8_pparg_adipogenic regulator and thiazolidinedione receptor

DIABETES, VOL. 47, APRIL 1998 507

Perspectives in DiabetesP PA R - : Adipogenic Regulator andThiazolidinedione ReceptorB.M. Spiegelman

The past several years have seen an explosive increasein our understanding of the transcriptional basis ofadipose cell differentiation. In particular, a key rolehas been illustrated for PPA R - , a member of thenuclear hormone receptor superfamily. PPA R - has alsobeen recently identified as the major functional recep-tor for the thiazolidinedione class of insulin-sensitizingdrugs. This review examines the evidence that hasimplicated this transcription factor in the processes ofadipogenesis and systemic insulin action. In addition,several models are discussed that may explain how a sin-gle protein can be involved in these related but dis-tinct physiological actions. I also point out severalimportant areas where our knowledge is incompleteand more research is needed. Finally, I discuss howadvances in our understanding of nuclear receptorfunction, particularly the docking of cofactors in aligand-dependent fashion, should lead to improveddrugs that utilize the PPA R - system for the treatmentof insulin resistance. D i a b e t e s 47:507514, 1998

A dipose tissue has been viewed historically as apassive player in the regulation of energy home-ostasis, storing energy in times of nutritionalexcess and releasing that energy when needed intimes of nutritional deprivation. However, the last decade orso has seen a rather startling revision of this view. It is nowappreciated that adipose cells secrete a large number ofbioactive molecules, including adipsin, angiotensinogen,tumor necrosis factor- ( T N F - ), leptin, ACRP30/AdipoQ,and plasminogen activator inhibitor 1 (1). Although the rolesof these molecules (and probably more to come) remain tobe fully elucidated, it is clear that TNF- and leptin are bonafid e signaling molecules, involved in insulin resistance and thecontrol of food intake, respectively. This new appreciation ofthe interactive role of the adipose cell in energy metabolism

and insulin sensitivity has added a further dimension to ourunderstanding of this cell type. It has also added urgency toour need to understand adipose cell differentiation and genee x p r e s s i o n .

This review will describe the identification and role ofP PA R - as a central regulator of fat cell differentiation. It willalso review the discovery and function of this protein as themajor receptor for the thiazolidinedione (TZD) class ofinsulin-sensitizing drugs. With the introduction of troglita-zone (under brand-names such as Rezulin, Romozin), thefirst PPA R - ligand of the TZD class, into clinical practice in1997 for the treatment of NIDDM, it is especially importantto point out where the major gaps in our understanding of thisr e c e p t o r-ligand system lie. Finally, I will describe somepotential new therapeutic opportunities that have beenopened up by the recent progress.

P PA R - AND ADIPOGENESIS

The cloning of mammalian PPA R - and its link with adipo-genesis came from our analysis of the adipose-specificenhancer from the aP2 gene, an abundant adipocyte-specificfatty-acid binding protein. Mutational analysis and study ofprotein binding to this 500base-pair piece of DNA identifie da key nuclear factor, termed ARF6, that bound to two sites(ARE6 and ARE7) in this enhancer (2). This DNA bindingactivity was observed only in extracts of fat cells. Cloning ofthis factor (3) revealed it to be a member of the peroxisomeproliferator activated receptor (PPAR) subfamily of nuclearhormone receptors and, in particular, the mammalianhomolog of Xenopus PPA R - , which had been cloned earlier(4). Several other labs also cloned mammalian PPA R - i n d e-p e n d e n t l y, through homology screens that sought new mem-bers of the PPAR family (5,6). It is now accepted that thereare three related but quite distinct PPAR proteins, PPA R - ,P PA R - (also called PPA R - , Nuc-1, or FAAR), and PPA R - .P PA R - is expressed in an adipose-selective fashion in bothrodents and humans, being 10- to 30-fold higher in fat than inmost other tissues (3). Interestingly, two forms of the protein,

1 and 2, exist as products of alternative promoter usage.The two forms differ in that PPA R - 2 has an NH2- t e r m i n a lextension of 30 amino acids. In addition, PPA R - 2 is foundselectively in fat tissue, whereas 1 is expressed at low lev-els in many tissues (7,8). The functional meaning of thesesplice variants is not yet clear.

Our early work indicated that PPA R - , like other PPA R - s ,heterodimerizes with the retinoid X receptor (RXR) andactivates the aP2 enhancer in fibroblasts, a cell type in whichthis enhancer ordinarily has little or no activity (3). The

From the Department of Cancer Biology, Dana-Farber Cancer Institute,and the Department of Cell Biology, Harvard Medical School, Boston,M a s s a c h u s e t t s .

Address correspondence and reprint requests to Dr. Bruce M. Spiegel-man, Dana-Farber Cancer Institute, 44 Binney St., Boston, MA 02115. E-mail:b r u c e _ s p i e g e l m a n @ d f c i . h a r v a r d . e d u .

Received for publication 30 December 1997 and accepted in revisedform 5 January 1998.

B.M.S. has received honoraria from Parke Davis-Warner Lambert. ADD1/SREBP1, adipocyte determination and differentiation factor

1/sterol response element binding protein 1; C/EBP, CAAT/enhancer bind-ing protein; ETYA, 5,8,11,14-eicosatetraynoic acid; Kd, dissociation constant;P PA R - , peroxisome proliferator activated receptor- ; RXR, retinoid Xreceptor; TNF- , tumor necrosis factor- ; TZD, thiazolidinedione.

