0021-972X/91/7201-0096$02.00/0 Journal of Clinical Endocrinology and Metabolism Copyright© 1991 by The Endocrine Society
Vol. 72, No. 1 Printed in U.S.A.
The Role of Free Fatty Acid Metabolism in the Pathogenesis of Insulin Resistance in Obesity and Noninsulin-Dependent Diabetes Mellitus* LEIF C. GROOPf, CAROLA SALORANTA, MYRON SHANK, RICCARDO C. BONADONNAJ, ELEUTERIO FERRANNINI, AND RALPH A. D E F R O N Z O Fourth Department of Medicine, Helsinki University Hospital (L.C.G., C.S.), Helsinki, Finland; and the C.N.R. Institute of Clinical Physiology and the Second Medical Clinic, University of Pisa (E.F.), Pisa, Italy; Department of Medicine, The University of Texas Health Science Center and Veterans Administration Hospital, San Antonio, Texas (R.C.B., R.A.D., M.S.)
ABSTRACT. To investigate the mechanisms of insulin resistance in obesity and noninsulin-dependent diabetes mellitus (NIDDM), we examined oxidative and nonoxidative pathways of free fatty acid (FFA) and glucose metabolism in 14 lean and 17 obese (with normal oral glucose tolerance) nondiabetic subjects and in 8 lean and 8 obese subjects with NIDDM. FFA and glucose metabolism were measured using the sequential insulin clamp technique in combination with indirect calorimetry and infusion of [3-3H]glucose and [l-14C]palmitate. Obesity was characterized by enlarged fat mass, which correlated positively with the plasma FFA concentration (r = 0.62; P < 0.01). FFA metabolism was less sensitive to insulin in obese than in lean nondiabetic subjects, but this defect could be overcome by increasing the plasma insulin concentration. NIDDM patients showed normal sensitivity to the inhibitory action of insulin on FFA metabolism; however, maximal suppression by insulin was impaired. The combination of obesity and NIDDM was associated with a further enhancement of reesterification of FFA than
observed in either condition alone. In both obesity and NIDDM, the dose-response curve for suppression of hepatic glucose production by insulin was impaired. While obesity was primarily characterized by reduced sensitivity to the stimulatory action of insulin on oxidative and nonoxidative pathways of glucose metabolism, resistance to the effect of insulin on glucose metabolism in NIDDM was characterized by a reduced maximal response. The combination of obesity and NIDDM further impaired the sensitivity of liver glucose output and glucose oxidation to insulin. The hypothesis is advanced that in uncomplicated obesity, increased availability and oxidation of FFA leads, by the FFA/ glucose cycle, to the impairment in glucose utilization. In NIDDM, on the other hand, the defect in glucose utilization is primary, and the enhanced rate of FFA oxidation may represent a compensatory phenomenon. (J Clin Endocrinol Metab 72: 96107, 1991)
R
in adipocytes of subjects with obesity and NIDDM (12, 14). In vivo studies measuring the fall in plasma FFA concentration in response to insulin have demonstrated both normal (14-17) and impaired (13, 18, 19) suppression of FFA by insulin. The plasma FFA concentration is determined by their rate of appearance from adipose tissue (lipolysis) and their rate of disappearance from plasma. The latter can occur by two pathways: oxidation to carbon dioxide and water, and reesterification to triglycerides. There is little information available as to which extent these pathways are altered by obesity and NIDDM. Increased plasma FFA concentrations are associated with an enhanced rate of FFA oxidation, and this, in turn, has been related to defects in both the oxidative and nonoxidative pathways of glucose metabolism (2025). A potential link between elevated plasma FFA concentrations and impaired glucose metabolism was ini-
ESISTANCE to the effect of insulin on glucose metabolism is a characteristic feature of both obesity (1-6) and noninsulin-dependent diabetes mellitus (NIDDM) (6-9) and involves both oxidative and nonoxidative pathways (10-11). There are, however, controversial data on whether FFA metabolism is resistant to insulin in these conditions (12-15). Most in vitro studies have demonstrated normal antilipolytic effects of insulin Received September 6, 1989. Address all correspondence and requests for reprints to: Leif C. Groop, M.D., Fourth Department of Medicine, Helsinki University Hospital, Unioninkatu 38, SF-00170 Helsinki, Finland. * Work in the U.S. was supported in part by NIH Grant AM-24092, CRC Grant RR-125, and a V.A. Merit Award (to R.A.D.). Work in Finland was supported by grants from the Sigrid Juselius Foundation, Nordisk Insulinfond, and the Finnish Medical Society (Finska Lakaresallskapet). t Recipient of a Fogarty International Fellowship (F05-TWO-3451). $ Recipient of a fellowship from the Italian Ministry of Public Education.
