| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
PHYSIOLOGICAL REVIEWS Vol. 78 No. 3 July 1998, pp. 783-809
Copyright ©1998 by the American Physiological Society
Department of Nutritional Sciences, University of California, Berkeley, California
I. INTRODUCTION
II. ADIPOSE TISSUE AS A SECRETORY ORGAN
III. ORIGIN OF ADIPOSE CELLS AND ADIPOSE TISSUE
IV. IN VITRO MODELS OF ADIPOCYTE DIFFERENTIATION
V. PROCESS OF ADIPOCYTE DIFFERENTIATION
A. Growth Arrest
B. Clonal Expansion
C. Early Changes in Gene Expression
D. Late Events and Terminal Differentiation
VI. FACTORS THAT MODULATE ADIPOCYTE DIFFERENTIATION
A. Hormones and Signal Transduction Pathways Regulating Adipocyte Differentiation
B. Pref-1, an EGF Repeat-Containing Inhibitor of Adipocyte Differentiation
C. Extracellular Matrix Components
VII. TRANSCRIPTION FACTORS CRITICAL FOR ADIPOCYTE DIFFERENTIATION
A. PPAR Family
B. C/EBP Family
C. bHLH Family
VIII. CONCLUSIONS
REFERENCES
 |
ABSTRACT |
|---|
 Top References
|
|---|
Gregoire, Francine M., Cynthia M. Smas, and Hei Sook Sul. Understanding Adipocyte Differentiation. Physiol. Rev. 78: 783-809, 1998. 
The adipocyte plays a critical role in energy balance. Adipose tissue growth involves an increase in adipocyte size and the formation of new adipocytes from precursor cells. For the last 20 years, the cellular and molecular mechanisms of adipocyte differentiation have been extensively studied using preadipocyte culture systems. Committed preadipocytes undergo growth arrest and subsequent terminal differentiation into adipocytes. This is accompanied by a dramatic increase in expression of adipocyte genes including adipocyte fatty acid binding protein and lipid-metabolizing enzymes. Characterization of regulatory regions of adipose-specific genes has led to the identification of the transcription factors peroxisome proliferator-activated receptor-
(PPAR-) and CCAAT/enhancer binding protein (C/EBP), which play a key role in the complex transcriptional cascade during adipocyte differentiation. Growth and differentiation of preadipocytes is controlled by communication between individual cells or between cells and the extracellular environment. Various hormones and growth factors that affect adipocyte differentiation in a positive or negative manner have been identified. In addition, components involved in cell-cell or cell-matrix interactions such as preadipocyte factor-1 and extracellular matrix proteins are also pivotal in regulating the differentiation process. Identification of these molecules has yielded clues to the biochemical pathways that ultimately result in transcriptional activation via PPAR- and C/EBP. Studies on the regulation of the these transcription factors and the mode of action of various agents that influence adipocyte differentiation will reveal the physiological and pathophysiological mechanisms underlying adipose tissue development.
White adipose tissue (WAT) is the major energy reserve in higher eukaryotes, and storing triacylglycerol in periods of energy excess and its mobilization during energy deprivation are its primary purposes. Recently, there has been a dramatic increase in the incidence of obesity resulting from an excess of WAT. Obesity is a prevalent health hazard in industrialized countries (145, 298) and is closely associated with a number of pathological disorders, including non-insulin-dependent diabetes, hypertension, cancer, gallbladder disease, and atherosclerosis. With regard to this wide range of health implications, the need to develop new and effective strategies in controlling obesity has become more acute. Recently, progress has been made in understanding the process of adipocyte differentiation and in the cloning of genes mutated in several monogenic mouse models of obesity. This has not only allowed us to begin to understand the cellular and molecular basis of adipose tissue growth in physiological and pathophysiological states, but has also provided means to develop therapeutic strategies for the treatment and prevention of obesity. This review addresses regulation of WAT development at the cellular and molecular levels. We summarize information gained from studies in preadipocyte cell lines and attempt to incorporate results from often neglected primary preadipocyte studies. Mature adipocytes, the main cellular component of WAT, are uniquely equipped to function in energy storage and balance under tight hormonal control. However, with the recent realization that adipocytes secrete factors known to play a role in immunological responses, vascular diseases, and appetite regulation, a much more complex and dynamic role of WAT has progressively emerged. Leptin, the obese (ob) gene product, is a hormone that is primarily made and secreted by mature adipocytes and binds to its receptor in the hypothalamus. Studies indicate leptin may function in regulating body fat mass. Loss of fat stores decreases leptin levels and increases neuropeptide Y levels; this leads to increased food intake. Conversely, weight gain increases leptin levels leading to decreased food intake; melanocyte-stimulating hormone is necessary for this response (31, 201). Leptin levels are elevated in human obesity and in animal models of obesity. More recently, the leptin receptor has been detected in peripheral tissues. This suggests additional roles for leptin, including modulation of insulin action in liver (45), production of steroids in the ovary (310), and direct effects on adrenocortical steroidogenesis (23). Leptin also has a role in reproductive physiology (39, 178) and is involved in hematopoietic and immune system development (18, 173). This rapidly growing list, clearly not exhaustive, indicates that leptin has a much broader range of action than initially perceived. Immune system-related proteins produced by adipocytes include adipsin, acylation stimulation protein (ASP), adipocyte complement-related protein (Acrp30/AdipoQ), tumor necrosis factor- Vascular function-related proteins that are secreted by adipocytes include angiotensinogen and plasminogen activator inhibitor type 1 (PAI-1). White adipose tissue contains all the main components of the renin-angiotensin system such as angiotensinogen, angiotensin converting enzyme, angiotensin II, and angiotensin receptors (130). Angiotensinogen could play a role in regulating adipose tissue blood supply and fatty acid efflux from fat (73). Angiotensin II, the cleavage product of angiotensinogen, has been implicated in adipose tissue growth by stimulating production of prostacyclin by mature fat cells and thereby promoting adipocyte differentiation via a paracrine/autocrine mechanism (53). Because angiotensin II increases lipogenesis in both human and 3T3-L1 adipocytes, it may also be involved in the control of adiposity through regulation of lipid synthesis and storage in adipocytes (127). Plasminogen activator inhibitor type 1 is produced by adipose tissue, and treatment of 3T3-L1 adipocytes with transforming growth factor- Taken together, these studies clearly establish that the adipocyte behaves as an endocrine as well as a paracrine/autocrine cell. Along with its active role in regulating energy balance, WAT has the potential to play a dynamic role in a variety of other physiological processes, including the autoregulation of adipose tissue growth and development. The origins of the adipose cell and adipose tissue are still poorly understood, and the molecular events leading to the commitment of the embryonic stem cell precursor to the adipocyte lineage remain to be characterized. In most species, WAT formation begins before birth, as assessed by morphological studies performed on human, pig, mouse, and rat embryos (59, 204, 205, 245). The chronology of WAT appearance, however, is strictly dependent on the species as well as the adipose depot (59, 218, 245). White adipose tissue expansion takes place rapidly after birth as a result of increased fat cell size as well as an increase in fat cell number. Even at the adult stage, the potential to generate new fat cells persists. It has been demonstrated that fat cell number can increase when rats are fed a high-carbohydrate or high-fat diet (67, 68, 176). Increase in fat cell number is also observed in severe human obesity. However, the relative contribution of fat cell size and fat cell number to human adipose tissue growth on nutritional stimulation remains to be clarified. Regardless, early differentiation markers of adipocyte differentiation can be detected even in adipose tissue derived from very old mice (138). Moreover, fat cell precursors isolated from adult WAT of various species, including humans, can be differentiated in vitro into mature adipocytes (21, 58, 98, 104, 160, 213, 264). The potential to acquire new fat cells from fat cell precursors throughout the life span is now undisputed. Although the developmental origin of fat cells is not known, several studies on multipotent clonal cell lines have suggested that the adipocyte lineage derives from an embryonic stem cell precursor with the capacity to differentiate into the mesodermal cell types of adipocytes, chondrocytes, osteoblasts, and myocytes. Treatment of the murine embryonic cell line C3H10T1/2 with a demethylating agent generates loci of muscle, cartilage, and fat cells (143, 