DEHP deregulates adipokine levels and impairs fatty acid storage in human SGBS-adipocytes

DEHP is a plasticizer which has been used in plastic products of everyday use for decades. Studies in mice and murine cell culture models identified DEHP as an endocrine disruptor that may also act as an obesogen. As this is of high concern in respect of the worldwide obesity epidemic, our aim is the translation of these findings into a human model system. On the basis of DOHaD, we investigated the influence of an environmentally relevant dose of DEHP [50 µg/ml] on adipogenesis in the human cell culture model SGBS. Pre-adipocytes were exposed to DEHP and differentiated into mature adipocytes. At different stages of differentiation, markers of adipogenesis like GLUT4, FABP4, LPL and PPARs, and of signaling pathways like AMPK/ACC2, JAK/STAT and MAPK were analyzed. Functional markers like adipokine secretion and triglyceride content as well as ROS production were measured in mature adipocytes. We found significantly lower expression levels of adipogenic markers, a reduction in lipid accumulation, higher leptin- and reduced adiponectin levels in the supernatant of treated adipocytes. Moreover, ROS production was significantly elevated after DEHP-exposure. In conclusion, DEHP led to lower grade of adipogenic differentiation in human SGBS-adipocytes under the chosen conditions.

During the last 2-3 decades, environmental pollution and concomitantly also the prevalence of obesity and diseases of the metabolic syndrome increased tremendously 13 . In the OECD (Organisation for Economic Co-operation and Development) area a majority of the population is overweight (BMI: [25][26][27][28][29][30] or obese (BMI: 30 ≥ 40).
In Germany 14% of the adult population is obese (OECD data from 2012), and one in five children (5-17 years) is overweight or obese (OECD data from 2010) 14 . Obesity in turn is a major risk factor for diabetes mellitus type II and cardiovascular diseases. A number of studies have shown a link between environmental pollution with industrial chemicals and diseases of civilization [15][16][17][18] .
On the base of the Developmental-Origins-of-Health-and-Disease-Hypothesis (DOHaD), our aim was to investigate the influence of an early, time-limited and environmentally relevant exposure to DEHP [50 µg/ml] on development, metabolism and function of differentiating pre-adipocytes in a human in vitro model (SGBS). That means, the main question behind this study was, if a time-limited exposure of differentiating SGBS-adipocytes to DEHP (d0-d4), simplifying the situation of the fetus in the womb of the mother, may have an impact on the further differentiation process and with that on the establishment of an obese phenotype in the future. The human Simpson-Golabi-Behmel syndrome (SGBS) pre-adipocyte cell strain is a well characterized in vitro model for studies of human adipocyte biology. The cells are neither transformed nor immortalized, and retain their capacity for adipogenic differentiation up to the 50 generations 19 . To our knowledge, this is the first study investigating the impact of DEHP on adipogenesis in a human in vitro model. A recent study by Kessler and colleagues on the bioavailability of DEHP has revealed that after oral dosage dose-normalized DEHP blood concentrations over time (area under the curve, AUC) are 50 to 100 times higher in humans compared to rats and marmosets, respectively 20 . In regard to the DEHP metabolite mono-(2-ethylhexyl)-phthalate (MEHP) dose-normalized AUCs in blood were 2 to 8 times higher. Because MEHP is assumed to be the active metabolite of DEHP in endocrine disruption 21 , we also tested if DEHP is hydrolyzed to MEHP in our SGBS culture setting. Furthermore, we were also interested in the influence of DEHP, and its downstream metabolite MEHP, on the oxidative stress level in SGBS cells after 4 days of exposure.

Results
DEHP is metabolized to MEHP in SGBS culture. MEHP is assumed to be the biologically active metabolite of DEHP [22][23][24] . For that reason the concentration of MEHP was measured in cell supernatants and cell lysates of SGBS cells exposed to DEHP from d0-d14. MEHP could be detected in both specimen of treated and untreated SGBS cells at d4 and d14 of adipogenesis (Fig. 1). MEHP was significantly increased in all treated samples compared to the corresponding DMSO controls. Furthermore, a significant 4-5 times higher concentration of MEHP was found, comparing DEHP treated supernatants and cell lysates of d4 to d14. In cell supernatants of the DMSO control, a significant reduction of MEHP concentration was found comparing d4 to d14. Nevertheless, in the cell supernatant MEHP concentrations in DMSO controls were very low with 3-8 ng/ml compared to 50-250 ng/ ml in DEHP treated samples, and 1-1.3 ng/ml versus 1.7-15 ng/ml in the cell lysate samples. We were also able to detect oxidized metabolites (5OH-MEHP, 5oxo-MEHP and 5cx-MEPP) of DEHP/MEHP in DEHP treated supernatants and cell lysates, but only at concentration levels roughly a factor of 100 lower than MEHP (data not shown).

