Specific Inhibitors of p38 Mitogen-activated Protein Kinase Block 3T3-L1 Adipogenesis*

SB203580 and SB202190, pyridinyl imidazoles that selectively inhibit p38 mitogen-activated protein (MAP) kinase, are widely utilized to assess the physiological roles of p38. Here, we demonstrate that treatment of 3T3-L1 fibroblasts with these p38 MAP kinase inhibitors prevents their differentiation into adipocytes as judged by an absence of lipid accumulation, a lack of expression of adipocyte-specific genes, and a fibroblastic morphological appearance. In 3T3-L1 fibroblasts and developing adipocytes, p38 is active. p38 activity decreases dramatically during later stages of differentiation. In accordance with the time course of p38 activity, p38 inhibitor treatment during only the early stages of differentiation is sufficient to block adipogenesis. In addition, we constructed a 3T3-L1 cell line harboring an inducible dominant negative p38 mutant. The induction of this dominant negative mutant of p38 prevents adipocyte differentiation. Thus, it is likely that the antiadipogenic activity of SB203580 and SB202190 is indeed due to inhibition of p38 MAP kinase. This study points out that CCAAT/enhancer-binding protein β (C/EBPβ), a transcription factor critical for the initial stages of 3T3-L1 adipogenesis, bears a consensus site for p38 phosphorylation and serves as a substrate for p38 MAP kinase in vitro. Although the induction of C/EBPβ is not significantly altered in the presence of the p38 inhibitor, the amount of in vivo phosphorylated C/EBPβ is reduced by SB203580. The transcriptional induction of PPARγ, a gene whose expression is induced by C/EBPβ, and a factor critically involved in terminal differentiation of adipocytes, is impaired in the presence of p38 inhibitors. Thus, transcription factors such as C/EBPβ that promote adipocyte differentiation may be p38 targets during adipogenesis. Collectively, the data in this study suggest that p38 MAP kinase activity is important for proper 3T3-L1 differentiation.

MAP 1 kinase and are widely utilized as tools to probe p38 MAP kinase function in vitro and in vivo (1)(2)(3)(4)(5). They are extremely specific, as SB203580 has no inhibitory activity on all kinases tested in vitro, including the other MAP kinases, at concentrations greater than 10-fold the concentration used in our treatment of 3T3-L1 cells (6). In addition, the human homolog of p38 MAP kinase was discovered because it was specifically photoaffinity-labeled by a derivative of SB202190. This compound inhibits LPS-stimulated IL-1 and TNF␣ production, and p38 is its molecular target (7). Recently, a crystal structure of p38 in complex with SB203580 revealed that the inhibitor binds within the ATP pocket of the kinase (8,9). Single amino acid substitutions within the ATP binding pocket of p38 alter the sensitivity of p38 to the inhibitors (10) and offer an explanation for the specificity of this class of inhibitors.
To date, three groups of MAP kinases have been identified in mammalian cells: 1) extracellular signal-regulated kinases that are also referred to as p42/44 MAPK; 2) c-Jun N-terminal kinases that are also termed stress-activated protein kinases; and 3) p38 MAP kinases. p38 MAP kinase, initially identified in macrophages that were stimulated with lipopolysacharides (11), is homologous to Saccharomyces cerevisiae HOG1 kinase (12). The p38 pathway is not commonly activated by mitogens; instead, it is usually induced by stressing cells with endotoxins, osmotic shock, or metabolic inhibitors.
However, although much has been learned regarding the importance of p38 MAP kinase in inflammatory and stress responses, relatively little is known regarding its role in differentiation processes. Recently, it was shown that p38 MAP kinase activities are elevated in association with the onset of hypertrophy and programmed cell death in cardiac myocytes. Indeed, upstream activators of p38 MAP kinase, MKK3 and MKK6, are sufficient to promote these phenotypic changes (3,13). Wang et al. (13) utilized SB202190 and dominant negative p38␣ and p38␤ mutants to demonstrate that these upstream activators indeed require p38 activity to exert these effects.
In this study, we assess the importance of p38 MAP kinase activity in the differentiation of 3T3-L1 fibroblasts into adipocytes. The conversion of preadipocytes to mature adipocytes is a paradigm for terminal differentiation. This process can be studied in cell culture (14,15). Markers for the predifferentiated and fully differentiated states have been characterized, and some factors that drive this process have been identified (for recent reviews, see Refs. 16 -19). In our own work, we have primarily focused on 3T3-L1 fibroblasts, which can be differentiated to mature adipocytes within a period of approximately 8 days. The sequence of events that leads to the accumulation of triglycerides and the induction of adipocyte-specific markers includes transient activation of key factors. The C/EBP family of transcription factors participates at all stages of the differentiation program. In particular, induction of C/EBP␤ and C/EBP␦ is observed early in the differentiation process. The standard differentiation protocol includes a mixture of insulin, dexamethasone and 3-isobutyl-1-methylxanthine (20). Apparently, 3-isobutyl-1-methylxanthine governs induction of C/EBP␤, and dexamethasone is responsible for the induction of C/EBP␦ in 3T3-L1 cells (21). Overexpression of C/EBP␤ in 3T3-L1 cells abrogates the need for 3-isobutyl-1-methylxanthine and dexamethasone during differentiation (22). Ectopic expression of C/EBP␤ in NIH3T3 fibroblasts cells together with appropriate hormonal stimulation is sufficient to drive these cells through adipogenesis (22). Similarly, the expression of C/EBP␣ in NIH-3T3 cells stimulates adipocyte differentiation (23). In addition, the constitutive expression of another factor, PPAR␥, promotes conversion of NIH-3T3 cells into adipocytes in the presence of the appropriate activating ligand (24,25). In fact, it is believed that C/EBP␤ induction and activation early in 3T3-L1 differentiation is directly responsible for the transcriptional activation of PPAR␥ (17,26,27).
