Regulating Adipogenesis*
- From the Department of Biological Chemistry, Johns Hopkins University School of Medicine, Baltimore, Maryland 21205
- § Supported by Research Grant DK 38418 from the NIDDK, National Institutes of Health. To whom correspondence should be addressed: Dept. of Biological Chemistry, The Johns Hopkins University School of Medicine, 512 Wood Basic Science Bldg., 725 N. Wolfe St., Baltimore, MD 21205. Tel.: 410-955-3554; Fax: 410-955-0903; E-mail: DAN.LANE{at}QMAIL.BS.JHU.EDU.
INTRODUCTION
The raison d'etre of the adipocyte is to store energy (in the form of triacylglycerol) for use during periods of caloric insufficiency. Adipocytes first appear late in fetal development preparatory to postnatal life when a substantial energy reserve is needed to survive periods of fasting. Considerable progress has been made during the past few years in our understanding of the adipocyte differentiation program. This review will focus on the roles of the CAAT/enhancer binding protein (C/EBP)1 and peroxisome proliferator-activated receptor (PPAR) families of transcription factors in the differentiation program. The reader is referred to other recent reviews (,1–3) for references too numerous to include in this Minireview.
Our understanding of adipocyte differentiation derives largely from studies with preadipose cell lines in culture, notably the C3H10T1/2 and NIH 3T3 fibroblastic cell lines and the 3T3-L1 and 3T3-F442A preadipocyte lines (1). Treatment of multipotent C3H10T1/2 cells with 5-azacytidine gives rise to cells committed to the myogenic, adipogenic, or chondrogenic lineages. This is consistent with the view that the adipose lineage arises from the same multipotent stem cell population of mesodermal origin that gives rise to the muscle and cartilage lineages (1). When appropriately induced with hormonal agents (e.g. glucocorticoid, insulin-like growth factor-1, and cyclic AMP or factors that mimic these agents) committed preadipocytes differentiate into adipocytes in culture. A large body of evidence shows that differentiation of 3T3 preadipocytes faithfully mimics the in vivo process giving rise to cells that possess virtually all of the biochemical and morphological characteristics of adipocytes (2). Following hormonal induction, confluent preadipocytes undergo mitotic clonal expansion, become growth arrested, and then coordinately express adipocyte gene products (1).
Several transcription factors have been identified, which act cooperatively and sequentially to trigger the terminal differentiation program (3–5). These include members of the C/EBP and PPAR families. The sequence of expression of certain of these transcription factors during differentiation is outlined in Fig. 1. It should be noted that these patterns differ somewhat depending upon the differentiation protocol and preadipocyte cell line employed, e.g. 3T3-L1 versus 3T3-F442A versus NIH 3T3 cells.
Expression of C/EBP and PPAR family members during differentiation of 3T3-L1 preadipocytes. The magnitude of expression is indicated by the intensity of the horizontal band corresponding to each transcription factor. The differentiation protocol is as described (4).
The C/EBP Family
Three members of the C/EBP family of transcription factors, i.e. C/EBPα, C/EBPβ, and C/EBPδ, have been implicated in the induction of adipocyte differentiation. These transcription factors possess C-terminal basic region/leucine zipper (bZIP) domains, which confer DNA binding ability (basic region) and the ability to dimerize (leucine zipper), either as homodimers or as heterodimers, with other family members. The structural features of the C/EBPs and cis-elements to which they bind have been reviewed (1 2).
C/EBPα
Convincing evidence shows that C/EBPα plays an important role in the transcriptional activation of adipose differentiation (1 2). C/EBPα is expressed just before the transcription of most adipocyte-specific genes is initiated. Moreover, the proximal promoters of many of these genes contain C/EBP binding sites, which have been shown to mediate the activation of reporter gene expression. Attempts to determine whether constitutive expression of C/EBPα could initiate the adipocyte differentiation program were complicated by the fact that C/EBPα blocks mitosis of preadipocytes. However, by placing the C/EBPα gene under the control of the Lac repressor (6) or using a retroviral vector (7), it was demonstrated that expression of C/EBPα is sufficient to induce differentiation of 3T3-L1 preadipocytes into adipocytes without the use of external hormonal inducers. Definitive proof that C/EBPα is required for adipocyte differentiation was obtained by showing that expression of antisense C/EBPα RNA in 3T3-L1 preadipocytes prevented differentiation (8). Consistent with this finding, disruption of the C/EBPα gene gave rise to mice that failed to develop white adipose tissue (9). Taken together these findings proved that C/EBPα is both required and sufficient to induce adipocyte differentiation.
