A Role for C/EBPβ in Regulating Peroxisome Proliferator-activated Receptor γ Activity during Adipogenesis in 3T3-L1 Preadipocytes*

The differentiation of 3T3-L1 preadipocytes is regulated in part by a cascade of transcriptional events involving activation of the CCAAT/enhancer-binding proteins (C/EBPs) and peroxisome proliferator-activated receptor γ (PPARγ) by dexamethasone (DEX), 3-isobutyl-1-methylxanthine (MIX), and insulin. In this study, we demonstrate that exposure of 3T3-L1 preadipocytes to DEX and insulin fails to induce adipogenesis as indicated by a lack of C/EBPα, PPARγ2, and adipose protein 2/fatty acid-binding protein expression; however, PPARγ1 is expressed. Treatment of these MIX-deficient cells with a PPARγ ligand, troglitazone, induces C/EBPα expression and rescues the block in adipogenesis. In this regard, we also show that induction of C/EBPα gene expression by troglitazone in C3H10T1/2 cells ectopically expressing PPARγ occurs in the absence of ongoing protein synthesis, suggesting a direct transactivation of the C/EBPα gene by PPARγ. Furthermore, ectopic expression of a dominant negative isoform of C/EBPβ (liver-enriched transcriptional inhibitory protein (LIP)) inhibits the induction of C/EBPα, PPARγ2, and adipose protein 2/fatty acid-binding protein by DEX, MIX, and insulin in 3T3-L1 cells without affecting the induction of PPARγ1 by DEX. Exposure of LIP-expressing preadipocytes to troglitazone along with DEX, MIX, and insulin induces differentiation into adipocytes. Additionally, we show that sustained expression of C/EBPα in these LIP-expressing adipocytes requires constant exposure to troglitazone. Taken together, these observations suggest that inhibition of C/EBPβ activity not only blocks C/EBPα and PPARγ2 expression, but it also renders the preadipocytes dependent on an exogenous PPARγ ligand for their differentiation into adipocytes. We propose, therefore, an additional role for C/EBPβ in regulating PPARγ activity during adipogenesis, and we suggest an alternative means of inducing preadipocyte differentiation that relies on the dexamethasone-associated induction of PPARγ1 expression.

RXR ligands such as 9-cis-retinoic acid as well as ligands for PPAR␥ (13). The latter includes polyunsaturated fatty acids and their derivatives as well as the thiazolidinedione family of insulin sensitizers such as troglitazone (14 -16). Transcription from the PPAR␥ gene has been detected in many tissues in which the ␥1 isoform is the predominant transcript (17). In contrast, transcription from the PPAR␥2 promoter is highly adipose tissue-selective giving rise to abundant production of the PPAR␥2 polypeptide in addition to the more ubiquitous PPAR␥1 isoform (18). Ectopic expression of PPAR␥2 or -␥1 in non-adipogenic fibroblasts under appropriate hormonal conditions results in potent induction of adipocyte differentiation (19).
Various mouse cell lines have been used to delineate the many different processes involved in regulating adipogenesis. Most notable are 3T3-L1 preadipocytes, which can be induced to differentiate into mature fat cells following exposure to a mixture of hormonal inducers including dexamethasone (DEX), isobutylmethylxanthine (MIX), insulin, and FBS. MIX and DEX induce expression of C/EBP␤ and C/EBP␦, respectively, which in turn activate C/EBP␣ and PPAR␥ expression (10,20,21). PPAR␥ and C/EBP␣ are then capable of cross-activating each other as well as governing expression of the mature adipocytic phenotype (22,23). The normal differentiation of preadipocytes in culture does not require addition of an exogenous PPAR␥ ligand. In contrast, non-adipogenic fibroblasts that ectopically express a C/EBP or PPAR␥ require exposure to a potent PPAR␥ ligand to undergo conversion into adipocytes (24,25). Preadipocytes have likely acquired the ability to produce an appropriate ligand of PPAR␥. The molecular mechanisms that regulate production of such molecules are not known. Earlier studies by others (26) have suggested a role for the sterol regulatory element-binding proteins (SREBPs).