96
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 13 March 2015. at 09:21 For personal use only. No other uses without permission. . All rights reserved.
FFA METABOLISM AND INSULIN RESISTANCE
97
Euglycemic insulin clamp. A five-step euglycemic insulin clamp was performed on 2 separate days in random order. In the first study the subjects received three different infusions of insulin (30, 144, and 720 pmol/m2-min); during the second they received insulin infusions of 72 and 288 pmol/m2-min. The studies were carried out in combination with indirect calorimetry and infusion of [3-3H]glucose and [l-14C]palmitate to allow quantitation of respiratory gas exchange, hepatic glucose production (HGP), and FFA turnover, respectively. For the twostep euglycemic insulin clamp (72 and 288 pmol/m2 • min) the subjcts were admitted to the Clinical Research Center on the morning of the study. For the three-step euglycemic insulin clamp (30, 144, and 720 pmol/m2-min) the subjects were admitted on the evening before the study. All studies were started at 0800 h after a 12-h overnight fast. After obtaining at least four basal samples for insulin, FFA, and glucose determinations, a primed constant infusion of crystalline porcine insulin (Lilly, Indianapolis, IN) was administered to acutely raise and maintain the plasma insulin concentration at the desired level Materials and Methods throughout the insulin clamp procedure by a variable infusion of 20% glucose (27). Plasma glucose was determined at 5-min Subjects intervals, and plasma insulin and FFA at 15-min intervals Four groups of subjects were studied (Table 1): 1) 14 lean during the clamp. In nondiabetic subjects each hyperinsulinemic step lasted 100 min. In the NIDDM subjects, no glucose healthy control subjects, 2) 17 age-matched obese subjects, 3) was infused until plasma glucose had declined to the desired 8 lean patients with NIDDM, and 4) 8 obese patients with level. During the 30 and 72 pmol/m2 • min insulin steps, it was NIDDM. None of the subjects had clinical or laboratory evinot possible to lower plasma glucose to the desired goal of 5.6 dence of cardiac, hepatic, renal, or endocrine disease other than mmol/L even though the time interval was extended to 130 diabetes. Seven of the lean and 6 of the obese diabetics were min. Therefore, from 90-130 min plasma glucose was clamped treated with sulfonylureas, while the remaining 3 patients were at the hyperglycemic level that was present at 90 min. At 130 controlled with diet alone. Sulfonylurea treatment was stopped min, the insulin space was reprimed for the second step of the 1 week before the studies. Before participating in the experiinsulin clamp {i.e. 144 and 288 pmol/m2-min), and plasma mental protocol, the purpose, nature, and potential risks of the glucose was allowed to decline to 5.6 mmol/L, at which level it study were explained, and informed written consent was then was clamped. The second (144 and 288 pmol/m2 • min) and third obtained from each subject. The study protocol was approved (720 pmol/m2 • min) steps of the insulin clamp each lasted 100 by the Human Investigation Committee of the Yale University min. The steady state plasma glucose and insulin concentraSchool of Medicine. tions during the different insulin clamp steps are shown in Table 2. The coefficient of variation in plasma glucose and Experimental protocol insulin concentrations during the last 40 min of the five insulin Oral glucose tolerance test (OGTT). All subjects were given a 75- clamp steps ranged from 3.6 ± 0.4% to 4.8 ± 0.4% (glucose) g oral glucose load at 0800 h after a 12-h overnight fast. Blood and from 7.8 ± 0.9% to 8.2 ± 0.6% (insulin). samples for determination of plasma glucose, insulin, and FFA Glucose turnover. One hundred and fifty minutes before the concentrations were drawn at 15- to 30-min intervals (Fig. 1).