DEHP does not affect proliferation of SGBS cells.
DEHP has been shown to alter the proliferation rate of different cell types [25][26][27] . To answer the question if DEHP affects proliferation of the SGBS cells, undifferentiated, proliferating cells were exposed to DEHP for a period of 24 h in a non-confluent stage. After this treatment, there was no significant change in the amount of the proliferating cell nuclear antigen (PCNA) compared to the corresponding DMSO controls (Supplementary Figure S1). The SGBS cells were exposed to DEHP [50 µg/ml] for the whole period of adipogenic differentiation (d0-d14). Samples were taken at d4 and d14 for analysis by HPLC. Statistics: a = DEHP versus corresponding DMSO control; b = DEHP d4 versus DEHP d14; c = DMSO d4 versus DMSO d14; ANOVA (Student-Newman-Keuls Method), significance: p > 0.05; N = 4, n = 1 (4 pooled wells each).
SCIEntIFIC REPoRtS | (2018) 8:3447 | DOI:10.1038/s41598-018-21800-4 DEHP leads to an accumulation of reactive oxygen species (ROS). It has been shown, that DEHP causes oxidative stress in different model systems by elevating ROS levels [28][29][30] . Though, ROS are not only a stress factor but also crucial for the fate of mesenchymal stem cells (MSC). They are multipotent cells that may differentiate into the osteogenic lineage, neuronal cells or into adipocytes. Among other things, differentiation into adipocytes needs a certain level of ROS, while under lower ROS conditions MSCs are more likely to differentiate into the osteogenic lineage (intensively reviewed by Atashi et al. 31 ). For those reasons, a H2DCFDA-assay was performed at d4 of adipogenesis, i.e. directly after 4 days of DEHP exposure. H 2 O 2 and an untreated sample were used as assay control. After DEHP exposure, the ROS level was significantly elevated compared to the corresponding DMSO control and the untreated, but not to the H 2 O 2 assay control (control: 137.49 ± 2.92; DMSO: 139.30 ± 3,59; DEHP: 171.37 ± 7.52; H 2 0 2 : 164.06 ± 3.76) ( Fig. 2A). To further investigate the effects of DEHP on oxidative stress, the ROS-detoxifying superoxide dismutase 2 (SOD2) and the enzyme glutathione peroxidase 1 (GPX1) were analyzed by western blot. The amount of both enzymes was not affected by DEHP at d4 and d8 of adipogenesis (Supplementary Figure S2). DEHP decreases the triglyceride content of SGBS-adipocytes. DEHP is known as an endocrine disruptor with obesogenic effects, mainly investigated in rodent models. To investigate this hypothesis in a human adipogenesis model, SGBS cells were exposed to DEHP in the induction phase from d0-d4. DEHP treatment caused a highly significant reduction of the triglyceride content in SGBS-adipocytes at d8 (DEHP 0.219 ± 0.0096 nmol/ng protein versus DMSO 0.414 ± 0.0246 nmol/ng protein) (Fig. 3A). This effect was also visible after Oil Red O staining. DEHP treated SGBS-adipocytes at d8 showed a lower accumulation of lipid droplets, indicated by less Oil Red O staining compared to the corresponding controls (Fig. 3B).