We demonstrate that inhibitors of p38 MAP kinases as well as a dominant negative p38 prevent conversion of 3T3-L1 fibroblasts to adipocytes. p38 is present throughout the adipocyte differentiation process. However, p38 kinase activity and phosphorylation are observed only during the early phases of differentiation. Accordingly, treatment of these cells with p38 inhibitors during only the initial stages of differentiation is sufficient to block adipogenesis, and inhibition of p38 during later stages of differentiation has little effect. Our data suggest that C/EBP␤ may be a direct target of p38 MAP kinase. Inhibition of p38 kinase leads to reduced levels of C/EBP␤ phosphorylation in vivo, and p38 phosphorylates C/EBP␤ in vitro, suggesting a possible direct link between p38 MAP kinase and adipogenic transcription factors.

EXPERIMENTAL PROCEDURES
Materials-Dulbecco's modified Eagle's medium (DMEM) was purchased from Cellgro Inc.; orthophosphate 32 P was purchased from NEN Life Science Products at a specific activity of 9000 Ci/mmol. DMEM lacking methionine, cysteine, and glutamate was purchased from ICN, and DMEM lacking phosphate and pyruvate was purchased from Specialty Media Inc.(Lavallette, NJ). SB203580, SB202190, SB202474, and PD98059 were purchased from Calbiochem and dissolved in Me 2 SO at concentrations of 10 and 50 mM, respectively and used at final concentrations of 10 and 50 M, respectively.
Cell Culture-3T3-L1 murine fibroblasts (a generous gift of Dr. Charles Rubin, Department of Molecular Pharmacology, Albert Einstein College of Medicine) were propagated and differentiated according to the protocol described by Student et al. (20). In brief, the cells were propagated in FCS (DMEM containing 10% fetal calf serum (JRH Biosciences) and penicillin/streptomycin (100 units/ml each)) and allowed to reach confluence (day Ϫ2). After 2 days (day 0), the medium was changed to DM1 (containing FCS and 160 nM insulin, 250 M dexamethasone, and 0.5 mM 3-isobutyl-1-methylxanthine). Two days later (day 2), the medium was switched to DM2 (FCS containing 160 nM insulin). After another 2 days, the cells were switched backed to FCS.
Oil Red O Staining-Staining was performed as described by Kasturi and Joshi (30).
Total RNA Isolation and Northern Blot Analysis-Isolation of total RNA from 3T3-L1 cells was performed with a Qiagen RNeasy purification kit. Agarose gel electrophoresis of 10 g of total RNA and its transfer to nitrocellulose membranes was described by Baldini et al. (31). Hybridizations were performed in ExpressHyb (CLONTECH) according to the manufacturer's instructions. The 32 P-labeled PPAR␥ probe was used at a concentration of 2 ϫ 10 6 cpm/ml. The filters were subsequently washed in 2ϫ SSC, 0.05% SDS and 0.1ϫ SSC, 0.05% SDS at 50°C before autoradiography.
Generation of a 3T3-L1 Cell Line Stably Expressing an Inducible Dominant Negative p38 Mutant-A full-length mouse p38␣ was obtained by polymerase chain reaction using a 3T3-L1 fibroblast library (31) as a template. The product was cloned into pCB7 as a HindIII fragment. This construct served as a template to generate the TGY 3 AGF mutant (residues 180 -182) described in Refs. 13, 32, and 33 by a standard polymerase chain reaction-based strategy. A carboxyl-terminal FLAG tag was introduced to distinguish between endogenous and dominant negative p38. The resulting fragment was cloned into vector pOPI3 (Stratagene) as a NotI fragment. The p38 insert was sequenced on both strands to confirm the mutation. This plasmid was co-transfected into 3T3-L1 fibroblasts with plasmid pCMVLacI (Stratagene), and stable transformants were isolated in the presence of hygromycin.
In Vivo Phosphorylation Experiments-Cells were washed twice in DMEM lacking phosphate and incubated for 3 h in DMEM lacking phosphate supplemented with 1 mCi [ 32 P]orthophosphate per 10-cm dish in the presence of 10 M SB203580. Control cells were treated with an equal volume of Me 2 SO. This medium was then removed, and cells were washed in ice-cold PBS and subsequently lysed in IP buffer (see below).