265); 25% of colonies contained myofibers, 7% adipocytes, and 1% chondrocytes. The C3H10T1/2 cells may represent multipotential stem cells that are blocked at the mesodermal pathway. That a single gene could convert C3H10T1/2 cells into myoblasts has been demonstrated with the identification and characterization of the MyoD family of regulatory genes (54). The lower clonal frequency of conversion of C3H10T1/2 cells to adipocytes as compared with myocytes may indicate that more genes are required to activate adipocyte development. However, these different frequencies of conversion may also somewhat reflect culture conditions that could selectively promote the differentiation of one cell lineage over another. There is also evidence that a common bone marrow stromal cell type may give rise to adipogenic or osteogenic cells with several indications of a reciprocal relationship in the differentiation of these two cell types (84). For example, bone morphogenetic proteins of the TGF- For the past 20 years, in vitro systems have been extensively used to study adipocyte differentiation. This has led to a dissection of the molecular and cellular events taking place during the transition from undifferentiated fibroblast-like preadipocytes into a mature round fat cells. Various preadipose cell lines and primary cultures of adipose-derived stromal vascular precursor cells have been used. Table 1 summarizes the characteristics of the most commonly employed preadipose cell models. Preadipose cell lines as well as primary preadipocytes are already committed solely to the adipocyte lineage, although they may represent different stages of adipocyte development. The most frequently employed cell lines are 3T3-F442A and 3T3-L1. These were clonally isolated from Swiss 3T3 cells derived from disaggregated 17- to 19-day mouse embryos (89, 90, 92). 3T3-C2 cells derive from the same source but are not preadipocyte in nature. They do not differentiate into adipocytes and can be used to compare the responses of differentiation-defective with differentiation-competent cell types. The TA1 cell line was established by treating CH310T1/2 mouse embryo fibroblast cells with the demethylating agent 5-azacytidine (37, 143, 265). Ob17 cells and their derivatives were generated from adipose precursors present in the epididymal fat pads of genetically obese (ob/ob) adult mice (187). In addition to these models, embryonic stem (ES) cells have recently been shown to differentiate into mature adipocytes in vitro (52) and will likely provide a useful system for investigating the initial steps of adipogenesis. In vitro-differentiated adipocytes have many characteristics of adipose cells in vivo. Subcutaneous injection of preadipose cells in nude mice leads to development of mature fat pads that are histologically indistinguishable from WAT; this is convincing evidence that adipose cell acquisition occurs by a similar mechanism in vivo (91, 278).
Because the stage of differentiation and the lineage of preadipocyte cell lines have not been well established, primary cultures have been particularly useful for validating results obtained in preadipose cell lines. Primary preadipocytes have been successfully cultured from a number of species including humans and have several advantages over preadipose cell lines (96-98, 104, 106, 138, 160, 191, 236). Primary cells are diploid and may therefore reflect the in vivo context better than aneuploid cell lines. Second, they can be derived from adipose tissue obtained from various species at different postnatal stages of development and from various adipose depots (21, 57, 58). The latter is of particular interest, since molecular and biochemical differences in fat pads have been observed (96, 150, 151, 169, 212, 308). Such differences may have major physiological consequences, since excessive centralized fat accumulation (upper body obesity) is associated with increased morbidity in humans (139). On the other hand, the potential cellular heterogeneity of the stromal vascular preadipose cell population is a drawback of primary cultures. This, however, may be a minor concern, since nearly 100% differentiation can be achieved when preadipocytes derived from young animals are adequately treated (57, 97, 236). For primary preadipocyte cultures, differentiation capacity is clearly donor dependent and decreases significantly with age (21, 58, 98, 138). The molecular basis for this reduced differentiation capacity is not known. It may, however, indicate that preadipose cells derived from aged donors represent a subpopulation of precursor cells arrested at distinct stage(s) of adipocyte development. These cells may therefore require a different subset of as yet to be characterized signals to undergo adipocyte differentiation (63, 98). During the growth phase, cells of preadipocyte lines as well as primary preadipocytes are morphologically similar to fibroblasts. At confluence, induction of differentiation by appropriate treatment leads to drastic cell shape changes. The preadipocyte converts to a spherical shape, accumulates lipid droplets, and progressively acquires the morphological and biochemical characteristics of the mature white adipocyte. The nature of the induction is dependent on the specific cell culture model system employed. A variety of differentiation protocols have been developed for preadipose cell lines and for primary preadipocytes. Embryonic stem cells can be differentiated into adipocytes at high efficiency in vitro if early-developing ES cell-derived embryoid bodies are exposed to retinoic acid (RA) for a precise time, followed by treatment with standard adipogenic hormones (52). 3T3-L1 preadipocytes spontaneously differentiate over a period of several weeks into fat-cell clusters when maintained in culture with fetal calf serum. This can be accelerated by the inducing agents dexamethasone and methylisobutylxanthine (MIX), a phosphodiesterase inhibitor. High concentrations of insulin have been used in combination with these inducing agents. The identification of agents that accelerate differentiation, but that are not required to maintain the differentiated phenotype, provides insights to the biochemical pathways that may function during adipocyte differentiation. Although preadipocytes from various sources are similar in many regards, their responsiveness to inducing agents varies considerably. This may represent differences in the stage of maturation at which the preadipocytes were originally harvested during cloning of cell lines or the age and/or source of tissue from which primary cells are isolated. The effects of the various inducing agents on adipocyte differentiation are discussed in detail in section VI. In most cases, primary preadipocytes derived from young animals require only either low concentrations (1-10 nM) of insulin in the presence of fetal calf serum, or high concentrations (1-10 µM) in serum-free medium (57, 97, 104, 264). Depending on the species, the age of the donor, and/or the adipose depot source, agents such as glucocorticoids and MIX are either necessary to trigger the differentiation program or act only to accelerate it (57, 58, 96, 97, 104, 264, 293, 309). For established cell lines as well as for primary cells, the development of serum-free culture conditions has confirmed the positive role of these inducers and the involvement of insulin-like growth factor I (IGF-I), glucocorticoid, and cAMP signaling pathways. For preadipose cell lines, addition of insulin at supraphysiological concentrations does not affect the number of differentiated cells but does serve to accelerate lipid accumulation. In contrast, for most primary preadipocytes, which have a very low degree of spontaneous differentiation, insulin increases the number of differentiated cells; few lipid-containing cells are observed in the absence of insulin (95, 104, 264). Depending on the culture system, the insulin/IGF-I signaling pathway may either be critical for differentiation or necessary only to achieve the maximal rate of triacylglycerol accumulation that accompanies adipocyte differentiation. An overview of the stages of adipocyte differentiation is presented in Figure 1. The committed preadipocyte maintains the capacity for growth but has to withdraw from the cell cycle before adipose conversion. During adipocyte differentiation, acquisition of the adipocyte phenotype is characterized by chronological changes in the expression of numerous genes. This is reflected by the appearance of early, intermediate, and late mRNA/protein markers and triglyceride accumulation. These changes take place primarily at the transcriptional level, although posttranscriptional regulation occurs for some adipocyte genes (180, 295). In addition to the activation of genes, those genes that are inhibitory to adipogenesis or simply unnecessary for adipose cell function are repressed. Changes in gene expression during the early and late stages of adipocyte maturation have been characterized mainly through the use of preadipose cell lines. These have been extensively reviewed by others (47, 166, 249) and are summarized below. The limited information that is available on the details of the molecular events occurring during primary preadipocyte differentiation is also incorporated.