DEHP affects adipokine secretion. Adipokines like leptin and adiponectin are key markers of adipocyte function and metabolism. Given that DEHP led to a significant reduction in lipid droplets and triglyceride content, the potential mechanisms underlying this observation were further analyzed in SGBS cells exposed to Figure 2. Analysis of ROS-production after DEHP-exposure: SGBS cells were exposed to DEHP from d0-d4 and subsequently differentiated into adipocytes. For the ROS-measurement (A), the medium was changed to DMEM/F12 with 10 µM H2DCFDA at d4 and incubated for 30 min. As the positive assay control H 2 O 2 (25 µM) was used, as well as an untreated control. Statistics: a = DEHP/H 2 O 2 versus control; b = DEHP/H 2 O 2 versus DMSO control; ANOVA (Student-Newman-Keuls Method); N = 2, n = 9. DEHP from d0-d4, during the induction of adipogenesis. At d8 of adipogenesis, a significant decrease in adiponectin secretion (DEHP: 168.84 ± 13.93 ng/ml versus DMSO: 231.21 ± 11.14 ng/ml) was found by ELISA. This decrease in adiponectin levels was accompanied by a significant increase in leptin secretion (DEHP: 1.58 ± 0,09 ng/ml versus DMSO: 1.26 ± 0,09 ng/ml) (Fig. 4A,B). On the basis of these findings, the paracrine effects on the adipocytes were determined by measuring the mRNA expression of the adiponectin receptor 2 (ADIPOR2) and the leptin receptor (LEPR). The ADIPOR2 expression showed no changes, while the LEPR expression was significantly reduced at d4 (DEHP: 43.33 DEHP partially alters adipokine signaling. The phosphorylation of AMPK (AMP-activated protein kinase), ACC2 (acetyl-CoA carboxylase), ERK1 and ERK2 (extracellular-signal regulated kinases) and STAT3alpha (signal transducer and activator of transcription 3) is presented as a ratio of phosphorylated to non-phosphorylated protein comparing DMSO controls and treated samples. Single proteins like SOCS3 (suppressor of cytokine signaling 3) are expressed as relative amount of protein versus β-actin. The ratio of pAMPK/ AMPK was not altered by DEHP-exposure, whereas the ratio of pACC2/ACC2 was significantly increased at d8 (DEHP: 69.27% ± 4.80 versus DMSO: 45.86% ± 3.88) (Fig. 5A,B). If pACC2 is analyzed singly, it was significantly reduced at d8 (DEHP: 269.04% ± 22.28 versus DMSO: 338.13% ± 13.45) (Fig. 5C). Also the amount of ACC2 was significantly reduced a d4 and d8 (d4: DEHP: 176.8 2% ± 18.48 versus DMSO: 306.89% ± 28.09; d8: DEHP: 390.32% ± 25.97 versus DMSO: 760.12% ± 91.26) (Fig. 5D). Furthermore the mRNA expression of ACC2 showed a significant decrease at d4 (DEHP: 65.76 ± 3.45 mol./mol. TBP versus DMSO: 93.32 ± 8.22 mol./mol. TBP) after DEHP treatment compared to the corresponding DMSO control (Fig. 5E). However, the amount of the ACC2 downstream target CPT1 (carnitine palmitoyltransferase I) was not altered (Fig. 5F). The analysis of the activation of the JAK/STAT pathway showed no alterations in the ratio of pSTAT3/STAT3, nor in the amount of SOCS3 (Fig. 6A,B). Besides the JAK/STAT pathway, leptin is also known to activate the MAPK (ERK1/2) pathway via its receptor (LepRb) 32,33 . Moreover, both ERK isoforms play distinct but important roles during adipogenesis 34,35 . However, the phosphorylation of ERK2 was significantly reduced at d8 (DEHP: 86.37% ± 19.62 vs. DMSO: 152.27% ± 20.58) with ERK1 showing a strong tendency to a decrease at d8 (p = 0.08), too (DEHP: 151.05% ± 62.54 versus DMSO: 559.92% ± 183.38) (Fig. 6C,D).    (Fig. 7). Also the mRNA expression of GLUT4 (glucose transporter 4) was unaffected by DEHP treatment (Fig. 8A). Immunohistochemical staining against GLUT4 showed that it was mainly localized in the cytoplasm of the adipocytes of the DMSO controls and also the DEHP treatments ( Fig. 9). Nevertheless, the mRNA expression of CD36 (fatty acid translocase), FABP4 (fatty acid binding protein 4), LPL (lipoprotein lipase) LIPE (lipase E, hormone sensitive type) as well as ATGL (adipose triglyceride lipase) were significantly decreased after DEHP-exposure if compared to the corresponding DMSO controls (Fig. 8B-F    Direct exposure with DEHP leads to a higher amount of free glycerol. Lipolysis requires three specific enzymes that hydrolyze ester bonds in triglycerides in a three-step process. Those enzymes are ATGL, LIPE, and MGL (monoglyceride lipase). Free fatty acids and free glycerol are the end products of lipolysis and can directly be measured in the supernatant of cells or tissue cultures by commercial assay kits (Cultured Human Adipocyte Lipolysis Assay Kit, Zenbio). To measure the impact of DEHP exposure on lipolysis in SGBS cells, three different treatments have been tested: treatment a: exposure from d0-d4 and sampling at d4; treatment b: exposure at d4 for 3 h and sampling at d4 (as recommended for substance testing in the kit manual); treatment c: exposure from d0-d4 and sampling at d8. The amount of free glycerol was significantly elevated after DEHP exposure in the treatment b, the direct and short-term exposure scenario, whereas the longterm exposure (d0-d4) in treatment a and treatment c did not show any changes (Fig. 10).

