Point-Counterpoint SEE LEBOVITZ AND BANERJI (P. 2322)

Counterpoint: Visceral Adiposity Is Not Causally Related to Insulin Resistance

t has been known for over half a cen- given waist circumference. Waist circum- and/or variability in the relationship be- tury that accumulation of fat in the up- ference reflects total truncal fat, which tween the size of a fat depot and its lipo- I per body is associated with metabolic consists of both subcutaneous and intra- lytic activity (23). The latter phenomenon complications of obesity. The mechanism abdominal (predominantly mesenteric could occur, for example, if a given fat responsible for this association is widely and omental, also referred to as visceral) depot contained fewer adipocytes of believed to be insulin resistance, which is fat. It has been suggested that adipose tis- larger size in one subject than in another. frequently present in individuals with up- sue in the visceral compartment has One would expect greater lipolysis and per-body obesity (1) and is associated greater lipolytic potential than subcutane- greater resistance to the antilipolytic effect with increased cardiovascular risk (2). ous adipose tissue (17). Although the of insulin when fat cells are larger and The upper-body obesity phenotype is ac- visceral fat depot is relatively small com- fewer in number (24). companied by increased free fatty acid pared with subcutaneous fat mass, the Other studies that have been cited as (FFA) flux and concentration, whereas in- possibility that it is more lipolytically ac- evidence for a key role of visceral fat in dividuals with predominantly lower- tive on a per-unit-mass basis and the fact insulin resistance have ambiguities that body obesity tend to have normal FFA that the release of FFA directly into the make their interpretation difficult. Surgi- kinetics (3). Increased FFA concentra- portal vein creates a “first pass” effect in cal removal of perinephric fat in obese rats tions, in turn, are thought to be a major the liver means that that organ may be produced improvement in insulin- mediator of insulin resistance (4) and exposed to higher FFA concentrations mediated glucose disposal (25). This have been shown to cause endothelial than peripheral tissues, especially in those should not be interpreted as an indication dysfunction (5), impair pancreatic ␤-cell with visceral obesity. A recently pub- that visceral lipolysis has an impact on function (6), and acutely raise blood pres- lished study found that the amount of systemic insulin action, however, because sure (7). Increased delivery of FFA to FFAs released by visceral fat does in fact the venous drainage of perinephric fat in skeletal muscle can result in insulin resis- correlate with visceral fat mass (18). the rat is caval, not portal. In humans, tance in that tissue, either directly or via There is a strong correlation between the individuals undergoing vertical-banded accumulation of increased intramyocellu- amount of visceral fat present and circu- gastroplasty and omentectomy simulta- lar triglyceride stores (8). Similarly, in- lating VLDL triglyceride concentrations neously had a greater improvement in in- creased delivery of FFA to the liver may be (19), which are thought to correlate in sulin action than those who had only responsible for hepatic insulin resistance turn with the delivery of FFA to the liver vertical-banded gastroplasty (26). How- (9), accumulation of intrahepatocellular (11). Increased visceral fat mass is associ- ever, the subjects who underwent omen- triglyceride (10), and increased synthesis ated with glucose intolerance (19), and tectomy (n ϭ 19) had a 9-kg greater and secretion of VLDL (11). Hormone- several studies have shown that insulin weight loss than those who did not (n ϭ sensitive lipase in adipocytes is exquis- sensitivity correlates with visceral fat mass 18), a difference that, although not statis- itely sensitive to the antilipolytic effect of in normal subjects (20) and people with tically significant with this relatively small insulin in lean healthy individuals (12). type 2 diabetes (21). number of subjects, is likely important. However, impaired insulin suppression of adipose tissue lipolysis is a feature of, Relationship between visceral The contribution of visceral and and may even be the primary cause of, obesity and insulin resistance subcutaneous fat to portal and insulin resistance (13). To our knowl- It seems logical that an increase in visceral peripheral FFA edge, insulin resistance has