Immunoprecipitations-Cells were lysed in IP buffer (150 mM NaCl, 50 mM Tris pH 8.0, 1% Nonidet P-40, 0.5% deoxycholate, 0.1% SDS, 50 mM NaF, 30 mM sodium pyrophosphate, 100 M sodium orthovanadate, 0.1 g/ml okadaic acid, and protease inhibitors). Lysates were precleared by addition of 50 l of a 1:1 slurry of protein A-Sepharose (Amersham Pharmacia Biotech) in TNET buffer (1% Triton X-100, 150 mM NaCl, 2 mM EDTA, 20 mM Tris, pH 8.0) containing 1 mg/ml bovine serum albumin. After 30 min at 4°C, samples were centrifuged for 5 s at 15,000 ϫ g, the supernatant was transferred to a fresh tube, and 50 l of protein A-Sepharose was added together with the corresponding antiserum. Samples were then incubated for 3 h at 4°C. Immunoprecipitates were washed six times in IP buffer lacking okadaic acid and analyzed by SDS-PAGE.
Immunoblotting-After SDS-PAGE, proteins were transferred to BA83 nitrocellulose (Schleicher & Schuell). Nitrocellulose membranes were blocked in PBS or Tris-buffered saline with 0.1% Tween 20 and 5% nonfat dry milk. Primary and secondary antibodies were diluted in PBS or Tris-buffered saline with 0.1% Tween 20 and 1% bovine serum albumin. Bound antibodies were detected by enhanced chemiluminescence according to the manufacturer's instructions (NEN Life Science Products).
Preparation of GST Fusion Proteins-Glutathione S-Transferase fusion proteins to rat C/EBP␤ (residues 153-297) were generated by polymerase chain reaction in vector pGEX4T-1 (Amersham Pharmacia Biotech). The fusion proteins were isolated by conventional methods (34,35) and used for in vitro kinase assays.
In Vitro Kinase Assay-Active p38 MAP kinase was immunoprecipitated in IP buffer from 3T3-L1 confluent fibroblasts, which had been incubated in the presence of TNF␣ (5 ng/ml) for 10 min. Kinase assays were performed essentially as described previously (36). Briefly, immunoprecipitates were washed six times in IP buffer followed by two washes in kinase buffer (25 mM Hepes, pH 7.4, 25 mM ␤-glycerophosphate, 25 mM MgCl 2, 1 mM dithiothreitol, and 0.1 mM Na 3 VO 4 ). The reactions were initiated by the addition of substrate and 50 M ATP (15 Ci) in a final volume of 40 l of kinase buffer. Samples were either treated with 10 M SB203580 or with an equal volume of Me 2 SO (control). Kinase reactions were incubated at 30°C for 30 min followed by the addition of SDS-PAGE sample buffer. Samples were run on a 10% polyacrylamide gel and visualized by autoradiography.
Preparation of Cell Lysates from 3T3-L1 Cells during Adipocyte Differentiation-To determine p38 and p42/44 MAP kinase activity at different stages of adipogenesis, lysates of 3T3-L1 cells at 2-day intervals of differentiation were washed twice with cold PBS, lysed in boiling 2ϫ SDS-PAGE sample buffer, and boiled again for another 3-5 min. Samples were then mildly sonicated to reduce viscosity.
To determine the effects of the inhibitors on the expression levels of the various adipocyte marker proteins, samples were prepared from day 8 adipocytes that had been treated with or without MAP kinase inhibitors. Cells were washed twice with cold PBS and lysed in TNET buffer containing 60 mM octyl glucoside in the presence of protease inhibitors for 30 min at 4°C. Samples were centrifuged for 5 min at 15,000 ϫ g, and the supernatant was transferred to a fresh tube. Protein determinations were made on the postnuclear supernatants using the BCA assay (Pierce).
Two-dimensional Gel Analysis-Isoelectric focusing followed by SDS-PAGE was performed as described previously in Ref. 37 on a Hoefer GT-1 tube gel apparatus.
Other Methods-Separation of proteins by SDS-PAGE, fluorography, and immunoblotting were performed as described previously (38). Automated DNA sequencing was performed by the custom DNA sequencing facilities at Research Genetics and by the Einstein DNA Sequencing Facility.

RESULTS
The Expression Patterns of p38 and p42/44 MAP Kinases during 3T3-L1 Adipogenesis-The state of p38 MAP kinase activation during 3T3-L1 adipogenesis was determined. Protein extracts were prepared from 3T3-L1 cells every 2 days after initiation of differentiation. Western blots were probed with antibodies that bind to the activated, phosphorylated p38 MAP kinase. p38 MAP kinase was most active on day 0, and its activity gradually declined during the first 4 days of differentiation (Fig. 1A). In contrast, p38 protein decreased only slightly during differentiation. The blots were also probed with antibodies against an isoform of GDI, a protein whose expression is relatively constant over the course of differentiation to control for protein loading (39), and ACRP30, a secretory pro-tein whose expression is dramatically induced during adipogenesis (29) (Fig. 1A). Analysis of p42/44 MAP kinase revealed a pattern that was similar to that of p38 MAP kinase. p42/44 MAP kinase protein declined only moderately over the course of differentiation, but the activated form was evident only during the early stages of adipogenesis.