Discussion
On the basis of several studies showing an obesogenic effect of DEHP in murine model systems [36][37][38] and in human epidemiological studies 39,40 , we tested this hypothesis in a broader scope in the human SGBS fat cell model. In the present study we demonstrate that the plasticizer DEHP (or its downstream metabolite MEHP) significantly elevated leptin levels and interfered with fatty acid metabolism and lipid storage in a human pre-adipocyte cell model (SGBS). So far, analyses on DEHP-exposure and obesity have mostly been studied in rodent models in-vivo [36][37][38] and murine cell lines like 3T3-L1 in-vitro 41,42 . To date, there are just a few studies that analyzed DEHP/ MEHP effects on human adipocyte cell models. For instance, a publication by Campioli and colleagues investigated the impact of MEHP on the relationship between the testicular translocator protein (TSPO) and the PPARs in a human liposarcoma cell line (SW 872). They found an overall induction of adipogenic differentiation after Figure 9. Measurement of free glycerol in SGBS derived adipocytes: The cells have been differentiated as described before and have been treated with DEHP and DMSO as control as followed: treatment a: exposure from d0-d4 and sampling at d4; treatment b: exposure at d4 for 3h and sampling at d4; treatment c: exposure from d0-d4 and sampling at d8. Statistics: Student's t-test, N = 4, n = 3. Figure 10. Immunohistochemistry for GLUT4 in SGBS-adipocytes: SGBS cells were exposed to DEHP from d0-d4 and subsequently differentiated into adipocytes. For the analysis of GLUT4 localization within the adipocytes, the cells have been fixed with PFA at d8, stained for GLUT4 (HRP-conjugated antibody -brown staining) and counterstained with hematoxylin (blue staining). The figure shows representative micrographs at × 200 magnification of DEHP treated adipocytes and DMSO controls, as well as the negative control without GLUT4 but secondary antibody. The scale bar indicates 50 µm.
SCIEntIFIC REPoRtS | (2018) 8:3447 | DOI:10.1038/s41598-018-21800-4 4 days MEHP-treatment with a significant reduction of PPARg mRNA expression, while it had no effect on triglyceride content. However, this study cannot be compared to our study, because it (i) used a tumor cell line (ii) the cells are cultured in media containing 10% undefined fetal calf serum (iii) the cells are treated with MEHP directly (this is possible due to the FCS used in the media) (iiii) the cells are additionally treated with PMA 43 . Another publication by Ellero-Simatos and colleagues investigated the influence of MEHP on primary cultures of human pre-adipocytes from the subcutaneous adipose tissue fraction. They treated differentiated adipocytes on day 11 with the relatively high concentration of 100 µM MEHP, for 24 and 48 hours, and performed a combined transcriptomic -H NMR metabonomic analysis. The MEHP-treatment led to an activation of the PPAR signaling pathway and other processes involved in lipid metabolism 44 . Although all together excellent studies, the multitude of differences in e.g. cell culture models, culture protocols and study design (among others: primary adipocytes versus SW 872 liposarcoma cells versus SGBS, FCS versus no serum in the media, MEHP-versus DEHP-exposure during different stages of adipogenesis, different duration of exposure) exacerbates a discussion of data among studies. To partly circumvent this fact, the current study design -DEHP-exposure during the induction phase (d0-d4) -was based on data by Biemann and colleagues, who tested 4 different exposure intervals (mixed exposures and single exposures with DEHP, TBT and BPA) in the murine MSC line C3H10T1/2 concerning their impact on different adipocyte endpoints 10,45 . The most effective interval for DEHP-exposure was the induction phase with higher adipocyte numbers and triglyceride levels. As it is known, that mice and men differ especially in regard to the PPAR-family in their ability to be activated by MEHP between isotypes and cell lines 41,42,46 , it is of high importance to transfer murine studies into human models. Using SGBS cells, we could not detect any changes in the amount of PPARa and PPARg. This may be due to several reasons. A likely one are the culture conditions, with rosiglitazone added to the medium between d0 and d4. While on the one hand rosiglitazone is a PPARg agonist and may lead to a maximum in expression and activity and by that masks any possible effect of DEHP, it is known on the other hand that humans have a lower sensitivity towards peroxisome proliferators than rodents 47,48 . Furthermore, we have used a serum-free medium and no fatty acids have been present. Nevertheless, mRNA expression or protein amount does not equal receptor activity. For that reason transactivation assays could reveal actual PPAR activation through DEHP/MEHP in SGBS-adipocytes in future experiments.