P PA R - /RXR heterodimer binds to direct repeats of hor-mone response elements separated by one base, so-calledDR-1 sites. While no bona fide ligands for PPA R - w e r eknown at this time, a diverse group of activators, mainlyfatty-acid derivatives such as 5,8,11,14-eicosatetraynoic acid( E T YA), were known to activate PPA R - (4); high levels ofthis compound will also activate PPA R - . Strong activationof the aP2 enhancer by PPA R - and RXR requires the appli-cation of a PPAR activator, an RXR ligand, or both. PPA R -and RXR also transactivated another bona fide fat enhancer,that from the PEPCK gene (9). This led to the most ambitiousquestion: were PPA R - expression and activation sufficientto give a full adipocyte differentiation response in fib r o-blasts? In fact, retroviral expression of PPA R - in manyfibroblast cell lines, followed by application of an activatorsuch as ETYA, gave abundant differentiation that includedlipid accumulation, changes in cell morphology, and expres-sion of most if not all of the genes that characterize theadipocyte phenotype (10). With the identification of rela-tively high-affinity ligands (see below) it is now clear thatP PA R - , expressed at or below the levels seen in fat tissue,can convert nearly every fibroblastic cell in a given cultureinto a fully differentiated adipocyte.

P PA R - interacts with other transcription factors in poten-tially important ways (11), as illustrated in Fig. 1.C A AT/enhancer binding protein (C/EBP)- , thoughexpressed in many tissues, is induced in adipogenesis anditself has adipogenic action when expressed at high levels(12). When expressed at levels equivalent to those seen in fat,it can cooperate powerfully with PPA R - , even allowing as i g n i ficant adipogenic response in the absence of addedP PA R - ligands. Precisely how these two transcription factorsinteract remains an important mechanistic problem. Thisability of PPA R - and C/EBP- to promote differentiation isnot limited to fibroblasts, as simultaneous expression of bothof these factors and activation of PPA R - can cause transd-ifferentiation of cultured myoblasts to adipocytes (13).

Another factor that can cooperate with PPA R - i sadipocyte determination and differentiation factor 1/sterolresponse element binding protein 1 (ADD1/SREBP1). Thistranscription factor, a member of the basic helix-loop-helixf a m i l y, was independently identified as a potential regulatorof adipogenesis and fatty-acid metabolism (14), and as a keyfactor in cholesterol homeostasis (15). Coexpression ofADD1/SREBP1 with PPAR- increases the transcriptionalactivity of PPAR- , even without adding a PPAR- ligand(16). Because ADD1/SREBP1 can increase the expression ofseveral key genes of fatty-acid metabolism, such as fatty acidsynthetase and lipoprotein lipase, it seems plausible that thiscooperation is through the generation of endogenousligands for PPA R - . Other key players in the transcriptionalcontrol of adipogenesis are C/EBP- and - . These factorsappear to be very important in the induction of PPAR- inadipogenic differentiation. Indeed, conditional expressionof C/EBP- and - has been shown to yield expression lev-els of PPAR- equivalent to normal fat cells; a very strongdifferentiation response can be seen in these cells uponapplication of PPAR- ligands (17). This data has led to amodel of a transcriptional cascade that is dependent on theexpression of C/EBP- and - to turn on PPAR- . The bind-ing of ligands, perhaps generated by the action ofADD1/SREBP1, triggers the full differentiation response.

C / E B P - both cooperates with PPA R - to yield a more pow-erful differentiation and may be needed to maintain PPAR-

expression at high levels. It may be noteworthy that abinding site for C/EBP family members has been noted inboth the PPAR- 1 and - 2 promoters (18).

T Z Ds AS LIGANDS FOR PPA R -

A key report that led to identification of TZDs as ligands forPPAR- came from Harris and Kletzien (19), who showedthat pioglitazone increased transcriptional activity from theaP2 enhancer and apparently did so through the differenti-ation-linked DNA site (ARE6) described above. When wecloned and identified the ARE6 binding factor as the PPA R -

/RXR heterodimer, two groups independently askedwhether the TZD drugs were acting as direct agonists forP PA R - . These studies identified BRL49653 and pioglita-zone as direct ligands of PPAR- with positive activity ongene transcription (20,21). Importantly, the TZDs were alsoshown to be highly selective for PPAR- , as they had veryminimal activity toward PPAR- or PPAR- . As expectedfrom the data cited above, the TZDs were potent and effec-tive at stimulating adipogenesis in cells containing endoge-nous or ectopically expressed PPA R - (20,21). This dataalso made sense of earlier reports in which the TZDs hadbeen shown to stimulate adipogenesis in preadipocyte celllines, though the mechanism of action of the drugs in theseearly studies was not known.

508 DIABETES, VOL. 47, APRIL 1998

PPAR- : ADIPOGENIC REGULATOR AND THIAZOLIDINEDIONE RECEPTOR

FIG. 1. Transcriptional cascade in adipogenesis. PPA R - and theC/EBP family interact to control adipose-cell diff e r e n t i a t i o n .C / E B P- and - are involved in the transcriptional control of PPA R -

. Adipogenesis is stimulated upon PPA R - activation by ligands.C / E B P - is activated later in the differentiation process but canfunctionally synergize with PPA R - and may also be involved inmaintaining high levels of PPA R - expression. ADD1/SREBP1 hasbeen implicated in the control of several key genes of fatty-acidmetabolism and can promote the transcriptional activity of PPA R - ,probably through the formation of endogenous ligands.

The evidence that PPA R - is the major receptor mediatingthe antidiabetic activity of the TZDs is now very strong, basedon the following multiple lines of pharmacological evidence.