SB203580, a Specific p38 MAP Kinase Inhibitor, Blocks 3T3-L1 Differentiation-SB203580, a potent and highly specific inhibitor of p38 MAP kinase activity (6), was added to 3T3-L1 cells at day 0, day 2, or day 4 of the differentiation protocol (see "Experimental Procedures"). The cells were then incubated with the inhibitor for the remainder of the differentiation protocol. Cells that did not receive the inhibitor were treated with an equal volume of vehicle (Me 2 SO). Cells incubated with SB203580 throughout the differentiation protocol did not accumulate lipid (indicated by the absence of oil red O stain) or adopt morphological characteristics of adipocytes (Fig.  2, A and B). Instead, they retained the appearance of 3T3-L1 fibroblasts. When the inhibitor treatment was initiated on day 2 of differentiation, we also observed a reduction in adipocytes and lipid accumulation. However, extensive adipocyte conversion occurred when the inhibitor regimen was begun on day 4, although there was still a marginal effect. This may be due to residual active p38 still present at this stage (Fig. 1). These results are consistent with the observation that p38 MAP kinase activity is significantly reduced by day 4 of adipogenesis ( Fig. 1A).
The effect of another highly specific p38 MAP kinase inhibitor, SB202190, was also investigated. Similar to our observations with SB203580, 3T3-L1 fibroblasts failed to differentiate into adipocytes in the presence of SB202190 (Fig. 2C). As a critical control for these experiments, 3T3-L1 cells were differentiated in the presence of SB202474, a control compound of similar chemical structure that does not inhibit p38 MAP kinase activity (7). SB202474 did not inhibit adipogenesis (Fig. 2C).
As a control for proper p38 MAP kinase inhibition by SB203580 in 3T3-L1 cells, the phosphorylation status of MAP kinase-activated protein kinase-2, a well established p38 substrate, was examined in the presence or absence of the p38 inhibitor (46). As shown in Fig. 2E, incubation of [ 32 P]orthophosphate-labeled 3T3-L1 cells with SB203580 resulted in a strong reduction (Ͼ80%) in the amount of phosphorylated MAP kinase-activated protein kinase-2. When total protein extracts from these cells were analyzed by two-dimensional gel electrophoresis, no global effects on the phosphorylation state of proteins were observed. A representative example of the acidic region of a gel is shown in Fig. 2F. This result underscores the specificity of SB203580. Although it inhibits the phosphorylation of a p38 MAP kinase substrate in vivo, it does not affect the overall pattern of protein phosphorylation. A, p38 and p42 MAP kinase activity peaks early during adipocyte conversion. 3T3-L1 cells were differentiated according to a standard protocol. On the indicated days after initiation of differentiation, the cells were lysed as described under "Experimental Procedures." 50 g of protein were analyzed by SDS-10% PAGE and transferred to nitrocellulose. The filters were then probed with the indicated polyclonal antibody preparations that included an adipocyte differentiation marker (Acrp30), a marker that remains constant throughout differentiation (GDI), p38 and p42 MAP kinase, and antibodies specific for the activated forms of p38 and p42 (antiphospho-p38 and antiphospho-p42, respectively). B, p38 can be activated by TNF␣ in 3T3-L1 fibroblasts and adipocytes. Subconfluent 3T3-L1 cells and fully differentiated day 8 adipocytes were serum-starved overnight and then treated for 10 min with DM1 (see "Experimental Procedures") or with 5 ng/ml TNF␣. The cells were subsequently lysed in boiling 2ϫ SDS-PAGE sample buffer and analyzed by SDS-10% PAGE and Western blotting with antiphospho-p38.
