Inhibition of p38MAPK Increases Adipogenesis From Embryonic to Adult Stages

Mice were housed on a 12-h light/dark schedule and had free access to water and food. High-fat (45% fat, 35% carbohydrates, and 20% proteins) and standard diets (10% fat, 70% carbohydrates, and 20% proteins) were purchased from SAFE (Epinay/Orge, France). The experiments were conducted following the standard ethical guidelines (European Union guidelines on animal laboratory care) and approved by the ethical committee of the Nice School of Medicine.

Mouse embryonic fibroblasts (MEFs) were prepared on day 14 postcoitum as described previously (16). Preadipocytes from adult tissue were isolated from inguinal and axillary fat pads of 8-week-old mice (17). Two days postconfluent, cells were treated for 48 h with the adipocyte differentiation cocktail containing 85 nmol/l insulin, 0.25 μmol/l dexamethasone, and 0.5 mmol/l 3-isobutyl-1-methylxanthine. The medium was then replaced for 6 days by Dulbecco’s modified Eagle’s medium with insulin only. Oil red O staining was carried out as described by Dani et al. (18). To perform the quantification of triglyceride accumulation, oil red O was extracted with isopropanol, and the optical density was measured at 490 nm.

The MEFs knockout for p38α isoform gene and their wild-type counterparts were obtained from Dr. M. Karin (San Diego, CA). Knockout MEFs and 3T3-L1 preadipocyte cells were induced to differentiate as described for MEFs.

Mouse enbryonic stem cells CGR8 (19) were grown on gelatin-coated plates and maintained in Dulbecco’s modified Eagle’s medium (Cambrex, Verviers, Belgium), supplemented with nonessential amino acids, 2 mmol/l l-glutamine, 100 μg/ml streptomycin, 100 units/ml penicillin, 1 mmol/l sodium pyruvate, 10% (vol/vol) fetal bovine serum (Invitrogen, Grand Island, NY) and 143 μmol/l β-mercaptoethanol (Sigma, St. Louis, MO), and 100 units/ml of leukemia inhibitory factor. For differentiation, cells were maintained in culture media without leukemia inhibitory factor, as described previously (20). Briefly, this protocol included two key steps: 1) between days 0 and 7, embryoid bodies were formed in suspension and treated daily between days 2 and 5 with 0.1 μmol/l all-trans retinoic acid (Sigma) and 2) from day 7 to the end, cells were seeded onto Petri dishes and treated with 85 nmol/l insulin (Lilly France SAS, Suresnes, France) and 2 nmol/l 3,3′,5-tri-iodotyronine (Sigma). For p38MAPK inhibition, 10 μmol/l PD169316 (Calbiochem, Darmstadt, Germany) were added to the culture media, which was renewed everyday from day 7 to the end of the differentiation protocol. Adipocyte colonies were analyzed 21 days after the formation of embryoid bodies. Area of adipocyte outgrowths was measured and normalized to the area of each embryoid body using IP Lab software (Image J 1.29x; National Institutes of Health, Bethesda, MD). The data were expressed as a histogram showing the percentage of embryoid bodies with 0–10%, 0–20%, 20–40%, 40–60%, 60–80%, or 80–100% of their area forming adipocytes.

The DBAp38C69 embryoid body cell line (obtained from Dr C.A. Gabel, Pfizer Laboratory, Groton, CT) knockout for the p38α isoform gene and the DBA-252 wild-type cells were induced to differentiate according to the same protocol. Teratocarcinomas were induced by subcutaneous injections of 5 × 10⁶ cells of each cell line in DBA1/LacJ mice (21).

Tissue extracts were prepared using lysis buffer (20) and were analyzed by Western blotting using antibodies against p38MAPK phosphorylated on Thr180/Tyr182, p38MAPK, JNK phosphorylated on Thr183/Tyr185, C/EBPβ phosphorylated on Thr 188, and C/EBPβ (Cell Signaling Technology, Danvers, MA).