1 . Each of the TZD drugs binds to and activates PPA R - in thesame concentration range that has antidiabetic activity (22).

2 . Among many TZDs surveyed, the rank order of potency oftheir antidiabetic activities closely matches the rank orderof their affinities for PPA R - ( 2 2 ) .

3 . Potent and selective ligands for PPA R - outside of theTZD class have now been developed on the basis of theiractivation of PPA R - . These have antidiabetic actions inpreclinical models of insulin resistance and diabetes (T.Willson, personal communication).

4 . Ligand stimulation of RXR, the heterodimeric partner ofP PA R - , also improves insulin sensitivity in vivo (23).

5 . No other receptor for the TZD drugs has been identifie d .

Taken together, these data make a compelling case thatP PA R - is the major functioning receptor for the commonTZD actions in diabetes. However, it is possible that individ-ual drugs of this class may have additional targets that con-tribute to their therapeutic actions or to their side effects.

One concern surrounding the clinical use of TZDs is thattheir adipogenesis-promoting effects could be detrimentalto patients with NIDDM, who all too often are overweightto begin with. In therapeutic doses, it is clear that the TZDsdo promote weight gain and increase fat deposition inrodent models (e.g., 24). Whether this is primarily due toincreased insulin sensitivity, more fat cell differentiation, ora combination of the two is not clear. However, clinical usein humans has not shown these drugs to induce significantweight gain. This may reflect the fact that increased adipo-genesis per se would not necessarily cause obesity; recentevidence indicates that TZDs may increase fat cell numberwhile simultaneously decreasing fat cell size (24). Alterna-t i v e l y, it may indicate that preadipocytes in adult humans arerelatively resistant to the differentiative effects of the TZDs.Three other side effects observed with the TZDs deservemention. A small increase in plasma volume is consistentlyobserved in patients undergoing treatment with troglita-zone (25). If and how this relates to PPA R - activation is notclear, but this phenomenon does not seem to limit clinicalutility. Of more concern is an increased adipose cell for-mation observed in the bone marrow of rodents treatedwith certain TZDs. Fatty transformation of bone marrow isthought to be a very serious condition, and the considerablepublished data linking PPAR- and adipogenesis, includingstromal cells of the bone marrow (26), suggest that thismay well be a consequence of PPAR- activation. On theother hand, this is only observed at high doses of TZDs, soit is likely that a reasonable therapeutic window is availablebetween beneficial effects in NIDDM and the potentiallydeleterious effects on stromal elements in the bone marrow.Most recently, liver toxicity has been observed in a small per-centage of patients taking troglitazone (27), leading to awithdrawal of this drug from the market in England. It isunclear whether the hepatatoxic effect is mediated by acti-vation of PPA R - or whether it represents a nonspecificeffect of troglitazone. In the U.S., the drug is now givenwith recommendations that liver enzyme levels be periodi-cally monitored.

Natural ligands. Radiolabeled TZD ligands have enableddevelopment of a displacement assay that allowed a search fornatural ligands of PPA R - . This question is interesting in twoways. First, the identity of natural ligands may provideinsights into new therapeutic approaches. Second, knowl-edge of the endogenous ligand will allow investigation ofwhether some insulin-resistant states are due to a defic i e n c yin the endogenous ligand for PPA R - . This seems plausiblewhen one considers that much endocrine pathologysuch ashypothyroidism and adrenal insufficiency (Addisons disease)or excess (Cushings Syndrome)results from dysregulationof the endogenous ligands for other nuclear receptors.

The initial screening for natural ligands examined manyfatty acids and fatty-acid derivatives for binding activity. Thefirst natural ligand described was an unusual prostanoid 15-d e o x y 1 2 , 1 4PG J2 (21,28). More recently, several polyunsatu-rated fatty acids, such as linoleic acid, have also been foundto bind directly to PPA R - (29). Although all of these mole-cules must be considered as potentially important ligands invivo, it should be emphasized that their affinities are relativelyl o w, in the range of 250 mol/l. This contrasts with thehigher affinities that most nuclear receptors have for theirendogenous ligands (dissociation constants [Kd s] in the lownmol/l range). Although 250 mol/l is not necessarily out ofthe concentration range at which fatty acids such as polyun-saturated fatty acids circulate in vivo, it is not clear that thesemolecules can reach the nucleus of relevant tissues (fat, mus-cle, liver) at these concentrations. Fatty-acid levels insidecells are usually tightly regulated through the actions of bind-ing proteins in the cell membrane and cytoplasms, as well asby the fatty-acid acylation machinery. Still, the notion ofwhether PPA R - is a promiscuous receptor for many fattyacids with low affinity or has a more limited number of spe-c i fic, high-affinity ligands has been the subject of much spec-ulation and interesting debate. More research is required.