Further on, we tested if DEHP-exposure had an influence on proliferation, but could not detect any, which was contradictory to results seen before in different cell types 27,49,50 . Having a closer look at adipocyte morphology, we could show a clearly lower accumulation of lipid droplets by Oil Red O staining and a significant decrease of triglyceride content after DEHP-exposure. This is in high concordance with data from Klöting and colleagues: They exposed mature 3T3-L1 adipocytes to DEHP and found a lower lipid content. Furthermore, they fed obesity-resistant inbred mice with DEHP and found lower circulating adiponectin levels and adiponectin protein in subcutaneous adipose tissue as well as in 3T3-L1 adipocytes 27,49,50 . This was also true in our SGBS-adipocytes at d8 with a significant reduction in adiponectin levels in the cell supernatant of DEHP-treated adipocytes, accompanied by a significant increase of leptin. In principle, a higher leptin level goes together with a higher fat mass, which was the opposite in the present model. But a study by Harris and colleagues supports the present findings by demonstrating, that administration of leptin to leptin-responsive animals reduces body fat mass. They concluded, that leptin inhibits the accumulation of lipids in adipocytes by increasing the turnover of triglycerides, inhibiting basal and insulin-stimulated de novo lipogenesis, but stimulating oxidation of glucose and free fatty acids 51 . Furthermore, Frühbeck and others showed that leptin has an autocrine-paracrine lipolytic effect on isolated murine adipocytes [52][53][54][55] . However, leptin is also known to phosphorylate and activate AMPK via its receptor 56 . Activated AMPK in turn phosphorylates and deactivates ACC2, the key enzyme of lipogenesis, which leads to an inhibition of malonyl-CoA formation from acetyl-CoA. In the absence of malonyl-CoA CPT1 is active and supports ß-oxidation 57 . In our study we found a significantly higher ratio of pAAC2/ACC2 after DEHP-exposure, with no changes in AMPK phosphorylation or CPT1 expression. Interestingly, if analyzed singly, the amount of ACC2 and pACC2 was significantly reduced compared to DMSO, as was the mRNA expression. The LEPR was significantly down-regulated at d4 with a similar tendency at d8, probably due to a negative feedback mechanism caused by the elevated leptin level. Unfortunately, the paracrine functions of leptin on the adipocyte itself and the LEPR are not well understood. Further on, we analyzed the phosphorylation of STAT3 as a commonly used marker of leptin receptor activation 58 and SOCS3, a silencer for STAT3, revealing no significant changes in their protein amount. The MAPK (ERK1/2)-pathway is known to be activated by leptin via its receptor and is important for general adipogenesis [32][33][34][35] . As Bost and colleagues described different functions of the ERK isoforms in adipogenesis, we analyzed the phosphorylation of ERK1 and ERK2 separately. We found that ERK2 phosphorylation was significantly reduced at d8 with a strong trend of a decreased phosphorylation of ERK1 at d8, too. This result seems not to be very conclusive if one thinks of the elevated leptin levels in the present study. But recent literature on how inhibition of ERK phosphorylation is associated with sevoflurane toxicity in the developing murine brain demonstrated that elevated ROS levels lead to a decreased ERK phosphorylation 59 . As ROS levels were significantly elevated after DEHP-exposure in the present study, this is considered a good reason for the reduction in ERK phosphorylation. However, the mechanisms behind remain unclear and need further investigations. In turn, ERK is responsible for the phosphorylation of Ser660 of LIPE, which together with PKA-activity leads to an activation of LIPE-induced lipolysis 60 . LIPE is a hormone-sensitive lipase that hydrolyzes stored triglycerides from lipid droplets to free fatty acids. The expression of LIPE has been shown to be positively correlated with the expression of adiponectin in obese women 61 . Although we measured a significantly lower amount of triglycerides in SGSB adipocytes after DEHP-treatment, we also observed a significantly decreased adiponectin level accompanied by a significant reduction in LIPE expression. Quantitatively, LIPE and ATGL are the most important lipases in the process of lipolysis and by that regulate the catabolism of lipid droplets (intensively reviewed by Zechner et al. and Frühbeck et al. 62,63 ). In the current study the expression of both enzymes of lipolysis (LIPE and ATGL) was downregulated at d8. Nevertheless, there was a significantly lowered amount of triglycerides after DEHP-exposure in SGBS-adipocytes of d8. Therefore, a direct measurement of lipolysis was conducted to improve the interpretation of these results, as mRNA expression does not necessarily reflect the