In view of its anti-mitotic activity, C/EBPα may serve to terminate the mitotic clonal expansion that occurs early in the differentiation program. The point in the program at which mitosis ceases coincides with the onset of expression of C/EBPα and the transcriptional activation of adipocyte genes. Ectopic expression of C/EBPα was recently shown to stimulate the expression of two growth arrest-associated proteins, GADD45 (10) and p21(WAF-1/CIP-1/SDI-1) (11), the latter of which is a cyclin-dependent kinase inhibitor directly involved in growth arrest.
Another probable function of C/EBPα in the differentiation program is maintenance of the terminally differentiated state through autoactivation of its own gene. Several lines of evidence support this role. The murine C/EBPα gene promoter possesses a specific C/EBP binding site, which can mediate activation of reporter gene transcription by either C/EBPα or C/EBPβ (12 13). Strong support for autoactivation by C/EBPα derives from the fact that expression of C/EBPα antisense RNA not only blocks expression of the C/EBPα but also the transcription of the C/EBPα gene (8). Furthermore, inducible or ectopic expression of C/EBPα is sufficient to activate expression of the endogenous C/EBPα gene (6 7). It is also possible that C/EBPβ, which is expressed much earlier in the differentiation program than C/EBPα, transcriptionally activates the C/EBPα gene through the C/EBP binding site (see below).
C/EBPβ and C/EBPδ
An increasing body of evidence indicates that C/EBPβ and possibly C/EBPδ function early (i.e. prior to C/EBPα) as transcriptional activators in the sequence of events leading to adipocyte differentiation (Fig. 1). Induction of differentiation by treating growth-arrested 3T3-L1 preadipocytes with exogenous hormonal inducers provokes an immediate, albeit transient, burst of expression of C/EBPβ and C/EBPδ (3). Concomitantly, the preadipocytes re-enter the cell cycle undergoing several rounds of mitotic clonal expansion and then again become growth-arrested as expression of C/EBPα and terminal differentiation ensues. Two of the differentiation inducers appear to be responsible for activating expression of C/EBPβ and C/EBPδ, i.e. cAMP for C/EBPβ and glucocorticoid for C/EBPδ (14). It should be noted that both C/EBPβ and C/EBPδ are also expressed at high levels in actively dividing preadipocytes (in the absence of exogenous differentiation inducers), and upon achieving growth arrest at confluence, expression of both decreases.2 Thus, it is uncertain whether expression of C/EBPβ and C/EBPδ following “induction” is the cause or the effect of re-entry into the cell cycle. Nevertheless, it is clear that conditions, which cause expression of either of these isoforms, accelerate adipogenesis in 3T3-L1 preadipocytes. In addition to the differentiation-associated induction of expression of C/EBPβ described above, this transcription factor may also be regulated by post-translational modification (Ref. ,15 and references therein) and by interaction with molecular chaperones (16). Whether these modifications and structural changes play a role in the activation of C/EBPβ during adipocyte differentiation is not known.
Ectopic expression of liver activator protein, the activating isoform of C/EBPβ, accelerates differentiation, whereas ectopic overexpression of liver inhibitory protein, the truncated dominant-negative isoform of C/EBPβ (which lacks the transactivation domain), blocks differentiation initiated by exogenous inducers (14). The fact that liver inhibitory protein interrupts the differentiation program at a point prior to the induction of C/EBPα suggests that it acts at an early stage, most likely by forming inactive heterodimers with liver activator protein. Ectopic expression of C/EBPβ in NIH 3T3 fibroblasts, which normally do not respond to the exogenous hormonal inducers, appears to cause commitment to the adipose lineage. When subsequently stimulated with hormonal inducers, the “committed” adipoblasts undergo differentiation into adipocytes (14 17 18). Curiously, however, PPARγ but not C/EBPα was detected in these cells (14 17), suggesting that in these cells, overexpression of C/EBPβ may substitute for C/EBPα as a transactivator for some adipocyte genes.