p38 Inactivation during the Early Stages of Differentiation Is Sufficient to Block Adipogenesis-To determine whether p38 inhibition during only the early stages of differentiation is sufficient to block adipogenesis, we treated the 3T3-L1 cells with SB203580 for the initial 2 or 4 days of differentiation and then continued the differentiation regimen in the absence of inhibitor. There was significantly less triglyceride accumulation in cells treated with inhibitor for the first 2 days of differ-entiation, and almost no fat droplets were observed in cells that were incubated with inhibitor for days 0 -4 (Fig. 3A). Modest decreases in fatty acyl-CoA synthase and caveolin-1 induction were evident in cells treated with inhibitor for days 0 -2 (Fig.  3B). Induction of markers was severely blocked when SB203580 was added to the differentiation media during days 0 -4 (Fig. 3B). Thus, inhibition of p38 MAP kinase during the first 4 days of differentiation is sufficient to prevent 3T3-L1 On the eighth day of differentiation, cells were stained with oil red O and analyzed macroscopically. D, induction of adipocyte-specific markers is inhibited by SB203580. An identical set of plates as shown in a and b were lysed. 50 g of protein was analyzed by SDS-PAGE and Western blotting for constitutive markers (GDI), strongly induced markers (fatty acyl-CoA synthase; FACS), Glut4, Rab3D, caveolin-1, and the highly adipocytespecific markers Acrp30 and adipsin. E, SB203580 inhibits p38 MAP kinase in 3T3-L1 cells. Day 1 3T3-L1 fibroblasts were metabolically labeled with 32 PO 4 as described under "Experimental Procedures" for 3 h either in the presence or absence of 10 M of SB203580. The cells were then lysed in IP buffer and immunoprecipitated with antibodies to MAP kinase-activated protein kinase-2 (MAPKAPK-2). Immunoprecipitates were analyzed by SDS-10% PAGE followed by autoradiography. F, SB203580 does not have any global effects on the phosphorylation pattern in 3T3-L1 cells. Day 1 3T3-L1 fibroblasts were treated as described for D. Lysates were analyzed by isoelectric focusing followed by SDS-PAGE and autoradiography. This exact experiment was performed in duplicate and on three different days after the initiation of differentiation. We did not observe any reproducible changes in the pattern or the intensity of total phosphorylated proteins.
p38 MAP Kinase Activity Is Required for Adipogenesis adipocyte differentiation.
Reduced Levels of Adipogenesis in the Presence of a Dominant Negative p38 Mutant-To examine if the antiadipogenic effects of SB203580 and SB202190 are indeed due to the specific binding of the inhibitors to p38, we tested whether a dominant negative mutant of p38 can mimic the effects of the inhibitors. The mutant carrying a replacement of the two activating phosphorylation sites with alanine and phenylalanine (TGY 3 AGF; residues 180 -182 in mouse p38␣) has been widely used to assess p38 function in vivo (13,32,33). We generated a number of 3T3-L1 cell lines that stably express a FLAG-tagged dominant negative p38 mutant under the control of the lac operon (LacSwitchII system, Stratagene). This system allows relatively tight repression in the absence of the inducer (IPTG). Several independent clones were analyzed. Experiments with a representative clone are shown. IPTG indeed induces the expression of the FLAG-tagged p38 dominant negative mutant as judged by Western blot analysis (Fig. 4A). This clonal cell line was subjected to the standard differentiation protocol either in the absence or presence of IPTG (Fig.   4B). To control for any potential effects of IPTG on the differentiation process, a cell line lacking the dominant negative p38 mutant was analyzed in parallel. As shown in Fig. 4B, induction of the dominant negative p38 inhibits adipocyte differentiation. These cells do not accumulate lipid and retain their fibroblastic morphology following the complete differentiation protocol. Furthermore, cells differentiated in the presence of the dominant negative p38 failed to express fatty acyl-CoA synthase, a protein normally induced during the course of adipocyte differentiation (Fig. 4C). Importantly, treatment of the control cell line with IPTG had no effect on the ability of cells to differentiate (Fig. 4B, lower part). In the presence of dominant negative p38, we consistently found an ϳ80% reduction of differentiation as judged by the reduced levels of fatty acyl-CoA synthase and lipid accumulation. This is slightly less than the ϳ95% inhibition achieved by treatment with the SB203580 and SB202190. Very high expression levels of the dominant negative mutant may be necessary to completely outcompete endogenous p38 for its in vivo substrates (32). Nevertheless, the strong antiadipogenic effect of the dominant  (4) demonstrated that p38 MAP kinase phosphorylates CHOP (GADD153) in vitro and in vivo and that this phosphorylation is necessary for its biological activity. CHOP is a member of the C/EBP family of transcription factors, and its ectopic overexpression blocks 3T3-L1 differentiation. Three members of the C/EBP family, C/EBP␣, C/EBP␤, and C/EBP␦, are induced during adipogenesis and promote 3T3-L1 differentiation. It has been proposed that ectopic overexpression of CHOP inhibits adipogenesis by forming heterodimers with C/EBPs (47) and thereby obstructing their normal adipogenic activity. A comparison between CHOP and the C/EBP transcription factors indicates that the p38 MAP kinase phosphorylation site is conserved (Fig. 5A). It has been previously demonstrated that C/EBP␤ must be phosphorylated at this threonine to promote transcription (48).

