indicated concentration of VDR siRNA duplex 24 hours after reaching confluence. Twenty-four hours later differentiation was induced cells from this transfection were stained with oil red O 8 days after the addition of DIM. hours after the addition of differentiation cocktail and Western analysis was performed for VDR. The remainder of the cells were allowed to differentiate for 5 days and then fixed and stained with oil red O. Oil red O accumulation was quantified by extracting in isopropanol and measuring optical density (OD) in duplicate. B. Groups of 3T3-L1 cells were transfected with a variety of concentrations of VDR or control siRNA duplex. Parallel groups of cells were allowed to differentiate or lysed for nuclear extracts. C. One-day post-confluent 3T3-L1 cells were transfected with 100nM of a VDR siRNA duplex or a scrambled control. Nuclear extracts were isolated at the indicated time points and equal amounts were subjected to SDS-PAGE and analyzed by Western blot.