Given that C/EBPβ is an early participant in the differentiation program, the question of its site of action is raised. The results above suggest that C/EBPβ may be involved in the induction of PPARγ (see below). Since the C/EBPα gene promoter contains a C/EBP binding site (12 13) that can mediate transactivation by members of the C/EBP family (13),2 it is possible that C/EBPβ acts as a transcriptional activator of the C/EBPα gene. The role of C/EBPδ, which has much lower adipogenic potential than C/EBPβ, remains uncertain, but recent results suggest that it may act synergistically with C/EBPα in the induction of PPARγ (18).
CHOP-10/GADD153
Another member of the C/EBP family of transcription factors, CHOP-10 (GADD153), heterodimerizes avidly with other C/EBP isoforms, notably C/EBPα and C/EBPβ. However, the basic region responsible for DNA binding deviates considerably in sequence from that of other C/EBP proteins; hence, CHOP-C/EBP dimers bind only to a subset of C/EBP sites (19). Thus, CHOP-10 acts as a dominant-negative inhibitor of transcription activated by C/EBP dimers. Nevertheless, CHOP-10 has the capacity to activate transcription by binding to other response elements. When ectopically expressed in 3T3-L1 preadipocytes, CHOP-10 inhibits adipocyte differentiation. Expression of CHOP-10 by preadipocytes in culture can be induced by cellular stresses including glucose starvation (20). Such perturbations or the ectopic expression of CHOP-10 inhibit preadipocyte differentiation (20). Thus, although CHOP-10 has the capacity to abort or forestall adipocyte differentiation under certain stress conditions, it is uncertain whether CHOP-10 functions in this manner under normal circumstances. CHOP-10 might have a metabolic function in fully differentiated adipocytes. Thus, the dramatic rise in expression of CHOP-10 following glucose depletion may down-regulate the transcription of adipose genes regulated by C/EBPα through inactive heterodimer formation and, thereby, inhibit energy storage.
The PPAR Family
The PPARs, a subgroup of the nuclear hormone receptors, have recently been implicated in the transcriptional activation of adipocyte differentiation. It has been shown that PPAR activators can induce adipocyte differentiation, not only of preadipocytes (21–23) but also myoblasts (24) and multipotent C3H10T1/2 stem cells (25). Moreover ectopic expression of certain PPARs in the presence of activators can induce adipogenesis in NIH 3T3 fibroblasts (26 27). In addition, PPAR response elements (PPREs) have been identified in a number of adipocyte genes known to be transcriptionally activated during adipocyte differentiation. These include genes encoding adipose fatty acid-binding protein (28), phosphoenolpyruvate carboxykinase (29), lipoprotein lipase (30), and stearoyl-CoA desaturase 1 (31).
PPARα, the first member of the PPAR family to be identified, was cloned as an orphan receptor activated by agents that induce peroxisome proliferation, primarily in the liver. Subsequently, PPARδ (also referred to as PPARβ, NUC1, and FAAR) and PPARγ were cloned by low stringency screening (32). These subtypes share a high degree of amino acid sequence similarity, both within their DNA- and ligand-binding domains (33). This is reflected in functional similarities in that the receptors are activated by structurally related compounds and bind as heterodimers with a retinoid X receptor partner to a PPRE (imperfect 6-base pair direct repeats of the sequence RGGTCA spaced by 1 base pair and referred to as DR-1; reviewed in Ref. 34). Nevertheless, the three PPAR subtypes appear to serve different roles in vivo. They exhibit markedly different tissue distributions (35 36), have differing affinities for different PPREs (27), and are activated to different extents by different activators/ligands (26 27 33 37).
Members of the PPAR family are activated by long chain fatty acids, fatty acid metabolites, and certain hypolipidemic drugs and plasticizers (32 37). While a few of these agents have recently been identified as true ligands of one or more of the PPARs, other activators may act indirectly. Certain thiazolidinediones, as well as prostaglandin J2 derivatives (e.g. 15-deoxy-Δ12,14-prostaglandin J2), were shown to be ligands of PPARγ (25 26 38). Leukotriene B4 was identified as a PPARα ligand (39), and recently, carbaprostacyclin was reported to be a ligand of PPARδ and PPARα and 8(S)-HETE to be a ligand of PPARα.3 Notably, all naturally occurring ligands identified to date are derived from arachidonic acid. The diversity of the ligands so far identified suggests that other ligands are likely to be found.