We demonstrate that attenuation of C/EBP␤ activity by omitting MIX from the culture medium or ectopically expressing a dominant negative form of C/EBP␤ (LIP) renders 3T3-L1 preadipocytes dependent on an exogenous PPAR␥ ligand for their differentiation into adipocytes. These studies have also uncovered an alternative pathway of adipogenesis, which involves a glucocorticoid-associated induction of PPAR␥1 in the absence of C/EBP␤ activity. Furthermore, activation of PPAR␥1 in the LIP-expressing cells with troglitazone directly activates C/EBP␣ gene expression.
Plasmids and Stable Cell Lines-The BOSC 23 packaging cells (27), pBabe-Puro and pBabe-PPAR␥-Puro retroviral expression vectors (19), were kind gifts of Dr. Bruce Spiegelman (Dana Farber Cancer Institute, Harvard Medical School). The pBabe vector expressing either the LAP or LIP isoforms of C/EBP␤ were constructed by subcloning corresponding PCR products of the C/EBP␤ cDNA (20) into the BamHI and EcoRI sites of the pBabe-puro vector. The LAP PCR fragments were generated using the following primers: c␤Ϫ1 (5Ј-CGCGGATCCCCACCATG-GAAGTGGCCAACTT) and c␤-3 (5Ј-CCGGAATTCGCATCAAGTC-CCGAAACCCGGT), and the LIP PCR fragments were generated using c␤-2 (5Ј-CGCGGATCCCCACCATGGCGGCCGGCTT) and c␤-3 primers. Transfection of BOSC 23 packaging cells and subsequent infection of target cells were performed as described by others (19,27). Infected target cells were selected for 6 -10 days in medium containing 2.0 g/ml puromycin.
Cell Culture-Murine 3T3-L1 preadipocytes were cultured, main-tained, and differentiated as described previously (28,29). Briefly, cells were plated and grown for 2 days post-confluence in DMEM supplemented with 10% calf serum. Differentiation was then induced (Day 0) by changing the medium to DMEM containing 10% FBS, 0.5 mM 3-isobutyl-1-methylxanthine, 1 M dexamethasone, and 1.67 M insulin. After 48 h, cells were maintained in DMEM containing 10% FBS. 3T3-L1 cells expressing either C/EBP␤ LIP or control vector and 10T1/2 cells expressing PPAR␥ were differentiated by the same protocol for 3T3-L1 cells, except growth medium was DMEM containing 10% FBS and 2.0 g of puromycin, and the cells were differentiated and maintained in the presence or absence of 10 M troglitazone, except as noted.
Oil Red O Staining-Oil Red O staining was performed following the procedure described previously (29). The cells were then photographed using phase contrast microscopy.
Preparation of Whole Cell Extracts-At the indicated times, cultured cells grown in 10-cm dishes were rinsed with phosphate-buffered saline (140 mM NaCl, 2.7 mM KCl, 1.5 mM KH 2 PO 4 , 8.1 mM Na 2 HPO 4 , pH 7.4) and then harvested in 1 ml of ice-cold buffer containing 50 mM Tris (pH 7.4), 100 mM NaCl, 1% sodium deoxycholate, 4% Nonidet P-40, 0.4% SDS, 5 M aprotinin, and 50 M leupeptin. Lysates were vortexed for 1 min and centrifuged for 15 min at full speed (13,000 rpm) in a microcentrifuge. Pellets were discarded and supernatants stored at Ϫ80°C. Protein content of supernatants was determined using the BCA kit (Amersham Pharmacia Biotech).
Gel Electrophoresis and Immunoblotting-Proteins were separated in SDS-polyacrylamide (acrylamide from American BioAnalytical) gels as described previously (23) and transferred to polyvinylidene difluoride membrane (Bio-Rad) in 25 mM Tris, 192 mM glycine. After transfer, the membrane was blocked with 4% nonfat dry milk in PBST for 1 h at room temperature. After incubation with the primary antibodies specified above, horseradish peroxidase-conjugated secondary antibodies (Sigma) and an enhanced chemiluminescent (ECL) substrate kit (PerkinElmer Life Sciences) were used for detection.
