Lipotoxic Diseases of Nonadipose Tissues in Obesity

Home , Lipotoxicity

International Journal of Obesity (2000) 24, Suppl 4, S28±S32 ß 2000 Macmillan Publishers Ltd All rights reserved 0307±0565/00 $15.00 www.nature.com/ijo Lipotoxic diseases of nonadipose tissues in obesity

RH Unger1* and L Orci2

1Gifford Laboratories, Touchstone Center for Diabetes Research, Department of Internal Medicine, University of Texas Southwestern Medical Center, Dallas, Texas, USA; and 2Department of Morphology, University of Geneva Medical School, Geneva, Switzerland

It is proposed that an important function of leptin is to con®ne the storage of triglycerides (TG) to the adipocytes, while limiting TG storage in nonadipocytes. Excess TG deposition in nonadipocytes leads to impairment of functions, increased ceramide formation, which triggers nitric oxide-mediated lipotoxicity and lipoapoptosis. The fact that TG content in nonadipocytes normally remains within a very narrow range irrespective of excess caloric intake, while TG content of adipocytes rises, is consistent with a system of fatty acid (FA) homeostasis in nonadipose tissues. When leptin is de®cient or leptin receptors are dysfunctional, TG content in nonadipose tissues such as pancreatic islets, heart and skeletal muscle, can increase 10 ± 50-fold, suggesting that leptin controls the putative homeostatic system for intracellular TG. The fact that function and viability of nonadipocytes is compromised when their TG content rises above normal implies that normal homeostasis of their intracellular FA is critical for prevention of complications of obesity. FA overload of skeletal muscle, myocardium and pancreatic islets cause, respectively, insulin resistance, lipotoxic heart disease and adipogenic type 2 diabetes. All can be completely prevented by treatment with antisteatotic agents such as troglitazone. In diet-induced obesity, leptin signaling is normal initially and lipotoxic changes are at ®rst prevented; later, however, post-receptor leptin resistance appears, leading to dysfunction and lipoapoptosis in nonadipose tissues, the familiar complications of obesity. International Journal of Obesity (2000) 24, Suppl 4, S28±S32

