Induction of Adipocyte Differentiation by Polybrominated Diphenyl Ethers (PBDEs) in 3T3-L1 Cells

Polybrominated diphenyl ethers (PBDEs) are a class of brominated flame retardants that were extensively used in commercial products. PBDEs are ubiquitous environmental contaminants that are both lipophilic and bioaccumulative. Effects of PBDEs on adipogenesis were studied in the 3T3-L1 preadipocyte cell model in the presence and absence of a known adipogenic agent, dexamethasone (DEX). A PBDE mixture designed to mimic body burden of North Americans was tested, in addition to the technical mixture DE-71 and the individual congener BDE-47. The mixture, DE-71, and BDE-47 all induced adipocyte differentiation as assessed by markers for terminal differentiation [fatty acid binding protein 4 (aP2) and perilipin] and lipid accumulation. Characterization of the differentiation process in response to PBDEs indicated that adipogenesis induced by a minimally effective dose of DEX was enhanced by these PBDEs. Moreover, C/EBPα, PPARγ, and LXRα were induced late in the differentiation process. Taken together, these data indicate that adipocyte differentiation is induced by PBDEs; they act in the absence of glucocorticoid and enhance glucocorticoid-mediated adipogenesis.


Introduction
Polybrominated diphenyl ethers (PBDEs) are a class of brominated flame retardants that were extensively used as additives in commercial products including electronics, textiles, and polyurethane foam [1]. There are approximately 209 congeners of PBDEs that differ based on the degree and position of halogenations [1,2]. Widely produced commercial PBDE mixtures, including pentabrominated phenyl ether (DE-71) and octabrominated diphenyl ether (DE-79), leach readily into the environment as these chemicals were not chemically bound to the polymers of the products to which they were added [3,4]. Production of these chemicals was voluntarily halted in 2004; however, concern has been raised for the potential of PBDEs to bioaccumulate [5][6][7]. Indeed, PBDEs are lipophilic, extremely stable, and remain in products manufactured before the ban, as well as in new products made from recycled materials [1].
As PBDEs are lipophilic and accumulate in fat tissue, it is postulated that they may alter adipose tissue function and metabolism. Evidence supporting this hypothesis showed that adipocytes collected from rats exposed to daily doses of the DE-71 technical mixture exhibited increased lipolysis and decreased glucose oxidation, both hallmarks of obesity, insulin resistance, and type 2 diabetes [8]. Furthermore, a separate study showed that maternal exposure to the congener BDE-47 induced increases in body weights in offspring during early postnatal development [9]. Correlational studies in humans showed that there is a positive association between metabolic syndrome and serum levels of the congeners PBB-153 and PBDE-153 [10]. More recently, the presence of PBDEs and other persistent organic pollutants (POPs) in obese individuals was positively correlated with visceral fat and visceral/subcutaneous abdominal fat ratio [11]. Hence, the effects that PBDEs may have on adipose tissue function warrant further investigation.
In the current study, the 3T3-L1 cell model was used to study the process of preadipocyte differentiation in response to PBDEs. As a cocktail of cAMP enhancer, insulin, and glucocorticoid results in complete differentiation of preadipocytes, glucocorticoid was substituted with PBDEs in order to assess the potential of PBDEs to induce differentiation without high background. A mixture of PBDEs representative of body burden in North America was tested, along with the technical mixture DE-71, and the predominant congener in both of the mixtures, BDE-47. In addition, the potential for PBDEs to differentiate 3T3-L1 preadipocytes was tested in the presence of suboptimal concentrations of the synthetic glucocorticoid, dexamethasone. This study is the first to characterize the terminal differentiation of preadipocytes to mature adipocytes induced by these lipophilic contaminants.

Formulation of a PBDE mixture representative of body burden in North America
A literature review was conducted to determine the formulation of a mixture of PBDEs representative of body burden in North America. Our review was confined to papers that reported a broad range of congeners. Studies with breast milk provided the broadest range of PBDE congeners that were detected with .50% frequency and thus allowed the reliable determination of a median level of these among the populations studied. Three studies examining levels of PBDEs in breast milk from women in various locations across North America showed remarkable similarity in the relative proportions of individual congeners of total PBDE [12][13][14]. Therefore, the mixture was formulated to reflect the relative proportions of congeners observed in these studies (Table 1).
Cell culture 3T3-L1 mouse embryo fibroblasts were cultured in DMEM containing 10% calf serum, maintained at confluence for 2 days, and induced to differentiate as illustrated in Figure 1

Immunofluorescence and quantification of adipocytes
Cells were seeded and differentiated in black, clear-bottomed 96 well plates. Cells were then fixed with 4% paraformaldehyde and stained with Nile red (1 mg/mL) to identify lipid droplets and DAPI (1 mg/mL) to stain nuclei. Nile red fluorescence was viewed at 530 nm and DAPI at 405 nm on a fluorescent microscope. Cells were photographed at 1006 magnification. The number of mature adipocytes in the population was calculated as percentage of Nile red stained cells from the total number of cells. At least 300 cells were counted per well.