RNA Analysis-Total RNA was harvested according to the procedure of Chomczynski and Sacchi (30). Cells were lysed in buffer containing 4 M guanidinium isothiocyanate. Lysates were extracted with acid phenol/chloroform, and RNA was precipitated in 50% isopropyl alcohol overnight at Ϫ20°C. Northern blot analysis was performed on 20 g of each sample RNA as described. cDNA probes for C/EBP␣ and PPAR␥ were labeled using Klenow fragment of DNA polymerase I and [␣-32 P[dCTP by random priming.
Preparation of Nuclear Protein Extracts-Nuclear protein extracts were prepared essentially as described. Cells were washed twice with ice-cold phosphate-buffered saline and then lysed in nuclear lysis buffer (10 mM Tris (pH 7.6), 10 mM NaCl, 3 mM MgCl 2 , 0.5% Nonidet P-40). Samples were spun at low speed in a clinical centrifuge. Supernatants were discarded, and nuclei were lysed in nuclear extraction buffer (20 mM HEPES (pH 7.9), 350 mM NaCl, 1.5 mM MgCl 2 , 0.2 mM EDTA (pH 8.0), 25% glycerol). Nuclear extracts were incubated on ice for 15 min and centrifuged at full speed (13,000 rpm) at 4°C. The resulting supernatants were stored at Ϫ80°C. Protein concentrations were determined using the BCA protein assay kit (Amersham Pharmacia Biotech).

RESULTS
To understand the roles of various inducers and the C/EBPs in regulating adipogenesis, we generated 3T3-L1 cell lines expressing vector DNA alone (designated L1-V cells) or the dominant negative C/EBP␤, LIP (designated L1-LIP cells). To determine the effect of the different inducers on expression of PPAR␥ and C/EBP␣, we stimulated L1-V cells to differentiate by exposure to different combinations of insulin, dexamethasone (DEX), and isobutylmethylxanthine (MIX) in the presence or absence of 10 M troglitazone. Cells were maintained according to the procedure described under "Experimental Procedures," and total protein was harvested 4 days after induction. Omission of DEX and/or MIX from the culture medium significantly attenuates differentiation of these preadipocytes into mature fat cells as indicated by the barely detectable expression of C/EBP␣ and aP2 in each case (lanes 1, 3, and 5). Interestingly, PPAR␥1 is abundantly expressed in cells exposed to either DEX or MIX alone but not to the same extent as that when the two inducers are used together (compare lane 7 with lanes 3 and 5). PPAR␥2, however, appears to be more highly expressed in cells exposed to DEX compared with those exposed to MIX (compare lane 5 with lane 3). Activation of PPAR␥1 in these differentiation-compromised cells (cells deprived of DEX or MIX) by exposure to troglitazone resulted in an extensive induction of C/EBP␣ and aP2 (lanes 4 and 6) and a corresponding conversion of the preadipocytes into morphologically distinct adipocytes (data not shown).
Earlier studies by us and others (10,21,29) have shown that one of the mechanisms by which DEX and MIX induce adipogenesis is to enhance expression of C/EBP␤ and C/EBP␦ during the initial few hours of adipogenesis in 3T3-L1 cells, which in turn activate expression of PPAR␥ and C/EBP␣. Fig. 1 suggests that these effectors may in fact be regulating two separate pathways in which DEX alone can induce PPAR␥1 expression, but MIX is required along with DEX for C/EBP␣ expression. Consequently, since MIX has been shown to regulate C/EBP␤ (20), we questioned whether the induction of C/EBP␣ by troglitazone in cells deprived of MIX was due to a corresponding troglitazone-associated induction of C/EBP␤. In the experiment presented in Fig. 2, 3T3-L1 cells were exposed to insulin and DEX in the presence or absence of MIX or troglitazone, and total protein extracts harvested at the indicated times were subjected to Western blot analysis. At 4 h post-induction, C/EBP␤ is expressed in cells treated with the complete set of inducers (lanes 3 and 4), whereas very little C/EBP␤ is produced in the absence of MIX (Fig. 2, lanes 1 and 2). Addition of troglitazone has no significant effect on C/EBP␤ expression in the presence or absence of MIX (compare lane 1 with lane 2, and lane 3 with lane 4). This figure also shows that expression of PPAR␥1 occurs much earlier (at 24 h) in cells exposed to mixture lacking MIX than cells cultured in the complete mixture (Fig. 1, compare lanes 5 and 7). By 72 h after treatment, PPAR␥1 is abundantly expressed in both populations of cells (minus or plus MIX); however, PPAR␥2 and C/EBP␣ are expressed to any significant extent only in cells exposed to DEX and MIX (compare lanes 9 and 11). The presence of troglitazone resulted in an extensive induction of C/EBP␣ and PPAR␥2 in MIX-deprived cells but also enhanced expression of C/EBP␣ in cells exposed to MIX.