tially suggested by Randle et al. (26). According to the Randle cycle an increased rate of FFA oxidation, by altering the redox potential of the cell and by inhibiting several key enzymes in the glycolytic and citric acid cycles, leads to inhibition of glycolysis and glucose oxidation. Despite these potentially important interactions between glucose and FFA metabolism, the role of the FFA/glucose cycle in the genesis of impaired glucose metabolism in obesity and NIDDM is still unclear. The current study was designed to examine the influence of obesity, NIDDM, and their combination on the different pathways of glucose and FFA metabolism. If insulin resistance also involves FFA metabolism, this could be compatible with the concept of an operative FF A/glucose cycle as a cause of impaired insulin action on glucose metabolism in NIDDM and obesity.
TABLE 1. Clinical characteristics of subjects Nondiabetic subjects Lean Obese No. (females/males) Age (yr) BW (kg) LBM (kg) Body mass index (kg/m2) Fat mass (%) Fasting plasma glucose (mmol/L) Fasting plasma insulin (pmol/L) Plasma FFA (Mmol/L)
14 (9/5) 48 ± 5 68 ± 2 50 ± 3 22.9 ± 0.6 25 ± 2 4.9 ± 0.1 58 ± 7 652 ± 77
17 (8/9) 53 ± 3 90 ± 3" 56 ± 3 30.5 ± 1.1° 38 ±2° 5.0 ± 0.1 101 ± 14° 812 ± 47
Diabetic subjects Lean
Obese
8 (5/3) 62 ± 2" 69 ± 4 52 ± 4 23.4 ± 0.8 26 ± 2 9.5 ± 1.0° 58 ± 7 857 ± 70
8 (5/3) 57 ± 2 87 ± 4 " 54 ± 3 32.9 ± 1.3° 38 ±3° 9.1 ± 0.6° 137 ± 29° 859 ± 70
Values are the mean ± SEM. " P < 0.05 or less vs. nondiabetic lean subjects (by analysis of variance).
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 13 March 2015. at 09:21 For personal use only. No other uses without permission. . All rights reserved.
GROOP ET AL.
98 18 16
Vol. 72, No. 1 Printed in U.S.A.
The Role of Free Fatty Acid Metabolism in the Pathogenesis of Insulin Resistance in Obesity and Noninsulin-Dependent Diabetes Mellitus* LEIF C. GROOPf, CAROLA SALORANTA, MYRON SHANK, RICCARDO C. BONADONNAJ, ELEUTERIO FERRANNINI, AND RALPH A. D E F R O N Z O Fourth Department of Medicine, Helsinki University Hospital (L.C.G., C.S.), Helsinki, Finland; and the C.N.R. Institute of Clinical Physiology and the Second Medical Clinic, University of Pisa (E.F.), Pisa, Italy; Department of Medicine, The University of Texas Health Science Center and Veterans Administration Hospital, San Antonio, Texas (R.C.B., R.A.D., M.S.)