activity of an enzyme. Yet, studies in mice observed a correlation between mRNA expression of LIPE and ATGL with the grade of lipolysis 62 , while others showed that in fasted human white adipose tissue (WAT) ATGL expression is reduced with increased ATGL protein concentration and increased lipase activity. According to Schweiger and co-workers, the reason for that is posttranscriptional regulation 64 . This correlates with our observation of a significantly elevated amount of free glycerol 3 hours after a direct exposure to DEHP, significantly reduced expression levels of ATGL and LIPE and a reduced triglyceride accumulation in lipid droplets. The finding that no changes were seen in free glycerol levels in treatment a and c (DEHP exposure: d0-d4) may be probably due to the fact, that lipolysis has taken place at the beginning of exposure and was finished due to a lack of TGs and changed metabolic conditions at the time point of measurement at d4 and d8. This needs to be further validated in future studies. As a reduced amount of triglycerides could also be due to a reduced glucose or fatty acid uptake, we also investigated the expression of FABP4, CD36, LPL, and GLUT4. The cytoplasmic protein FABP4 binds long-chain fatty acids and other hydrophobic ligands and is involved in fatty acid uptake, transport, and metabolism. Interestingly, FABP4 bound to fatty acids physically interacts with phosphorylated LIPE and by that can form a FABP4:pLIPE complex at the lipid droplet to facilitate LIPE-dependent lipolysis 63,65 . In our studies, both, LIPE and FABP4, were significantly down-regulated after DEHP-exposure. Moreover, our results on FABP4 and also leptin expression are in concordance with a study by Gan and colleagues who found that FABP4 was down-regulated by leptin assuming FABP4 and leptin to play opposite roles in the regulation of fatty acid oxidation 66 . The fatty acid translocase CD36 is another key player in adipose function that also interacts with FABP4 [67][68][69] . According to Coburn and others, CD36 deficiency is associated with an impaired fatty acid uptake and an increased basal lipolysis 63,70 . In our studies the expression of the fatty acid translocase CD36 and FABP4 was significantly decreased after DEHP exposure, pointing towards less fatty acid uptake and transport. This is further supported by the significantly down-regulated expression of LPL in our study. LPL functions as triglyceride hydrolase and is responsible for receptor-mediated lipoprotein uptake. In studies with ob/ob mice leptin decreased LPL-activity 71 which is in accordance with our findings. GLUT4 is a protein that functions as an insulin-regulated facilitative glucose transporter. Within minutes of insulin stimulation, the protein translocates to the cell surface and transports glucose across the cell membrane. DEHP did not change the mRNA expression of GLUT4, nor did it alter the localization of GLUT4 if compared to the control.
A likely candidate for DEHP-interference with adipogenesis are reactive oxygen species (ROS). It has been shown before that DEHP causes oxidative stress 25,28,29 and that it plays a role in adipogenic differentiation 72,73 . Moreover, ROS are associated with a decreased secretion of adiponectin in 3T3-L1 cells 74 and also in human studies 75 , which is in line with our findings. Furthermore, as mentioned above, ROS seem to lead to a decreased ERK phosphorylation 59 and thus may influence adipogenesis in general. In the present study, the H2DCFDA-assay revealed a significant elevation of the ROS level at d4 in DEHP-treated SGBS cells and a significant decrease in adiponectin in the cell supernatant. However, no changes could be observed in the protein amounts of the ROS-detoxifying enzymes GPX1 and SOD2. Wang and colleagues had similar results, when they investigated the influence of DEHP [1, 10 and 100 µg/ml] on cultured antral follicles. Although they measured a significant elevation in ROS levels, they could not detect any changes in GPX1 and CAT expression, but a decrease in SOD1 expression in the 10 µg/ml dose group. However, the question why in the present study no changes could be observed at least in the amount of SOD2 remains unanswered.

Conclusion
We could not confirm an obesogenic effect of DEHP in SGBS-adipocytes under the chosen conditions. In contrast, our results clearly indicate that DEHP-exposure during the induction phase of adipogenesis led to a significantly lower degree of adipogenic differentiation, a higher lipolytic activity and a reduced uptake and transport of fatty acids into and within the adipocytes. Besides this, it is likely that ROS play a more important role in the effects of DEHP than so far expected. Furthermore, we showed that there are species differences between human and murine models that must be considered when translating murine data into human context.