C/EBP␤ Is a p38 MAP Kinase Target both in Vitro and in Vivo-A study by Wang and Ron
Since p38 is active early in the course of differentiation (Fig.  1A), we investigated whether p38 MAP kinase phosphorylates C/EBP␤, which is also expressed early in differentiation. To test whether C/EBP␤ could serve as a direct substrate for p38, we immunoprecipitated p38 from confluent 3T3-L1 fibroblasts and performed a kinase assay on a glutathione S-transferase fusion protein comprising the carboxyl-terminal domain of rat C/EBP␤ (starting at residue 153). As shown in Fig. 5B, the glutathione S-transferase-C/EBP␤ serves as a phosphoacceptor in this in vitro kinase assay (center lane). This phosphorylation was inhibited by treatment of the p38 immunoprecipitate with SB203580 (left lane). Additionally, glutathione S-transferase alone did not serve as a substrate under these conditions (right lane). To determine whether this effect is also observed in vivo, 3T3-L1 cells on the second day of differentiation were [ 32 P]orthophosphate-labeled and incubated either in the presence or absence of SB203580. A strong reduction in the amount of immunoprecipitated, phosphorylated C/EBP␤ was observed in 3T3-L1 cells incubated in the presence of the p38 MAP kinase inhibitor (Fig. 5C, top part). The same amount of C/EBP␤ was recovered from both inhibited and control cells (Fig. 5C, bottom part). Incorporation of [ 32 P]orthophosphate into total protein was similar in cells treated with and without inhibitor (data not shown). These results suggest that inhibition of p38 MAP kinase activity reduces the amount of phosphorylated C/EBP␤ in vivo.
To determine inhibition of the p38 MAP kinase pathway prevents the induction of C/EBP␤, we probed for C/EBP␤ levels after 2 days of inhibitor treatment. Under standard conditions, C/EBP␤ is most strongly induced during the first 2 days of differentiation (21,22). Fig. 6A shows the presence of the inhibitor does not significantly block the induction of C/EBP␤ during the first 2 days of differentiation. After 3T3-L1 cells reach confluence, induction of the differentiation program leads to a brief period of clonal expansion in the course of which the cells reenter the cell cycle and undergo an additional one or two cell divisions (49). While this does not seem essential for differentiation per se (50), we did want to see whether p38 inhibitors have an effect on cell proliferation during that phase. We did not observe any difference in cell number between inhibitor-treated and control cells after 4 days of differentiation (data not shown) and therefore conclude that the p38 inhibitors do not have a significant effect on the clonal expansion phase of 3T3-L1 cells.
If one of the major functions for p38 MAP kinase is indeed the post translational activation of C/EBP␤, we would predict that induction of a downstream target of C/EBP␤ would be affected. Expression of PPAR␥, a member of the peroxisome proliferator-activated receptor subfamily of nuclear hormone receptors (51), has been demonstrated to be induced by C/EBP␤ (17,26,27) and is in turn responsible for the transcriptional activation of many adipocyte-specific markers (24). Northern blot analysis on RNA isolated on day 4 (Fig. 6B) reveals a Ͼ80% reduction of PPAR␥ mRNA in cells differentiated in the presence of inhibitor. Taken together, these data suggest that p38 may promote adipogenesis at least in part through its effect on C/EBP␤ activity. DISCUSSION In this study, we have focused on the role of p38 MAP kinase in 3T3-L1 adipocyte conversion. p38 MAP kinase is present throughout differentiation, but its activity decreases by day 4 of the differentiation program. When 3T3-L1 cells are differentiated in the presence of p38 inhibitors, adipogenesis, as measured by (i) lipid accumulation, (ii) cell morphology, and (iii) induction of several adipocyte-specific marker proteins, is severely inhibited. In agreement with the observation that p38 MAP kinase activity is observed during only the initial stages of differentiation, the presence of p38 inhibitors during these early stages is sufficient to block adipogenesis. In addition, expression of a dominant negative p38 mutant strongly inhibits the differentiation process, suggesting that the antiadipogenic effects of SB203580 and SB202190 are indeed due to inactivation of p38 and are less likely to be due to an effect on other kinases.
The inhibitors used for this study are highly specific, since they have no inhibitory activity on all other kinases tested in vitro, including other MAP kinases, at concentrations greater FIG. 3. p38 inactivation during the first 2-4 days is sufficient to block differentiation. A, morphological appearance. 3T3-L1 cells were incubated with 10 M SB203580 only for the indicated period of time. The inhibitor was then removed, and the cells were differentiated according to the standard protocol. On the eighth day of differentiation, cells were stained with oil red O and analyzed microscopically at 200ϫ magnification. B, induction of adipocyte-specific markers is inhibited by early treatment with SB203580. An identical set of plates as shown in A was lysed and analyzed by SDS-PAGE and Western blotting for constitutive markers (GDI) and for the strongly induced markers (fattyacyl CoA synthase; FACS) and caveolin-1. than 10-fold the concentration we used for treatment of the 3T3-L1 cells (6,7). Specifically, they also have no significant effect (Յ10%) on phosphatidylinositol 3-kinases, a family of enzymes particularly relevant in the context of 3T3-L1 adipocytes. 2 The p38 inhibitors are widely employed to probe p38 function in many different systems (1)(2)(3)(4)(5).