ABSTRACT. To investigate the mechanisms of insulin resistance in obesity and noninsulin-dependent diabetes mellitus (NIDDM), we examined oxidative and nonoxidative pathways of free fatty acid (FFA) and glucose metabolism in 14 lean and 17 obese (with normal oral glucose tolerance) nondiabetic subjects and in 8 lean and 8 obese subjects with NIDDM. FFA and glucose metabolism were measured using the sequential insulin clamp technique in combination with indirect calorimetry and infusion of [3-3H]glucose and [l-14C]palmitate. Obesity was characterized by enlarged fat mass, which correlated positively with the plasma FFA concentration (r = 0.62; P < 0.01). FFA metabolism was less sensitive to insulin in obese than in lean nondiabetic subjects, but this defect could be overcome by increasing the plasma insulin concentration. NIDDM patients showed normal sensitivity to the inhibitory action of insulin on FFA metabolism; however, maximal suppression by insulin was impaired. The combination of obesity and NIDDM was associated with a further enhancement of reesterification of FFA than
observed in either condition alone. In both obesity and NIDDM, the dose-response curve for suppression of hepatic glucose production by insulin was impaired. While obesity was primarily characterized by reduced sensitivity to the stimulatory action of insulin on oxidative and nonoxidative pathways of glucose metabolism, resistance to the effect of insulin on glucose metabolism in NIDDM was characterized by a reduced maximal response. The combination of obesity and NIDDM further impaired the sensitivity of liver glucose output and glucose oxidation to insulin. The hypothesis is advanced that in uncomplicated obesity, increased availability and oxidation of FFA leads, by the FFA/ glucose cycle, to the impairment in glucose utilization. In NIDDM, on the other hand, the defect in glucose utilization is primary, and the enhanced rate of FFA oxidation may represent a compensatory phenomenon. (J Clin Endocrinol Metab 72: 96107, 1991)
R
in adipocytes of subjects with obesity and NIDDM (12, 14). In vivo studies measuring the fall in plasma FFA concentration in response to insulin have demonstrated both normal (14-17) and impaired (13, 18, 19) suppression of FFA by insulin. The plasma FFA concentration is determined by their rate of appearance from adipose tissue (lipolysis) and their rate of disappearance from plasma. The latter can occur by two pathways: oxidation to carbon dioxide and water, and reesterification to triglycerides. There is little information available as to which extent these pathways are altered by obesity and NIDDM. Increased plasma FFA concentrations are associated with an enhanced rate of FFA oxidation, and this, in turn, has been related to defects in both the oxidative and nonoxidative pathways of glucose metabolism (2025). A potential link between elevated plasma FFA concentrations and impaired glucose metabolism was ini-
ESISTANCE to the effect of insulin on glucose metabolism is a characteristic feature of both obesity (1-6) and noninsulin-dependent diabetes mellitus (NIDDM) (6-9) and involves both oxidative and nonoxidative pathways (10-11). There are, however, controversial data on whether FFA metabolism is resistant to insulin in these conditions (12-15). Most in vitro studies have demonstrated normal antilipolytic effects of insulin Received September 6, 1989. Address all correspondence and requests for reprints to: Leif C. Groop, M.D., Fourth Department of Medicine, Helsinki University Hospital, Unioninkatu 38, SF-00170 Helsinki, Finland. * Work in the U.S. was supported in part by NIH Grant AM-24092, CRC Grant RR-125, and a V.A. Merit Award (to R.A.D.). Work in Finland was supported by grants from the Sigrid Juselius Foundation, Nordisk Insulinfond, and the Finnish Medical Society (Finska Lakaresallskapet). t Recipient of a Fogarty International Fellowship (F05-TWO-3451). $ Recipient of a fellowship from the Italian Ministry of Public Education.
96
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 13 March 2015. at 09:21 For personal use only. No other uses without permission. . All rights reserved.
FFA METABOLISM AND INSULIN RESISTANCE
97