Material and Methods
Culture conditions and DEHP-exposure. The SGBS pre-adipocyte cell strain was obtained from the laboratory of Prof. Martin Wabitsch et al. 76 . SGBS pre-adipocytes were grown to near confluence and then incubated (d0) in a serum-free differentiation medium [2 μmol/l rosiglitazone (Cayman #714740), 25 nmol/l dexamethasone (Sigma Aldrich #D-1756), 0.5 mmol/l methylisobuthylxantine (Sigma Aldrich #I-5879), 0.1 μmol/l cortisol (Sigma Aldrich #H-0888), 0.01 mg/ml transferrin (Sigma Aldrich #T-2252), 0.2 nmol/l triiodothyronine (Sigma Aldrich #T-6397), and 20 nmol/l human insulin (Sigma Aldrich #19278)]. After 4 days the medium was changed, and the cells were further cultured in medium supplemented with 0.1 μmol/l cortisol, 0.01 mg/ml transferrin, 0.2 nmol/l triiodothyronine, and 20 nmol/l human insulin. The SGBS pre-adipocytes were exposed to DEHP [50 µg/ ml] (Sigma Aldrich) from d0-d4. DEHP had been solved in DMSO (Sigma Aldrich) in a 1000-fold stock solution. The maximum concentration of DMSO in the culture media was 0.1%. Controls were run as vehicle controls with 0.1% DMSO in the media. The dose used is comparable to DEHP doses found in neonates undergoing clinical procedures, such as transfusion or extracorporeal membrane oxygenation as well as in whole blood and blood components [77][78][79] . All experiments have been finalized at d8 of differentiation.

HPLC analysis of MEHP in supernatants and cell lysates. MEHP is negatively charged and by that
cannot be taken up into the cells in an in vitro model cultured without fetal calf serum (FCS) 80,81 . Moreover, the systemic availability of DEHP is 50-100 times higher in humans than in marmosets and rodents 20 . Therefore we used DEHP as the primary agent and tested whether DEHP is metabolized to the active metabolite MEHP in SGBS cells. Only in the case of the HPLC-analysis of DEHP-metabolisation to MEHP, SGBS cells were cultured until d14 and exposed to DEHP from d0 to d14 of adipogenic differentiation. The data obtained from this part of the study was basic for the whole study design, as it aimed to reveal DEHP-degradation in the SGBS cell model in general and the grade of accumulation of e.g. MEHP in the cells and in the media over time. Due to that, samples of media and cell lysates were obtained at d4 and d14. The assays used to determine the DEHP metabolite MEHP (and other oxidized DEHP metabolites) were performed in the medium and cell lysates using high-performance liquid chromatography tandem mass spectrometry with quantification via isotope dilution (Koch et al., 2003). The selectivity and specificity of such an HPLC-MS/MS method (including isotope labelled internal standards) can be considered as a gold standard for a robust, sensitive and specific determination of such biomarkers. The respective LOQs for MEHP (and other phthalate metabolites) are 0.5 µg/l for MEHP and 0.2 µg/l for the oxidized metabolites. The extensive chromatographic separation and the specific mass transitions ensure that "cross-reactivity" (in LC-MS/MS better termed interference) with or by other phthalates or phthalate metabolites can be excluded. Sample measurements were carried out by the IPA laboratory in Bochum (Germany), which has successfully participated as a reference laboratory for phthalate metabolite analyses in the quality assurance programme of the European Union financed Consortium to Perform Human Biomonitoring on a European Scale (COPHES) 82 .

Analysis of cell proliferation by Proliferating-Cell-Nuclear-Antigen (PCNA) western blot. SGBS
cells were seeded on 12-well plates and cultured in basal medium 19 for 2 days. Afterwards the medium was changed, now containing 50 µg/ml DEHP for 24 h. The exposed SGBS cells were harvested with RIPA buffer [10% 10 X PBS, Nonidet NP40, 10% SDS, 0.5% sodiumdeoxycholate, protease inhibitors, phosphatase inhibitors (Roche, Germany)] and the protein was used for western blot analysis with the PCNA antibody 83 .