The induction of C/EBP␣ by troglitazone suggests that PPAR␥ may be capable of directly transactivating the C/EBP␣ gene. To test this idea, we ectopically expressed PPAR␥2 in C3H10T1/2 mesenchymal stem cells to create a cell line (10T-P␥) whose differentiation into adipocytes was dependent on an exogenous PPAR␥ ligand. The Northern blot in Fig. 3 shows that exposure of these cells to 10 M troglitazone results in the induction of C/EBP␣ mRNA expression that also occurs in the absence of ongoing protein synthesis (ϩ cycloheximide). These data are consistent with the notion that PPAR␥ is interacting directly with the C/EBP␣ gene to enhance C/EBP␣ mRNA production.
To determine if the selective effect of omitting MIX on C/EBP␣ compared with PPAR␥1 expression was primarily due to its role in regulating C/EBP␤, we generated a 3T3-L1 cell line ectopically expressing a dominant negative isoform of C/EBP␤ (designated L1-LIP cells). To gain insight into the effect of LIP and troglitazone on the differentiation of preadipocytes, L1-LIP cells were induced to differentiate as in Fig. 1, where MIX and/or DEX were omitted from the culture medium. Extracts of total protein were then harvested at day 4 and subjected to Western blot analysis. Fig.  5 shows that expression of LIP greatly attenuates adipogenesis under all hormonal conditions as indicated by a lack of C/EBP␣ and aP2 expression (lanes 1, 3, 5, and 7). DEX is capable of enhancing PPAR␥1 expression above the low basal levels pro-  1 and 3). Addition of troglitazone along with DEX induces adipogenesis as indicated by the expression of C/EBP␣, PPAR␥2, and aP2. It is worth noting that DEX has a similar effect on gene expression in the presence or absence of LIP (compare Fig. 5, lanes 3 and   4 with Fig. 1, lanes 3 and 4). In contrast, exposure of LIP cells to MIX and insulin only slightly enhances PPAR␥1 with no PPAR␥2 expression, and when troglitazone is added under these conditions it does not induce adipogenesis (i.e. minimal aP2 and C/EBP␣ expression). This pattern of gene expression differs significantly from that observed in the absence of LIP. Specifically, MIX induces both PPAR␥1 and -2 in the L1-vector cells, and consequently, exposure of these cells to troglitazone promotes adipogenesis (compare Fig. 1, lanes 5 and 6, with Fig.  4, lanes 5 and 6). Taken together, these data are consistent with a model in which DEX is capable of priming the preadipocytes to be responsive to troglitazone even in the absence of C/EBP␤; this likely involves induction of PPAR␥1 expression. MIX, however, is only capable of a similar priming process if C/EBP␤ is actively expressed in the absence of LIP. These data also strongly suggest that inhibiting C/EBP␤ activity blocks production of an endogenous activator of PPAR␥, which renders the 3T3-L1 preadipocytes dependent on an exogenous PPAR␥ ligand for their differentiation into adipocytes.