Keywords: lipotoxicity; lipoapoptosis; nonadipose tissues; obesity

Introduction Lipotoxic diseases Long chain fatty acids (FA) provide the building Caloric preloading, a temporary increase in caloric blocks for the phospolipid bilayers of cell membranes intake, was intended to result in temporary increase in and for the intracellular phospholipid messengers fat stores to meet the anticipated environmental chal- upon which normal cell function depends. It is lenges mentioned. However, in the latter portion of likely, therefore, that all cells maintain a small stock- the twentieth century in the United States, radical pile of FA, stored as triacylglycerol (TG), for these environmental changes took place: a permanent essential housekeeping functions. But this small inter- increase in caloric intake, attributable to the aggres- nal reserve is too precious to be used as a fuel. Not sive marketing of high-energy, high-fat foods, coupled until the evolution of the adipocyte, with its unlimited with the advent of immobilizing technological inno- storage capacity for TG and its ability to redistribute vations, such as television, computers and automo- the stored FA on demand to nonadipose tissues, could biles, drastically increased the daily caloric balance. this most ef®cient of fuels be fully exploited. The The result was an unprecedented epidemic of obesity. ability to preload calories and to store them as TG may be one of the most transforming events in all of evolution, improving survival during famine and permitting migratory activity of many species including FA over¯ow hypothesis man, as proposed in the `thrifty gene' hypothesis of Neel.1 Lipotoxic complications of obesity have been attri- buted to an over¯ow of FA from adipocytes to nonadipocytes and deposition of the unneeded FA excess as TG.2 Circulating free fatty acids (FFA), higher in obese individuals than in lean persons,3 `¯ip- ¯op' across the plasma membranes of all cells.4 Unoxidized FA will be esteri®ed to TG. While the *Correspondence: RH Unger, Center for Diabetes Research, enlarged stores of TG are probably inert and therefore University of Texas Southwestern Medical Center, 5323 Harry Hines Blvd, Dallas, TX 75235-8854, USA. in themselves harmless to the cells, the TG excess E-mail: [email protected] may be hydrolyzed back to fatty acyl CoA at a rate Lipotoxicity in obesity RH Unger and L Orci S29 that exceeds the oxidative needs of the cell, driving blocked by transgenic overexpression of a functional the unoxidized excess into deleterious pathways of FA wild-type OB-Rb in the islets.10 In the absence of a metabolism, such as de novo ceramide formation.5 functional OB-R, these abnormalities can also be completely prevented by measures that lower the ectopic accumulation of FA in nonadipose tissues (troglitazone treatment)9 or which block nitric oxide Normal FA homeostasis in (NO) formation (nicotinamide or aminoguanidine nonadipocytes treatment).11 This implies that lipoapoptosis requires FA-induced overproduction of NO. Recent studies in rodents indicate that FA homeostasis Leptin normally protects nonadipocytes from stea- in nonadipocytes is regulated by leptin, which oper- tosis by raising the rate of b oxidation of long chain ates as an antisteatotic hormone.6 Four lines of evi- FA above the level required for their own energy dence support this hypothesis: (1) on the very ®rst day needs. It does this by permitting FA to upregulate of a high-fat diet, plasma leptin levels rise progres- PPARa,12 which in turn increases the transcription of sively and continue to increase in parallel with the the oxidative enzymes, acyl CoA oxidase (ACO) of accumulation of fat in adipocytes; (2) during such peroxisomes and carnitine palmitoyl transferase-1 overfeeding, the TG content of nonadipose tissues (CPT-1) of mitochondria. The unneeded energy remains at low levels, while TG content in adipocytes thereby generated is presumably dissipated as heat, rises progressively; (3) in rodent models lacking leptin as a consequence of upregulation of the uncoupling activity as a result of either leptin de®ciency or proteins (UCP) 1 and 2.13 Ectopic accumulation of TG defective leptin receptors, TG content of nonadipo- in nonadipose tissues is thus minimized as long as the cytes is very high, even on a relatively normal fat leptin system is operative (Figure 1A). intake; and (4) if normal leptin receptors are The loss-of-function mutation in the leptin receptor overexpressed in the islets of leptin receptor-defective that occurs in fa=fa ZDF rats completely inactivates rats, the ectopic accumulation of TG is reduced. this system. Fatty acids no longer upregulate PPARa, Leptin, therefore, appears to meet the criteria for an ACO and CPT-1. Rather, PPARa is expressed at antisteatotic hormone. subnormal levels and, in its stead, PPARg, a lipogenic transcription factor, is expressed at high levels.12 In addition, the lipogenic enzymes, acetyl CoA carbox- Lipotoxicity secondary to abnormal ylase (ACC) and fatty acid synthase (FAS), are expressed at increased levels. This promotes increased leptin receptors (OB-R) lipogenesis from glucose and increased esteri®cation of excess FA to TG14 (Figure 1B). In the obese fa=fa ZDF rat a Glu 269? Pro mutation involving the extracellular domain of all OB-R isoforms appears to block leptin action.7,8 These rats exhibit widespread steatosis involving skeletal Mechanism of lipoapoptosis muscle, heart, pancreas, pancreatic islets and liver.6 In b-cells, this is associated with extensive alterations The lipoapoptotic pathway resembles in its distal of the mitochondria and apoptosis,9 which can be portion the autoimmune apoptotic pathway through

Figure 1 (A) Normal intracellular FA homeostasis. Plasma FFA increases only when tissues require FA for oxidation. (B) In obesity with impaired FA homeostasis, plasma FFA and TG are increased. FA traverse plasma membranes freely and enter all cells. There, they must bind to a fatty acid-binding protein and=or be activated to FA CoA. The intracellular FA CoA will therefore exceed the oxidative requirements for the cell and will enter other metabolic pathways, causing some of the complications of obesity. We postulate that increased TG content of certain tissues, such as smooth muscle, may reduce their compliance and give rise to some of the complications of obesity.