Gene expression analysis was performed with the ABI Prism 7000 (Applied Biosystem) and SYBR green reagents (Eurogentec, Seraing, Belgium). cDNAs were synthesized from 2 μg total RNA using Superscript II reverse transcriptase (Invitrogen). Primers set were designed according to the manufacturer software. Samples contained 1 × SYBR green master mix, 0.5 μmol/l primers, and synthesized cDNA in a 25-μl volume. The PCR conditions were as follows: 10 min at 95°C, then 40 cycles of 15 s at 94°C, 30 s at 60°C, and 1 min at 72°C. We used 36B4 as an internal control. Primer sequences are as follows: CEBP/α forward 5′ GACCATTAGCCTTGTGTGTACTGTATG 3′, reverse 5′ TGGATCGATTGTGCTTCAAGTT 3′; aP2: forward 5′ TTCGATGAAATCACCGCAGA 3′, reverse 5′ GGTCGACTTTCCATCCCACT 3′; PPARγ: forward 5′ CTGTTTTATGCTGTTATGGGTGAAA 3′, reverse 5′ CTGTTTTATGCTGTTATGGGTGAAA 3′; leptin: forward 5′ GACACCAAAACCCTCAT 3′, reverse 5′ CAGTGTCTGGTCCATCT 3′; adiponectin: forward 5′ TCATGCCGAAGATGACGTTACT 3′, reverse 5′ CCATCCAACCTGCACAAGTTC 3′; 36B4: forward 5′ TCCAGGCTTTGGGCATCA 3′, reverse 5′ CTTTATCAGCTGCACATCACTCAGA 3′.

Total RNA was prepared using Trizol reagent (Invitrogen, Carlsbad, CA). For Northern blot analysis, 10–15 μg total RNA per lane were blotted onto Hybond-N nylon membrane (Amersham Biosciences, Buckinghamshire, U.K.), cross-linked with UV, and hybridized with radiolabeled DNA probes. Hybridization signals were quantified using a Molecular Dynamics radioimager and normalized to S26 signal.

MEFs were transfected using Superfect reagent (Qiagen, Valencia, CA) with 1 μg of each of the various plasmids, according to the manufacturer’s instructions. Forty-eight hours after transfection, in presence or absence of thiazolidinedione, cells were lysed and luciferase activity was assessed as previously described (16). The promoter driving luciferase is composed of the minimal thymidine kinase promoter and of three copies of the PPARγ responsive element (PPRE)-binding site; therefore, the luciferase activity of this construct is a reliable marker of PPARγ activity. PPARγ cDNA was under the control of β-actin gene promoter (CAG-PPARγ construct) (22), and PD169316 did not interfere with its expression.

To determine whether p38MAPK activity is altered during development of white adipose tissue, we examined its activity in different fat depots of dietary (high-fat diet [HFD]) and genetically ob/ob models of obesity compared with their respective lean control mice (Fig. 1A). p38MAPK protein level was not significantly different in the four groups of mice. However p38MAPK activity depended on the fat depot location, being from 2.5- to 10-fold higher in inguinal and axillary (I + A) depots than in the epididymal and perirenal depots (Fig. 1A, white bars). Strikingly, whereas obesity did not significantly affect p38MAPK activity in epididymal and perirenal tissues, we observed an 80% decrease in p38MAPK activity in inguinal and axillary depots in HFD (obese-HFD) and ob/ob obese mice compared with lean animals (Fig. 1A, black bars). By contrast, in the same tissues, we observed a significant increase in JNK activity in response to HFD (online appendix [available at http://diabetes.diabetesjournals.org]), as previously described (23). These results suggest that p38MAPK activity specifically decreased in white adipose tissue of obese mice.