not been re- fat mass could result in greater FFA re- Insulin-resistant states are marked by in- ported in individuals with normal lease into the portal circulation, and the sulin resistance in both skeletal muscle sensitivity to the antilipolytic effects of association between visceral fat mass and and liver (27). It has been suggested that insulin. insulin sensitivity seems clear. However, increased delivery of FFA to the liver pro- whether increased visceral fat mass plays duces hepatic insulin resistance by stim- What is visceral obesity? a true causal role in insulin resistance is ulating gluconeogenesis (9). Since the The degree of upper-body obesity can be not as obvious. Some of the same studies rate of visceral lipolysis correlates with estimated by measurement of waist cir- that demonstrate a strong correlation be- visceral fat mass (18), and in view of the cumference, which is predictive of cardio- tween insulin sensitivity and visceral fat “first pass” effect, it seems logical to spec- vascular risk (14). However, nonobese mass also show that insulin sensitivity is ulate that increased visceral fat could be individuals exhibit considerable variabil- strongly associated with subcutaneous fat responsible for hepatic insulin resistance. ity in insulin sensitivity (15) that is not mass (20). There is actually considerable However, if the majority of FFAs deliv- necessarily explained by differences in variability in results regarding the rela- ered to the liver were to come from sub- waist circumference (16). It is possible tionship between insulin sensitivity and cutaneous fat, then abnormalities in that differences in the distribution of up- regional fat depots in humans (22). This subcutaneous lipolysis could be more im- per-body fat could explain variability in could be due to technical issues related to portant than visceral lipolysis as a cause of insulin sensitivity at a given BMI and a measurement of the visceral fat depot (22) hepatic insulin resistance, especially con-

2326 DIABETES CARE, VOLUME 28, NUMBER 9, SEPTEMBER 2005 Miles and Jensen sidering that subcutaneous fat stores tend cerally obese. A potential model for iso- However, direct evidence to support this to be increased in individuals who have lated visceral obesity is the individual idea is lacking. Moreover, a recent study increased visceral fat mass, even when all who is “metabolically obese” but of nor- in type 2 diabetic patients who were given of the subjects studied are of normal mal weight (MONW), as described by Ru- acipimox for 7 days to lower plasma FFA weight (28). Relatively little information derman et al. (31). Unfortunately, no concentrations makes this notion less on the contribution of visceral lipolysis to studies are available that relate FFA me- attractive. In that study, FFA concentra- FFA delivery to the liver is available. tabolism to insulin sensitivity in normal- tions were reduced and a significant im- The contribution of visceral fat to sys- weight individuals. Dvorak et al. (28) provement in insulin action was observed temic FFA availability also has been un- reported body composition data in 13 in the absence of changes in adiponectin, certain. Recently, Nielsen et al. (18) MONW women and 58 insulin-sensitive resistin, interleukin-6, and tumor necro- measured regional fat depots and visceral control subjects. They found that MONW sis factor-␣ levels (33). This observation and systemic lipolysis in 44 obese and 24 women had higher visceral and upper- does not necessarily mean that adipocyto- lean subjects using combined isotope di- body subcutaneous fat than control sub- kines are not mediators of insulin resis- lution and arteriovenous sampling (leg, jects; moreover, the MONW individuals tance. It does suggest, however, that FFAs splanchnic) techniques. Visceral lipolysis had slightly higher BMIs than control are important. Additional research is was responsible for only 5–10% and 20– subjects (22.5 vs. 21.5 kg/m2, P ϭ 0.08). needed to delineate the relative impor- 25% of total FFA delivery to the liver in A study in nonobese Japanese subjects tance of FFAs and adipocytokines in this lean and obese subjects, respectively, al- found insulin resistance (using the eugly- regard. though the contribution of visceral fat was cemic clamp) and greater visceral fat in proportional to the size of the visceral fat MONW compared with lean control sub- depot. Importantly, however, the contri- jects (32). However, the MONW subjects Summary bution of visceral lipolysis to