Peroxisome Proliferator Activated Receptor γ, CCAAT/ Enhancer-binding Protein α, and Cell Cycle Status Regulate the Commitment to Adipocyte Differentiation*

Terminal differentiation of stem cells is characterized by cessation of cell proliferation as well as changes in cell morphology associated with the differentiated state. For adipocyte differentiation, independent lines of evidence show that the transcription factors peroxisome proliferator activated receptor γ (PPARγ) and CCAAT/enhancer-binding protein α (C/EBPα) as well as the tumor suppressor retinoblastoma (Rb) protein are essential. How these proteins promote adipocyte conversion and how they function cooperatively during the differentiation process remain unclear. We have used retinoic acid (RA) inhibition of adipogenesis to investigate these issues. RA blocked adipogenesis of 3T3-L1 cells induced to differentiate by ectopic expression of PPARγ and C/EBPα independently or together. However, under these circumstances RA was only effective at preventing adipogenesis when added prior to confluence, suggesting that factors involved in regulation of the cell cycle might play a role in establishing the commitment state of adipogenesis that is insensitive to RA. During differentiation of wild type 3T3 L1 preadipocytes, we found that Rb protein is hyperphosphorylated early in adipogenesis, corresponding to previously quiescent cells re-entering the cell cycle, and later becomes hypophosphorylated. The data suggest that, together with the coexpression of PPARγ and C/EBPα, permanent exit from the cell cycle establishes the irreversible commitment to adipocyte differentiation.

The molecular mechanisms relating to cell proliferation and cell differentiation are inadequately understood. Adipocyte conversion provides an excellent model system to study terminal differentiation. In the case of 3T3-L1 cells, differentiation is induced upon exposure of cells to a mixture of hormonal stimulants including dexamethasone, isobutylmethylxanthine, insulin, and fetal calf serum (1,2). These pharmacological stimuli, or alternatives such as thiazolidinediones or other activators of peroxisome proliferator activated receptors (PPARs) 1 (3)(4)(5)(6), are necessary for adipocyte differentiation of 3T3-L1 cells. During adipocyte conversion, a variety of transcription factors are induced, including C/EBP␤, PPAR␥, and C/EBP␣ (reviewed in Ref. 7). Enforced expression of PPAR␥ (8), C/EBP␣ (9 -11), or C/EBP␤ (10,12,13) stimulates adipogenesis in NIH 3T3 fibroblasts, suggesting the essential roles of these transcription factors in regulating adipogenesis. Furthermore, combined expression of PPAR␥ and C/EBP␣ has synergistic effects on promoting fat cell conversion in myoblasts (14). Therefore, it is likely that PPAR␥ and C/EBP␣ function cooperatively to establish terminal adipocyte differentiation.
In addition to the expression of differentiated marker genes, terminal differentiation is characterized by permanent withdrawal of cells from the cell cycle. One protein that is involved in cell cycle progression is the retinoblastoma susceptibility gene product, Rb (15). Hypophosphorylation of Rb inhibits cell cycle progression, and this inhibitory effect of Rb is lost upon phosphorylation of the protein (16,17). The involvement of Rb in adipocyte differentiation is suggested by the observation that ectopic expression of protein kinase C in quiescent NIH 3T3 cells induces hypophosphorylation of Rb and promotes adipogenesis (18), whereas Rb binding by SV40 large T antigen interferes with adipogenic differentiation (19). Moreover, Rb Ϫ/Ϫ mouse embryonic lung fibroblasts failed to undergo adipocyte differentiation under appropriate conditions, and ectopic expression of Rb restored the adipogenic phenotype of the Rb Ϫ/Ϫ cells (20). Together, these observations suggest an essential role of Rb in adipocyte differentiation.
To explore early events in adipogenesis, we have used retinoic acid (RA), which normally inhibits adipocyte differentiation of 3T3-L1 cells (21)(22)(23). Liganded RAR blocks C/EBPstimulated gene transcription, and RA also prevents adipogenesis due to ectopic expression of C/EBP␣ or -␤ (13). However, during normal adipogenesis RA exerts its inhibitory function only when added in the first 24 -48 h after exposure to differentiating stimuli that are applied postconfluence. The inhibitory function of RA is mediated by RA receptors (RARs), which are down-regulated early in adipocyte differentiation (23). However, ectopic expression of RAR in 3T3 L1 cells only extends the period during which RA is effective in preventing adipogenesis by an additional 24 -48 h. Interestingly, at this time, although PPAR␥ is already induced, the level of PPAR␥ protein is diminished upon RA treatment. This could be explained by the fact that RA/RAR inhibits C/EBP function, which may be responsible for maintaining the level of PPAR during differentiation (23).
