Osteoblasts and adipocytes are derived from mesenchymal stem cells, and many studies have suggested the presence of plasticity between these two cell types (1, 2). Aging and estrogen deficiency are also known to produce an increase in the bone marrow adipocity as well as decrease in the ability of mesenchymal precursor to progress toward osteogenesis (2, 3). Several studies have also suggested the occurrence of both osteoblastic and adipocytic phenotype in the same cell population (4). In fact, a preadipocyte cell line has been derived from osteoblastic cell population from mouse calvaria (5).

One of the factors that regulate adipogenesis is peroxisome proliferator-activated receptor-gamma (PPARγ), which belongs to the class of nuclear hormone receptors that are affected by specific ligands such as fatty acids as well as by cross talk with other molecules. PPARγ is both sufficient and necessary for conversion of preadipocytes into adipocytes and is a master regulator of adipogenesis (6).

Vitamin D is an important bone anabolic agent, and several bone specific genes are directly regulated by this hormone. 1,25-Dihydroxyvitamin D3, the active form of vitamin D functions through its receptor, vitamin D receptor (VDR) a ligand dependent transcription factor, belonging to the same class of nuclear receptor family of transcription factors as PPARs. The similarity between these two factors is that they mediate their action by heterodimerizing with retinoic acid receptor in the regulation of gene expression. Studies on vitamin D effects on osteoblasts are mixed. While some studies show vitamin D to inhibit adipocyte differentiation, other studies have suggested adipocytic conversion of bone marrow osteoblasts by vitamin D (6–8). While vitamin D has been shown to inhibit adipocytic differentiation of preadipocytes, there are studies that demonstrate a stimulatory effect on other mesenchymal derived cells (7, 8).

There is also cross talk between PPARγ and VDR as VDR elements have been discovered in the PPAR genes demonstrating an influence of vitamin D over PPAR system (9). PPARγ also appears to bind to VDR and inhibit vitamin D dependent transactivation (10).

Retinoblastoma tumor suppressor gene (Rb1) function is important for controlling the cell cycle progression and differentiation of several cell types (11). Loss of Rb1 gene function predisposes individuals to retinoblastoma and osteosarcomas which are commonly seen as secondary tumors in the pediatric population. Recent animal studies have suggested a role for pRb in the regulation of mesenchymal cell fate towards osteoblasts, by potentiating the effect of Runx2 a master regulator for osteoblast differentiation (12). pRb1 is also known to be important for adipocytic differentiation as it is required for activating several of the transcription factors involved in adipocyte differentiation (13). However, complete loss of Rb1 expression from mesenchymal stem cells predisposes them to an adipocytic cell fate (14). While the role is complex and yet to be clearly elucidated, Rb1 is generally considered to regulate cell fate and in the absence of Rb1 cells default to an adipogenic phenotype (14, 15).

In this study, we attempted to determine the effect exposure to adipogenesis promoting (AP) media on MC3T3-E1 preosteoblasts. We also determined how the presence of vitamin D might affect the expression of adipocytic phenotype. The effect of Rb1 deficiency in osteoblasts and their transdifferentiation ability to adipocytes was also studied.

To elucidate if exposure of AP media to the preosteoblast cell line MC3T3-E1 produced transdifferentiation to an adipocytic phenotype, we exposed cells for 6 days to the media and analyzed RNA expression for several transcription factors that regulate adipogenic differentiation. The transcriptional program regulating adipogenesis is well understood and proceeds through sequential expression of different transcription factors (16). Our data on MC3T3- E1 cells followed the expected changes that occur during adipogenesis where C/EBPβ and C/EBPδ expression precedes expression of PPARγ and C/EBPα. As seen in Figure 1A, the two transcription factors (C/EBPβ and C/EBPδ) were increased within a day of treatment to about fourfold and a steady increase that continued through the treatment (P < 0.01 for C/EBPβ, P < 0.05 for C/EBPδ). CEPBβ showed a larger increase (15-fold) than C/EBPδ (5-fold). This was followed by an increase in PPARγ after 3 and 6 days of treatment (P < 0.01). An increase in C/EBPα was observed on Day 6 on AP media. The changes in these transcription factors were accompanied by an increase in expression of several adipocyte markers such as FABP4, UCP1, and SREFBP1 (Figure 1B). Positive Oil Red O staining confirmed the presence of lipid droplets, which were observed in MC3T3-E1 cells exposed to AP media but not regular media (Figure 1C).