systemic were selected for visceral obesity, and The data on the role of increased visceral (i.e., extrahepatic) FFA availability was they had both significantly higher BMIs fat as a cause of insulin resistance are con- very small—generally Ͻ5% (18). and greater subcutaneous fat than the flicting. There are probably several rea- It remains possible that isolated vis- control group (both P Ͻ 0.01). Thus, iso- sons for the confusion. Methodologic ceral “hyperlipolysis” could be a factor in lated visceral obesity is not known to oc- imprecision in the measurement of body systemic insulin resistance. For this to be cur in humans. fat depots and real biological variation in the case, however (considering that the The RISC (Relationship between In- lipolytic activity of fat depots may both contribution of visceral fat to the periph- sulin Sensitivity and Cardiovascular dis- play a role. It appears that visceral fat has eral FFA pool is so small and assuming ease risk) study, now underway in 20 very little if any role in the oversupply of that FFAs are the chief mediators of insu- European centers, provides preliminary FFAs to extrahepatic tissues in insulin- lin resistance), systemic insulin resistance observations of potential interest. Among resistant states. Instead, subcutaneous fat, would have to result from FFA-mediated 625 individuals with BMI values Ͻ25 kg/ especially in the upper body, is the major hepatic insulin resistance in the absence m2, there was marked variation in insulin contributor to this known mediator of in- of insulin resistance in peripheral tissues. sensitivity (measured by the euglycemic sulin resistance. Increases in visceral fat Parenthetically, in this scenario, one clamp) and waist circumference. Distinct mass, even among normal-weight indi- would not expect to see the lipotoxicity groups of individuals were identified with viduals, are accompanied by increases in that may be responsible for decline in insulin resistance alone, increased waist subcutaneous fat mass. Thus, the weight ␤-cell function when systemic FFAs are circumference alone, both, and neither of evidence suggests that visceral fat may increased (29). A recent study in dogs re- (E. Ferrannini, personal communica- be responsible for hepatic insulin resis- ceiving a moderately high-fat diet demon- tion). Although FFAs and regional fat tance in some circumstances but is no strated an increase in both visceral and mass were not measured in the RISC more than a marker for insulin resistance subcutaneous fat, with insulin resistance study, these findings underscore the po- in extrahepatic tissues. Future studies primarily due to impaired insulin sup- tential complexity of the relationship be- should focus on the relationship between pression of endogenous glucose produc- tween the size of body fat depots and upper-body subcutaneous and visceral fat tion (30). There was no change in plasma insulin resistance. depots, regulation of FFA metabolism, FFA concentrations. Does isolated vis- and the relative contribution of insulin re- ceral obesity or isolated hepatic insulin Non-FFA mediators of insulin sistance in liver versus that in skeletal resistance occur in humans? There are no resistance muscle to systemic insulin resistance. data to suggest that this is the case. More- Increased delivery of FFAs is thought over, it seems unlikely considering that to be the primary effector of insulin re- JOHN M. MILES, MD increased fat stores are the common de- sistance in extrahepatic tissues, primar- MICHAEL D. JENSEN, MD nominator for abnormally high rates of ily skeletal muscle (4,13). Based on the lipolysis, and subcutaneous fat mass study of Nielsen et al. (18), it seems un- From the Endocrine Research Unit, Division of En- tends to be increased when visceral fat likely that visceral fat could contribute in docrinology, Metabolism and Nutrition, Mayo mass is increased (28). If isolated visceral any significant way to increased FFA sup- Clinic, Rochester, Minnesota. obesity exists, individuals with it would ply to skeletal muscle. It is possible that Address correspondence and reprint requests to John M. Miles, MD, Endocrine Research Unit, be expected to be at normal or near- visceral fat could mediate peripheral in- Mayo Clinic, Rochester, MN 55905. E-mail: miles. normal weight because the visceral fat sulin resistance via another mechanism, [email protected]. depot is relatively small, even in the vis- such as the release of adipocytokines. © 2005 by the American Diabetes Association.