INSULIN SENSITIZATION BY TZDs: TISSUE TARGETS ANDM E C H A N I S M S

The work described above identifying PPA R - as the pri-mary target for the antidiabetic action of the TZDs has alsopresented two paradoxes. How can a receptor expressed pre-dominantly in fat tissue improve insulin sensitivity in all ofthe major insulin-sensitive tissues? Additionally, given thewell-established connection between obesity and insulinresistance, how can we reconcile the enhancement ofinsulin sensitivity by a receptor known to promote adipo-genesis? There are no clear answers to these questions yet,but it is important to recognize that adipose cell differenti-ation is not identical to obesity. Obesity is primarily a disor-der of energy balance, where energy intake exceeds energyexpenditure. Although an increase in fat cell number mayaccompany great obesity, an increased fat cell number perse without an increase in total energy stored would not nec-essarily lead to insulin resistance. There are several possiblecellular and/or molecular explanations for how the TZDsmight work in vivo (Fig. 2).Direct effects on fat, muscle, and liver. Although PPA R -

levels are 1030 times higher in fat than in muscle or liver,this receptor is expressed in these latter tissues. Pharmaco-logical doses of TZDs may be sufficient to stimulate PPA R -at all of these sites and hence achieve alterations in geneexpression that can reduce insulin resistance. In addition,


B.M. SPIEGELMAN

there is evidence from experiments in cell culture that acti-vation of PPA R - by ligands can increase the expression ofthis receptor in a positive-feedback loop. It is entirely possi-ble that PPA R - levels are initially low in insulin-resistantpatients but become significantly elevated in muscle andliver during TZD treatment. Most recently, troglitazone hasbeen shown to improve insulin sensitivity in an experimen-tal model of lipodystrophy in mice, suggesting that importantregulation of PPA R - can occur in the absence or nearabsence of fat (30).E ffects on fat cell diff e r e n t i a t i o n . As described above,activation of PPA R - by TZDs promotes adipose-cell differ-entiation. In the absence of increased energy storage, thiswould be expected to produce more fat cells of a smaller aver-age size. Because smaller adipose cells are usually more sen-sitive to insulin, such a differentiative response would beexpected to produce greater insulin-dependent glucoseuptake (24). In addition, because insulin is a powerfulantilipolytic agent, smaller fat cells with increased insulin

sensitivity would be expected to have lower relative rates oflipolysis. Because high levels of free fatty acids may becausally involved in the induction of insulin resistance, thiscould affect insulin sensitivity at distant sites such as muscleand liver (e.g., 31,32). This mechanism would also be com-patible with the so-called Randle Effect, where the tendencyof muscle to utilize glucose as an energy source is inverselycorrelated with the use of fatty acids as an energy source.There are also reports that TZD administration to rodentsgreatly increases the amounts of brown adipose tissue(33,34). Considering that this tissue functions to dissipateenergy through the oxidation of fatty acids, it could alsoreduce circulating lipid levels and have a beneficial effect oninsulin sensitivity, as discussed above.Control of adipose-cell signaling. Since it is now appre-ciated that adipose cells send molecular signals to other tis-sues participating in energy metabolism, it is possible thatP PA R - activation controls one or more genes that regulatesystemic insulin sensitivity. Two interesting candidate genes



FIG. 2. Tissue targets for TZD drugs. In A, theprimary tissue target for the TZDs is PPA R - i nfat. Effects on insulin action in other tissueswould then occur as a consequence of alter-ations in signaling molecules produced by fat,such as free fatty acids, TNF- , leptin, or oth-ers. Alternatively (B), TZDs may have directactions on PPA R - in fat, muscle, and liver andcontrol the expression of genes that influenceinsulin action in these tissues.

A

B

in this regard are TNF- and leptin. A large and increasingbody of data suggests that TNF- is produced by adiposecells and is overexpressed in obesity and insulin resistance(1). In experimental obese animals, neutralization of TNF-with a soluble receptor-IgG fusion protein showed a signifi-cant improvement in systemic insulin sensitivity (35). Mostr e c e n t l y, experiments using genetic knockouts of the TNF-

ligand or the p55 TNF receptor definitively demonstrate acrucial role for this cytokine in insulin resistance in vivo(36). The role of TNF- in human insulin resistance remainsto be determined, though a small study using an antiTNF-antibody in established NIDDM did not ameliorate hyper-glycemia or hyperinsulinemia (37). TZDs apparently haves i g n i ficant effects on two branches of the TNF- s y s t e m .Treatment of obese mice with pioglitazone for 2 weeksdecreased adipose TNF- mRNA by 50% (38). Most recently,TZDs have been shown to block the ability of TNF- to inter-fere with the most proximal events of insulin signaling (39).Although TNF- treatment of cultured adipocytes blocksinsulin-stimulated tyrosine phosphorylation of the insulinreceptor and insulin receptor substrate 1, these tyrosinephosphorylations occur normally in the presence of TNF- i fcells are pretreated for several hours with TZDs. This effectof the TZDs has an apparent specificity for the insulin-sig-naling cascades because TNF- is able to induce the tran-scription factor NF B whether or not TZDs are added. In anindependent series of experiments, similar effects of TZDshave also been observed in vivo, as application of troglitazoneto rats prevented the induction of insulin resistance causedby an infusion of TNF- (40). TZDs have also been impli-cated in the regulation of leptin expression. Application ofTZDs in vivo or to cultured adipose cells can cause a reduc-tion in the expression of leptin mRNA and protein (41,42).Although the role of leptin in insulin resistance is a contro-versial one, some reports indicate that leptin might interferewith insulin signaling in certain cell types (e.g., 43). Hence, arole of the TZDs through leptin must be considered possible.