We first examined the effect of RA on adipocyte differentiation due to incubation of wild type 3T3-L1 cells with the PPAR␥ ligand BRL49653, as well as in 3T3-L1 cells that ectopically express PPAR␥. RA inhibited PPAR␥ activator-mediated fat cell conversion, suggesting that RA blocks adipogenesis due to endogenous PPAR␥. The inhibitory effects of RA were also dominant over ectopic co-expression of both PPAR␥ and C/EBP␣ when the cells were maintained in RA at the time of gene transduction and thereafter. However, if added after gene transduction but rather at the time of confluency, RA was no longer effective at blocking differentiation of 3T3-L1 cells that ectopically expressed these adipogenic transcription factors. Thus, co-expression of PPAR␥ and C/EBP␣ in cells was not sufficient to commit cells to undergo differentiation in the presence of RA until the cells became postconfluent. Since confluency is associated with withdrawal from the cell cycle, we hypothesized that the stage of irreversible commitment to adipocyte differentiation required exit from the cell cycle as well as the co-expression of both PPAR␥ and C/EBP␣. Indeed, during adipocyte differentiation of wild type 3T3-L1 cells, confluent cells undergo clonal expansion followed by permanent withdrawal from the cell cycle that occurs at about the time RA loses effectiveness in preventing differentiation. During this process Rb protein shifts from a highly phosphorylated state to its hypophosphorylated form. We conclude that the state of RA-resistant commitment to adipocyte differentiation involves not only expression of PPAR␥ and C/EBP␣ but also hypophosphorylation of Rb and withdrawal from the cell cycle.

MATERIALS AND METHODS
Cell Culture and Differentiation-3T3-L1 cells were obtained from the American Type Culture Collection (ATCC, Rockville, MD). Cells were cultured in growth medium (GM) containing 10% iron-enriched fetal bovine serum in Dulbecco's modified Eagle's medium. For standard adipocyte differentiation, 2 days after cells reached confluency (referred as day 0), cells were exposed to differentiation medium (DM) containing 10% fetal bovine serum, 10 g/ml of insulin, 1 M dexamethasone, and 0.5 M isobutylmethylxanthine, for 48 h. Cells then were maintained in postdifferentiation medium containing 10% fetal bovine serum, and 10 g/ml of insulin. RA was dissolved in ethanol and used at a concentration of 10 M. For BRL49653-induced adipogenesis, cells were maintained in GM. 1 M BRL49653 was added to cells on day 0. Cells were exposed to BRL49653 constantly for 7-10 days until fat cells were seen. For experiments involving RA, retrovirally infected cells were studied in the following two protocols: 1) cells were maintained in the constant presence of 10 M RA from the time of infection and then exposed to the presence or absence of 1 M BRL49653 at day 0, or 2) cells were grown in the absence of RA until day 0 and then exposed to RA and various conditions. The first of these two conditions corresponds to that described previously (13).
Construction of Plasmids and Retroviral Infection-pLXSN-PPAR␥2 was generated by insertion of a 1.7-kilobase pair SalI fragment of mouse PPAR␥2 cDNA into the XhoI site of pLXSN (neo r ) (24). pTS13-C/EBP␣ was generated by insertion of a 1.26 kilobase pair EcoRI-BamHI fragment of C/EBP␣ into the BamHI site of the TS13 vector (HgmB r ). Standard calcium phosphate-DNA transfections were performed. To generate retrovirus-producing packaging cells, 293T cells were transfected with 7.5 g of plasmid DNA and viral gag and pol plasmids (25). 48 h post-transfection, filtered viral supernatants from the ecotropic packaging cell line were used to infect 3T3 L1 cells. Two days after infection, cells were selected in G418 (400 g/ml; Life Technologies, Inc.) or hygromycin B (200 g/ml) for 10 -14 days. For double infection, pLXSN-PPAR␥ cells were infected with virus containing TS13-C/EBP␣ for 48 h. Cells then were selected in G418 plus hygromycin growth medium.