PPARγ
PPARγ is the most adipose-specific of the PPARs. It is expressed at the highest level in adipose tissue and adipocyte cell lines and at low levels, or not at all, in other tissues and cell lines (26 32 35). Two isoforms of PPARγ, i.e. PPARγ1 and PPARγ2, are generated by alternative splicing and alternate translation initiation (28 40 41). Although PPARγ2 appears to be more adipose-specific than PPARγ1 (42), functional differences between the two isoforms have not been detected. Expression of PPARγ is induced concomitantly with (or perhaps preceding (43)) and prior to the transcriptional activation of most adipocyte genes.
PPARα and PPARδ
PPARδ is expressed in many tissues and cell lines including adipose tissue and adipocyte cell lines (33 35 44). There are conflicting reports concerning changes in the expression of PPARδ during adipocyte differentiation, e.g. it has been reported that the level of PPARδ remains relatively constant during differentiation (43)4 and also that its expression is activated early in the differentiation program (,44). These discrepancies may be due to the fact that the various cell lines used in these studies were arrested at different stages of preadipocyte development when they were originally cloned (2) and therefore might differ in their ability to express the different PPAR genes.
It has been reported that the level of PPARα increases during adipocyte differentiation (22). Relative to the high level of PPARα in brown adipose tissue, however, the level in white adipose tissue and adipocyte cell lines is extremely low (33 35) suggesting that this isoform plays a minor role in adipocytes from white adipose tissue. Consistent with this notion, targeted disruption of the mouse PPARα gene revealed that its major role is the regulation of peroxisomal β-oxidation rather than adipose development (45).
Adipogenicity of the Different PPARs
The adipogenic potential of the different PPARs has been assessed by transfection into a variety of cell lines that have the capacity to undergo adipogenesis (26 27 43 44 46). When compared directly by ectopic expression in NIH 3T3 cells, PPARγ was by far the most adipogenic, PPARα was less so, and PPARδ was inactive (27). The finding that PPARγ is the most adipogenic is consistent with its adipose-specific expression and with the fact that thiazolidinediones, which are high affinity ligands of PPARγ, are potent inducers of adipocyte differentiation (25). Furthermore, a prostaglandin derivative (15-deoxy-Δ12,14-prostaglandin J2), which is a ligand of PPARγ, stimulates adipogenesis of the pluripotent C3H10T1/2 cell line (38) and of NIH 3T3 cells when PPARγ is expressed ectopically (26). Precursors of 15-deoxy-Δ12,14-prostaglandin J2 also activate adipocyte differentiation, but to a lesser extent (26 38). However, it has not been determined, as yet, whether this prostaglandin is present in vivo.
Although ectopic expression of PPARδ fails to induce adipogenesis in NIH 3T3 cells, this receptor appears to up-regulate expression of a subset of adipose-specific genes in these and certain other cell types (23 27 44). The fact that PPAR activators stimulate adipogenesis of 3T3-L1 preadipocytes (22) as well as myoblasts (47), which express PPARδ constitutively (but not detectable levels of other PPARs), suggests that PPARδ may have the capacity to initiate adipocyte differentiation in at least some cell culture models. Antisense RNA and/or gene knockout experiments to disrupt expression of PPARδ and PPARγ would be useful to determine whether these isoforms actually play the roles tentatively ascribed to them in adipose determination and differentiation.
Cross-talk with the Retinoic Acid Signaling Pathway?
Retinoic acid (RA) has long been recognized as a potent inhibitor (effective concentration ranging from 3 nM to 10 mM) of adipocyte differentiation in several different cell lines (22 41 48–53). The effects of retinoids are known to be mediated by two types of nuclear hormone receptors, retinoic acid receptors (RARs) and retinoid X receptors (RXRs) (reviewed in Ref. 54). All-trans-RA is a ligand for RAR (Kd in the 0.5 nM range), whereas its isomerization product, 9-cis-RA, is a ligand for both RAR and RXR (55). Studies with RAR- and RXR-specific agonists and antagonists indicate that inhibition of adipocyte differentiation by RA is mediated by an RAR(s) (41 53). Like a number of other non-steroid nuclear hormone receptors, members of the RAR subfamily bind, as heterodimers with an RXR partner, to imperfect repeats of a hexad consensus. Depending upon the spacing between repeats, the RXR-RAR dimer acts either as a RA-dependent transcriptional activator or as a RA-independent repressor (34).