A. Growth Arrest
In preadipose cell lines as well as in primary preadipocytes, growth arrest and not cell confluence or cell-cell contact per se appears to be required for adipocyte differentiation. Confluent 3T3-F442A cells shifted to methylcellulose-stabilized suspension culture still undergo differentiation (196). This indicates that although confluence in these cells leads to growth arrest, cell-cell contact is not a prerequisite for adipocyte conversion. Primary rat preadipocytes plated at low density in serum-free medium can also differentiate in the absence of cell-cell contact (277). Two transcription factors, CCAAT/enhancer binding protein B. Clonal Expansion
After growth arrest at confluence, preadipocytes must receive an appropriate combination of mitogenic and adipogenic signals to continue through subsequent differentiation steps. Studies on preadipose cell lines have shown that growth-arrested cells undergo at least one round of DNA replication and cell doubling. This has been proposed to lead to the clonal amplification of committed cells (196). For 3T3-F442A and Ob17 cells, an increase in DNA synthesis precedes expression of late mRNA markers, and inhibition of DNA synthesis prevents the formation of fat cells (9, 147). However, primary preadipocytes derived from human adipose tissue do not require cell division to enter the differentiation process (63). In these cells, inhibition of mitosis with cytosine arabinoside does not impair adipocyte development, indicating that clonal amplification of committed cells is not a critical step. These cells may have already undergone potential critical cell divisions in vivo and may therefore correspond to a later stage of adipocyte development. Examination of growth-related proteins indicates potential differences between clonal expansion and preconfluent cell growth in 3T3-L1 preadipocytes. Retinoblastoma proteins pRB, p107, and p130 bind the E2F/DP complex to inactivate growth-promoting transcriptional activities. A recent report indicates changes in the expression of retinoblastoma proteins during differentiation of 3T3-L1 cells, including a transient increase in p107. This appears specific to clonal expansion, since this change is not detected during serum-stimulated cell growth (217). The role of E2F and RB family members during adipogenesis is unknown and needs further examination. Similarly, another group of growth arrest-specific (gas) genes shows a distinct expression pattern during clonal expansion. Gas6 appears to be preferentially expressed during clonal expansion of postconfluent preadipocytes, whereas gas1 and gas3 are expressed in serum-starved preadipocytes (244). Combined, these observations suggest differential regulation of the cell cycle in preconfluent proliferation versus postconfluent hormonally stimulated clonal expansion. An intracellular receptor for the immunosuppressive drug FK506, FKBP51, also transiently accumulates during the clonal expansion phase. Interestingly, whereas FK506 itself does not have any effect on clonal expansion and adipocyte differention, at high concentrations it reverses the inhibitory effect of rapamycin, another immunosuppressant that shares intracellular receptors (FKBPs) with FK506 (305, 307) Regardless of its mechanism, inhibition of proliferation by rapamycin is accompanied by inhibition of mitogen-activated serine-threonine S6 kinase, indicating that clonal expansion is necessary for subsequent adipocyte differentiation of 3T3-L1 cells. This also suggests that a phosphorylation-dephosphorylation mechanism may be involved during the clonal expansion period. Interestingly, a transient increase in levels of the phosphatase inhibitor HA2 occurs during the clonal expansion phase. Constitutive expression of HA2 blocks differentiation only during the clonal expansion period but not during later stages of differentiation. Moreover, inhibition of adipocyte differentiation by HA2 can be overcome by treatment with the phosphatase inhibitor vanadate, suggesting that a critical tyrosine dephosphorylation is necessary during clonal expansion (157). C. Early Changes in Gene Expression
Although it is helpful to schematize the stages of adipocyte differentiation into a hierarchy of molecular events, an accurate chronology of the earliest steps in adipocyte differentiation has not been elucidated. Growth arrest and clonal expansion are accompanied by complex changes in the pattern of gene expression that can differ with the cell culture models and the specific differentiation protocols employed. Expression of lipoprotein lipase (LPL) mRNA has often been cited as an early sign of adipocyte differentiation (3, 47, 98, 123, 166). Lipoprotein lipase is secreted by mature adipocytes and plays a central role in controlling lipid accumulation (48, 86). However, LPL expression occurs spontaneously at confluence and is independent of the addition of agents required for adipocyte differentiation (8, 9, 277). This suggests that LPL expression may reflect the growth-arrest stage rather than being an early differentiation step. It is also synthesized and secreted by other mesenchymal cell types including cardiac muscle cells and macrophages (49, 267). Because LPL expression is not adipocyte specific and it is independent of the additional agents required for adipocyte differentiation, classification of LPL as an early marker of adipocyte differentiation remains somewhat questionable. At least two families of transcription factors, C/EBP and PPAR, are induced early during adipocyte differentiation. The early expression of C/EBP and PPAR is logical given their subsequent involvment in terminal differentiation by transactivation of adipocyte-specific genes, described in section VII. Peroxisome proliferator-activated receptor- During adipocyte differentiation, cells convert from a fibroblastic to a spherical shape, and dramatic changes occur in cell morphology, cytoskeletal components, and the level and type of extracellular matrix (ECM) components. Many of the studies on the effect of cytoskeletal and ECM components in adipocyte differentiation predate the characterization of adipocyte transcription factors. It is likely that these changes could influence the expression and action of PPARs and/or C/EBPs during adipocyte differentiation. Decrease in actin and tubulin expression is an early event in adipocyte differentiation that precedes overt changes in morphology and the expression of adipocyte-specific genes (255). These changes in cell shape reflect a distinct process in differentiation and are not the result of accumulated lipid stores. 3T3 preadipocytes can undergo biochemical and morphological differentiation even in conditions when triglyceride accumulation is blocked by deprivation of biotin, a cofactor for fatty acid synthesis (148), or by addition of lipolytic agents (258). A switch in collagen gene expression is also an early event of adipocyte differentiation. The relative concentrations of fibroblast-expressed type I and type III procollagen mRNA decline by 80-90% during 3T3-L1 differentiation, and secretion of type IV collagen and entactin/nidogen increases (13, 290). Expression of D. Late Events and Terminal Differentiation