F i n a l l y, as mentioned above, the release of fatty acids dur-ing lipolysis has taken on potentially great significance, withthe recognition that these substrates may be important sig-naling molecules themselves. Improvement of insulin sig-naling in fat by the TZDs would be expected to reduce lipol-ysis and diminish whatever impact elevated circulating fattyacids have in systemic insulin resistance. Taken together,these and other secreted fat-derived signaling molecules aregood candidates to play some considerable role in systemicinsulin resistance. Regulation of these processes by TZDsthrough PPA R - seems likely to contribute to their actions.The quantitative nature of this contribution in different ani-mal models and patient subsets remains to be determined.U l t i m a t e l y, the contribution of PPA R - in individual tissues tothe overall antidiabetic effects of the TZDs will be deter-mined through tissue-specific genetic ablation in mice. Thisis well under way in several laboratories.Target genes for PPA R - . Of course, it is not possible todetermine which potential target genes for PPA R - are mostrelevant for the antidiabetic action of the TZDs until the tis-sue(s) directly affected by these drugs are clarified. Most ofthe work done to date has been on adipose cells; and evenhere, the target genes relevant for amelioration of insulinresistance have not been clearly determined. Similarly, thedownstream targets that trigger adipogenesis per se are not

known. In differentiated cells and tissues, TNF- and leptinexpression are reduced by PPA R - activation. In addition, twogenes of fatty-acid metabolism, lipoprotein lipase and thefatty-acid binding protein aP2, appear to be direct targets ofP PA R - activation. Increased levels of lipoprotein lipase in fathave been shown to occur as a consequence of T Z D t r e a t m e n tin vivo (44). This would be expected to increase uptake oftriglycerides by fat and thereby improve insulin signaling inmuscle and liver. Increased aP2 could affect many aspects offatty-acid metabolism, though a knock-out of aP2 functionin mice has shown a decreased tendency to develop insulinresistance (45). The resolution of this potential paradoxawaits further studies on the exact function of aP2. Anotherpotentially important target gene for PPA R - is GLUT4. Theexpression of this gene is increased in cultured adipocytes andfat tissue through PPA R - activation by TZDs (46,47); pre-s u m a b l y, this could contribute to reduced hyperglycemia. Ofcourse, because muscle is the major sink for insulin-depen-dent glucose disposal, it will be important to determine ifGLUT4 is induced under these conditions in muscle.

FUTURE DEVELOPMENTS: BETTER PPA R - L I G A N D S ?

The TZDs hold enormous promise for the treatment ofpatients with NIDDM. Because these drugs were originallydeveloped without any knowledge of their molecular tar-gets, there is hope that the research described above willlead to the development of new and better PPA R - l i g a n d s .This development is important because for all the promise ofthese drugs, they do not normalize the glucose levels of mostdiabetic patients when used alone, and many patients do notrespond well clinically even when TZDs are used in combi-nation with insulin.

The word better can be thought of in at least two ways:potency and efficacy. Given that the potency of a particularligand will ultimately depend on its affinity for PPA R - , thereis probably much room for improvement. Most of the antidia-betic TZDs have affinities for PPA R - between 30 nmol/l(BRL49653) and 700 nmol/l (troglitazone). With the use ofmodern drug-screening procedures, large chemical libraries,and combinatorial chemistry methods, it is very likely thatagents with higher affinity for PPA R - can be found. Presum-a b l y, this will include many agents from outside the TZD class.

Efficacy may be a trickier issue. Although the affinity oftroglitazone for PPA R - is relatively low, normal therapeuticdoses of this drug result in blood levels that exceed its Kd f o rthe receptor. Hence, further increases in receptor occu-pancy may not be possible. Indeed, clinical studies suggestthat efficacy of troglitazone is maximal in the 400600mg/day range, with little or no benefit resulting from furtherdose escalation (25).

Improvements in efficacy may reside in two areas: ligandsthat have more full agonist activity on PPA R - or agents thatstimulate the PPA R - /RXR heterodimer in ways that trogli-tazone or the other TZDs do not. In this regard, a recentreport that RXR agonists could have antidiabetic action inobese/diabetic rodents carries exciting promise (23). Thesedrugs working together with a PPA R - ligand could theoret-ically lower glucose more than either one alone. However, itis not clear from the studies done to date whether maximalcostimulation of this heterodimeric receptor pair gives alarger glucose-lowering effect than does maximal stimulationof PPA R - alone. Another key issue will be whether stimula-


B.M. SPIEGELMAN

tion of RXR, a heterodimeric partner of many other nuclearreceptorssuch as the other PPAR-s, the thyroid hormoner e c e p t o r, vitamin D receptor, and retinoic acid receptorswillresult in unacceptable side effects. Although none have beenobserved in the early studies in rodents, the chronic natureof diabetes suggests that careful long-term studies will berequired to assess the issue of side effects.

Several recent studies show that the transcriptional activ-ity of PPA R - stimulated by TZDs can be sharply reduced asa result of phosphorylation by the enzyme mitogen-activatedprotein kinase (48,49). Studies to date also suggest that undermost common culture conditions, some significant portion ofP PA R - is in the less active, phosphorylated state. The bio-chemical basis for this decreased activity is not yet known,but it could theoretically involve a decreased affinity forligand by the phosphorylated receptor. Alternatively, phos-phorylation could alter interactions with important (and as yetu n d e fined) protein factors of PPA R - , such as corepressorsor coactivators. In fact, it is now known that nuclear recep-tors function as ligand-gated platforms for the assembly ofthese cofactors into large protein complexes on specific DNAsequences (5052) (Fig. 3). Some of these coactivator proteins(CBP/p300, SRC1, pCAF) have histone acetyltransferaseactivity that functions to open the configuration of chro-

matin, allowing more efficient transcription. One theoreticalproblem is that essentially none of the nuclear receptor coac-tivators or corepressors identified to date are selective for par-ticular receptors. Hence, it is not clear which cofactors aremore important for the function of any particular receptor. Itis also not obvious how the tremendous specificity of bio-logical actions of the individual nuclear receptors are gener-ated. It is extremely likely that many more cofactors will bei d e n t i fied, including some that function selectively for indi-vidual receptors, including PPA R - . Once this issue is clari-fied, the appropriate cofactor molecules may then serve in thedevelopment of assays that could yield greater efficacy in theimprovement of insulin sensitivity.