International Journal of Obesity Lipotoxicity in obesity RH Unger and L Orci S30 which cytokines are believed to bring about destruc- normal OB-Rb into these islets will reduce the TG tion of b cells in type 1 diabetes. Only the proximal content and correct SPT overexpression. Ceramide segments of the pathways appear to differ. In auto- formation is thereby reduced and apoptosis comple- immune destruction of b cells the increased ceramide tely prevented.9 In other words, when a large excess of is thought to be the result of sphingomyelin break- FA is trapped within pancreatic islets (steatosis), down as a consequence of enhanced sphingomyeli- ceramide may reach cytotoxic levels. This will rid nase activity.15 Ceramide, in turn, activates nuclear the islets of b cells that can no longer carry out their factor Kappa B, which upregulates inducible nitric normal function. As long as formation of new func- oxide synthesis (iNOS). In autoimmune diabetes the tional b cells keeps pace with the rate of lipo-apop- increase in nitric oxide (NO) is a major factor in the tosis, islet function may remain adequate; when, complete destruction of all b cells that takes place. however, apoptosis exceeds the rate of replacement, The lipoapoptotic pathway differs from the auto- diabetes will ensue. immune pathway in that the increase of ceramide is the result of de novo ceramide synthesis. Each mole- cule of ceramide contains two molecules of long- chain FA; the ®rst step in ceramide synthesis is the Prevention of lipoapoptosis in vitro condensation of palmitoyl CoA and serine to form and in vivo dihydrosphingosine. This reaction is catalyzed by the enzyme serine palmitoyl transferase (SPT). In islets of One can reduce lipoapoptosis in vitro by blocking obese ZDF rats with defective OB-R, expression of ceramide synthesis, either with the competitive inhi- this enzyme is increased,16 as is the ability to synthe- bitor of SPT, L-cycloserine, or the fungal derivative, 16 size [3H]-ceramide from either [3H]-serine or [3H]- fumonisin B1. In addition, inhibitors of NO produc- palmitate5 (Figure 2). Transgenic overexpression of tion, such as nicotinamide and aminoguanidine, effec- tively block the lipotoxic and lipoapoptotic effects of FA in vitro. The thiazolidine dione compound, troglitazone, also protects islets in vitro, presumably through its antisteatotic effect.17 Both troglitazone9 and the iNOS inhibitors, aminoguanidine and nicotinamide,11 have been used thera- peutically to prevent diabetes in ZDF rats. When administered from the age of 6 weeks (before the onset of diabetes) until 12 weeks of age, at which time all untreated animals have become diabetic, all three agents are remarkably effective in preventing the diabetes and preserving the normal architecture and functional characteristics of the pancreatic islets. Troglitazone lowers the islet TG content and prevents the loss of b cells that otherwise occurs.9 The ®brosis and deformation of the islets is averted, together with the loss of GLUT-2 and glucose-respon- sive insulin secretion (Figure 3). Most strikingly, the profound alterations of mitochondria observed in the b cells of untreated ZDF rats do not occur9 (Figure 4). The mechanism of this dramatic action of troglitazone is not known. The lipogenic effect of the drug in islets is paradoxical inasmuch as this agent has lipogenic activity in adipocytes. The iNOS inhibitors aminoguanidine and nicotinamide also completely prevent the b cell destruction, loss of b cell function and the hyperglycemia that Figure 2 (A) Effect of transgenic expression of normal leptin 11 receptors on prevention of FA-induced apoptosis. In normal occurs in sham-treated controls (Figure 5). How- islets (not shown) DNA laddering, an index of apoptosis, is below ever, they appear to work at a more distal level in the 1%. In fat-laden islets of prediabetic ZDF rats laddering is over 7% lipoapoptotic pathway by blocking NO production. and rises to 15% in the presence of FA. Leptin has no effect on this. If leptin receptors (OB-Rb) are expressed, leptin prevents the FA- Thus, the pancreatic islets, because of their sensi- induced increase in laddering. (B) Lowering by FA of Bcl2 protein tivity to apoptotic factors, appear to be vulnerable to expression in islets of normal lean ( = ) and obese (fa=fa) ZDF rats. the ectopic accumulation of TG resulting from failure Leptin blocks the lowering effect in = but not fa=fa rats, but when fa=fa islets are transgenically altered to express the normal leptin of the normal system of intracellular FA homeostasis. receptor (Ob-Rb), leptin blocks Bcl2 lowering. Since Bcl2 is an It seems possible that other complications of obesity antiapoptotic factor, this may account for the antiapoptotic action involving nonadipose tissues may be the result of a of leptin shown in panel (A). (Reprinted with permission of Proc Natl Acad Sci USA). similar mechanism.

International Journal of Obesity Lipotoxicity in obesity RH Unger and L Orci S31

Figure 3 Immuno¯uorescent staining for insulin (A and B), glucagon (C and D) and glucose transporter 2 (GLUT-2) (E and F) in consecutive sections of an islet of an untreated ZDF rat (A, C, E) and one of a troglitazone-treated (B, D, F) ZDF rat. Compared with the treated condition, the untreated islet shows a reduced number of dispersed insulin-staining cells (A), glucagon-staining cells that have been displaced from their characteristic peripheral location in the islet by the topographical changes (C), and a marked eduction in the GLUT-2 ¯uorescence in b-cells (E). The inset in (E) enhances the staining in a few cells by confocal microscopy. The bars represent 50 mm; in the insert, 20 mm. (Reprinted with permission of Proc Natl Acad Sci USA).

Figure 4 (A) Thin section of an insulin cell from an untreated ZDF fa=fa rat, showing alterations of mitochondria (m) and reduction in the number of dense core secretory granules; areas of dense particles representing glycogen (gl) can also be detected with high frequency. (B) Thin section of a b-cell from a troglitazone-treated ZDF fa=fa rat. Both mitochondria and dense core secretory granules appear normal in aspect and number. G Golgi complex. (Reprinted with permission of Proc Natl Acad Sci USA).