During the terminal phase of differentiation, adipocytes in culture markedly increase de novo lipogenesis and acquire sensitivity to insulin. The activity, protein, and mRNA levels for enzymes involved in triacylglycerol metabolism including ATP citrate lyase, malic enzyme, acetyl-CoA carboxylase, stearoyl-CoA desaturase (SCD1), glycerol-3-phosphate acyltransferase, glycerol-3-phosphate dehydrogenase, fatty acid synthase, and glyceraldehyde-3-phosphate dehydrogenase increase 10- to 100-fold (200, 256, 291). Glucose transporters (80), insulin receptor number, and insulin sensitivity increase. During adipocyte differentiation, there is a loss of Experiments with BALB/c 3T3 mesenchymal stem cells indicate that cells that have progressed beyond a specific stage in the differentiation process are committed to subsequent terminal differentiation and can neither dedifferentiate nor reenter mitosis (287, 294). However, recent evidence indicates that the precise stage beyond which adipocytes can be considered terminally differentiated is not clearly defined. Partially differentiated human preadipocytes, evidenced by substantial cytoplasmic lipid accumulation, are still capable of cell division as assessed histologically and by flow cytometry (210). The progressive dedifferentiation of primary mature adipocytes followed by cell division has also been reported (262). Tumor necrosis factor- The growth and differentiation of animal cells are controlled by communication between individual cells or between cells and the extracellular environment. Adipocyte differentiation therefore requires the cell to process a variety of combinatorial inputs during the decision to undergo differentiation. Hormones and growth factors with a role in adipocyte differentiation act via specific receptors to transduce external growth and differentiation signals through a cascade of intracellular events. Identification of agents or molecules that modulate the process in either a positive or negative manner provides insight into the signal transduction pathways involved. Extracellular matrix proteins may play an important role in modulating adipocyte differentiation by permitting the morphological changes and adipocyte-specific gene expression that accompany differentiation. The combination of hormones and growth/differentiation factors that trigger or potentiate adipocyte differentiation has been extensively examined and is summarized in Table 2. All preadipose cell lines exhibit some degree of spontaneous conversion when grown in the presence of fetal calf serum, indicating serum is integral to the differentiation process. However, serum components are neither well characterized nor easily controllable. The development of chemically defined serum-free media suitable for the differentiation of preadipose cell lines and primary preadipocytes has aided the assessment of the precise hormonal requirements for differentiation. Although the full complement of inducing agents required for differentiation varies with each cell culture model, IGF-I, cAMP, and glucocorticoids are generally considered necessary for the induction of differentiation either in serum-containing or in serum-free media. In addition to stimulatory factors, agents that suppress the differentiation of preadipocytes have also been identified. These may act either by their mitogenic properties, as for various growth factors, or by other independent mechanisms, as in the case of pref-1. Recent work has begun to clarify how these exogenous signals may influence transcription factors such as C/EBP and PPAR-
A. Hormones and Signal Transduction Pathways Regulating Adipocyte Differentiation
1. Growth hormone and IGF-I
I. INTRODUCTION
Top
Next
References
II. ADIPOSE TISSUE AS A SECRETORY ORGAN
Top
Previous
Next
References
(TNF-), and macrophage migration inhibitory factor (MIF). With the exception of TNF-, their physiological function remains to be elucidated. Nonetheless, these adipocyte-derived factors might also be involved in either the control of energy homeostasis or insulin resistance. Acylation stimulation protein is generated by the interaction of complement factor D (identical to adipsin), factor B, and complement C3, components of the alternate complement pathway (41); ASP may be involved in regulating energy storage by stimulating triacylglycerol synthesis and glucose transport (44, 168, 252). Acrp30/AdipoQ displays sequence homology with C1q, the first component of the classical complement activation pathway. Like adipsin, Acrp30/AdipoQ is abundant in normal serum, and its secretion is enhanced by insulin. This suggests that it could function as a signaling molecule from adipocytes and might therefore regulate energy homeostasis (119, 229). As addressed in detail in section VIA, TNF- not only inhibits adipocyte differentiation, but treatment of mature adipocytes with TNF- reduces the expression of adipocyte genes (202, 259, 260, 273, 274). In addition, TNF- may contribute to the insulin resistance that accompanies obesity and non-insulin-dependent diabetes mellitus; TNF- levels are elevated in WAT of obese rodents and humans. This may contribute to insulin resistance by inhibiting insulin-stimulated tyrosine kinase activity of the insulin receptor (114-117, 135). Treatment of adipocytes with TNF- was recently reported to increase secretion of MIF, a proinflammatory cytokine. Therefore, MIF may mediate the above-mentioned contribution of TNF- to insulin resistance (112).
(TGF-) significantly increases PAI-1 production. Higher PAI-1 levels have been reported in omental fat compared with subcutaneous fat. This may be correlated with increased PAI-1 levels noted in central obesity and may be involved in the development of vascular diseases associated with abdominal obesity (5, 164, 241).
III. ORIGIN OF ADIPOSE CELLS AND ADIPOSE TISSUE
Top
Previous
Next
References
superfamily act as potent osteogenic agonists (83), whereas TGF- inhibits adipocyte differentiation, as discussed in section VIA. Furthermore, the antidiabetic compounds thiazolidinediones, recently discovered to act through peroxisome proliferator-activated receptor- (PPAR-), an adipogenic transcription factor described in section VIIA, reduce bone marrow density and increase bone marrow adipocytes (84). Teratocarcinoma-derived C1 cells also behave as progenitor cells with osteoblast, chondroblast, or adipoblast cell fates (206). This suggests a close ontogenetic relationship between these connective tissue cell types. Osteogenic, chondrogenic, and adipogenic cells may arise from sclerotomal cells. The recent cloning of the basic-helix-loop-helix (bHLH) transcription factors twist and scleraxis has provided hints about the initial events that might precede and/or lead to the formation of the adipose lineage. In mouse embryo, these genes appear to play a significant role in the development of mesodermal tissues. Twist is expressed in early somites, and its expression is restricted to the sclerotome and excluded from myotome upon somite compartmentalization. Twist may be essential for the establishment of mesodermal cell fate and may be involved in the subdivision of the mesoderm lineage later in development (85). Significant levels of scleraxis can only be detected after somite compartmentalization, when it is expressed in sclerotome but not in myotome. Scleraxis could be a regulator of gene expression within the mesenchymal cell lineages that give rise to connective tissues (50). The precise role of these genes in adipocyte determination remains to be established.
IV. IN VITRO MODELS OF ADIPOCYTE DIFFERENTIATION
Top
Previous
Next
References
View this table:
TABLE 1.
In vitro models of adipocyte differentiation
V. PROCESS OF ADIPOCYTE DIFFERENTIATION
Top
Previous
Next
References
View larger version (32K):
[in a new window]
FIG. 1.