To determine the effect of vitamin D on osteoblasts in the presence of adipogenic differentiation factors, we exposed cells to 20 nM vitamin D for a week with media change every other day. Real-time PCR analyses of adipocyte transcription factors revealed a large upregulation of expression with vitamin D treatment. Figure 2 shows changes seen with and without vitamin D treatment of MC3T3-E1 cells. The overall sequence of expression of the transcription factors followed what we saw in cells without vitamin D. However, there was a dramatic statistically significant (P < 0.001) interaction between media and vitamin D in the upregulation of several factors. The only transcription factor that did not appear to be affected by vitamin D was C/EBPβ, which showed a change similar to that of cells treated only with adipogenic media. C/EBPδ showed about a 50-fold increase, PPARγ a 200-fold increase, and C/EBPα a 300-fold increase in activity when compared to cells treated only with adipogenic media.

To confirm that this was a result of vitamin D receptor mediated action, we measured expression of vitamin D receptor (VDR) during vitamin D treatment. We treated MC3T3-E1 cells with adipogenic media with 20 nM 1,25 dihydroxyvitamin D with frequent changes of media as described in Section “Materials and Methods.” Control cells received the same amount of vehicle (EtOH). Addition of vitamin D produced a maximal increase of 600-fold 4 days (P < 0.001) after treatment with media with other days showing several fold elevation of expression (Figures 3A,B), demonstrating that the action of vitamin D did indeed result in an increase in its cognate receptor levels. Western blots for PPARγ (not shown) and C/EBPα (Figure 3C) also confirmed the changes seen in RNA expression.

Adipocytes and osteoblasts share a common precursor in the bone marrow, and it is known that there is a high degree of plasticity between the two cell lineages (1). MC3T3-E1 is a clonal preosteoblast cell line that was originally isolated from new born mouse calvaria (17). To demonstrate plasticity in this line, investigators have generally employed genetic manipulations that have included overexpression of adipogenic transcription factors (18), treatment with drugs (19), etc. In this study, we attempted to understand the influence of adipogenic media with and without vitamin D or the effect of reduction in Rb1 expression on MC3T3-E1cells. Our studies show that exposure of these cells to media generally utilized to promote adipogenesis in preadipocyte cell lines such as 3T3-L1 can produce gene expression characteristic of adipocytes in determined preosteoblasts. However, we were surprised to find a synergistic enhancement of adipocyte differentiation factors and associated markers on treatment with vitamin D. While the role of vitamin D receptor and vitamin D on adipogenesis has not been clearly elucidated, it is well known that vitamin D receptors are necessary and present in adipocytes (20, 21). In fact, studies have demonstrated an inhibition by vitamin D of adipocyte specific expression of PPARγ and other factors in preadipocytes (22). None of the other transcription factors involved in adipocyte differentiation are also known to be regulated by vitamin D. Since VDR is expressed in response to vitamin D treatment, it appears that some of the effect could be mediated at the level of ligand receptor interaction. However, it is also known that vitamin D might have non-genomic actions (23) that may not involve the classical steroid hormone mediated stimulation of gene expression. Irrespective of the mechanism, treatment of preosteoblasts with vitamin D did not inhibit but promoted adipocyte differentiation. We cannot exclude the fact that adipogenic media components 3-isobutyl-1-methylxanthine and insulin may have played a role in influencing the adipocytic phenotype of these cells (20, 21).