Of course, learning more about the components of theP PA R - signaling systemendogenous ligands and theenzymes that produce them, receptor levels and modific a-tions, coactivators and corepressors, downstream transcrip-tional targetsmay also lead to a better understanding of thepathogenesis of NIDDM. A genetic defect in any of theseaspects of this receptor system could result in reducedinsulin action, and further progress identifying these com-ponents will allow scrutiny in diabetic patients.

In summary, the last several years have seen a remarkableimprovement in our understanding of the role of PPA R - a s



FIG. 3. Docking of cofactors for PPA R - through ligand binding. The PPA R - /RXR heterodimer binds to distinct DNA sequence elements calledDR-1 sites. The binding of ligands to nuclear receptors is now understood to stimulate release of negative factors (corepressors) and triggerbinding of positive cofactors (coactivators). Many of these proteins modulate the acetylation state of histones and serve to open chromatinfor more efficient transcription.

a receptor for the TZD drugs and as a key regulator ofadipocyte differentiation. Considerable gaps in our knowledgestill exist, particularly with regard to PPA R - cofactors andthe direct tissue targets and downstream-effector genes ofthese drugs. However, our knowledge has advanced suffi-ciently to make it likely that the full therapeutic potential ofP PA R - ligands will be revealed by the combination of mod-ern cell and molecular biology, coupled with the newest tech-niques of drug development. That improvements beyond thecurrent generation of TZD drugs can be made in potencyand/or safety is, in my view, a virtual certainty. It is also highlylikely that more efficacious drug regimens that target PPA R -

and a better understanding of the role of this system in thepathogenesis of NIDDM will also be developed.

A C K N O W L E D G M E N T S

I am grateful to several colleagues for critical comments andsuggestions on this manuscript, including Drs. Evan Rosen,Stephen Farmer, Timothy Willson, Juergen Lehmann, andAlan Saltiel.

R E F E R E N C E S1 . Spiegelman BM, Flier JS: Adipogenesis and obesity: rounding out the big pic-

ture. Cell 87:377389, 19962 . Graves RA, Tontonoz P, Spiegelman BM: Analysis of a tissue-specific

enhancer: ARF6 regulates adipogenic gene expression. Mol Cell Biol12:12021208, 1992

3 . Tontonoz P, Hu E, Graves RA, Budavari AB, Spiegelman BM: mPPA R - 2: tis-s u e - s p e c i fic regulator of an adipocyte enhancer. Genes Dev 8:12241234, 1994

4 . Dreyer C, Krey G, Keller H, Givel F, Helftenbein G, Wahli W: Control of the per-oxisomal beta-oxidation pathway by a novel family of nuclear hormonereceptors. Cell 68:879887, 1992

5 . Zhu Y, Alvares K, Huang Q, Rao MS, Reddy JK: Cloning of a new member ofthe peroxisome proliferator-activated receptor gene family from mouse liver.J Biol Chem 268:2681726820, 1993

6 . Kliewer SA, Forman BM, Blumberg B, Ong ES, Borgmeyer U, Mangelsdorf DJ,Umesono K, Evans RM: Differential expression and activation of a family ofmurine peroxisome proliferator-activated receptors. Proc Natl Acad Sci USA91:73557359, 1994

7 . Tontonoz P, Graves RA, Budavari AB, Erdjument-Bromage H, Lui M, Hu E,Tempst P, Spiegelman BM: Adipocyte-specific transcription factor ARF6 is aheterodimeric complex of two nuclear hormone receptors, PPA R 2 andR X R . Nucleic Acids Res 22:56285634, 1994

8 . Vidal-Puig AJ, Considine RV, Jimenez-Linan M, Werman A, Pories WJ, Caro JF,Flier JS: Peroxisome proliferator-activated receptor gene expression inhuman tissues: effects of obesity, weight loss, and regulation by insulin andglucocorticoids. J Clin Invest 99:24162422, 1997

9 . Tontonoz P, Hu E, Devine J, Beale EG, Spiegelman BM: PPA R 2 regulates adi-pose expression of the phosphoenolpyruvate carboxykinase gene. Mol CellB i o l 15:351357, 1995

1 0 . Tontonoz P, Hu E, Spiegelman BM: Stimulation of adipogenesis in fib r o b l a s t sby PPA R 2, a lipid-activated transcription factor. Cell 79:11471156, 1994

1 1 . Brun RP, Kim JB, Hu E, Spiegelman BM: Peroxisome proliferator- a c t i v a t e dreceptor gamma and the control of adipogenesis. Curr Opin Lipidol8:212218, 1997

1 2 . MacDougald OA, Lane MD: Transcriptional regulation of gene expressionduring adipocyte differentiation. Ann Rev Biochem 64:345373, 1995

1 3 . Hu E, Tontonoz P, Spiegelman BM: Transdifferentiation of myoblasts by theadipogenic transcription factors PPA R and C/EBP . Proc Natl Acad SciU S A 92:98569860, 1995

1 4 . Tontonoz P, Kim JB, Graves RA, Spiegelman BM: ADD1: a novel helix-loop-helix protein associated with adipocyte determination and differentiation. M o lCell Biol 13:47534759, 1993