To gain more insight into the ligand dependence of these LIP-expressing preadipocytes, we analyzed the temporal pattern of gene expression following exposure to troglitazone as well as determining the optimum dose of troglitazone required to induce adipogenesis. In the experiment shown in Fig. 6, confluent L1-LIP cells were exposed to DEX, MIX, and insulin in the presence or absence of troglitazone, and total cellular proteins were subjected to Western blot analysis. The combination of DEX, MIX, and insulin is capable of initiating the early phase of adipogenesis in these LIP-expressing cells as indicated by induction of C/EBP␤ as well as PPAR␥1 (compare lane 2 and 4 with lane 1). Exposure of these cells to troglitazone appears to have no significant effect on this pattern of gene expression during the first 2 days. After this time, however, troglitazone is essential for the induction of C/EBP␣ and aP2 expression. Taken together, the studies shown above demonstrate that culturing LIP cells in troglitazone for 6 days, along with an initial priming with DEX, MIX, and insulin, results in their conversion into adipocytes based on accumulation of lipid droplets in Ͼ95% of the cells (Fig. 4B) and the abundant expression of PPAR␥2, C/EBP␣, and aP2 (Fig. 6). To establish the troglitazone dose dependence of LIP cells, both L1-LIP and L1-V cells were exposed to differentiation medium containing DEX, MIX, insulin, and increasing concentrations of troglitazone. Total protein samples were harvested 6 days later and subjected to Western blot analysis of the indicated proteins. Fig. 7 demonstrates that expression of both C/EBP␣ and PPAR␥2 increased substantially with increasing doses of troglitazone. Expression of aP2 also seemed proportionate to tro- FIG. 3. PPAR␥ induces C/EBP␣ mRNA expression in the absence of ongoing protein synthesis. 10T1/2 cells ectopically expressing PPAR␥2 were exposed to DMEM containing 1 M DEX, 0.5 mM MIX, 1.67 M insulin, and 10% FBS for 48 h. Cells were maintained in 10% FBS for an additional 24 h and then treated with or without cycloheximide (CHX, 5 g/ml), in the presence or absence of 5 M troglitazone (Trog). Total RNA was harvested at the indicated times post-treatment and analyzed by Northern blot for C/EBP␣ and PPAR␥ mRNA expression.

FIG. 4. Ectopic expression of a dominant negative isoform of C/EBP␤ (LIP) inhibits adipogenesis in 3T3-L1 preadipocytes.
A, total proteins harvested from proliferating preadipocytes, which ectopically express either vector (V), LAP, or LIP were subjected to Western blot analysis using an anti-C/EBP␤ antibody. CRM, cross-reacting material. B, confluent L1-vector and L1-LIP cells were induced to differentiate in the presence or absence of troglitazone (Trog). At day 7, cells were fixed and stained for neutral lipids with Oil Red O. glitazone concentration, correlative to the number of cells accumulating lipid droplets (data not shown). LIP expression was unaffected by the PPAR␥ ligand.
The LIP polypeptide retains the C-terminal basic leucine zipper region of the full-length C/EBP␤ protein, and therefore, it can dimerize with other C/EBP␤ isoforms and bind to DNA. It was important, therefore, to determine what effect LIP and/or troglitazone may have on the DNA binding activity of the different C/EBPs during adipogenesis. Consequently, L1-V and L1-LIP cells were treated to differentiate in the presence or absence of troglitazone, and nuclear proteins were harvested at day 1 and day 6. The electrophoretic mobility shift assay presented in Fig. 8 shows binding of nuclear protein complexes to an oligonucleotide corresponding to the C/EBP regulatory element within the promoter of the C/EBP␣ gene (31). The profile and intensity of binding observed at day 1 is very similar in the LIP and vector cells, with one exception. There is a faster migrating species in the LIP cells that likely corresponds to LIP-LIP homodimers (Fig. 8A). Troglitazone has no significant effect on the overall binding activity in either cell line. Fig. 8A also shows that the complexes present in L1-V samples at day 6 migrate with a slightly larger mass than the day 1 complexes, which is probably due to the presence of C/EBP␣. This same shift in migration is observed in the LIP samples following exposure to troglitazone for 6 days. To examine the composition of these DNA-protein complexes, a series of supershift assays were performed using antibodies corresponding to C/EBP␣, C/EBP␤, and C/EBP␦. Fig. 8B, lane 4, demonstrates that a proportion of the