Stephen Lillioja

Insulin Resistance and Insulin Secretory Dysfunction as Precursors of Non-Insulin-Dependent Diabetes Mellitus: Prospective Studies of Pima Indians

BACKGROUND The relative roles of obesity, insulin resistance, insulin secretory dysfunction, and excess hepatic glucose production in the development of non-insulin-dependent diabetes mellitus (NIDDM) are controversial. We conducted a prospective study to determine which of these factors predicted the development of the disease in a group of Pima Indians. METHODS A body-composition assessment, oral and intravenous glucose-tolerance tests, and a hyperinsulinemic--euglycemic clamp study were performed in 200 non-diabetic Pima Indians (87 women and 113 men; mean [+/- SD] age, 26 +/- 6 years). The subjects were followed yearly thereafter for an average of 5.3 years. RESULTS Diabetes developed in 38 subjects during follow-up. Obesity, insulin resistance (independent of obesity), and low acute plasma insulin response to intravenous glucose (with the degree of obesity and insulin resistance taken into account) were predictors of NIDDM: The six-year cumulative incidence of NIDDM was 39 percent in persons with values below the median for both insulin action and acute insulin response, 27 percent in those with values below the median for insulin action but above that for acute insulin response, 13 percent in those with values above the median for insulin action and below that for acute insulin response, and 0 in those with values originally above the median for both characteristics. CONCLUSIONS Insulin resistance is a major risk factor for the development of NIDDM: A low acute insulin response to glucose is an additional but weaker risk factor.

Skeletal Muscle Triglyceride Levels Are Inversely Related to Insulin Action

In animal studies, increased amounts of triglyceride associated with skeletal muscle (mTG) correlate with reduced skeletal muscle and whole body insulin action. The aim of this study was to test this relationship in humans. Subjects were 38 nondiabetic male Pima Indians (mean age 28 ± 1 years). Insulin sensitivity at physiological (M) and supraphysiological (MZ) insulin levels was assessed by the euglycemic clamp. Lipid and carbohydrate oxidation were determined by indirect calorimetry before and during insulin administration. mTG was determined in vastus lateralis muscles obtained by percutaneous biopsy. Percentage of body fat (mean 29 ± 1%, range 14–44%) was measured by underwater weighing. In simple regressions, negative relationships were found between mTG (mean 5.4 ± 0.3 μmol/g, range 1.3–1.9 μmol/g) and log10M (r = −0.53, P ≤ 0.001), MZ (r = −0.44, P = 0.006), and nonoxidative glucose disposal (r = −0.48 and −0.47 at physiological and supraphysiological insulin levels, respectively, both P = 0.005) but not glucose or lipid oxidation. mTG was not related to any measure of adiposity. In multiple regressions, measures of insulin resistance (log10M, MZ, log10[fasting insulin]) were significantly related to mTG independent of all measures of obesity (percentage of body fat, BMI, waist-to-thigh ratio). In turn, all measures of obesity were related to the insulin resistance measures independent of mTG. The obesity measures and mTG accounted for similar proportions of the variance in insulin resistance in these relationships. The results suggest that in this human population, as in animal models, skeletal muscle insulin sensitivity is strongly influenced by local supplies of triglycerides, as well as by remote depots and circulating lipids. The mechanism(s) underlying the relationship between mTG and insulin action on skeletal muscle glycogen synthesis may be central to an understanding of insulin resistance.