Real-time PCR
Cells were harvested on days 1, 2, 4, 5, 6, and 8. RNA was extracted using the RNeasy Mini kit (Qiagen, Missisauga, Canada) and then reverse transcribed using iScript cDNA Synthesis Kit (BioRad). The expressions of the following genes were determined using SsoFast EvaGreen supermix: aP2, perilipin, CCAATenhancer binding protein a (C/EBPa), peroxisome proliferatoractivated receptor c (PPARc), and liver X receptor a (LXRa) ( Table S1). PCR efficiency was examined using standard curves for each gene using pooled cDNA from all control samples. Relative gene expression was normalized to b-actin gene expression.

Statistical analyses
Statistical analyses were performed with SigmaPlot 12.5 software. One way analysis of variance was conducted, followed by Holm-Sidak post-hoc test. For non-normal data, a Kruskal-Wallis one way ANOVA on ranks was performed, followed by Dunnett's post-hoc test. For dexamethasone co-treatment, two way ANOVA was performed followed by Holm-Sidak post-hoc test to compare all groups. When normality failed, the natural logarithm of each data value was determined. Data are presented as mean 6 SEM.

Cell viability
Treatment of undifferentiated 3T3-L1s with the PBDE mixture, DE-71, or BDE-47 did not induce cytotoxicity as indicated by MTT assay ( Figure S1). Moreover, no evidence of cytotoxicity was observed in any treated cultures during the 8 days of differentiation.

PBDE exposure induced protein and mRNA expression of adipogenic markers
To assess the potential for PBDEs to induce differentiation of 3T3-L1 preadipocytes, cells were treated with a mixture of PBDEs representative of body burden in North America, the technical mixture DE-71, and the congener BDE-47. The extent of differentiation was assessed by measurement of protein and transcript levels of adipogenic markers from all cells collected from the culture well at specific time points to give average expression changes across the cellular mixture consisting of fibroblasts and adipocytes. Treatment of the 3T3-L1 cells with PBDEs using the regiment illustrated in Figure 1A resulted in significant increases in aP2 (at 25.5 mM) and perilipin (at 12.8 and 25.5 mM) protein levels ( Figure 2A). The technical mixture DE-71 induced a dose dependent increase in perilipin expression which reached significance at 12 mM, and an increasing trend was observed for aP2 ( Figure 2B). Treatment with 6 or 12 mM BDE-47 induced significant, dose dependent increases in aP2 and perilipin ( Figure 2C). The positive control of 250 nM DEX induced robust increases in aP2 and perilipin expression (40-fold and 97-fold over control, respectively) (data not shown). aP2 and perilipin RNA expression was measured throughout the differentiation process ( Figure 3). 250 nM DEX increased aP2 and perilipin expression levels significantly from controls (DMSO) at days 5, 6, and 8. Perilipin mRNA levels were significantly increased by the PBDE mixture at days 5, 6, and 8 of differentiation, reaching a 15-fold increase at day 8. aP2 mRNA expression levels were also significantly increased by PBDE treatment at day 8 compared to controls (6-fold). The congener BDE-47 (12 mM) significantly increased perilipin expression at days 6 and 8 of differentiation reaching 8-fold over MI controls, and DE-71 (12 mM) significantly increased perilipin mRNA levels at day 8 (7-fold).

PBDE exposure increases staining of lipids
Nile red staining was performed to visualize lipid accumulation on day 8 of differentiation. Increased lipid accumulation was observed with DEX, 25.5 mM PBDE mixture, 12 mM DE-71, and 12 mM BDE-47 on day 8 (Figure 4). Adipocyte counts were significantly increased following exposure to DEX and PBDE mixture (78% for DEX and 27% for PBDEs). A modest increase was also observed in DE-71 and BDE-47 treated cultures.