complexes expressed at day 6 in L1-V cells consist of C/EBP␣ homodimers. As expected, there is a significant increase in C/EBP␣ binding activity in LIP cells following exposure to troglitazone for 6 days (Fig. 8B, compare lane 12  with lane 8). In fact, the C/EBP␣ binding activity is slightly higher in the LIP cells plus troglitazone compared with that expressed in the vector cells. Furthermore, ectopic expression of LIP has not affected the ability of C/EBP␣ to bind to the C/EBP regulatory element. This figure also shows the existence of LIP-LIP homodimers binding to the C/EBP oligonucleotide since the faster migrating complexes present in the LIP cells can be supershifted selectively with an anti-C/EBP␤ antibody (lanes 6 and 10). C/EBP␦ is minimally expressed at 6 days under all conditions since there is no detectable supershift with an anti-C/EBP␦ antibody. A, at 1 and 6 days after induction (d1 and d6, respectively), nuclear proteins were harvested and subjected to electrophoretic mobility shift assay, as described under "Experimental Procedures." B, supershift analysis of C/EBP DNA binding activity in L1-V and L1-LIP cells induced in the presence or absence of troglitazone. Day 6 nuclear protein samples from L1-V and L1-LIP cells induced to differentiate in the standard inducers, with or without troglitazone, were analyzed by supershift analysis using antibodies to C/EBP␤ (C␤), C/EBP␦ (C␦), C/EBP␣ (C␣), or IgG (Ϫ).
Our observation that LIP cells require an exogenous PPAR␥ ligand such as troglitazone for their complete conversion into adipocytes suggests that they are not capable of producing the endogenous ligand(s) or activators. Alternatively, it is possible that they express PPAR␥1 at levels below a threshold required for activation by the endogenous activator(s). To determine the reason for the ligand dependence of the LIP cells, we induced PPAR␥ and C/EBP␣ to fully differentiated levels by exposing LIP cells to DEX, MIX, insulin, and 10 M troglitazone for 6 days. We then questioned whether these LIP adipocytes still require the exogenous PPAR␥ ligand to maintain normal adipocyte gene expression. This was achieved by withdrawing troglitazone from half the cultures at day 6 and measuring expression of PPAR␥, C/EBP␣, and aP2 in these and a control set of cultures that were maintained in troglitazone for the entire experiment. The Western blot in Fig. 9 shows abundant levels of PPAR␥, C/EBP␣, and aP2 following 7 days of exposure of LIP cells to troglitazone. Withdrawal of the exogenous ligand at day 6, however, results in an extensive dedifferentiation as indicated by a drop in expression of C/EBP␣ and aP2 to virtually undetectable levels by day 10 (4 days of withdrawal). Interestingly, the abundance of both PPAR␥1 and -␥2 remains constant throughout this period even in the absence of troglitazone. Notably, expression of LIP does not increase when troglitazone is withdrawn; on the contrary, it appears to decrease (Fig. 9).

DISCUSSION
The differentiation of 3T3-L1 cells into mature adipocytes requires their exposure to a mixture of hormonal inducers including DEX, MIX, insulin, and FBS. These effectors have been shown to activate a cascade of transcriptional events that culminate in expression of the mature adipocytic phenotype. Most notably, they facilitate the induction of C/EBP␤ and C/EBP␦, which together activate expression of PPAR␥ and C/EBP␣ (10,20,21,29). The data presented in this study suggest an additional role for C/EBP␤, and the effectors that control its expression, in regulating the production of PPAR␥ ligands. These studies further show that adipogenesis can be induced in 3T3-L1 preadipocytes in the absence of C/EBP␤ by exposing the cells to an exogenous PPAR␥ ligand. This alternative mechanism appears to depend on the ability of insulin and DEX to induce PPAR␥1 expression, which is then capable of inducing C/EBP␣ and PPAR␥2 expression following exposure to troglitazone. Induction of C/EBP␣ gene expression in the absence of an exogenous PPAR␥ ligand depends on MIX and/or C/EBP␤ expression. It appears, therefore, that expression of C/EBP␣ during adipogenesis can be regulated by at least two independent pathways. One pathway involves a MIXassociated induction of C/EBP␤, which transactivates a C/EBP regulatory element within the promoter of the C/EBP␣ gene (32). The other mechanism can occur in the absence of C/EBP␤ due to a DEX-associated induction of PPAR␥1, which is also capable of transactivating the C/EBP␣ gene in the presence of troglitazone.