Little is known concerning signaling pathways implicated in vivo in the recruitment of new adipocytes from precursor cells (preadipocytes). To determine whether the p38MAPK pathway is involved in recruitment of new adipocytes, we studied p38MAPK activity in adipocytes and preadipocytes isolated from white adipose tissue of lean and obese animals. Tissues were dissected and dissociated with collagenase. After centrifugation, floating adipocytes were collected, and the pellet containing the stroma vascular cells was resuspended and filtrated to separate and isolate the preadipocytes from the other cells (17). Following isolation, cells were incubated for 24 h in medium with 10% serum in order to overcome the immediate stress response due to the collagenase treatment (24), and cell lysates were prepared. While we noticed no difference in p38 protein expression level, p38MAPK activity was threefold lower in adipocytes than in preadipocytes isolated from both lean and obese inguinal fat pads (Fig. 1B). Interestingly, this decrease in p38MAPK activity is also observed in adipocytes isolated from the epididymal fat depot of lean and obese animals, suggesting that decreased p38MAPK activity correlates not only to pathological adipogenesis but also more generally to in vivo adipocyte differentiation.

We then determined whether there is a causal relationship between p38MAPK activity decrease and adipocyte maturation during in vitro adipocyte differentiation. Preadipocytes from inguinal and axillary depots were isolated and induced to differentiate into adipocytes for 10 days in the presence or absence of a specific inhibitor of p38MAPK, PD169316. This inihibtor, used at 10 μmol/l, is highly specific for p38MAPK and has no inhibitory activity on the other MAPKs, ERK, or JNK (online appendix). Adipocyte differentiation was evaluated after oil red O staining of triglycerides and quantified by measurement of optical density at 490 nm. Inhibition of p38MAPK by PD169316 increased triglyceride accumulation by 1.5 ± 0.3-fold compared with nontreated cells (Fig. 2A). Moreover, the expression of specific adipocyte markers such as aP2, PPARγ, adiponectin, leptin, and C/EBPα was two- to threefold higher in the presence of p38MAPK inhibitor (Fig. 2B). We investigated the effect of PD169316 in another cellular model of adipocyte precursor cells: the primary cultures of MEFs, which can be differentiated in adipocytes following the same protocol. As in adult preadipocytes, the expression of the aP2, PPARγ, adiponectin, and leptin adipocyte markers was increased in MEFs by up to twofold after the inhibition of p38MAPK by PD169316 (online appendix).

We then asked whether p38MAPK plays a role during the early stages of adipocyte differentiation by analyzing the effect of p38MAPK inhibition on adipogenesis in the CGR8 embryonic stem cell line. Adipocyte differentiation of embryonic stem cells can be achieved after a 3-week period, including the formation of cell aggregates during 7 days (embryoid bodies), a critical early treatment with retinoic acid followed by the application of classical adipogenic inducers from day 7 of the differentiation (18,20).

To determine whether p38MAPK activity varies during adipocyte differentiation in embryonic stem cells, we performed a Western blot analysis using anti-phospho p38MAPK antibodies. Two peaks of activation were detected, between days 4 and 5 and between days 13 and 16 (Fig. 3), while no change in p38 expression was observed.

We then tested the specificity of the p38MAPK inhibitor in embryonic stem cells (online appendix) and added PD169316 at a different time. The addition of the inhibitor during the early peak (between days 0 and 7) did not affect adipogenesis (20). PD169316 was then concomitantly added to the adipogenic inducers from day 7. Adipocyte differentiation was analyzed by measuring the area of adipocyte outgrowth in each embryoid body. Addition of the p38MAPK inhibitor increased by twofold adipocyte areas within each embryoid body compared with untreated cells (Fig. 4A). We then determined the number of embryoid bodies with 0–20%, 20–40%, 40–60%, 60–80%, or 80–100% of their surface covered by adipocytes. All the embryoid bodies treated with the inhibitor had between 40 and 80% of their surface covered by adipocytes, whereas in the untreated conditions the maximum area occupied by the adipocytes was 60% (Fig. 4B). Interestingly, inhibition of p38MAPK activity between days 13 and 16 (second peak of activation) increased adipogenesis as well as a treatment from day 7 to 21 (online appendix). We then analyzed the expression of adiponectin, leptin, and C/EBPα by real-time quantitative RT-PCR and aP2 and PPARγ by Northern blot. PD169316 alone did not affect the expression of these markers compared with no retinoic acid–treated cells. Retinoic acid–treated cells presented a specific expression of these markers that was significantly increased by PD169316 treatment (Fig. 5A and B). Altogether, these data showed that PD169316 potentiates adipogenesis in embryonic stem cells.