PBDEs do not affect glucocorticoid receptor-mediated transcription on the MMTV promoter
To test the ability of PBDEs to induce MMTV promoter activity via glucocorticoid receptor (GR), luciferase assays were performed. Treatment of transfected 3T3-L1 cells with PBDEs (25.5 mM), DE-71 (12 mM), or BDE-47 (12 mM) did not affect

Glucocorticoids enhance PBDE-induced adipocyte differentiation
To assess PBDE interaction with glucocorticoids, 3T3-L1 cells were differentiated in the presence of a low dose of DEX (5 nM). The results are summarized in Table 2. Treatment with 5 nM DEX alone resulted in a 7-fold increase in aP2 expression and 15fold increase in perilipin expression when compared to the MI control. When the preadipocytes were treated with 3.2 mM PBDE mixture and DEX, perilipin expression was further increased by 4fold compared to 3.2 mM PBDE mixture alone, and enhanced DEX-mediated increases by 1.4-fold. BDE-47 co-treatment with DEX enhanced aP2 expression by 3-fold compared to 12 mM BDE-47 alone and 4-fold compared to DEX, although these changes did not reach statistical significance. Both 1.5 mM and 12 mM BDE-47 with DEX co-treatment induced significant increases in perilipin expression compared to 1.5 and 12 mM BDE-47 alone, and induced 2-fold and 5-fold increases over DEX alone. 12 mM DE-71 co-treatment with DEX enhanced DEXmediated aP2 and perilipin expression by 4.5-fold and 4-fold, respectively.

PBDE exposure induces RNA expression of transcription factors and nuclear receptors required for adipogenesis
To characterize the transcriptional cascade involved in PBDEinduced differentiation, PPARc, C/EBPa, and LXRa mRNA levels were determined by RT-PCR following exposure to the highest doses of flame retardants tested ( Figure 5). PPARc mRNA levels were induced by DEX beginning at day 2, and this increase was sustained throughout the differentiation process. Treatment with the PBDE mixture, DE-71, or BDE-47 throughout the differentiation process caused increases in PPARc transcript levels at day 8, although only the treatment with the PBDE mixture reached statistical significance.
C/EBPa RNA expression was increased by DEX as early as day 4. The PBDE mixture induced a slight increase in C/EBPa levels at day 5, although a more marked increase was evident only at day 8. DE-71 and BDE-47 increased C/EBPa RNA levels starting at day 5; however, these changes were more modest.
As expected, LXRa RNA expression was increased by DEX significantly at days 5, 6, and 8. The PBDE mixture (25.5 mM) induced significant increases in LXRa RNA levels as early as day 5, which was maintained until day 8. DE-71 and BDE-47 increased LXRa transcript levels at days 6 and 8, although these changes were not statistically significant.