Oil Red O Staining-Dishes were washed three times with phosphate-buffered saline, fixed by 10% formalin in phosphate buffer for 1 h at room temperature. After fixation, cells were washed once with phosphate-buffered saline and stained with a filtered oil red O stock solution (0.5 g of oil red O (Sigma) in 100 ml of isopropyl alcohol) for 15 min at room temperature. Cells then were washed twice with water for 15 min each and visualized.
Western (Immunoblot) Analysis-3T3-L1 cells were lysed in cell lysate buffer (500 l for the 10-cm dish and 150 l for the 60-mm dish), and cells were incubated on ice for 30 min, followed by centrifugation at 17,000 rpm at 4°C for 30 min. Supernatant was collected, and protein concentration was determined by Bio-Rad protein assay. 60 -100 g of protein was subjected to 10% polyacrylamide gel electrophoresis. Proteins were then transferred to nitrocellulose membrane, and Ponceau-S (Sigma) staining was performed to verify equal loading/transfer. The membrane was incubated first with primary antibody (anti-PPAR␥, 1:1500 (23); anti-C/EBP␣, 1:300 (Santa Cruz Biotechnology, Inc.)) for 2 h, followed by secondary antibody (horseradish peroxidase-conjugated) for 1 h. Blots were developed using ECL chemiluminescence detection reagent (Amersham Life Sciences, Inc.) and visualized by exposure to autoradiography film.
BrdUrd Labeling Assay-Cells were incubated with BrdUrd (Amersham Corp.) for 24 h and then trypsinized, washed twice in phosphatebuffered saline, and fixed with 70% ethanol on ice for 30 min or longer. Fixed cells were incubated with anti-BrdUrd and then fluorescein isothiocyanate-conjugated secondary antibody. BrdUrd-positive cells were sorted by flow cytometry.
[ 3 H]Thymidine Incorporation Assay-Cells were incubated with 1 Ci/ml of [ 3 H]thymidine for 24 h and then harvested, washed, and resuspended in 0.5 ml of 0.3 N NaOH and incubated on ice for 30 min prior to the addition of 0.5 ml of 20% trichloroacetic acid. Precipitated DNA was filtered through a Whatman GFC filter, washed, dried, and counted in a liquid scintillation counter.

RA Inhibits PPAR␥ Ligand-mediated Adipogenesis in 3T3 L1
Cells-During DM-induced 3T3-L1 differentiation, RA is only effective in preventing adipogenesis when added during the first 24 -48 h (day 0 and day 1), when PPAR␥ has not yet been induced (7). Indeed, under circumstances when PPAR␥ is already expressed, RA can lead to the loss of expression of PPAR␥, consistent with its ability to block the transcriptional activity of C/EBP␤, which induces PPAR␥ during adipogenesis (12,23). To further the use of RA as a tool for understanding the stages of adipocyte differentiation, we first tested the ability of RA to affect the adipogenesis of 3T3-L1 cells that is induced by the PPAR␥2 activator, BRL49653. BRL49653 induces adipogenesis of post-confluent 3T3-L1 cells, typically causing 20 -70% of cells to differentiate into adipocytes within 7-10 days (26). This is presumably caused by the activation of an endogenous low level of PPAR␥ in the preadipocytes (23) and subsequently the activation of the differentiation program and fat cell conversion. These results were confirmed in Fig oil red O staining. RA was able to inhibit BRL49653-induced differentiation (Fig. 1, column 1, row 3). Fig. 2 shows that BRL49653 treatment induced both PPAR␥ and C/EBP␣ and that, consistent with the oil red O staining, this induction was blocked by RA. Note that adipocytes express two forms of C/EBP␣, referred to in this paper as C/EBP␣-LAP and C/EBP␣-LIP using nomenclature derived from similar alternative translation products for C/EBP␤ (27).