To rule out an effect of the chemical inhibitor unrelated to p38MAPK, we investigate the role of p38MAPK in adipogenesis by a more direct genetic approach. The capacity of p38MAPKα^−/− embryonic stem cells (25) to differentiate into adipocyte was analyzed. According to the protocol described for CGR8 cells, the capacity of p38MAPKα^−/− DBA-252 embryonic stem cells to differentiate after retinoic acid treatment into adipocytes was evaluated and compared with their p38MAPKα^+/+ counterparts. Real-time PCR analysis of the expression of adipocyte specific markers showed that deletion of p38MAPKα strongly increased the expression of aP2, PPARγ, adiponectin, leptin, and C/EBPα genes (Fig. 6A).

Upon subcutaneous injection in animals, embryonic stem cells are tumorigenic, inducing teratocarcinoma containing various differentiated tissues derived from the injected cells. To analyze the capacity of p38MAPKα^+/+ and p38MAPKα^−/− embryonic stem cells to differentiate in vivo after subcutaneous injection, six animals for each cell line were injected in two sites per mouse. After 15–20 days, tumors were removed and analyzed for the expression of adipocyte markers. Teratocarcinomas from p38MAPKα^−/− displayed a significantly higher expression of PPARγ, adiponectin, and leptin than tumors from wild-type cells (Fig. 6B). By contrast, no change was found in markers of other lineages, such as for example Map2, a neuronal marker (data not shown).

Altogether, our data demonstrated that inhibition or knockout of p38MAPK increases adipogenesis both in vivo, in teratocarcinomas, and in vitro in adipose precursor cells, adult preadipocytes and embryonic fibroblasts, and in embryonic stem cells.

To investigate the molecular mechanisms implicated in p38MAPK regulation of adipogenesis, we analyzed the effect of PD169316 on C/EBPβ and PPARγ transcription factor activities in MEFs. Several reports (6,26,27) have demonstrated that expression and transcriptional activity of C/EBPβ are induced and that C/EBPβ proteins are activated by phosphorylation on Thr-188 within the two 1st days of the differentiation protocol. C/EBPβ phosphorylation on Thr-188 on day 2 of differentiation was analyzed in the presence or absence of PD169316. While C/EBPβ protein expression level was not affected by the inhibition of p38MAPK, addition of PD169316 significantly increased its phosphorylation in MEFs and preadipocytes (Fig. 7A and B). C/EBPβ activation results in an increase in PPARγ expression (28); accordingly, we found that inhibition of p38MAPK increased PPARγ mRNA levels (Fig. 2B) and protein expression (Fig. 8A). JNK and, to a lesser extent p38MAPK, phosphorylates PPARγ1 on Ser-82 in vitro (29), and this phosphorylation negatively regulates PPARγ1 transactivation. However, no role in adipogenesis has been associated with the phosphorylation of this serine. We have investigated the phosphorylation status of PPARγ on Ser-82 in wild-type and p38MAPKα^−/− MEFs treated or not treated with PD169316. Despite their incapacity to undergo terminal adipocyte differentiation due to a defect of contact inhibition (30), we analyzed PPARγ expression 2 days after the induction of differentiation in these cells. PPARγ was higher in p38MAPKα^−/− and cells treated with the inhibitor than in p38^+/+ cells (Fig. 8B). Furthermore, we show that the stronger signal obtained with anti-phospho PPARγSer-82 is due to the increase of PPARγ expression rather than the increase of PPARγ phosphorylation (Fig. 8B). We then determined whether PPARγ transcriptional activity is regulated by p38MAPK. PPARγ binds to the PPRE and activates the transcription of genes implicated in adipogenesis, such as aP2 (31). We analyzed PPARγ activity in MEF cells treated or not treated with PD961396 by transitory transfections of a PPRE-luciferase (PPRE-luc) reporter plasmid (22) and in the presence or absence of thiazolidinedione (TZD), a PPARγ agonist. Consistent with the positive effect of the inhibitor on PPARγ protein expression, inhibition of p38MAPK increased by twofold the endogenous PPARγ activity, with or without TZD (Fig. 8C). To know whether inhibition of p38MAPK also increased PPARγ activity itself, we repeated these experiments by cotransfecting the reporter with a PPARγ expression vector. Introducing the PPARγ expression vector increased by twofold the basal PPARγ activity, and, in the presence of TZD, the resulting activity was stimulated fivefold by a PD169316 treatment (Fig. 8D). These results strongly suggested that, in addition to its effect on C/EBPβ, inhibition of p38MAPK also directly activates PPARγ proteins.