While the final outcome of MC3T3-E1 cells exposed to adipogenic media was the same with and without vitamin D, there were important differences in expression of C/EBP factors. In the case exposure to just the media, the changes in gene expression agreed with previously published observations during preadipocyte differentiation to adipocytes (24). With vitamin D, our results show a dramatic increase in the case of C/EBPα, which went up several hundred fold and C/EBPδ which was increased 20-fold to 40-fold over just the media alone. While the extent of changes seen in C/EBPβ after exposure to AP media was the highest, with addition of vitamin the effect remained the same without any enhancement we saw in CEBPα and C/EBPδ (Figure 2). In a preadipocyte cell line 3T3-L1, liganded vitamin D appears to inhibit CEBPβ while unliganded VDR is seen necessary for adipogenesis (21). Our results may represent an inhibition of CEBPβ expression by vitamin D treatment, but the fact that upregulation provided by the media alone was still retained, might explain the overall enhancement in adipogenesis.

Previous work has suggested that the tumor suppressor gene Rb1 might be critical for cell fate determination and it has been shown that loss of pRb1 in mesenchymal stem cells cause them to default to an adipocytic cell lineage (14). We therefore used a cell line that has stable reduction in Rb1 expression to demonstrate that a similar phenotype of enhanced adipogenesis was seen. Interestingly, in these cells, there was a steady state increase in PPARγ expression and agrees with previous work that suggests that pRb inhibits PPARγ expression in mesenchymal stem cells (25). C/EBPα and C/EBPβ expression was generally increased with Rb1 loss, but C/EBPδ expression was unaffected. This was unlike our observations with vitamin D and might relate to independent roles of Rb1 in regulating these transcription factors. It is also known the Rb1 regulation of adipogenesis might depend on several factors, since pRb regulates not only the CEBPs but also cell cycle control through E2Fs. In addition, loss of Rb1 promotes brown adipose tissue conversion as evidenced by markers such as UCP1 by these cells (Figure 5C).

Our studies generally reveal that uninduced preosteoblasts maintain their plasticity toward adipocytes and factors in the adipogenic media appear to be sufficient to induce MC3T3-E1 cells to express the adipocyte phenotype. While this is not surprising, what is surprising is the fact that treatment with vitamin D in the presence of adipogenic factors dramatically enhances gene expression in preosteoblasts toward adipocytes. While this result was unexpected, studies that have looked at genomic binding of transcription factors in differentiated and undifferentiated cells demonstrate that this might relate to the plasticity of the epigenome (26). Both histone modifications and transcription factor binding appear to be different in mature versus immature osteoblasts (26, 27). While there may be a default pathway or preference for a particular cell fate, it appears to be easily perturbed at the level of transcription factor binding (26). In fact, these investigators show that transdifferentiation of preosteoblasts to adipocytes is associated with rapid changes to chromatin in regions containing the adipocyte specific genes in MC3T3-E1 cells. It is also known that unliganded VDR remains bound to a number of target genes with the genome and undergo modifications with vitamin D treatment. It is possible that vitamin D treatment might not have directly contributed to the activation of adipocyte specific gene expression but may have altered the transcription factor binding sites of VDR and essentially mitigated the inhibition of VDR targets. It is known that vitamin D exerts both positive and negative effects on osteoblast differentiation (25) and is dependent on timing. In immature osteoblasts, vitamin D downregulated the master regulator of osteoblast differentiation Runx2 while increasing PPAR activity and adipocyte differentiation (28). While our studies have not investigated in parallel how osteoblast differentiation may be affected in the presence of adipogenic media and vitamin D, it is likely that competition of adipogenic and osteogenic transcription factors for binding on DNA might have influenced the response seen. There is much to be learned about vitamin D targets as several of the sites where VDR bound on chromatin is not known targets of vitamin D action (27). Our results suggest the need for further functional studies to uncover vitamin D-dependent pathways that govern adipocyte–osteoblast transdifferentiation.

EP designed and performed experiments, analyzed data, and helped write the paper; NC designed experiments, analyzed data, and wrote the paper.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.