Previous studies have shown that an important role for MIX and DEX is to induce expression of C/EBP␤ and C/EBP␦, respectively, which in turn activate C/EBP␣ and PPAR␥2 expression through C/EBP regulatory elements in the promoters of the corresponding genes (10,(31)(32)(33)(34). Our data are consistent with these observations since inhibition of C/EBP␤ activity by either omitting MIX or expressing LIP blocks both C/EBP␣ and PPAR␥2 expression. Of interest is the observation that DEX can induce PPAR␥1 expression in the absence of C/EBP␤ activity, which may be due to a DEX-associated induction of C/EBP␦. Other studies, however, have shown that ectopic expression of C/EBP␦ alone in 3T3 fibroblasts does not induce C/EBP␣, PPAR␥1, or -␥2 expression (10,29). Despite the abundant expression of PPAR␥1 in the C/EBP␤-deficient preadipocytes, they are incapable of expressing the adipogenic program unless exposed to an exogenous PPAR␥ ligand (i.e. troglitazone) along with the normal mixture of hormonal inducers. This observation suggests that C/EBP␤ may play a role in regulating processes that lead to production of PPAR␥ ligands/ activators. Further support for this notion are the data in Fig.  9 showing a continuing requirement of the LIP cells for troglitazone to maintain adipogenic gene expression even after they have completely converted into adipocytes by a 6 -day exposure to the PPAR␥ ligand.
Mechanisms by Which C/EBP␤ May Regulate PPAR␥ Activity-The most likely determinant of PPAR␥ activity is the availability of ligands within the preadipocyte. Even though the natural cellular ligand for PPAR␥ has not been identified, evidence suggests that derivatives of polyunsaturated fatty acids are potent activators of PPAR␥ both in in vitro assays as FIG. 10. Transcriptional control of PPAR␥ ligand production is a central component of the signaling cascade that regulates adipogenesis. Initiation of adipogenesis involves induction of C/EBP␤, C/EBP␦, and PPAR␥1 in response to exposure of preadipocytes to a variety of physiological effectors including insulin, glucocorticoids, and agonists that elevate cAMP. C/EBP␤ and C/EBP␦ activate expression of C/EBP␣ and PPAR␥2 as well as stimulate a pathway that leads to production of PPAR␥ ligands. Ligand-activated forms of PPAR␥1 and PPAR␥2 can directly induce expression of C/EBP␣ to establish a positive feedback loop in which C/EBP␣ maintains expression of the PPARs. The synergistic activity of C/EBP␣ and PPAR␥ ensures expression of the entire adipogenic gene program.
FIG. 9. L1-LIP cells require exposure to troglitazone throughout the differentiation process in order to maintain adipogenic gene expression. L1-LIP cells were induced to differentiate and maintained in the presence of 10 M troglitazone (Trog). Six days after induction, troglitazone was either withdrawn (WD) from the media or maintained at a concentration of 10 M (ϩ). Total protein extracts were harvested at 1 (day 7, d7), 2 (day 8, d8), 3 (day 9, d9), and 4 (day 10, d10) days later. Samples were subjected to Western blot analysis for PPAR␥, C/EBP␣, aP2, and LIP expression. well as in vivo (13,16) Mechanisms that control the cellular production of polyunsaturated fatty acids or their derivatives may play an important role in regulating adipogenesis. In this regard, studies have shown that ADD1/SREBP-1, a transcription factor that is linked to processes controlling fatty acid production, appears to be involved in the production of endogenous PPAR␥ ligands (26). In fact, ADD1/SREBP-1 is induced early during adipogenesis, and its ectopic expression in nonadipogenic cells can enhance fat cell formation by directly activating the PPAR␥2 gene as well as stimulating production of PPAR␥ ligands (26,35,36). It is conceivable, therefore, that a role for C/EBP␤ in regulating PPAR␥ activity may involve induction and/or activation of ADD1/SREBP-1.