Quantitative real time PCR (qRT-PCR). Total RNA of SGBS cells has been purified with the RNeasy Lipid
Tissue Mini Kit (Qiagen) and was quantified by UV/VIS-spectrometry at 260 nm using the NanoVue system (GE Healthcare). The reverse transcription reaction was performed with 1 µg of total RNA using the RevertAid ™ H Minus Reverse Transcriptase (Fermentas, Germany), following the instructions of the Fermentas manual. The samples were analyzed by qRT-PCR using an iQ5 Optical System (Bio-Rad Laboratories, Herts, UK) and SYBR green master mix with fluorescein (Eurogentec, Germany) as an intercalating fluorescent dye specific for double stranded DNA. Primer sequences are listed in Table 1. All primers used were tested for specificity by sequencing of the PCR products. Absolute mRNA expression was calculated by target specific plasmid standards using serial dilutions. Each assay was run with duplicates of each cDNA sample as well as a no template control (NTC) and the target specific standard which also served as a positive control. The expression of the housekeeping gene TATA-box binding protein (TBP) was used to normalize samples for the amount of cDNA used per reaction. To confirm the amplification of only one single target, the resulting qRT-PCR products were analyzed by a melting curve. If possible, exon-spanning primers have been used. For   cells have been seeded into black 24-well plates and exposed and differentiated into adipocytes as described before (2,1.). At d4 and d11 the medium was changed to DMEM/F12 (Gibco ® # 31330-038) with 10 µM H 2 DCFDA (Thermo Fisher Scientific), dissolved in DMSO and incubated for 30 min at 37 °C in the absence of light. Afterwards the cells have been washed with medium. Measurement was carried out with fresh media in a plate reader at Ex/Em: 495 nm/535 nm. The dye is oxidized in the cells and through that gives a detectable fluorescence signal. As the positive assay control H 2 O 2 (25 µM) was used as well as an untreated control.

Western blot analysis.
Immunohistochemistry. The glucose transporter 4 is a transporter that translocates to the membrane, if activated by insulin. Without insulin stimulation, GLUT4 is inactive and located in the cytoplasm of the adipocyte. By immunohistochemical staining we aimed to analyze the localization of GLUT4 after DEHP-exposure. For the culture of the SGBS cells, 4-well chamber slides (Sarstedt) were used. The cells were seeded in a density of 0.5 × 10 4 per well. DEHP-exposure and differentiation were as described before (2). At d8 the SGBS-adipocytes were fixed with 4% paraformaldehyde for 25 min followed by three washing steps with PBS and one washing step with PBS + Tween 20 [0.1%] (PBST) for 5 min. Afterwards the endogenous peroxidases were inactivated with 3% H 2 O 2 in methanol for 20 min at room temperature in the absence of light. This was again followed by washing steps with H2O (2x , 5 min) and PBST (2x , 10 min). Afterwards the adipocytes were blocked with 10% goat serum for 60 min at room temperature. Next, the primary antibody (GLUT4, 1:300, 4048, Sigma Aldrich) in 3% (wt/vol) bovine serum albumin (BSA)/PBS was added and incubated overnight at 4 °C. The cells were washed three times with PBST and incubated with the secondary antibody Dako EnVision + System-HRP labeled Polymere anti mouse (1:1 in PBS). Diaminobenzidine (DAB; WAK-Chemie Medical) was used for detection. The DAB reaction was stopped in water after 10 min. The cells were counterstained with hematoxylin and subsequently embedded on Superfrost slides (Menzel Gläser) using 4.8 g of Mowiol reagent (Calbiochem) dissolved in 12.0 g of glycerol (Merck). All steps were performed within the same experiment, examined microscopically during the same session, using identical microscope and camera settings (BZ 8100, Keyence).
Statistical Analyses. All data obtained from qRT-PCR and western blotting were run in at least four independent biological experiments (N ≥ 4), and presented as mean ± standard error of the mean (SEM). Four wells have been pooled within one biological experiment (N = 1, n = 1 (4 pooled wells)) to obtain sufficient material for mRNA or protein extraction. For ROS-detection and the adipogenesis assay the independent biological experiments were run two times (N = 2) with at least 6 individual replicates (n = 6) per experiment (N). Experiments for the adipokine ELISAs were run in 6 independent biological replicates (N = 6) with 1 individual replicate (n = 1). In the case of qRT-PCR, ELISA, H2DCFDA-and adipogenesis assay the samples have been measured in two technical replicates and the mean has been calculated and used for statistical analyses. For statistical analysis Student's t-test or, if necessary, the Wilcoxon rank-sum test was performed. In the case of multiple comparison procedures, the one-way analysis of variance (ANOVA) or, if necessary, the Student-Newman-Keuls Method was performed. The tests were done using Sigma Plot software and are mentioned in figure legends, too. The datasets generated during and/or analysed in the current study are available from the corresponding author on reasonable request.