We then analyzed PPARγ activity in p38MAPKα^−/− MEFs (32) compared with wild-type cells. To normalize for transfection efficiency differences, MEFs were transiently cotransfected with PPRE-luc and RSV-βgalactosidase (RSV-βGal) plasmids and luciferase activities were expressed relatively to βGal activities. Endogenous PPARγ activity was threefold higher in p38MAPKα^−/− cells compared with control wild-type cells (Fig. 8E), demonstrating that the knockout of p38MAPK is sufficient to activate this transcription factor. MEF cultures were then cotransfected with RSV-βGal, PPRE-luc, and CAG-PPARγ plasmids. In these conditions, PPARγ activity was increased ninefold in p38MAPKα^−/− MEFs compared with wild-type cells (Fig. 8E). To further demonstrate the specific role of p38MAPK pathway in regulating PPARγ, p38MAPKα^−/− MEFs were transiently transfected with a RSV-p38MAPKα expression vector. As shown in Fig. 8F, reexpression of p38MAPKα in knockout MEFs significantly decreased PPARγ activity. Interestingly, this effect can be blocked by the treatment of p38MAPKα^−/− MEFs with PD169316. Whereas luciferase activity of control cells treated by this compound was not affected, the PD169319 restored the PPARγ transcriptional activity in cells cotransfected with RSV-p38MAPKα (Fig. 8F). We demonstrate here that the p38MAPKα isoform is specifically involved in the regulation of PPARγ transcriptional activity.

To clarify the role of p38MAPK in normal and pathological adipogenesis, we determined the function of this pathway by complementary biochemical and genetic approaches in various experimental models, from embryonic to adult stages.

In vivo, in dietary and genetic (ob/ob) models of obesity, we found a strong decrease of p38MAPK activity in subcutaneous (inguinal and axillary) fat depots compared with the same tissue in lean mice. By contrast to the activations of the JNK and ERK pathways observed in obesity (33), our results suggest an unsuspected negative role of p38MAPK in adipogenesis. No change in p38MAPK activity was detected in epididymal and perirenal fat depots. These differences between fat depots are in agreement with the well-known regional variations in growth, cellularity, metabolism, and gene expression of the white adipose tissue depots in mice and humans (34–36). Interestingly, we found that p38MAPK activity is significantly higher in preadipocyte precursor cells, from both inguinal and epididymal fat depots, than in mature adipocytes, showing an inverse correlation between p38MAPK activation and adipocyte differentiation.

By using the p38MAPK-specific inhibitor PD169316, and analyzing p38MAPKα^−/− cells, we then demonstrated that p38MAPK plays a direct negative role in adipocyte differentiation. Inhibition of p38MAPK increased adipogenesis in several primary cell cultures, from the embryonic to the adult stages (MEFs and preadipocytes), and in two different embryonic stem cell lines (CGR8 and DBA 252). We found that this effect is mediated by the regulation of the adipogenic factors C/EBPβ and PPARγ, whose transcriptional activities were activated by PD169316. C/EBPβ has been found to regulate PPARγ expression and is activated by phosphorylation on Thr-188. For example, ERK and GSK3 have been shown to phosphorylate Thr-188 and increase adiponectin expression in 3T3-L1 (7). We found an opposite regulation by p38MAPK; the p38MAPK pathway inhibits C/EBPβ proteins and decreases expression of adipocyte markers. A similar negative role on C/EBPβ function has been described for the phosphatidylinositol 3-kinase pathway (7). These results support the hypothesis that the C/EBPβ phosphorylation site Thr-188 is a crucial convergent target for opposite regulations by several transduction pathways. On one hand, ERK and GSK3 activate C/EBPβ proteins, and, on the other hand, p38MAPK and phosphatidylinositol 3-kinase inhibit them.