Euglycemic insulin clamp. A five-step euglycemic insulin clamp was performed on 2 separate days in random order. In the first study the subjects received three different infusions of insulin (30, 144, and 720 pmol/m2-min); during the second they received insulin infusions of 72 and 288 pmol/m2-min. The studies were carried out in combination with indirect calorimetry and infusion of [3-3H]glucose and [l-14C]palmitate to allow quantitation of respiratory gas exchange, hepatic glucose production (HGP), and FFA turnover, respectively. For the twostep euglycemic insulin clamp (72 and 288 pmol/m2 • min) the subjcts were admitted to the Clinical Research Center on the morning of the study. For the three-step euglycemic insulin clamp (30, 144, and 720 pmol/m2-min) the subjects were admitted on the evening before the study. All studies were started at 0800 h after a 12-h overnight fast. After obtaining at least four basal samples for insulin, FFA, and glucose determinations, a primed constant infusion of crystalline porcine insulin (Lilly, Indianapolis, IN) was administered to acutely raise and maintain the plasma insulin concentration at the desired level Materials and Methods throughout the insulin clamp procedure by a variable infusion of 20% glucose (27). Plasma glucose was determined at 5-min Subjects intervals, and plasma insulin and FFA at 15-min intervals Four groups of subjects were studied (Table 1): 1) 14 lean during the clamp. In nondiabetic subjects each hyperinsulinemic step lasted 100 min. In the NIDDM subjects, no glucose healthy control subjects, 2) 17 age-matched obese subjects, 3) was infused until plasma glucose had declined to the desired 8 lean patients with NIDDM, and 4) 8 obese patients with level. During the 30 and 72 pmol/m2 • min insulin steps, it was NIDDM. None of the subjects had clinical or laboratory evinot possible to lower plasma glucose to the desired goal of 5.6 dence of cardiac, hepatic, renal, or endocrine disease other than mmol/L even though the time interval was extended to 130 diabetes. Seven of the lean and 6 of the obese diabetics were min. Therefore, from 90-130 min plasma glucose was clamped treated with sulfonylureas, while the remaining 3 patients were at the hyperglycemic level that was present at 90 min. At 130 controlled with diet alone. Sulfonylurea treatment was stopped min, the insulin space was reprimed for the second step of the 1 week before the studies. Before participating in the experiinsulin clamp {i.e. 144 and 288 pmol/m2-min), and plasma mental protocol, the purpose, nature, and potential risks of the glucose was allowed to decline to 5.6 mmol/L, at which level it study were explained, and informed written consent was then was clamped. The second (144 and 288 pmol/m2 • min) and third obtained from each subject. The study protocol was approved (720 pmol/m2 • min) steps of the insulin clamp each lasted 100 by the Human Investigation Committee of the Yale University min. The steady state plasma glucose and insulin concentraSchool of Medicine. tions during the different insulin clamp steps are shown in Table 2. The coefficient of variation in plasma glucose and Experimental protocol insulin concentrations during the last 40 min of the five insulin Oral glucose tolerance test (OGTT). All subjects were given a 75- clamp steps ranged from 3.6 ± 0.4% to 4.8 ± 0.4% (glucose) g oral glucose load at 0800 h after a 12-h overnight fast. Blood and from 7.8 ± 0.9% to 8.2 ± 0.6% (insulin). samples for determination of plasma glucose, insulin, and FFA Glucose turnover. One hundred and fifty minutes before the concentrations were drawn at 15- to 30-min intervals (Fig. 1).
tially suggested by Randle et al. (26). According to the Randle cycle an increased rate of FFA oxidation, by altering the redox potential of the cell and by inhibiting several key enzymes in the glycolytic and citric acid cycles, leads to inhibition of glycolysis and glucose oxidation. Despite these potentially important interactions between glucose and FFA metabolism, the role of the FFA/glucose cycle in the genesis of impaired glucose metabolism in obesity and NIDDM is still unclear. The current study was designed to examine the influence of obesity, NIDDM, and their combination on the different pathways of glucose and FFA metabolism. If insulin resistance also involves FFA metabolism, this could be compatible with the concept of an operative FF A/glucose cycle as a cause of impaired insulin action on glucose metabolism in NIDDM and obesity.
TABLE 1. Clinical characteristics of subjects Nondiabetic subjects Lean Obese No. (females/males) Age (yr) BW (kg) LBM (kg) Body mass index (kg/m2) Fat mass (%) Fasting plasma glucose (mmol/L) Fasting plasma insulin (pmol/L) Plasma FFA (Mmol/L)
14 (9/5) 48 ± 5 68 ± 2 50 ± 3 22.9 ± 0.6 25 ± 2 4.9 ± 0.1 58 ± 7 652 ± 77
17 (8/9) 53 ± 3 90 ± 3" 56 ± 3 30.5 ± 1.1° 38 ±2° 5.0 ± 0.1 101 ± 14° 812 ± 47
Diabetic subjects Lean
Obese
8 (5/3) 62 ± 2" 69 ± 4 52 ± 4 23.4 ± 0.8 26 ± 2 9.5 ± 1.0° 58 ± 7 857 ± 70
8 (5/3) 57 ± 2 87 ± 4 " 54 ± 3 32.9 ± 1.3° 38 ±3° 9.1 ± 0.6° 137 ± 29° 859 ± 70
Values are the mean ± SEM. " P < 0.05 or less vs. nondiabetic lean subjects (by analysis of variance).
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 13 March 2015. at 09:21 For personal use only. No other uses without permission. . All rights reserved.
GROOP ET AL.
98 18 16