The time window during which RA can inhibit adipocyte differentiation is limited to the period immediately following induction of differentiation with hormonal agents, preceding the expression of C/EBPα (41 50 51). RA does not interfere with the induction of expression of C/EBPβ but blocks the expression of C/EBPα and PPARγ and is able to repress expression of PPARγ even after it is induced (41). The mechanism by which RA inhibits adipocyte differentiation is not well understood. However, recent findings suggest that the inhibition of differentiation is not due to RXR-RAR-mediated stimulation of an inhibitory gene(s). Rather, liganded RAR appears to interfere with transcriptional activation by C/EBPβ and C/EBPα, possibly by competing with these transcription factors for a common and limiting cofactor (56). Thus, it is possible that RA blocks C/EBPβ-activated expression of the PPARγ and C/EBPα genes (see above) and, thereby, differentiation.
Cooperation among Members of the C/EBP and PPAR Families
The induction of adipocyte differentiation involves the cooperative interplay of members of the C/EBP and PPAR families. A model incorporating the best documented of these interactions is presented in Fig. 2. C/EBPβ and C/EBPδ are expressed immediately following hormonal induction of differentiation. Several lines of evidence implicate C/EBPβ in the activation of expression of both PPARγ and C/EBPα. While C/EBPα can also induce expression of PPARγ (56), C/EBPβ is probably the initial inducer since it is expressed before C/EBPα in the differentiation program. In addition to activating the expression of PPARγ, C/EBPβ is believed to be a transactivator of the C/EBPα gene. Once transcription of the C/EBPα gene has been initiated its continued expression is assured through transcriptional autoactivation. This mechanism is believed to be responsible, at least in part, for maintaining expression of adipocyte genes transactivated by C/EBPα in the terminally differentiated state.
Model illustrating the regulatory interplay among transcription factors involved in adipocyte differentiation. Arrows indicate activation of expression either directly or indirectly. Broken lines with arrowheads indicate activations where the mechanism is less well understood. The thickness of arrows depicts the relative importance of the activation. Inhibition by all-trans-retinoic acid mediated by RAR is indicated by a red arrow and an encircled “−” symbol.
C/EBPα and PPARγ appear to act synergistically by triggering the adipocyte differentiation program and reciprocally activating transcription of one another (Fig. 2). This is supported by the finding that ectopic expression of either C/EBPα or PPARγ in myoblasts and NIH 3T3 fibroblasts promotes only partial differentiation, while co-expression of both factors provokes “full” differentiation, i.e. the same rate and extent of differentiation produced by hormonal inducers (27 43 46). Cooperative interaction between C/EBPα and PPARγ is evidenced by the fact that ectopic expression of either transcription factor alone induces expression of the other (43 56). This may involve reciprocal gene activation, i.e. transactivation of the C/EBPα gene by PPARγ and vice versa. It is also of interest that C/EBPα and PPARγ act cooperatively in the transcriptional activation of other adipocyte genes. The 422/aP2 and the phosphoenolpyruvate carboxykinase gene promoters both possess C/EBP and PPAR binding sites, which mediate transactivation by the corresponding transacting factors (5). Thus, there appear to be two levels of cooperative interplay between members of the C/EBP and PPAR families during adipocyte differentiation, i.e. they reciprocally activate expression of one another and subsequently, they activate some adipocyte genes.
Footnotes
-
↵‡ Supported by a grant from the Danish Natural Science Research Council. Present address: Dept. of Molecular Biology, Odense University, Campusvej 55, DK-5320 Odense M, Denmark.
-
↵* This minireview will be reprinted in the 1997 Minireview Compendium, which will be available in December, 1997.
-
↵1 The abbreviations used are:
- C/EBP
-
CAAT/enhancer binding protein
- PPAR
-
peroxisome proliferator-activated receptor
- PPRE
-
PPAR response element
- RA
-
retinoic acid
- RAR
-
retinoic acid receptor
- RXR
-
retinoid X receptor.
-
↵2 P. Cornelius and M. D. Lane, unpublished results.
-
↵3 R. M. Evans, personal communication.
-
↵4 S. Mandrup, unpublished results.
- © 1997 by The American Society for Biochemistry and Molecular Biology, Inc.