International Journal of Obesity Lipotoxicity in obesity RH Unger and L Orci S32 Proc Natl Acad Sci USA 1998; 95: 2498 ± 2502. 6 Unger RH, Zhou YT, Orci L. Regulation of fatty acid homeostasis in cells: novel role of leptin. Proc Natl Acad Sci USA 1999; 96: 2327 ± 2332. 7 Iida M, Murakami T, Ishida K, Mizuno A, Kuwajima M, Shima K. Substitution at codon 269 (glutamine?proline) of the leptin receptor (OB-R) cDNA is the only mutation found in the Zucker fatty ( fa=fa) rat. Biochem Biophys Res Commun 1996; 224: 597 ± 604. 8 Phillips MS, Liu QY, Hammond HA, Dugan V, Hey PJ, Caskey CT, Hess JF. Leptin receptor missense mutation in the fatty Zucker rat. Nature Genet 1996; 13: 18 ± 19. 9 Higa M, Zhou YT, Ravazzola M, Baetens D, Orci L, Unger RH. Troglitazone prevents mitochondrial alterations, b-cell destruction and diabetes in obese prediabetic rats. Proc Natl Acad Sci USA 1999; 96: 11513 ± 11518. 10 Shimabukuro M, Wang MY, Zhou YT, Newgard CB, Unger RH. Protection against b-cell lipoapoptosis through leptin- dependent maintenance of Bcl-2 expression. Proc Natl Acad Figure 5 (A) Effects of iNOS inhibitors on b-cells of fa=fa ZDF Sci USA 1998; 95: 9558 ± 9561. prediabetic rats given sham treatment or treatment with nico- 11 Shimabukuro M, Ohneda M, Lee YH, Unger RH. Role of tinamide or aminoguanidine for 7 weeks. (B) Effect of nicotina- nitric oxide in obesity-induced b-cell disease. J Clin Invest mide or aminoguanidine treatment on blood glucose of the fa=fa 1997; 100: 290 ± 295. ZDF prediabetic rats. b-Cell function is preserved and diabetes 12 Zhou YT, Shimabukuro M, Wang MY, Lee Y, Higa M, prevented. (Reprinted with permission of J Clin Invest). Milburn JL, Newgard CB, Unger RH. Role of peroxisome proliferator-activated receptor a in disease of pancreatic b- Acknowledgements cells. Proc Natl Acad Sci USA 1998; 95: 8898 ± 8903. We acknowledge the grant support of the Department 13 Zhou YT, Wang ZW, Higa M, Newgard CB, Unger RH. Reversing adipocyte differentiation: implications for treatment of Veterans Affairs Institutional Support (SMI 821- of obesity. Proc Natl Acad Sci USA 1999; 96: 2391 ± 2395. 109), The National Institutes of Health (DK02700- 14 Zhou YT, Shimabukuro M, Lee Y, Koyama K, Higa M, 37), The National Institutes of Health=Juvenile Dia- Ferguson T, Unger RH. Enhanced de novo lipogenesis in the betes Foundation Diabetes Interdisciplinary Research leptin-unresponsive pancreatic islets of prediabetic Zucker Program, Novo-Nordisk Corporation, Sankyo Inc., diabetic fatty rats: role in the pathogenesis of lipotoxic diabetes. Diabetes 1998; 47: 1904 ± 1908. and Swiss National Science Foundation (to L Orci). 15 Katsuyama K, Shichiri M, Marumo F, Hirata Y. Role of nuclear factor kB activation in cytokine-and sphingomyeli- nase-stimulated inducible nitric oxide synthase gene expres- References sion in vascular smooth muscle cells. Endocrinology 1998; 1 Neel JV. The `thrifty genotype' in 1998. Nutr Rev 1999; 139: 4506 ± 4512. 57(5Pt2): S2 ± S9. 16 Shimabukuro M, Higa M, Zhou YT, Wang MY, Newgard CB, 2 Unger RH. How obesity causes diabetes in Zucker diabetic Unger RH. Lipoapoptosis in b-cells of obese prediabetic fa=fa fatty rats. Trends Endocrinol Metab 1997; 7: 276 ± 282. rats: role of serine palmitoyl transferase overexpression. J Biol 3 Campbell PJ, Carlson MG, Nurghan NM. Fat metabolism in Chem 1998; 273: 32487 ± 32490. human obesity. Am J Physiol 1994; 266: E600 ± E605. 17 Shimabukuro M, Zhou YT, Lee Y, Unger RH. Troglitazone 4 Hamilton JA. Fatty acid transport: dif®cult or easy? J Lipid lowers islet fat and restores b-cell function of Zucker diabetic Res 1998; 39: 467 ± 481. fatty rats. J Biol Chem 1998; 273: 3547 ± 3550. 5 Shimabukuro M, Zhou YT, Levi M, Unger RH. Fatty acid- induced b-cell apoptosis: a link between obesity and diabetes.

International Journal of Obesity