Reduced rate of energy expenditure as a risk factor for body-weight gain

The contribution of reduced energy expenditure to the development of obesity has been a point of controversy. We measured 24-hour energy expenditure (adjusted for body composition, age, and sex), in a respiratory chamber, in 95 southwestern American Indians. Energy expenditure correlated with the rate of change in body weight over a two-year follow-up period (r = -0.39, P less than 0.001). The estimated risk of gaining more than 7.5 kg in body weight was increased fourfold in persons with a low adjusted 24-hour energy expenditure (200 kcal per day below predicted values) as compared with persons with a high 24-hour energy expenditure (200 kcal per day above predicted values; P less than 0.01). In another 126 subjects, the adjusted metabolic rate at rest at the initial visit was also found to predict the gain in body weight over a four-year follow-up period. When the 15 subjects who gained more than 10 kg were compared with the remaining 111 subjects, the initial mean (+/- SD) adjusted metabolic rate at rest was lower in those who gained weight (1694 +/- 103 vs. 1764 +/- 109 kcal per day; P less than 0.02) and increased to 1813 +/- 134 kcal per day (P less than 0.01) after a mean weight gain of 15.7 +/- 5.7 kg. In a group of 94 siblings from 36 families, values for adjusted 24-hour energy expenditure aggregated in families (intraclass correlation = 0.48). We conclude that a low rate of energy expenditure may contribute to the aggregation of obesity in families.

Skeletal muscle capillary density and fiber type are possible determinants of in vivo insulin resistance in man.

We have compared the capillary density and muscle fiber type of musculus vastus lateralis with in vivo insulin action determined by the euglycemic clamp (M value) in 23 Caucasians and 41 Pima Indian nondiabetic men. M value was significantly correlated with capillary density (r = 0.63; P less than or equal to 0.0001), percent type I fibers (r = 0.29; P less than 0.02), and percent type 2B fibers (r = -0.38; P less than 0.003). Fasting plasma glucose and insulin concentrations were significantly negatively correlated with capillary density (r = -0.46, P less than or equal to 0.0001; r = -0.47, P less than or equal to 0.0001, respectively). Waist circumference/thigh circumference ratio was correlated with percent type 1 fibers (r = -0.39; P less than 0.002). These results suggest that diffusion distance from capillary to muscle cells or some associated biochemical change, and fiber type, could play a role in determining in vivo insulin action. The association of muscle fiber type with body fat distribution may indicate that central obesity is only one aspect of a more generalized metabolic syndrome. The data may provide at least a partial explanation for the insulin resistance associated with obesity and for the altered kinetics of insulin action in the obese.

Impaired glucose tolerance as a disorder of insulin action. Longitudinal and cross-sectional studies in Pima Indians.

Impaired glucose tolerance often presages the development of non-insulin-dependent diabetes mellitus. We have studied insulin action and secretion in 24 Pima Indians before and after the development of impaired glucose tolerance and in 254 other subjects representing the whole spectrum of glucose tolerance, including subjects with overt non-insulin-dependent diabetes. The transition from normal to impaired glucose tolerance was associated with a decrease in glucose uptake during hyperinsulinemia, from 0.018 to 0.016 mmol per minute (from 3.3 to 2.8 mg per kilogram of fat-free body mass per minute) (P less than 0.0003). Mean plasma insulin concentrations increased during an oral glucose-tolerance test, from 1200 to 1770 pmol per liter (from 167 to 247 microU per milliliter). In 151 subjects with normal glucose tolerance, the insulin concentration measured during an oral glucose-tolerance test correlated with the plasma glucose concentration (r = 0.48, P less than or equal to 0.0001). This relation was used to predict an insulin concentration of 1550 pmol per liter (216 microU per milliliter) in subjects with impaired glucose tolerance (actual value, 1590 pmol per liter [222 microU per milliliter]; P not significant), suggesting that these subjects had normal secretion of insulin. In contrast, plasma insulin concentrations in the diabetics decreased as glucose concentrations increased (r = -0.75, P less than or equal to 0.0001), suggesting deficient secretion of insulin. This relative insulin deficiency first appears at the lower end of the second (diabetic) mode seen in population frequency distributions of plasma glucose concentrations. Our data show that impaired glucose tolerance in our study population is primarily due to impaired insulin action. In patients with non-insulin-dependent diabetes mellitus, by contrast, impaired insulin action and insulin secretory failure are both present.