RA Inhibits PPAR␥-induced Adipogenesis-Since RA blocked PPAR␥ induction by BRL49653, we next examined whether expression of PPAR␥ would be sufficient to establish insensitivity to RA by testing the efficacy of RA on cells that are ectopically expressing PPAR␥2 (designated as L1-PPAR␥2). Fig. 3A shows that these cells constitutively express PPAR␥2 protein. In a typical experiment, approximately 20 -50% of L1-PPAR␥2 cells spontaneously differentiated into fat cells in the absence of any adipogenic stimuli in 7-10 days postconfluency, due either to a low level of endogenous ligand or constitutive activity of the ectopic PPAR␥ (see Ref. 28 and Fig. 1, column 2, row 1). The addition of the PPAR␥ ligand BRL49563 accelerated and enhanced the differentiation process, with nearly 100% of cells differentiating into fat cells. This is shown by oil red O staining in Fig. 1 (column 2, row 2). In addition, Fig. 3 shows that the adipocyte markers C/EBP␣-LAP and C/EBP␣-LIP were markedly induced (lane 5). However, when maintained in the presence of RA, cells that were grown in GM with or without the PPAR␥ ligand BRL49653 did not convert into fat cells, documented both by oil red O staining (Fig. 1, column 2, row 4) and C/EBP␣ expression (Fig. 3, lanes 6 and 7). The presence of RA had no effect on the ectopic expression of PPAR␥ (Fig. 3, compare lanes 6 and 7 to lanes 4 and 5). These results suggest that RA is able to block adipogenesis induced by ectopic PPAR␥ protein expression, and expression of PPAR␥ is not able to override the inhibitory effect of RA on adipogenesis.
RA Is Ineffective When Added to Postconfluent Cells That Ectopically Express PPAR␥ or C/EBP␣-Above we showed that RA blocks the function of PPAR␥ to induce fat cell conversion, suggesting that additional factor(s) may be necessary to be insensitive to RA. Previously we have shown that adipogenesis in cells that ectopically express either C/EBP␣ or C/EBP␤ is also blocked when cells are maintained in the presence of RA from the onset of ectopic gene expression ( Fig. 1 and Ref. 13). This was confirmed in Fig. 1 (column 3, row 3, where the ectopic C/EBP␣-expressing cells are referred to as L1-C/EBP␣ cells). Interestingly, when RA was added not at the onset of ectopic expression of either PPAR␥ or C/EBP␣, but rather at a later time when cells were confluent, it was no longer effective at blocking differentiation (Fig. 1, row 5). Constitutive expression of C/EBP␣ in the L1-C/EBP␣ cells is demonstrated in Fig.   4 (lanes 5 and 7). The vector used to ectopically express C/EBP␣ does not express the smaller translation product of C/EBP␣, called C/EBP␣-LIP (23). Thus, C/EBP␣-LIP expression serves as a marker of adipogenesis in these experiments (29). Note that C/EBP␣-LIP was undetectable in the L1-C/EBP␣ cells at day 0 (Fig. 4, lane 5), consistent with their lack of adipocyte phenotype. In contrast, continued incubation of the L1-C/EBP␣ cells for 9 days postconfluency did induce adipogenesis ( Fig. 1  (column 3, row 1) and Fig. 4 (note C/EBP-LIP expression in lane 6)). Consistent with the cell morphology shown in Fig. 1, this adipose conversion of L1-C/EBP␣ cells was blocked when cells were grown in media containing RA (Fig. 1, column 3, row  3; Fig. 4, lane 8).
PPAR␥ and C/EBP␣ Are Mutually Regulated-Note that while allowing the cells to reach confluency prior to the addition of RA appeared to prevent the effects of RA on the ectopic C/EBP␣and PPAR␥-expressing cells, RA normally prevents adipocyte differentiation of wild type 3T3-L1 cells. Those cells however, do not express PPAR␥ or C/EBP␣ until they have begun to differentiate. Thus, the state of refractoriness to RA seems to require both confluency and expression of PPAR␥ and C/EBP␣. Indeed, during normal 3T3-L1 differentiation, RA loses effectiveness at times when PPAR␥ and C/EBP␣ are simultaneously expressed in the cells (23). Since C/EBP␣ (13) and C/EBP␤ induce PPAR␥ (12,13), we considered the possibility that PPAR␥ and C/EBP␣ may induce each other, leading after some time to a state that is refractory to RA inhibition of adipogenesis. To determine whether PPAR␥ is able to activate C/EBP␣ gene expression, L1-PPAR␥ cells were collected prior to differentiation and subjected to Western analysis. Fig. 3 shows that a low but detectable level of C/EBP␣ protein was present in day 0 L1-PPAR␥ preadipocytes (lane 3). This is notably different from wild type 3T3-L1 cells, which do not express C/EBP␣ on day 0 (e.g. lane 1). Thus, ectopic PPAR␥ expression induced a low level of expression of C/EBP␣. The C/EBP␣ expression was even greater on day 7, consistent with the adipocyte phenotype of these cells. Interestingly, RA blocked expression of C/EBP␣ despite continued expression of the ectopic PPAR␥, suggesting that RA inhibits PPAR␥ induction of C/EBP␣. This would be consistent with the results of others indicating that liganded RAR can interfere with PPAR-

FIG. 2. RA inhibits BRL-mediated adipogenesis of 3T3 L1 cells.