We found that the inhibitory role of p38MAPK in adipocyte differentiation is also mediated by the regulation of the crucial adipogenic transcription factor PPARγ. Biochemical inhibition of the p38MAPK pathway induces both PPARγ expression and activity. By transient transfection experiments, this effect was evidenced on transfected PPARγ proteins, independently of the increase in endogenous proteins, suggesting that it is also effective on PPARγ activity itself. Furthermore, PPARγ activity and expression is significantly higher in MEFs p38α^−/− compared with wild-type cells, and reexpression of p38MAPKα in knockout cells represses PPARγ activity. Therefore, by both biochemical and genetic approaches, we show that p38MAPK activity negatively regulates PPARγ.

In contrast to our results, a positive role for p38MAPK in adipogenesis has been described in preadipocyte cell lines (10,37,38). They showed that, during the differentiation of 3T3-L1 preadipocyte cells, addition of p38MAPK inhibitors decreased adipocyte formation. They attributed this effect to a lower phosphorylation of C/EBPβ protein in the presence of the inhibitors. By using PD169316 instead of SB20358, we confirmed a positive role of p38MAPK in the adipogenesis of 3T3-L1. Addition of the p38MAPK inhibitor decreased the formation of adipocytes in 3T3-L1 and downregulated PPARγ transcriptional activity (data not shown). Taken together, these results show that p38MAPK inhibition decreases PPARγ activity in one cellular model, 3T3-L1 cells, and has the opposite effect in the other cellular models analyzed. Because the latter are primary cultures from embryos to adults, they represent more physiological experimental models than the established cell line 3T3-L1.

Altogether, these observations strongly suggest that the regulation of PPARγ by the p38MAPK pathway involves molecular partners that are different in 3T3-L1 and in the other cellular models. One can hypothesize that p38MAPK is acting on PPARγ coactivators. It is known, for example, that one of them, PGC-1, is phosphorylated by p38MAPK. However, PGC-1 is not implicated in white adipocyte differentiation (39,40). Although we do not know whether they are phosphorylated by p38MAPK, other PPARγ coactivators, such as steroid receptor coactivator, CREB-binding protein, or p300, involved in adipogenesis, represent interesting potential targets (41). Alternatively, PPARγ corepressors could also be involved in p38MAPK regulation, especially because they regulate the PPARγ transcriptional activity in adipocytes (42). The comparison between 3T3-L1 and the other cellular models for activity of the various PPARγ partners should help to understand the opposite response of PPARγ transcriptional activity to p38MAPK regulation.

In conclusion, our results demonstrated that p38MAPK inhibition increased adipogenesis via C/EBPβ and PPARγ activations, and further investigations would be of interest to elucidate molecular mechanism involved in this activation. These observations were obtained from in vivo and in vitro experimental models and from embryonic stages to adulthood. Because no activated form of p38MAPK has been described so far, we do not know yet whether specific activation of p38MAPK would lead to inhibition of the development of white adipose tissues and could have a therapeutic effect regarding obesity. Conversely, inhibition of p38MAPK by resulting in the enhancement of both the expression and the activity of PPARγ could represent an interesting approach in the treatment of type 2 diabetes. Analysis of the in vivo effects of p38MAPK inhibitors regarding insulin resistance parameters and obesity should be of great interest.

We thank Foundation Bettencourt-Schueller and Association pour la Recherche sur le Cancer (grant no. 4525) for support. M.A. is supported by a fellowship from INSERM-IRIS-PACA. M.P. was supported by Sidaction.

F.B. is a Centre National de la Recherche Scientifique investigator.

We thank Drs. C. Dani and J. Piedrafita for the gift of plasmids, Drs. C.A. Gabel and M. Karin for the gift of p38MAPK knockout cells, and P. Gual for critical reading of the manuscript. We are grateful to Dr. Jean-François Peyron for his help and L. Caron for her help.