DIABETES CARE, VOLUME 28, NUMBER 9, SEPTEMBER 2005 2327 Point-Counterpoint

very low density lipoprotein production for the defense. J Clin Invest 113:1530– Acknowledgments— Supported in part by in humans. J Clin Invest 95:158–166, 1532, 2004 grants from the USPHS (HL67933 and 1995 24. Ravussin E, Smith SR: Increased fat in- DK40484) and the Mayo Foundation. 12. Jensen MD, Caruso M, Heiling VJ, Miles take, impaired fat oxidation, and failure of We thank W. Isley for helpful comments. JM: Insulin regulation of lipolysis in non- fat cell proliferation result in ectopic fat diabetic and IDDM subjects. Diabetes 38: storage, insulin resistance, and type 2 di- ●●●●●●●●●●●●●●●●●●●●●●● 1595–1601, 1989 abetes mellitus. Ann N Y Acad Sci 967: References 13. Bays H, Mandarino L, DeFronzo RA: Role 363–378, 2002 1. Evans DJ, Murray R, Kissebah AH: Rela- of the adipocyte, free fatty acids, and ec- 25. Barzilai N, She L, Liu BQ, Vuguin P, Co- tionship between skeletal muscle insulin topic fat in pathogenesis of type 2 diabetes hen P, Wang J, Rossetti L: Surgical re- resistance, insulin mediated glucose dis- mellitus: peroxisomal proliferator-acti- moval of visceral fat reverses hepatic posal and insulin binding: effects of obe- vated receptor agonists provide a rational insulin resistance. Diabetes 48:94–98, sity and body fat topography. J Clin Invest therapeutic approach. J Clin Endocrinol 1999 74:1515–1525, 1984 Metab 89:463–478, 2004 26. Thorne A, Lonnqvist F, Apelman J, 2. Stern MP: Do non-insulin-dependent di- 14. Pouliot MC, Despres JP, Lemieux S, Hellers G, Arner P: A pilot study of long- abetes mellitus and cardiovascular disease Moorjani S, Bouchard C, Tremblay A, term effects of a novel obesity treatment: share common antecedents? Ann Intern Nadeau A, Lupien PJ: Waist circumfer- Med 124:110–116, 1996 ence and abdominal sagittal diameter: omentectomy in connection with adjust- 3. Jensen MD, Haymond MW, Rizza RA, best simple anthropometric indexes of ab- able gastric banding. Int J Obes Relat Metab Cryer PE, Miles JM: Influence of body fat dominal visceral adipose tissue accumu- Disord 26:193–199, 2002 distribution on free fatty acid metabolism lation and related cardiovascular risk in 27. DeFronzo RA: Lilly Lecture 1987: The tri- ␤ in obesity. J Clin Invest 83:1168–1173, men and women. Am J Cardiol 73:460– umvirate: -cell, muscle, liver: a collusion 1989 468, 1994 responsible for NIDDM. Diabetes 37:667– 4. Boden G: Fatty acids and insulin resis- 15. Cusi K, Maezono K, Osman A, Pender- 687, 1988 tance (Review). Diabetes Care 19:394– grass M, Patti ME, Pratipanawatr T, De- 28. Dvorak RV, DeNino WF, Ades PA, Poehl- 395, 1996 Fronzo RA, Kahn CR, Mandarino LJ: man ET: Phenotypic characteristics asso- 5. Steinberg HO, Tarshoby M, Monestel R, Insulin resistance differentially affects the ciated with insulin resistance in meta- Hook G, Cronin J, Johnson A, Bayazeed B, PI 3-kinase- and MAP kinase-mediated bolically obese but normal-weight young Baron AD: Elevated circulating free fatty signaling in human muscle. J Clin Invest women. Diabetes 48:2210–2214, 1999 acid levels impair endothelium-depen- 105:311–320, 2000 29. Unger RH, Zhou YT: Lipotoxicity of dent vasodilation. J Clin Invest 100:1230– 16. Cibula D, Skrha J, Hill M, Fanta M, Haa- ␤-cells in obesity and in other causes of 1239, 1997 kova L, VrbIkova J, Zivny J: Prediction of fatty acid spillover. Diabetes 50 (Suppl. 