Day 0 postconfluent 3T3-L1 preadipocytes maintained in the growth media were exposed to 1 M of BRL49653, in the absence or presence of 10 M RA. Cell extracts were prepared on the indicated days and subjected to Western immunoblot to detect the expression of PPAR␥ and C/EBP␣. The arrows indicate the two PPAR␥ isoforms (␥1 and ␥2) as well as the two translation products of C/EBP␣ (LAP and LIP). In a reciprocal experiment we examined the expression of endogenous PPAR␥ in L1-C/EBP␣ cells. Fig. 4 shows that endogenous PPAR␥ was expressed in the day 0 L1-C/EBP␣ preadipocytes (lane 7), whereas expression of PPAR␥ was undetectable in day 0 control preadipocytes (lane 1). Expression of PPAR␥ in the L1-C/EBP␣ cells was abolished by treatment with RA from the time of C/EBP␣ expression (lanes 7 and 8), consistent with the ability of RA to inhibit transcriptional activation by C/EBP␣ (13).

RA Inhibits Adipogenesis of Preadipocytes Coexpressing Ectopic PPAR␥ and C/EBP␣-The above results demonstrate that PPAR␥ and C/EBP␣ can activate each other's expression.
Therefore, it is likely that the inability of RA to prevent fat cell conversion in L1-PPAR␥ or C/EBP␣ cells when added postconfluency is due to the coexpression of PPAR␥ and C/EBP␣. We next tested whether the requirement for confluency could be overcome by forcing cells to express both PPAR␥ and C/EBP␣ at higher levels. For these experiments, we doubly infected 3T3 L1 cells with LXSN-PPAR␥2 (neo r ) and TS13-C/EBP␣ (hygromycin r ). These cells grew slowly, perhaps related to growth suppressive properties of C/EBP␣ in other cell lines (30); cells that express only ectopic C/EBP␣ grew somewhat slowly but closer to normal than those ectopically expressing both PPAR␥ and C/EBP␣, suggesting that PPAR␥ provided a second growthinhibitory signal in 3T3-L1 cells. In any case, the cells coexpressing C/EBP␣ and PPAR␥ spontaneously differentiated into adipocytes during 10 days of selection, even prior to confluency. However, when cells were selected and maintained in the presence of RA, they were able to grow to confluence at normal rates, and no fat cells were observed in the absence or presence of BRL49653 (Fig. 1, column 4, row 1). The ability of RA to prevent differentiation of cells that ectopically express both PPAR␥ and C/EBP␣ suggested that concomitant expression of these two proteins was not sufficient to establish a commitment state to adipogenesis that is no longer responsive to RA.