Discussion
In the current study, we show that a PBDE mixture, DE-71, and BDE-47 induce an increase in aP2 and perilipin expression in the mixture of preadipocytes and mature adipocytes at the end of the differentiation process. Cell counts of the cells that stained positively with Nile red, a dye specific for lipid, show a significant increase in adipocyte numbers following treatment with the PBDE mixture, more so than with DE-71 or BDE-47. As the content of BDE-47 in the PBDE mixture is comparable to the highest dose of BDE-47 tested, it is possible that the components of the mixture may be acting together to enhance the adipogenic effect. In humans, the most prevalent congeners found in fat depots are BDE-47, BDE-99, and BDE-153 [17]. Hence, future studies directed towards testing the other congeners on their own and in combination would provide valuable, environmentally-relevant data on PBDE-induced adipogenesis.
Although current literature regarding the adipogenic effects of PBDEs is limited, a recently published report tested the differentiation potential of BDE-47 by exposing 3T3-L1 preadipocytes to IBMX, insulin, and DEX (MID), in addition to BDE-47 [18]. These experiments showed approximately a 1.5-fold increase in differentiation with 12 mM BDE-47 compared to controls exposed to MID. In our system, we omitted DEX when testing the differentiation potential of the PBDEs in order to delineate the mechanism of action. It is important to note that in the absence of DEX, 3T3-L1 cells do not undergo significant differentiation. Therefore, it is particularly striking that in the absence of DEX, PBDEs induce the differentiation process in 3T3-L1s. Although these results suggest that PBDEs may be acting as glucocorticoids, neither the PBDE mixture, DE-71, or BDE-47 could induce glucocorticoid-mediated MMTV-luciferase activity in the presence or absence of DEX. This suggests that the PBDEs do not act on GR-mediated transcriptional activity, or at least they do not influence GR action on the MMTV promoter. It is also intriguing that in our experiments, co-treatment of the cells with the PBDEs and DEX resulted in an additive effect which further suggests that the PBDEs act on a different molecular target than the glucocorticoid receptor. This may have implications on the potency of the flame retardants in conditions where higher glucocorticoids are present, such as under stress [19].
The effects of PBDEs on the expression levels of key transcription factors were tested by measuring mRNA expression. Compared to DEX, exposure to the PBDE mixture, DE-71, or  BDE-47 resulted in the late induction of PPARc and C/EBPa on days 5-8, which in turn was sufficient to induce the adipogenic factor, LXRa [20][21][22]. It has been postulated that LXRa regulates lipogenesis and maintenance of the mature adipocyte phenotype, but not adipogenesis itself [23]. However, the role of this transcription factor in adipogenesis is controversial and it may be involved in PPARc regulation. A previous study demonstrated that treatment of the 3T3-L1s with an LXRa agonist induces the expression of PPARc [21]. As PPARc can also upregulate LXRa expression, these two transcription factors may form a positivefeedback loop that then pushes the adipocyte differentiation process forward [21,24]. It is therefore possible that the PBDEs are acting through LXRa or PPARc. The 3T3-L1 cell line is widely used as a model to study adipogenesis, and has been used to establish most of the molecular events that occur in the differentiation process; however, there are limitations to this model [25]. This cell line may not be completely reflective of adipogenesis in human preadipocytes as it was derived from clonal expansion of isolated rodent fibroblasts, and may not accurately represent the variations in micro-environments observed in different fat depots in vivo [26]. As observed in many other cell lines, the 3T3-L1s do not appear to have the capability for metabolism of some xenobiotics [27]. Differences may also exist between the delivery of PBDEs to the target tissue in vivo compared to cell culture, although the precise mechanism in both models is unknown. In vivo, PBDEs are thought to circulate by binding to serum proteins, as well as serum lipoproteins [28][29][30]. As 96% of PBDEs administered in culture remains serum-bound, PBDEs are possibly binding to proteins and/or lipoproteins present in the serum [30,31]. In addition, uptake of PBDEs by active transporters has been reported; however, it is unknown whether these are present in 3T3-L1 fibroblasts [32]. While the precise mechanisms by which delivery and uptake occur remain unknown in vitro and in vivo, different mechanisms may be involved, and that may result in comparatively different adipogenic responses. Nevertheless, the 3T3-L1 preadipocyte model remains a useful tool for identification of adipogenic agents, as well as characterization the molecular pathways involved in adipogenesis.
Here, we identified PBDEs as a potential adipogenic agent in the 3T3-L1 cell line and although these results were obtained in vitro, the molecular targets may be relevant in vivo.
Previously reported values regarding human body burden of PBDEs vary regionally. Highest levels were observed in occupational exposures, specifically at electronics dismantling plants, with serum values ranging from 3.8-8500 ng/g lipid for total PBDEs [33]. Assuming that the lipid content of serum is approximately 0.3-0.5%, the lowest dose of BDE-47 that induced an effect on aP2 or perilipin expression in the current study (1.5 mM) represents a concentration that is approximately 20-fold higher than the highest serum levels for total PBDEs measured in occupational exposures. While the nominal concentrations used in the current study are higher than what would be expected in serum samples from the general population, the actual concentration of our test substance at the cellular level in the current in vitro model is not clear. Transfer of BDE-47 from culture media to cells in vitro was previously shown to be inhibited by the addition of serum in the media, as only 3% of the applied PBDE was associated with the cells following an hour of incubation, and 96% remained in the media [30]. In addition, as PBDEs tend to accumulate in adipose depots, it is possible that human preadipocytes may be exposed to higher concentrations of PBDEs than found in serum due to the physiological proximity of these cells to white adipose cells. While the ranges of concentrations used in this study are higher than reported serum concentrations, their effects on adipogenesis may be physiologically relevant due to the inherent properties of cell culture, as well as bioaccumulative effects in vivo.
In this study, we present the novel finding that a PBDE mixture designed to mimic body burden in North America, the technical mixture DE-71, and the congener BDE-47 induces differentiation of 3T3-L1 preadipocytes in the absence of glucocorticoid. Examination of the transcriptional cascade indicated that C/ EBPa and PPARc are upregulated late in the differentiation process by the PBDEs, suggesting that the mechanism of action may be different from that of glucocorticoid. Future studies are required to examine these mechanisms, as the lipophilic and bioaccumulative nature of these chemicals warrant further examination on the effects of adipogenesis and health. Figure S1 Cell viability as assessed by MTT assay. Data is expressed as mean 6 SEM, n = 3. Experiments were performed in triplicate. (TIFF)