6. Zhou YP, Grill VE: Long-term exposure of insulin sensitivity in nonobese women 1):S118–S121, 2001 rat pancreatic islets to fatty acids inhibits with polycystic ovary syndrome. J Clin En- 30. Kim SP, Ellmerer M, Van Citters GW, glucose-induced insulin secretion and docrinol Metab 87:5821–5825, 2002 Bergman RN: Primacy of hepatic insulin biosynthesis through a glucose fatty acid 17. Ostman J, Arner P, Engfeldt P, Kager L: resistance in the development of the met- cycle. J Clin Invest 93:870–876, 1994 Regional differences in the control of li- abolic syndrome induced by an isocaloric 7. Stojiljkovic MP, Zhang D, Lopes HF, Lee polysis in human adipose tissue. Metabo- moderate-fat diet in the dog. Diabetes 52: CG, Goodfriend TL, Egan BM: Hemody- lism 28:1198–1205, 1979 2453–2460, 2003 namic effects of lipids in humans. Am J 18. Nielsen S, Guo Z, Johnson CM, Hensrud 31. Ruderman NB, Schneider SH, Berchtold Physiol 280:R1674–R1679, 2001 DD, Jensen MD: Splanchnic lipolysis in P: The “metabolically-obese,” normal- 8. Kelley DE, Simoneau JA: Impaired free human obesity. J Clin Invest 113:1582– weight individual. Am J Clin Nutr 34: fatty acid utilization by skeletal muscle in 1588, 2004 1617–1621, 1981 non-insulin-dependent diabetes mellitus. 19. Despres JP: Abdominal obesity as impor- 32. Katsuki A, Sumida Y, Urakawa H, Gaba- J Clin Invest 94:2349–2356, 1994 tant component of insulin-resistance syn- zza EC, Murashima S, Maruyama N, Mo- 9. Bergman RN: New concepts in extracellu- drome. Nutrition 9:452–459, 1993 rioka K, Nakatani K, Yano Y, Adachi Y: lar signaling for insulin action: the single 20. Goodpaster BH, Thaete FL, Simoneau JA, gateway hypothesis. Recent Prog Horm Res Kelley DE: Subcutaneous abdominal fat Increased visceral fat and serum levels of 52:359–385, 1997 and thigh muscle composition predict in- triglyceride are associated with insulin re- 10. Westerbacka J, Corner A, Tiikkainen M, sulin sensitivity independently of visceral sistance in Japanese metabolically obese, Tamminen M, Vehkavaara S, Hakkinen fat. Diabetes 46:1579–1585, 1997 normal weight subjects with normal glu- AM, Fredriksson J, Yki-Jarvinen H: 21. Banerji MA, Lebowitz J, Chaiken RL, Gor- cose tolerance. Diabetes Care 26:2341– Women and men have similar amounts of don D, Kral JG, Lebovitz HE: Relationship 2344, 2003 liver and intra-abdominal fat, despite of visceral adipose tissue and glucose dis- 33. Bajaj M, Suraamornkul S, Kashyap S, Cusi more subcutaneous fat in women: impli- posal is independent of sex in black K, Mandarino L, DeFronzo RA: Sustained cations for sex differences in markers of NIDDM subjects. Am J Physiol 273:E425– reduction in plasma free fatty acid con- cardiovascular risk. Diabetologia 47: E432, 1997 centration improves insulin action with- 1360–1369, 2004 22. Garg A: Regional adiposity and insulin re- out altering plasma adipocytokine levels 11. Lewis G, Uffelman K, Szeto L, Weller B, sistance. J Clin Endocrinol Metab 89:4206– in subjects with strong family history of Steiner G: Interaction between free fatty 4210, 2004 type 2 diabetes. J Clin Endocrinol Metab acids and insulin in the acute control of 23. Klein S: The case of visceral fat: argument 89:4649–4655, 2004

2328 DIABETES CARE, VOLUME 28, NUMBER 9, SEPTEMBER 2005