Overview of stages in adipocyte differentiation. Our current understanding of adipocyte differentiation indicates that a pluripotent stem cell precursor gives rise to a mesenchymal precursor cell with the potential to differentiate along mesodermal lineages of myoblast, chondroblast, osteoblast, and adipocyte. As discussed in text, given appropriate environmental and gene expression cues, preadipocytes undergo clonal expansion and subsequent terminal differentiation. Selected molecular events accompanying this process are indicated to right, with their approximate duration reflected by the solid line. PPAR- , peroxisome proliferator-activated receptor-; C/EBP, CCAAT/enhancer binding protein; pref-1, preadipocyte factor-1; ECM, extracellular matrix; FA, fatty acid.
(C/EBP-) and PPAR-, have been shown to transactivate adipocyte specific genes, and are discussed in section VII. Both C/EBP- and PPAR- also appear to be involved in the growth arrest that is required for adipocyte differentiation. McKnight and co-workers (276) have demonstrated the antimitotic activity of C/EBP- through the use of a C/EBP--estrogen receptor fusion protein. Activation by estrogen treatment results in cessation of cell growth as assessed by cell number and DNA synthesis (276). Darlington and co-workers (268) have reported that C/EBP- increases p21/SDI-1 mRNA and protein levels, but not other cell cycle components. Moreover, antisense p21/SDI-1 eliminates growth inhibition brought about by C/EBP- (268). This indicates that C/EBP- may function by increasing p21/SDI-1 levels. Spiegelman and colleagues (6) have showed that PPAR- is sufficient to induce growth arrest in fibroblasts and in adipogenic simian virus 40 large T antigen-transformed cells. In these PPAR--expressing cells, cell cycle withdrawal is accompanied by a decrease in the DNA binding and transcriptional activity of the E2F/DP-1 complex because of phosphorylation of DP-1 and a decrease in the expression of the serine/threonine phosphatase PP2A catalytic subunit. Therefore, C/EBP- and PPAR- may act cooperatively to bring about growth arrest (6). Although C/EBP- and PPAR- expression increases dramatically during adipocyte differentiation, the low level of these factors expressed in preadipocytes may be sufficient to mediate growth arrest that precedes differentiation. is largely adipocyte specific and is expressed at low but detectable levels in preadipocytes. Its expression rapidly increases after hormonal induction of differentiation. It is easily detectable during the second day of 3T3-L1 adipocyte differentiation, and maximal levels of expression are attained in mature adipocytes (27, 38). Induction of PPAR-
appears to precede that of PPAR-. Expression of PPAR-, however, is rather widespread. It is detected in a variety of tissues as well as in several cultured cell lines, including the CH310T1/2, 3T3-C2, and NIH 3T3 (7). A transient increase in the expression of C/EBP- and C/EBP- isoforms precedes the increase in PPAR- expression (27, 167, 299). The subsequent decrease of C/EBP- and C/EBP- in early to mid stages of differentiation is concomitant with the induction of C/EBP- mRNA. This increase in C/EBP- expression occurs slightly before the expression of adipocyte-specific genes (27, 153, 167). Another transcription factor induced very early during adipocyte differentiation is sterol regulatory element binding protein-1c (SREBP-1c)/adipocyte determination and differentiation factor 1 (ADD1), a bHLH-leucine zipper protein that is involved in cholesterol metabolism (26) and may also participate in adipocyte gene expression (136, 137). 2-collagen type VI mRNA is first detectable upon confluence in Ob1771 cells and rises sharply after confluence (51). It reaches maximal levels 4 days postconfluence and gradually decreases to 50% of maximal levels during differentiation (51). Increased production of soluble and cell-associated chondroitin sulfate proteoglycan-I (versican) has been reported during 3T3-L1 differentiation, and this may account for the observed increase in culture medium viscosity (29). The amount of pericellular fibronectin, as well as cellular synthesis of fibronectin, decreases by four- to fivefold during differentiation of 3T3-F442A cells (10). Preadipocyte factor-1 (pref-1), a recently described preadipocyte protein with epidermal growth factor (EGF)-like repeats, discussed in section VIB, has been hypothesized to be involved in maintaining the preadipose phenotype (246-248). A dramatic decrease in pref-1 expression accompanies adipocyte differentiation; it is abundant in preadipocytes and is not detectable in mature fat cells. It is the only known gene whose expression is completely downregulated during adipocyte differentiation. 1-adrenergic receptors and an increase in the 2- and the 3-subtypes; this results in an increase in total adrenergic receptor number (69, 70, 101, 152). In addition to increases in mRNAs for proteins directly related to lipid metabolism, adipocytes also synthesize other adipose tissue-specific products. These include the following: aP2, an adipocyte-specific fatty acid binding protein also identified as 422 (19, 256); FAT/CD36, a putative fatty acid transporter (125, 239); and perilipin, a lipid droplet-associated protein (94). In addition, adipocytes produce a number of secreted products, discussed in section II. These include the following: monobutyrin, an angiogenic agent; adipsin, a homolog of the serine protease complement factor D; Acrp30/AdipoQ; PAI-1; and angiotensinogen II (5, 41, 61, 119, 127, 229). Leptin is also increased during in vitro terminal differentiation of adipocytes, although its level is much lower than that detected in adipose tissue (165). Peroxisome proliferator activated receptor- and/or C/EBP- is implicated in the coordinate activation of several of these genes, including aP2, GLUT4, SCD1, phosphenolpyruvate carboxykinase (PEPCK), and leptin (113, 162, 175, 269, 270). The function of PPAR- and C/EBP- in adipocyte differentiation is presented in section VII. treatment of mature TA1 or 3T3-L1 adipocytes or newly differentiated primary human adipocytes results in decreased expression of adipocyte markers and loss of lipid, with the development of morphological changes resulting in long, spindle-shaped cytoplasmic extensions (202, 273, 274). These cells therefore come to resemble preadipocytes. However, recent evidence indicates that although TNF--treated adipocytes and preadipocytes appear to share many similar morphological characteristics and gene expression patterns, they likely differ considerably. As described in section VIB, pref-1 expression is easily detected in preadipocytes and is absent from mature adipocytes; pref-1 levels are not restored by TNF- of mature adipocytes. Pref-1 expression may reflect a fundamental difference between naive preadipocytes and those that result from TNF- treatment (301). True reversion to a preadipose phenotype, if it takes place in vivo, could play a role regulating adipocyte number and consequently adipose tissue mass. The molecular and cellular events occurring during this process warrant more detailed study.
VI. FACTORS THAT MODULATE ADIPOCYTE DIFFERENTIATION
Top
Previous
Next
References
that are critical for activation of adipocyte genes and adipocyte differentiation.
View this table:
TABLE 2.