Relationships between insulin secretion, insulin action, and fasting plasma glucose concentration in nondiabetic and noninsulin-dependent diabetic subjects.

The relationships between insulin secretion, insulin action, and fasting plasma glucose concentration (FPG) were examined in 34 southwest American Indians (19 nondiabetics, 15 noninsulin-dependent diabetics) who had a broad range of FPG (88-310 mg/100 ml). Fasting, glucose-stimulated, and meal-stimulated plasma insulin concentrations were negatively correlated with FPG in diabetics but not in nondiabetics. In contrast, fasting and glucose-stimulated plasma C-peptide concentrations did not decrease with increasing FPG in either group and 24-h urinary C-peptide excretion during a diet of mixed composition was positively correlated with FPG for all subjects (r = 0.36, P less than 0.05). Fasting free fatty acid (FFA) was correlated with FPG in nondiabetics (r = 0.49, P less than 0.05) and diabetics (r = 0.77, P less than 0.001). Fasting FFA was also correlated with the isotopically determined endogenous glucose production rate in the diabetics (r = 0.54, P less than 0.05). Endogenous glucose production was strongly correlated with FPG in the diabetics (r = 0.90, P less than 0.0001), but not in the nondiabetics. Indirect calorimetry showed that FPG was also negatively correlated with basal glucose oxidation rates (r = -0.61, P less than 0.001), but positively with lipid oxidation (r = 0.74, P less than 0.001) in the diabetics. Insulin action was measured as total insulin-mediated glucose disposal, glucose oxidation, and storage rates, using the euglycemic clamp with simultaneous indirect calorimetry at plasma insulin concentrations of 135 +/- 5 and 1738 +/- 59 microU/ml. These parameters of insulin action were significantly, negatively correlated with FPG in the nondiabetics at both insulin concentrations, but not in the diabetics although all the diabetics had markedly decreased insulin action. We conclude that decreased insulin action is present in the noninsulin-dependent diabetics in this population and marked hyperglycemia occurs with the addition of decreased peripheral insulin availability. Decreased peripheral insulin availability leads to increased FFA concentrations and lipid oxidation rates (and probably also increased concentrations of gluconeogenic precursors) that together stimulate gluconeogenesis, hepatic glucose production, and progressive hyperglycemia.

Familial Dependence of the Resting Metabolic Rate

Human obesity is known to be a familial disorder. We studied 130 nondiabetic adult southwestern American Indians (74 men and 56 women) from 54 families to determine whether the resting metabolic rate, as measured by indirect calorimetry, is a familial trait that is independent of individual differences in fat-free mass (estimated mass of metabolically active tissue), age, and sex. We found that most of the variance in the resting metabolic rate (83 percent, P less than 0.0001) was accounted for by three covariates--fat-free mass, age, and sex--and that fat-free mass was the most important determinant. Family membership accounted for an additional 11 percent (P less than 0.0001) of the variance in the resting metabolic rate. Thus, resting metabolic rate is a familial trait in this population, and it is independent of differences in fat-free mass, age, and sex. We also found that persons from families with lower resting metabolic rates were no more obese than persons from families with higher metabolic rates. This finding may be partly explained by the close correlation between fat-free mass and percentage of body fat (r = 0.81, P less than 0.0001), which indicates that the resting metabolic rate, as adjusted for fat-free mass, is already partly adjusted for obesity. Only prospective studies will elucidate whether the familial dependence of the resting metabolic rate is a contributing mechanism to the familial predisposition to obesity.

Correlation between muscle glycogen synthase activity and in vivo insulin action in man.