This report is the first demonstration that inhibitors of the p38 MAP kinase interfere with a differentiation process. A recent report (3) describes a role for the p38 mitogen-activated protein kinase pathway in cardiac hypertrophic growth in response to phenylephrine. Although not a genuine differentiation process per se, those observations hint at the pluripotent effects of p38 MAP kinase beyond conventional stress-induced phenomena observed in most cell types. Other systems amenable to a differentiation protocol in tissue culture will have to be tested to establish the generality of this phenomenon.
p38 MAP kinase was first identified by Han et al. (11) in studies demonstrating phosphorylation of a 38-kDa protein in cells stimulated with lipopolysacharides. This kinase was found to be the mouse homolog of the S. cerevisiae HOG1, a MAP kinase required for growth in conditions of high osmolarity (12). Human homologues of p38 MAP kinase were described by Lee et al. (7). They discovered that an inhibitor (similar to SB202190) prevents lipopolysacharide-stimulated production of inflammatory cytokines. This compound specifically photoaffinity-labeled the human homologs of p38 MAP kinase. Recently, novel members of the p38 family have been identified. The p38 MAP kinase family now consists of p38␣ (7,11), p38␤1, (52), p38␤2 (36), p38␥ (53), and p38␦ (54,55). Although we do not know which isoform(s) of p38 are responsible for promoting 3T3-L1 differentiation, it is unlikely to be p38␦ or p38␥, since p38␦ is not inhibited by SB2020190, and p38␥'s expression seems to be restricted chiefly to skeletal muscle (53). In addition, some studies report that p38␥ is not inhibited by SB203580 (1,36), but that finding is in conflict with the results of another study (53). Northern analysis of 3T3-L1 cells demonstrated the presence of p38␣ throughout differentiation (data not shown). In addition, reverse transcription-polymerase chain reaction with primers capable of amplifying both p38␣ and p38␤ with similar efficiencies produce amplicons corresponding to p38␣ 10 times more frequently than amplicons corresponding to p38␤ when used on first strand cDNA from 3T3-L1 fibroblasts (data not shown). This suggests that p38␣ is the most predominant p38 form in 3T3-L1 fibroblasts. However, we cannot rule out the possibility that p38␤1 or p38␤2 may also play a critical role in 3T3-L1 adipocyte conversion.
We have looked at the effect of p38 activation on a number of known key p38 targets. We show that MAP kinase-activated protein kinase-2 is present and is a p38 target in 3T3-L1 fibroblasts, since its phosphorylation is markedly reduced in the presence of p38 inhibitors (Fig. 2). We have found that another well described p38 target, ATF2, is present in this cell line and is a target for p38 (data not shown). The relative contributions of these and other, yet unknown, p38 targets toward adipogenesis remain to be determined. CHOP (GADD153) (56) is a member of the C/EBP family of transcription factors. Ectopic overexpression of CHOP inhibits adipogenesis, most likely by acting as a dominant negative inhibitor of C/EBP dimer-activated transcription of gene targets (4,18,57). A report by Wang and Ron (4) demonstrates that CHOP is phosphorylated much more efficiently by p38 MAP kinase than by extracellular signal-regulated kinase 2 or c-Jun N-terminal kinase. CHOP is phosphorylated by p38 MAP kinase on serine residues 78 and 81 in vitro and in vivo, and 2 U. Siddhanta and J. Backer, personal communication.

FIG. 4. An inducible dominant negative mutant of p38 MAP kinase inhibits adipogenesis.
A, induction of a dominant negative p38 mutant. 3T3-L1 cells were stably transfected with plasmid pOPI3 carrying the cDNA for the dominant negative p38. Plasmid pCMVLacI was co-transfected, and stable transformants were isolated in the presence of hygromycin. Subconfluent cells were then treated overnight with or without 5 mM IPTG and lysed according to the standard protocol. 50 g of protein lysates were analyzed by Western blotting for the presence of the FLAG-tagged mutant p38. B, the presence of a dominant negative p38 mutant inhibits adipogenesis. The cell line described in A and a control cell line carrying the identical set plasmids lacking the cDNA for the mutant p38 were differentiated either in the presence or absence of 5 mM IPTG according to the standard differentiation protocol. IPTG treatment was started 2 days prior to the addition of DM1. After 8 days of differentiation, cells were fixed and analyzed microscopically (magnification, ϫ 100). C, the presence of a dominant negative p38 mutant prevents induction of fatty acyl-CoA synthase, an adipocyte-specific marker. Cells carrying an inducible dominant negative p38 mutant were differentiated either in the presence or absence of IPTG. Lysates were prepared after 8 days of differentiation and analyzed by Western blotting for the presence of fatty acyl-CoA synthase (FACS). To control for equal protein loading, the same blot was probed for GDI. treatment with SB203580 blocks these phosphorylation events (4). Furthermore, mutation of residues 78 and 81 to alanine moderately mitigated the ability of CHOP overexpression to inhibit adipogenesis. However, our study suggests that when differentiated under endogenous conditions (i.e. without ectopic overexpression of CHOP), 3T3-L1 cells require p38 MAP kinase for proper adipocyte conversion. A sequence alignment of the C/EBP family revealed that the p38 MAP kinase recognition motif is conserved among all members of the family. Interestingly, another study demonstrated that the threonine in C/EBP␤ corresponding to serine 81 of CHOP is essential for its transcriptional activity (48). Here, we demonstrate that p38 can directly phosphorylate C/EBP␤ in vitro and that the amount of phosphorylated C/EBP␤ is decreased in 3T3-L1 cells incubated in the presence of SB203580. We analyzed C/EBP␤ because it is expressed early in differentiation at a time when p38 MAP kinase is active and because it strongly promotes adipocyte conversion. It is quite possible that p38 MAP kinase phosphorylates additional C/EBP factors to promote adipogenesis.