Modulation of adipocyte differentiation by hormones, cytokines, and growth factors
-I collagen. The role of GH is also recognized in studies employing Ob1771 preadipocytes; in serum-free medium, GH increases the extent of adipocyte differentiation (35). Growth hormone stimulates IGF-I gene transcription and, as discussed below, IGF-I is necessary for adipocyte differentiation. It has been suggested that IGF-I secreted from Ob1771 cells could act in an autocrine/paracrine fashion to induce differentiation of Ob1771 cells (133). In contrast to observations in preadipose cell lines, no stimulatory effect of GH is observed in primary preadipocyte cultures. Initial reports on the effect of GH on rat and human preadipocytes indicate a lack of GH action (283). Because these studies were performed under serum-containing conditions, the effects of GH may not have been detectable. This concern was addressed in serum-free studies wherein GH was demonstrated to inhibit differentiation of rat, porcine, and human preadipocytes (81, 107, 283, 284). Although GH markedly stimulates IGF-I production in rat preadipocytes, which in turn promoted cell proliferation, the antiadipogenic action of GH is not related to growth promotion mediated by IGF-I. Addition of an anti-IGF-I monoclonal antibody prevents the stimulatory effect on cell proliferation but not the reduction of differentiation (284). In addition to the known lipolytic activity of GH, inhibition of adipocyte differentiation may explain the effect of GH on adipose tissue mass in vivo. The conflicting results for the action of GH on the differentiation of preadipose cell lines and primary preadipocyte cultures highlight a substantial biological difference between these two models. Each may represent different stages of the adipocyte lineage; it has been proposed that the requirement for GH is obviated in primary cultures by their prior in vivo exposure to circulating GH. These cells may therefore be primed for subsequent sensitivity to other adipogenic agents (2, 283).
2. Other growth factors and cytokines
Unlike IGF-I, other growth factors and cytokines are generally considered as inhibitors of adipocyte differentiation. This is perhaps because of the their mitogenic effects, since cell growth and differentiation are usually mutually exclusive. As discussed in section VA, growth arrest is requisite for differentiation. Several studies indicate a role for those growth factors that function through the EGF receptor, such as EGF and TGF-, in adipose tissue development. Transforming growth factor- inhibits differentiation of 3T3-F442A and rat preadipocytes, and transgenic mice overexpressing TGF- have a 50% reduction of total body fat (163, 233). Epidermal growth factor inhibits differentiation of mouse, rat, and human preadipocytes (105, 233, 282), and subcutaneous administration of EGF to newborn rats results in a substantial decrease in fat pad weight, which suggests a delayed formation of adipocytes from preadipocytes (238). However, EGF is not always inhibitory. Differentiation of 3T3-L1 preadipocytes grown in serum-free medium has been reported to depend on EGF or platelet-derived growth factor (PDGF) (230). Chronic treatment of porcine preadipocyte cultures with EGF does not significantly alter their differentiation (Gregoire et al., unpublished data).
The role of basic fibroblast growth factor (bFGF) and PDGF in adipocyte differentiation is not clear. Basic fibroblast growth factor has been shown to have antiadipogenic effects in several preadipocyte cell lines under serum-containing conditions (109, 184, 185), whereas it has no effect in serum-free conditions (230). Moreover, in serum-free conditions, exposure of human preadipocytes to varying concentrations of bFGF has no effect on the number and morphology of differentiating cells (105) and either a modest or no stimulatory effect on rat preadipocytes (236, 282). Platelet-derived growth factor also has been reported to either inhibit (109, 184), have no effect (230), or promote (16) differentiation of preadipose cell lines. In addition, despite the obvious mitogenic activity seen in human preadipocyte cultures, PDGF did not substantially affect adipogenesis (105). Taken together, these reports suggest that the inhibitory effects of these growth factors may depend on the origin, the state of development of the target preadipocytes, and culture conditions. In most cell culture models, TGF- is a potent inhibitor of adipocyte differentiation (160, 203, 237, 253, 282). A possible mechanism for TGF- inhibition of adipocyte differentiation may be via increasing synthesis of ECM components. The ECM influences adipocyte differentiation, as discussed in section VIC, and TGF- increases synthesis of ECM components (110). Inhibition of in vitro adipocyte differentiation, assessed by triglyceride accumulation and expression of various marker mRNAs, is also reported for a number of cytokines. Interleukin-11 has a dose-dependent inhibitory effect for both 3T3-L1 differentiation and for adipose conversion of bone marrow stroma-derived H-1/A cells. Inhibition is dominant over the effect of standard inducing agents (193, 194). Interferon- and interleukin-1 inhibit the adipoconversion of 3T3-derived preadipocyte cell lines and primary rodent preadipocytes (95, 134, 199). Tumor necrosis factor- decreases LPL synthesis and inhibits adipocyte differentiation. As mentioned in section VD, when applied to mature adipocytes at fairly high doses for a long period, TNF- has been shown to cause loss of intracytoplasmic lipids, wherein the cells resemble a preadipocyte phenotype (273, 274). This TNF--mediated reversal of adipocyte differentiation has been shown to be associated with the downregulation of C/EBP- and induction of c-myc expression (189, 259, 297). In addition, TNF- treatment causes a rapid decrease in the levels of PPAR- mRNA and protein, as well as a parallel decrease in PPAR- DNA binding activity that precedes the decrease in C/EBP- and aP2. This suggests that the downregulation of PPAR- may be a mechanism whereby TNF- exerts its effects in the mature adipocyte (301, 311). Furthermore, it has been shown that PPAR- is a phosphoprotein that undergoes EGF-stimulated MEK and MAPK-dependent phosphorylation. This phosphorylated form is less active in transactivation of adipocyte genes and in promoting adipogenesis (1, 30, 118). This therefore suggests that some of the growth factors inhibitory to adipocyte differentiation might act through the MAPK pathway to phosphorylate PPAR-. However, insulin treatment, which is known to increase lipid accumulation during adipocyte differentiation, also results in PPAR- phosphorylation (311). Further studies are needed to address this apparent discrepancy.
3. Nuclear hormone superfamily
Members of the nuclear hormone superfamily, including glucocorticoids, 3,3',5-triiodothyronine (T3), and RA, influence adipocyte differentiation. Their action in adipocyte differentiation is not well characterized at the molecular level, but these hormones in general exert nuclear effects by binding to their respective intracellular hormone receptors. Although some general conclusions can be made, the variability of serum composition must be taken into account, since serum itself may provide various factors. Dexamethasone, a synthetic glucocorticoid, is a component of the dexamethasone/MIX differentiation cocktail established for 3T3-L1 cells by Rubin et al. and dexamethasone is also routinely used for the differentiation of other preadipocyte cell lines (37, 78, 99, 224). It is also employed for the differentiation of primary preadipocytes derived from different fat depots from various species, including rodents, rabbits, pigs, and humans (96, 104, 160, 213, 264). Furthermore, dexamethasone is used for optimal adipocyte differentiation even in studies where transfection of PPAR- and/or C/EBP- was employed to induce adipocyte differentiation of fibroblasts (272, 300). The role of PPAR- and C/EBP in adipocyte differentiation is discussed in detail in section VII. Depending on the origin of the cells and culture conditions, glucocorticoid treatment is either required for differentiation or acts to only accelerate this process. In serum-free medium, differentiation of porcine, rabbit, and human preadipocytes appears strictly dependent on the addition of glucocorticoids (104, 213, 264), whereas extensive differentiation of rat preadipocytes occurs in the absence of glucocorticoids (57, 97). It has been demonstrated in 3T3-L1 cells that glucocorticoids induce expression of C/EBP-. This increase may contribute to the formation of C/EBP--C/EBP- heterodimers, which in turn may lead to PPAR- expression (299). In studies with Ob1771 preadipocytes, glucocorticoid effects have been shown to be mediated through increased metabolism of arachidonic acid leading to an increase in production of prostacyclin, which in turn increases intracellular cAMP (4).