PPAR␥ Ligand-dependent Induction of C/EBP␣ Expression-Several investigations have demonstrated that activation of PPAR␥ in a variety of different fibroblast lines results in expression of many adipogenic genes including C/EBP␣ (19,(22)(23)(24)(25)37). Similarly, ectopic expression of C/EBP␣ in nonadipogenic cells can induce PPAR␥ expression (22,23). In fact, Spiegelman and co-workers (22) have suggested that crossregulation between C/EBP␣ and PPAR␥ is important in maintaining the differentiated state. The molecular mechanisms involved in such a cross-regulatory process are not known. It is very likely that C/EBP␣ directly transactivates PPAR␥2 gene expression through the C/EBP regulatory elements within the PPAR␥2 promoter (33,34). The data presented in Fig. 3 suggest that PPAR␥ may also be capable of a similar direct transactivation of the C/EBP␣ gene based on the observation that activation of PPAR␥ by troglitazone in 10T1/2 cells induces C/EBP␣ mRNA expression in the absence of ongoing protein synthesis. For most PPAR␥ target genes, PPAR␥ initiates transcription by binding to cognate PPAR regulatory elements at DR-1 sites within the promoter/enhancer regions of the genes (18,38). It seems likely that similar DR-1 sites exist within the C/EBP␣ gene. Analysis of sequences in the 5Ј-flanking region of the C/EBP␣ gene have identified DR-1 elements that bind strongly to particular COUP-TF proteins, but very weakly to PPAR␥. 2 We are presently determining whether this or any other elements facilitate the PPAR␥-dependent induction of C/EBP␣ gene expression.
What Are the Roles of C/EBP␤ and/or C/EBP␦ in Regulating Adipogenesis?-Several studies performed in a variety of cultured cell systems have led to a model for the transcriptional control of adipogenesis, which involves the sequential activation of C/EBPs and PPAR␥. The function of C/EBP␤ and C/EBP␦ in this process is to induce expression of both PPAR␥2 and C/EBP␣. Investigations using mice lacking C/EBP␤ and/or C/EBP␦ suggest an alternative role for these C/EBPs in regulating adipose tissue formation and function in the animal (39). Specifically, C/EBP␤(Ϫ/Ϫ)⅐C/EBP␦(Ϫ/Ϫ) mice express defects in lipid accumulation despite normal expression of C/EBP␣ and PPAR␥. However, primary embryonic fibroblasts derived from these knock-out animals have lost the potential to undergo adipogenesis and, in so doing, do not express PPAR␥ or C/EBP␣ in response to DEX, MIX, insulin, and FBS. Taken together, these observations suggest a role for C/EBP␤ and/or C/EBP␦ in regulating PPAR␥ and C/EBP␣ gene expression in cultured cells (cell lines or mouse embryonic fibroblasts) exposed to a restricted set of inducers. In preadipocytes in adipose tissue, however, PPAR␥ and C/EBP␣ may be activated by an alternative mechanism due to the presence of effectors not present in the culture system. As mentioned above, there appears to be a role for C/EBP␤ and C/EBP␦ in facilitating the formation and function of adipose tissue in vivo that is independent of C/EBP␣ and PPAR␥ expression, which may involve production of PPAR␥ ligands.
In summary, we propose an alternative model for the transcriptional control of adipogenesis (Fig. 10) that incorporates the conclusions drawn from these studies with those already presented by others (10,21,22). In this model, C/EBP␤ and C/EBP␦ regulate production of PPAR␥ ligands as well as PPAR␥2 and C/EBP␣ expression. Additionally, we suggest that physiological effectors can induce expression of PPAR␥1 in the absence of C/EBP␤ and C/EBP␦ as part of a default pathway. This event can then initiate a cascade of transcription factor expression, commencing with C/EBP␣, which in turn induces expression of the entire adipogenic program providing the preadipocyte is exposed to PPAR␥ ligands. Further dissection of the transcriptional events that regulate production of PPAR␥ ligands should provide a greater understanding of the processes controlling adipogenesis.