We have studied the relationship between in vivo insulin-mediated glucose disposal rates, muscle glycogen content, and muscle glycogen synthase activity in 25 southwest American Indians with normal glucose tolerance and with varying degrees of glucose intolerance. Insulin-mediated glucose disposal (M) was measured by using the hyperinsulinemic euglycemic clamp technique at plasma insulin concentrations of 134 +/- 7 and 1709 +/- 72 microU/ml, with simultaneous indirect calorimetry to assess glucose oxidation and storage rates. Muscle glycogen content and glycogen synthase activity were measured in percutaneous muscle biopsy samples obtained from the vastus lateralis muscle before and after the euglycemic clamp procedure. The results showed that muscle glycogen synthase activity at the end of the euglycemic clamp was well correlated with insulin-mediated glucose storage rates at both low (r = 0.50, P less than 0.02) and high (r = 0.78, P less than 0.0001) insulin concentrations; and also correlated with M (r = 0.66, P less than 0.001 and r = 0.76, P less than 0.0001). Similar correlations were observed between the change in muscle glycogen synthase activity and glucose storage rates and M. The change in muscle glycogen synthase activity correlated with the change in muscle glycogen content (r = 0.46, P less than 0.03) measured before and after the insulin infusions. The change in muscle glycogen content did not correlate with glucose storage rates or M. The data suggest the possible importance of glycogen synthesis in muscle in determining in vivo insulin-mediated glucose disposal rates in man.

Insulin resistance associated with lower rates of weight gain in Pima Indians.

UNLABELLED Insulin resistance is commonly associated with obesity and noninsulin-dependent diabetes. Whereas it predicts the development of diabetes, its effect on body weight change is unknown. We measured glucose disposal rates at submaximally- and maximally-stimulating insulin concentrations in 192 nondiabetic Pima Indians and followed their weight change over 3.5 +/- 1.8 y (mean +/- SD). RESULTS (a) Insulin-resistant subjects gained less weight than insulin-sensitive subjects (3.1 vs. 7.6 kg, P less than 0.0001). (b) The percent weight change per year correlated with glucose disposal at submaximally-(r = 0.19, P less than 0.01) and maximally-stimulating (r = 0.34, P less than 0.0001) insulin concentrations independent of sex, age, initial weight, and 24-h energy expenditure; the correlations were stronger for glucose oxidation than for glucose storage. (c) Weight gain was associated with an increase in insulin resistance more than four times that predicted from the cross-sectional data. We conclude that insulin resistance is associated with a reduced risk of weight gain in nondiabetic Pima Indians.

In Vivo Insulin Action Is Familial Characteristic in Nondiabetic Pima Indians

Non-insulin-dependent diabetes mellitus (NIDDM) is a genetic disorder characterized by two major pathogenic processes: reduced insulin action and a relative or absolute decrease in plasma insulin concentrations. We studied 116 nondiabetic siblings from 45 families to determine if in vivo insulin action showed any aggregation among siblings. Subjects were Pima Indians from the Gila River Indian Community in Arizona who, as a group, have the highest reported incidence and prevalence of NIDDM in the world. In vivo insulin action was determined by the euglycemic-clamp technique at two rates of insulin infusion in each subject with resulting mean plasma insulin concentrations of 119 and 1938 μU/ml. After adjustment for age, sex, and degree of obesity, there was significant aggregation among siblings of in vivo insulin action at the high insulin infusion rate (P < .0001). Family membership independently accounted for ∼34% of the variance in this measure of insulin action. Glucose uptake at the lower insulin infusion rate also showed familial aggregation (P < .01), with family membership independently accounting for ∼15% of the variance of this measurement. We conclude that in vivo insulin action is a familial characteristic. The familial component of insulin action occurs in addition to the effects of obesity, age, and sex on insulin action. Therefore it is not sufficient to simply know that an individual is lean or obese to predict his/her in vivo insulin resistance, because it must also be known whether he/she is from an insulin-resistant or insulin-sensitive family. An alteration of insulin action may be an underlying and potentially genetically determined abnormality in Pima Indians that could help explain the familial aggregation of diabetes in this population. Such a genetic mechanism might also operate in populations with Native American admixture (e.g., Mexicans) as well as other populations.