PPAR␥, C/EBP␣, and Withdrawal from the Cell Cycle Are Necessary for Commitment to Adipogenesis-We demonstrated that RA is able to inhibit 3T3-L1 adipogenesis due to C/EBP␣ and PPAR␥, alone or in combination, provided that RA is added at the time of gene transduction. In contrast, RA did not have an appreciable effect when added to cells postconfluency. Since confluency is associated with withdrawal of cells from the cell cycle and since cell proliferation and differentiation are often mutually exclusive events, we hypothesized that both cell cycle arrest and the expression of adipogenic genes are necessary for the commitment to fat cell differentiation. It is known that standard differentiation medium induces mitosis of quiescent 3T3-L1 cells prior to cell cycle withdrawal and completion of differentiation along adipogenic lineage. Although RA blocks differentiation, it does not prevent the mitosis and resultant increase in cell number due to the adipogenic stimulation (22). Fig. 5A shows the time course of the mitotic response to adipogenic stimulation in the absence and in the presence of RA, using BrdUrd incorporation as a measure of DNA synthesis. In both cases, a major increase in BrdUrd incorporation occurs on days 1 and 2 following adipogenic stimulation. Thus, the cells become postmitotic on day 3 and beyond. These data indicate that, although RA prevented adipocyte differentiation, it did not block the clonal expansion that occurs following adipogenic stimulation, although it appears to be quantitatively reduced. Note that in the presence or absence of RA, the adipogenically stimulated cells return to a quiescent state characterized by few mitotic events. However, these states are fundamentally different as shown in Fig. 5B. The RA-treated cells are quiescent due to contact inhibition, because reseeding them at low density allows them to re-enter the cell cycle. In contrast, Fig.  5B shows that adipocytes differentiated by the standard protocol in the absence of RA are permanently postmitotic and do not divide after reseeding at low density. These results suggest that after the clonal expansion that occurs in the early phase of adipogenesis, cells permanently exit from the cell cycle.
The Phosphorylation State of Rb as a Marker of the Commitment to Adipocyte Differentiation-Interestingly, the time after adipogenic stimulation at which the cells are permanently postmitotic corresponds to the time at which cells become refractory to the effects of RA, i.e. between 48 and 72 h after beginning the differentiation protocol (23). This further suggested the role of cell cycle events in adipocyte differentiation. The Rb protein plays a major regulatory role in the withdrawal of cells from the cell cycle, and recently it was shown that Rb is required for adipogenic differentiation in a less well characterized model involving embryonic lung fibroblasts (20). We therefore examined expression and phosphorylation of Rb protein in 3T3-L1 cells before and during differentiation in the presence and absence of RA. The immunoblot in Fig. 6 shows that confluent, quiescent 3T3-L1 cells (day 0) contained both hyperphosphorylated and hypophosphorylated forms of Rb protein.
The total amount of Rb protein was relatively constant during differentiation. Adipogenic stimulation led to an increase in hyperphosphorylation of Rb during the first 2 days, consistent with the process of clonal expansion that would be aided by hyperphosphorylation and, thereby, inactivation of Rb. As adipogenesis proceeded, as confirmed by expression of PPAR␥ and C/EBP␣, the Rb protein began to shift to its hypophosphorylated form (Fig. 6A, lane 3), corresponding to cells exiting the cell cycle. By day 7, when nearly 100% adipocyte conversion was achieved, Rb protein was almost entirely hypophosphorylated. The stage of RA insensitivity corresponds to the time of withdrawal from the cell cycle in addition to the expression of the adipogenic factors PPAR␥ and C/EBP␣ (day 3) .   Fig. 6B shows that the presence of RA at the time of adipogenic stimulation altered the phosphorylation state of Rb. The degree of hyperphosphorylation on day 1 was reduced, consistent with the reduced percentage of cells that underwent clonal expansion (Fig. 5A). Furthermore, in the presence of RA, Rb was present in both hyper-and hypophosphorylated forms on all days after exposure to adipogenic stimulation. Consistent with the lack of adipogenesis, there was no induction of PPAR␥ or C/EBP␣ in the presence of RA (Fig. 6B). These data suggested a correlation between terminal differentiation and the hypophosphorylation of Rb protein. Together with our earlier demonstration of the role of confluency in determining RA insensitivity of cells that ectopically express both PPAR␥ and C/EBP␣, we conclude that irreversible commitment to adipogenic differentiation requires both the expression of the adipogenic transcription factors and the cell cycle arrest associated with hypophosphorylation of Rb. DISCUSSION We have used the ability of RA to inhibit adipogenesis as a tool to explore the molecular events that occur during the adipogenic differentiation process. In particular, we were interested in pursuing the observation that at early times after exposure of confluent 3T3-L1 cells to adipogenic stimulation, the cells become committed to the differentiation pathway and no longer respond to RA. RA was able to block BRL49653mediated differentiation in wild type 3T3-L1 cells, as well as adipogenesis due to ectopic expression of C/EBP␣, PPAR␥, or both, even in the presence of the PPAR␥ ligand BRL49653. These results strongly suggest that the insensitivity to RA inhibition that occurs during normal differentiation is not due solely to the concomitant expression of PPAR␥ and C/EBP␣. However, we have noted that allowing C/EBP␣or PPAR␥expressing cells to reach confluence reproduces the committed, RA-irreversible state that occurs after about 48 h during wild type 3T3-L1 cell differentiation. Consistent with this notion, we found that during standard 3T3-L1 cell differentiation, at the point at which cells normally become refractory to the effects of RA, cells are not only expressing PPAR␥ and C/EBP␣ but are exiting the cell cycle following clonal expansion.