In addition, phosphorylation of C/EBP␤ by p38 MAP kinase may partially explain the previous observation that SB203580 blocked TNF␣-induced interleukin-6 synthesis (2). C/EBP␤ was initially termed NF-IL6 because it was identified as a transcription factor that bound to the interleukin-1-responsive element in the promoter of the interleukin-6 gene. Thus, in the presence of SB203580, it may not activate transcription of interleukin-6 in response to TNF␣ because it requires phosphorylation by p38 MAP kinase for activity.
Our results suggest that the p38 MAP kinase pathway is required for adipocyte differentiation. There have been contradictory reports in the literature concerning the importance of the p42/44 MAP kinase pathway. Sale et al. (58) reported that 3T3-L1 cells treated with antisense oligonucleotides to p42/44 MAP kinase depleted p42/44 MAP kinases and severely blocked adipocyte differentiation. While the experiments described in this paper were ongoing, Font de Mora et al. (59) published a report that sharply contrasts with the findings reported by Sale et al. (58). They demonstrated that PD98059 (60 -63), a selective MEK (and therefore p42/44 MAP kinase) inhibitor, does not inhibit differen- tiation. Furthermore, they show that cells overexpressing hyperactive MEK1 and MAPK do not differentiate well. We also found that treatment of 3T3-L1 cells with PD98059 has little effect on the ability of 3T3-L1 cells to round up and accumulate lipid (data not shown). Proteins harvested from day 8 3T3-L1 cells that were differentiated in the presence of the inhibitor were compared with protein lysates from control 3T3-L1 adipocytes. There were no significant differences in the accumulation of proteins normally induced during adipogenesis. Thus, while PD98059 effectively blocks activation of p42/44 MAP kinase in 3T3-L1 cells (data not shown), it does not affect 3T3-L1 adipogenesis per se. However, while the level of lipid accumulation was not affected in the presence of the inhibitor, we noticed that PD98059-treated cells were significantly bigger. This is consistent with a recent observation by Porras and colleagues (64), who reported that p42/44 activation is required for insulin-induced proliferation in rat fetal brown adipocytes. A possible explanation for the increased cell size that we observe in our 3T3-L1 system is that the MEK inhibitor inhibits the proliferative phase during differentiation, resulting in an increased cell volume. Additionally, we demonstrate that p42/44 MAP kinase activity sharply declines by day 4 of differentiation ( Fig. 1). It is possible that the inability of the hyperactive MEK1 and extracellular signal-regulated kinases (p42/44) overexpressing 3T3-L1 cell lines to differentiate (59) (an observation also made in the study by Porras et al. (64) in the rat fetal brown adipocyte system) may be partially due to the inappropriate expression of these kinases during the late stages of differentiation.
TNF␣ has a strong inhibitory effect on adipocyte differentiation and can even induce dedifferentiation of fully mature adipocytes (65). The inhibition of p42/44 MAP kinase allows cells to differentiate even in the presence of TNF␣, suggesting that TNF␣'s antiadipogenic effect requires activation of the p42/44 MAP kinase pathway (59). Due to the strong antiadipogenic effect of the p38 MAP kinase inhibitors, the analogous experiment cannot be performed with the these inhibitors. However, we can take advantage of the fact that the addition of p38 inhibitors after day 4 of differentiation does not significantly inhibit adipogenesis. TNF␣, however, even added after day 4 of differentiation, can still induce a marked inhibition of the adipocyte phenotype. When both p38 MAP kinase inhibitors and TNF␣ are added in combination on the fourth day of differentiation, TNF␣ is equally effective in blocking the differentiation process (data not shown), suggesting that the signal transduction cascades leading to the antiadipogenic effects of TNF␣ do not employ p38 MAP kinase. This explains the apparently conflicting observations that while p38 activation is an essential step for adipogenesis and TNF␣ can induce p38 activation in confluent fibroblasts, TNF␣ has a potent antiadipogenic activity. Future studies will have to more closely examine the role of the components of the various MAP kinase pathways and their cross-talk in differentiating as well as in mature adipocytes.
We have also determined if active p38 may simply reflect a cellular measuring gauge for confluency. Contact inhibition is absolutely required for the differentiation program to occur. However, we do not find significant differences between active p38 levels in preconfluent versus confluent 3T3-L1 fibroblasts (data not shown), thereby excluding this possibility. Thus, active p38 appears to be necessary for differentiation but does not seem to trigger it. It is noteworthy though that NIH-3T3 fibroblasts have significantly lower active p38 levels (approximately 3-4-fold) in comparison with 3T3-L1 cells (data not shown). Elevated active p38 MAP kinase levels could therefore be one of the distinguishing features of the 3T3-L1 cell line that predisposes these cells toward adipogenesis.