The ability of RA to affect various differentiation processes including the terminal events of the adipocyte differentiation program has been recognized for several years. When used at supraphysiological concentration, RA inhibits adipocyte differentiation of preadipocyte cell lines and primary porcine preadipocytes (60, 263). Retinoic acid addition either before or after treatment with inducing agents does not affect differentiation, indicating that RA acts at an early stage in differentiation. This finding is supported by the observation that RA treatment prevents induction of C/EBP- and interferes with the mechanisms that induce as well as maintain PPAR- expression. These actions of RA seem to be predominantly mediated by liganded RA receptors (RARs) rather than retinoid X receptors (RXRs) (38, 302). Moreover, recent evidence indicates that the inhibitory effects of RA occur before PPAR- expression by blocking C/EBP- induction (232). In contrast to the inhibitory effects observed for supraphysiological concentration of RA, concentrations close to the receptor dissociation constant act as potent adipogenic inducers for Ob17 cells and rat preadipocytes; this specifically involves the RAR- subtype (225, 226). A critical role of RA in adipocyte differentiation has been recently highlighted by the finding that pretreatment of differentiating ES cell-derived embryoid bodies with RA for a short period of time results in a high degree of adipogenesis. In this case, the role of RA in these very early events of the adipocyte differentiation program can be distinguished from that of RA on terminal differentiation described above, since neither adipogenic hormones nor potent activators of PPARs could substitute for RA (52).
3,3',5-Triiodothyronine also has been implicated in the terminal differentiation of Ob17 preadipocytes (82). The role of T3 appears to be restricted to the Ob17 preadipose cell line, since no clear requirement for T3 is observed in other preadipocyte culture models, including the 3T3-L1 and rat, porcine, or human primary preadipocyte cultures (57, 103, 104, 230, 236, 264, 293).
4. Prostaglandins
Several lines of evidence indicate that arachidonate metabolites may play an important physiological role in adipose tissue metabolism and development. Mature adipocytes and cultured preadipocytes produce significant amounts of prostaglandins (PGs), including PGF2
, PGE2 , PGD2 , and PGI2 (122, 215). Prostaglandin E2 is a strong antilipolytic compound, and PGF2 and PGI2 have been shown to modulate preadipocyte differentiation. Prostaglandin D2 and its 15-deoxy-J2 derivative may be endogenous ligands for PPAR- and therefore act as adipogenic signals (141). As presented in section VII, treatment with synthetic PPAR- agonists stimulates adipose conversion of 3T3-L1 preadipocytes (27).
Prostaglandin F2 inhibits differentiation of various preadipose cell lines and primary rat preadipocytes (122, 174, 188, 216, 235, 282). Prostanoid FP receptor agonists have been recently reported to be potent inhibitors of differentiation for 3T3-L1 cells and primary rat preadipocytes, confirming the involvement of a FP prostanoid receptor in the inhibition of adipocyte differentiation (34, 234). In 3T3-L1 cells, FP receptor stimulation causes a transient increase in intracellular calcium, activation of a calcium/calmodulin-dependent protein kinase (CaM kinase), and an increase in DNA synthesis that is associated with the inhibition of differentiation. The addition to differentiating cells of KN-62, a CaM kinase inhibitor, in the presence of the FP receptor agonist reverses the inhibition of differentiation and suggests a critical role for a CaM kinase in adipocyte differentiation (174). However, the precise role of CaM kinases in this process remains to be clarified, since temporal activation of CaM kinase type II has been recently reported to be an obligatory step for adipogenesis. Blockage of CaM kinase type II activation with either KN-62 or KN-93 prevents the conversion of 3T3-L1 fibroblasts to adipocytes. This effect was dependent on the timing of inhibitor addition (285). A different mode of action for PGF2 has been proposed for primary rat preadipocytes. In these cells, PGF2 stimulates mRNA expression and production of TGF-, both in undifferentiated and differentiated cells. Both PGF2 and TGF-, which are inhibitors of adipocyte differentiation, are produced locally in adipose tissue. Therefore, stimulation of TGF- expression by PGF2 could represent an amplification mechanism to modulate preadipocyte differentiation and adipocyte function within adipose tissue (156). In contrast to the inhibitory role of PGF2 , a potent and specific adipogenic role has been attributed to prostacyclin (PGI2). This prostanoid is one of the major metabolites of arachidonic acid both in preadipose and adipose cells, and it has been previously described as an autocrine/paracrine adipogenic effector for Ob1771 preadipose cells and primary rodent and human preadipocytes (35, 186, 282). Carbacyclin, a stable prostacyclin analog, has been shown to act by means of two intracellular signaling pathways known to synergize in inducing adipocyte differentiation, i.e., concomitant elevation of cAMP and free intracellular calcium (280, 281). Prostacyclin has also been reported to be an activator of the three known mammalian PPARs (, , and ) and to be the most effective activator for PPAR- described to date. This suggests that in addition to the biological effects of prostacyclin mediated by its cell-surface receptor, its ability to promote differentiation may also be mediated by PPARs (15, 28, 111). The paracrine adipogenic effect of PGI2 has also been reported to be controlled by angiotensin II. This effect is mediated through the AT2 subtype of the angiotensin II receptor. Ob1771 adipose cells challenged with this vasoactive peptide produce PGI2 . Prostacyclin released by differentiated Ob1771 cells is able to induce preadipose cells to differentiate into adipose cells (53). In vivo, this paracrine mode of action may represent a crucial biological signal in the hyperplastic development of adipose tissue known to occur once adipose cells reach their maximal size. The role of PGE2 and PGD2 in adipocyte differentiation is less clear. Depending on the cell culture model systems, they have been reported to either inhibit or have no effect on preadipocyte differentiation (34, 235).
5. cAMP, G proteins, and protein kinase C
Methylisobutylxanthine accelerates the differentiation of preadipose cell lines and primary preadipocytes. As with glucocorticoids, MIX is routinely used for the differentiation of a variety of preadipocytes. Methylisobutylxanthine has been shown to increase expression of C/EBP-, and this increase is required for subsequent PPAR- expression and adipocyte differentiation. The role of C/EBP- in adipogenesis is presented in detail in section VII. The precise mode of action of MIX is not resolved. Methylisobutylxanthine is known to inhibit phosphodiesterases and block A1 adenosine receptor in a competitive manner. It also stimulates adenylyl cyclase activity by blocking the inhibitory regulatory protein Gi (198). This indicates that MIX may function through increasing cAMP accumulation. However, contradictory results have been obtained in studies that address whether the cAMP elevating agents dibutyryl cAMP or forskolin can replace the effects of MIX in stimulating adipocyte differentiation (243, 293, 296, 304). In studies on 3T3-F442A preadipocytes, modulation of adipogenesis by cAMP has been reported to depend on the increase in intracellular cAMP levels achieved by forskolin treatment; forskolin concentrations in the nanomolar range promote differentiation, whereas concentrations above the micromolar range inhibit the differentiation process (304).