The ability of RA to block PPAR␥-induced adipocyte differentiation suggests that liganded endogenous RAR can interfere with the function of PPAR␥ in addition to its ability to block transcriptional activity of C/EBPs (13). Indeed, the ability of liganded RAR to block PPAR transactivation can be demonstrated in transfected cells (Ref. 31 and data not shown). However, during normal adipogenesis, RA is effective at times prior to expression of PPAR␥, suggesting that its natural target is C/EBP␤-mediated transcription (13).
We have now shown that RA blocks adipocyte differentiation of confluent 3T3-L1 cells when added prior to expression of adipogenic transcription factors and also prevents adipogenesis of cells that constitutively express C/EBP␣ and PPAR␥ when added prior to confluency. Thus, excluding the situation where RAR is limiting (23), preadipocytes enter a state of commitment to differentiation that is not reversible by RA when PPAR␥ and C/EBP␣ are expressed and growth cessation has occurred. In the absence of RA, cells that ectopically express PPAR␥ and C/EBP␣ undergo adipocyte differentiation before cells reach confluency, indicating that cell-cell contact may not be necessary for adipogenesis to occur. However, these cells grow very slowly, suggesting that expression of adipogenic genes may contribute to the cessation of cell growth that is also required for cells to undergo adipogenic differentiation.
Both PPAR␥ and C/EBP␣ activate several adipocyte-specific genes. Coexpression of PPAR␥ and C/EBP␣ has been shown to have a synergistic effect and strongly promote fat cell differen- tiation of fibroblastic (32) and myogenic cell lines (14). C/EBP␣ has known antimitotic effects, suggesting that its role in promoting differentiation is at least partially related to suppression of cell growth (30). Expression of C/EBP␣ in 3T3-L1 cells reduced their growth rate, and cells that ectopically expressed both PPAR␥ and C/EBP␣ grew extremely slowly and underwent adipocyte conversion even before they reached confluence. However, we found that co-expression of these two proteins was not sufficient to commit cells to undergo differentiation in a manner that was irreversible by RA unless the cells were growtharrested. These observations suggest that growth arrest played a positive role, along with the specific differentiation factors, in determining whether cells proliferate or differentiate.
During the normal differentiation process, we found that Rb protein is first hyperphosphorylated and then hypophosphorylated. This correlated well with the cell cycle status of the cells both during and after clonal expansion. While this work was in progress, another group also showed that Rb protein levels were stable throughout adipogenesis but reported that the Rb protein was hypophosphorylated throughout the differentiation process (33). The discrepancy appears to be due to the use of different Rb antibodies, and using the same antibody as we used, those investigators have more recently found changes in Rb phosphorylation similar to those reported here. 2 The role of Rb in adipogenesis has been predicted by other studies. In addition to being required for the adipogenesis of mouse lung fibroblasts (20), Rb is necessary for myoblastic differentiation (35). Indeed, there is evidence that RB interacts with adipogenic C/EBP proteins (34) as well as with the myogenic differentiation factor Myo D (35). However, the precise role of Rb in differentiation should be further pursued.
It is unlikely that cell cycle arrest is merely a downstream effect of adipogenic differentiation factors because expression of C/EBP and PPAR␥ complement but cannot replace the important effects of cell growth arrest on the ability of 3T3-L1 cells to irreversibly commit to adipocyte differentiation. Taken together, our results suggest that cell cycle arrest is as important as the expression of adipogenic genes in promoting cell differentiation and the irreversible commitment of preadipocytes to this process.