1 5 . Yokoyama C, Wang X, Briggs MR, Admon A, Wu J, Hua X, Goldstein JL,Brown MS: SREBP-1, a basic helix-loop-helix, loop-helix-leucine zipper pro-tein that controls transcription of the low density lipoprotein receptor gene.Cell 75:187197, 1993

1 6 . Kim JB, Spiegelman BM: ADD1/SREBP1 promotes adipocyte differentiationand gene expression linked to fatty acid metabolism. Genes Dev10:10961107, 1996

1 7 . Wu Z, Xie Y, Bucher NLR, Farmer SR: Conditional ectopic expression of

C / E B P - in NIH-3T3 cells induces PPA R and stimulates adipogenesis. G e n e sD e v 9:23502356, 1995

1 8 . Zhu Y, Qi C, Korenberg JR, Chen XN, Noya D, Rao MS, Reddy JK: Structuralorganization of mouse peroxisome proliferator-activated receptor gamma( m P PAR gamma) gene: alternative promoter use and different splicing yieldtwo PPAR gamma isoforms. Proc Natl Acad Sci USA 92:79217925, 1995

1 9 . Harris PK, Kletzien RF: Localization of a pioglitazone response element in theadipocyte fatty acid-binding protein gene. Mol Pharmacol 45:439445, 1994

2 0 . Lehmann JM, Morre LB, Smith-Oliver TA, Wilkison WO, Willson TM, KliewerSA: An antidiabetic thiazolidinedione is a high affinity ligand for peroxisomep r o l i f e r a t o r-activated receptor . J Biol Chem 270:1295312956, 1995

2 1 . Forman BM, Tontonoz P, Chen J, Brun RP, Spiegelman BM, Evans RM: 15-d e o x y - 1 2,1 4-Prostagalandin J2 is a ligand for the adipocyte determinationfactor PPA R . Cell 83:803812, 1995

2 2 . Willson TM, Cobb JE, Cowan DJ, Wiethe RW, Correa ID, Prakash SR, Beck KD,Moore LB, Kliewer SA, Lehmann JM: The structure-activity relationshipbetween peroxisome proliferator-activated receptor gamma agonism and theantihyperglycemic activity of thiazolidinediones. J Med Chem 39:665668, 1996

2 3 . Mukherjee R, Davies PJ, Crombi DL, Bischoff ED, Cesario RM, Jow L, BoehmM F, Modon CE, Nadzan AM, Paterniti JR Jr, Heyman RA: Sensitization of dia-betic and obese mice to insulin by retinoid X receptor agonists. N a t u r e386:407410, 1997

2 4 . Hallakou S, Doare L, Foufelle F, Kergoat M, Guerre-Millo M, Berthault MF,Dugail I, Morin J, Auwerx J, Ferre P: Pioglitazone induces in vivo adipocytedifferentiation in the obese Zucker fa/fa rat. Diabetes 46:13931399, 1997

2 5 . Henry RR: Thiazolidinediones. Endocrinol Metab Clin North Am26:95539573, 1997

2 6 . Gimble JM, Robinson CE, Wu X, Kelly KA, Rodriguez BR, Kliewer SA,Lehmann JM, Morris DC: Peroxisome proliferator-activated receptor- g a m m aactivation by thiazolidinediones induces adipogenesis in bone marrow stro-mal cells. Mol Pharmacol 50:10871094, 1996

2 7 . Food and Drug Administration: FDA Talk Paper. Washington, DC, U.S.Department of Health and Human Services, December 1, 1997

2 8 . Kliewer SA, Lenhard JM, Willson TM, Patel I, Morris DC, Lehmann JM: Aprostaglandin J2 metabolite binds peroxisome proliferator-activated receptorgamma and promotes adipocyte differentiation. Cell 83:813819, 1995

2 9 . Kliewer SA, Sundseth SS, Jones SA, Brown PJ, Wisely GB, Koble CS, Devc-hand P, Wahli W, Willson TM, Lenhard JM, Lehmann JM: Fatty acids andeicosanoids regulate gene expression through direct interactions with per-oxisome proliferator-activated receptors alpha and gamma. Proc Natl AcadSci USA 94:43184323, 1997

3 0 . Burant CF, Sreenan S, Hirano K, Tai T-AC, Lohmiller J, Lukens J, Davidson NO,Ross S, Graves RA: Troglitazone action is independent of adipose tissue. J ClinI n v e s t 100:29002908, 1997

3 1 . Rebrin K, Steil GM, Getty L, Bergman RN: Free fatty acid as a link in the reg-ulation of hepatic glucose output by peripheral insulin. D i a b e t e s44:10381045, 1995

3 2 . Rebrin K, Steil GM, Mittelman SD, Bergman RN: Causal linkage betweeninsulin suppression of lipolysis and suppression of liver glucose output in dogs.J Clin Invest 98:741749, 1996

3 3 . Tai TAC, Jennermann C, Brown KK, Oliver BB, MacGinnitie MA, Wilkison WO,Brown HR, Lehmann JM, Kliewer SA, Morris DC, Graves RA: Activation of thenuclear receptor peroxisome proliferator-activated receptor gamma pro-motes brown adipocyte differentiation. J Biol Chem 271:2990929914, 1996

3 4 . Foellmi-Adams LA, Wyse BM, Herron D, Nedergaard J, Kletzien RF: Inductionof uncoupling protein in brown adipose tissue: synergy between norepi-nephrine and pioglitazone, an insulin-sensitizing agent. Biochem Pharm52:693701, 1996