The G proteins Gs and Gi have been shown to mediate adipocyte differentiation in 3T3-L1 cells in a manner apparently independent of adenylyl cyclase (87, 288). Antisense oligodeoxynucleotides to Gs accelerate adipocyte differentiation, and agents that activate Gs block differentiation. Likewise, expression of the inhibitory subunit Gi-2 promotes lipid accumulation. These effects of Gs and Gi-2 are exerted at ambient or elevated intracellular cAMP levels, demonstrating that this critical role of G proteins in adipocyte differentiation is independent of adenylyl cyclase (288). Constitutive expression of Gi-2 and Gs chimeras has been used in 3T3-L1 cells to define the specific regions of these proteins responsible for repression of adipogenesis. The domain of Gs requisite for regulation of adipogenesis maps to a region that includes switch domains I and II. These have been found to be spatially distinct from the domains that regulate adenylyl cyclase; this is consistent with the inability of cAMP to influence the adipocyte differentiation process (286). A potential role for the G protein Gq in the control of adipose tissue physiology and/or preadipocyte differentiation has been recently reported. Ablation of Gq by antisense RNA under the control of PEPCK promoter, which directs expression to WAT and liver in transgenic mice, causes increased body mass and hyperadiposity that persists through adult life (79). Absence of Gq abolishes the A1 adrenergic regulation of lipolysis, apparently predisposing the mice to fat accumulation. However, because the basis for hyperadiposity in these Gq-deficient mice remains to be established, Gq may also control adipogenic conversion in vivo (79). Rat primary subcutaneous preadipocytes in culture display a higher capacity to differentiate than epididymal preadipocytes (97). This site specificity of differentiation capacity was recently shown to be related at least in part to the differences in Gq subunit expression (56). During adipogenesis, no major site differences are found in the amount of Gs and Gi subunits. However, Gq dramatically decreases in subcutaneous-derived cells while it remains constant in epididymal-derived cells. The Gq subunit mediates phospholipase C- activation leading to diacylglycerol formation and protein kinase C (PKC) activation (65), an event known to downregulate adipoconversion of preadipose cell lines as well as rat primary preadipocytes (184, 240, 242). Therefore, compared with epididymal preadipocytes, the higher capacity of subcutaneous preadipocytes to differentiate into adipocytes seems to correlate with a decrease in Gq expression and a decrease in Gq-mediated PKC activation (56).
Although the negative effect of PKC activation on adipocyte differentiation appears well established, the role played by the different PKC isoforms in this process remains unsettled. Protein kinase C is a family containing at least 11 isoforms divided into three major groups based on their structure and mode of activation (55). Several PKC isoforms are expressed in mature isolated rodent adipocytes as well as in 3T3-L1 preadipocytes and adipocytes. Differentiation-dependent changes in isoform expression occur during adipogenesis (66, 74, 149, 171). Although total PKC activity is reduced during 3T3-L1 differentiation, not all isoforms have lowered expression. Protein kinase C-
expression occurs only with differentiation, and upregulation of this isoform does not occur in cells treated with the inhibitory agent TNF-. During TNF--induced inhibition of adipocyte differentiation, downregulation of the PKC- isoform was blocked; this suggests that downregulation of PKC- is a discrete step in the 3T3-L1 differentiation program (171). Protein kinase C- and PKC-
also decrease during rat preadipocyte differentiation, and the possible involvement of PKC- in the postreceptor signaling pathway of insulin is suggested by studies in these primary cultures (149). Ectopic expression of PKC-
is also reported to alter the expression of cyclins and cdk inhibitors and induce adipogenesis in NIH 3T3 fibroblasts (161). These findings indicate that the PKC pathway of signal transduction is part of a highly complex system that likely exerts negative as well as positive effects on the adipocyte differentiation process.
B. Pref-1, an EGF Repeat-Containing Inhibitor of Adipocyte Differentiation
Comparison of the genes that are regulated by the adipogenic inducing agents dexamethasone/MIX in 3T3-L1 cells and the closely related but differentiationdefective 3T3-C2 cells has resulted in the identification of preadipocyte factor-1 (pref-1). 3T3-C2 cells, which do not undergo differentiation in response to the adipogenic inducing agents dexamethasone/MIX, express approximately threefold higher pref-1 levels than 3T3-L1 preadipocytes. Recent evidence has demonstrated that pref-1 is a EGF repeat-containing transmembrane protein that inhibits adipocyte differentiation and suggests that this molecule may link adipocyte differentiation signals from the extracellular environment to the cell interior (246-248, 250). Expression of pref-1 in 3T3-L1 preadipocytes decreases to undetectable levels during their differentiation to mature adipocytes. Fetal calf serum, an essential component for in vitro adipocyte differentiation, dramatically downregulates pref-1 levels. Taken together, these observations indicate that pref-1 may not only be regulated during adipocyte differentiation, but suggest a regulatory role for pref-1 in adipocyte differentiation. Preadipocyte factor-1 does not appear to have the mitogenic effects (C. Li, C. M. Smas, and H. S. Sul, unpublished data) normally associated with the inhibitory actions of growth factors and cytokines described in section VIA2. This suggests that pref-1 may function via a unique mechanism to inhibit adipocyte differentiation.
1. Inhibitory action in adipogenesis
The inhibitory action of pref-1 on adipocyte differentiation has been demonstrated by two approaches: 1) interfering with the normal downregulation of pref-1 that occurs during adipocyte differentiation by constitutive expression of pref-1 in 3T3-L1 preadipocytes, and 2) addition of a soluble form of the pref-1 ectodomain to culture media during and after the dexamethasone/MIX differentiation treatment. To determine the effect of persistent pref-1 expression on adipocyte differentiation, two independent pools of several hundred clones transfected with the pref-1 expression construct have been examined for their ability to differentiate in response to the adipogenic agents dexamethasone/MIX. Although a high degree of differentiation is observed in controls of nontransfected 3T3-L1 preadipocytes or stable pools of 3T3-L1 harboring the reverse orientation of the pref-1 cDNA, differentiation is drastically reduced in the pref-1-transfected cultures. These cultures maintain fibroblast morphology and contain few fat cells. Oil Red O staining reveals little to no lipid accumulation and markedly lower expression of aP2 and SCD1 mRNAs. Examination of the cultures by light microscopy reveals that the decrease in lipid staining in the pref-1-transfected cells is because of a decrease in the total number of cells that differentiate to adipocytes, not to a generally lower amount of lipid per cell (248). The inhibitory effects of the pref-1 ectodomain have been shown by addition of soluble pref-1-glutathione-S-transferase (GST) fusion protein to 3T3-L1 preadipocytes during the differentiation period. In the presence of soluble pref-1, only 10% of cells undergo adipocyte differentiation in response to the dexamethasone/MIX treatment, in comparison with the high degree of differentiation in GST-treated and untreated control cultures. The levels of mRNA for fatty acid synthase, SCD1, and aP2 in the pref-1-treat