3 5 . Hotamisligil G, Shargill N, Spiegelman BM: Adipose expression of tumornecrosis factor-alpha: direct role in obesity-linked insulin resistance. S c i e n c e259:8791, 1993

3 6 . Uysal KT, Wiesbrock SM, Marino MW, Hotamisligil GS: Protection from obe-sity-induced insulin resistance in mice lacking TNF-alpha function. N a t u r e389:610614, 1997

3 7 . Ofei F, Hurel S, Newkirk J, Sopwith M, Taylor P: Effects of an engineeredhuman anti-TNF- antibody (CD571) on insulin sensitivity and glycemic con-trol in patients with NIDDM. Diabetes 45:881885, 1996

3 8 . Hofman C, Lorenz K, Braithwaite SS, Colca JR, Palazuk BJ, Hotamisligil GS,Spiegelman BM: Altered gene expression for tumor necrosis factor-a and itsreceptors during drug and dietary modulation of insulin resistance.Endocrinology 134:264270, 1994

3 9 . Peraldi P, Xu M, Spiegelman BM: Thiazolidinediones block TNF- -induced inhi-bition of insulin signaling. J Clin Invest 100:18631869, 1997

4 0 . Miles PDG, Romeo OM, Higo K, Cohen A, Rafaat K, Olefsky JM: TNF- -induced insulin resistance in vivo and its prevention by troglitazone. D i a b e t e s


B.M. SPIEGELMAN

46:16781683, 19974 1 . Kallen CB, Lazar MA: Antidiabetic thiazolidinediones inhibit leptin (ob) gene

expression in 3T3-L1 adipocytes. Proc Natl Acad Sci USA 93:57935796, 19964 2 . De Vos P, Lefebvre AM, Miller SG, Guerre-Millo M, Wong K, Saladin R,

Hamann LG, Staels B, Briggs MR, Auwerx J: Thiazolidinediones repress obgene expression in rodents via activation of peroxisome proliferator- a c t i-vated receptor gamma. J Clin Invest 98:10041009, 1996

4 3 . Muller G, Ertl J, Gerl M, Preibisch G: Leptin impairs metabolic actions of insulinin isolated rat adipocytes. J Biol Chem 272:1058510593, 1997

4 4 . Lefebvre AM, Peinado-Onsurbe J, Leitersdorf I, Briggs MR, Paterniti JR,Fruchart JC, Fievet C, Auwerx J, Staels B: Regulation of lipoprotein metabo-lism by thiazolidinediones occurs through a distinct but complementarymechanism relative to fibrates. Arterioscler Thromb Vasc Biol 1 7 : 1 7 5 6 1 7 6 4 ,1 9 9 7

4 5 . Hotamisligil GS, Distel RJ, Johnson RS, Papaioannou V, Spiegelman BM:Uncoupling of obesity from insulin resistance through a targeted mutation inthe fatty acid binding protein aP2. S c i e n c e, 274:13771379, 1996

4 6 . Young PW, Cawthorne MA, Coyle PJ, Holder JC, Holman GD, Kozka IJ,Kirkham DM, Lister CA, Smith SA: Repeat treatment of obese mice with BRL49653, a new potent insulin sensitizer, enhances insulin action in white

adipocytes: association with increased insulin binding and cell-surfaceGLUT4 as measured by photoaffinity labeling. Diabetes 44:10871092, 1995

4 7 . Wu Z, Xie Y, Bucher NLR, Farmer SR: PPAR(gamma) induces the insulin-depen-dent glucose transporter, GLUT4, in the absence of C/EBP(alpha) during theconversion of 3T3 fibroblasts into adipocytes. J Clin Invest 101:2232, 1998

4 8 . Hu E, Kim JB, Sarraf P, Spiegelman BM: Inhibition of adipogenesis throughMAP kinase-mediated phosphorylation of PPA R . Science 2 7 4 : 2 1 0 0 2 1 0 3 ,1 9 9 6

4 9 . Adams M, Reginato MJ, Shao D, Lazar MA, Chatterjee VK: Transcriptional acti-vation by peroxisome proliferator-activated receptor gamma is inhibited byphosphorylation at a consensus mitogen-activated protein kinase site. J BiolC h e m 272:51285132, 1997

5 0 . Shibata H, Spencer TE, Onate SA, Jenster G, Tsai SY, Tsai MJ, OMalley BW:Role of co-activators and co-repressors in the mechanism of steroid thyroidreceptor action. Recent Prog Horm Res 52:141164, 1997

5 1 . Torchia J, Rose DW, Inostroza J, Kamei Y, Westin S, Glass CK, Rosenfeld MG:The transcriptional co-activator p/CIP binds CBP and mediates nuclear- r e c e p-tor function. Nature 387:677684, 1997

5 2 . Pazin MJ, Kadonaga JT: Whats up and down with histone deacetylation andtranscription? Cell 89:325328, 1997



Author Queries (please see Q in margin and underlined text)

Q1: Please spell out ETYA.>Q2: Please spell out C/EBP.>Q3:Please spell out ADD1/SREBP1, if possible.>Q4:Do you mean kilodaltons or Kd? Please specify what KDstands for.>Q4a: Should this read expected to result not expectedin result as it is now?Q5:Please clarify what Kd stands for.>

Ref 27: Please add more information for ref. 27. Can the FDAbe considered the author of the paper? What is the title of the paper?>Ref 30: Can you update ref. 30 now?>Ref 40: Please add the journal title to ref. 40.>Ref 47: Can you update ref. 47 now?>

8_pparg_adipogenic regulator and thiazolidinedione receptor

Documents