Diabetes Care 1

Steven K. Malin,1,2 Sangeeta R. Kashyap,3 Adjusting Glucose-Stimulated Jeff Hammel,1 Yoshi Miyazaki,4 CARE/EDUCATION/NUTRITION/PSYCHOSOCIAL CLIN Ralph A. DeFronzo,4 and Secretion for Adipose John P. Kirwan1,2,5 : An Index of b-Cell Function in Obese Adults DOI: 10.2337/dc13-3011

OBJECTIVE The hyperbolic relationship between insulin secretion and sensitivity has been used to assess in vivo b-cell function (i.e., the disposition index). The disposition index emphasizes the importance of taking into account both skeletal muscle and hepatic insulin resistance to depict insulin secretion. However, we propose that adipose tissue insulin resistance also needs to be accounted for when character- izing glucose-stimulated insulin secretion (GSIS) because elevated plasma free fatty acids (FFAs) impair b-cell function.

RESEARCH DESIGN AND METHODS To characterize the adipose disposition index, we used [1-14C] palmitate infusion to determine basal FFA turnover rate/adipose insulin resistance and an oral glu- cose tolerance test to characterize the first (i.e., 0–30 min) and second phase (i.e., 60–120 min) of GSIS. We validated a simplified version of the tracer infusion calculation as the product of (1/plasma FFA concentration 3 plasma insulin con- centration) 3 GSIS in 44 obese insulin-resistant subjects.

RESULTS The plasma FFA and palmitate tracer infusion calculations of the first- and second- phase disposition index were strongly correlated (r = 0.86, P < 0.000001 and r = 0.89, P < 0.000001, respectively). The first- and second-phase adipose disposition 1Department of Pathobiology, Lerner Research index derived from plasma FFA also was tightly associated with fasting hypergly- Institute, Cleveland Clinic, Cleveland, OH 2 cemia (r = 20.87, P < 0.00001 and r = 20.89, P < 0.00001, respectively) and 2-h Department of Nutrition, School of Medicine, r 2 P < r 2 P < Case Western Reserve University, Cleveland, OH glucose concentrations ( = 0.86, 0.00001 and = 0.90, 0.00001). 3Department of Endocrinology, Diabetes, and Metabolism, Cleveland Clinic, Cleveland, OH CONCLUSIONS 4University of Texas Health Science Center, Di- Adjusting GSIS for adipose insulin resistance provides an index of b-cell function in abetes Division, San Antonio, TX 5 obese subjects across the glucose spectrum. Plasma FFA–derived calculations of Metabolic Translational Research Center, Endo- b crine and Metabolism Institute, Cleveland Clinic, -cell function may provide additional insight into the role of adipose tissue in Cleveland, OH glucose regulation. Corresponding author: John P. Kirwan, kirwanj@ ccf.org, and Ralph A. DeFronzo, albarado@uthscsa .edu. Prevention of b-cell dysfunction is critical for avoiding the development and pro- gression of (1). Assessing in vivo insulin secretion is complex because Received 23 December 2013 and accepted 30 July 2014. the release of insulin from the b-cell is influenced by multiple factors including © 2014 by the American Diabetes Association. systemic insulin sensitivity, hepatic insulin extraction, plasma free fatty acid (FFA) Readers may use this article as long as the work concentrations, and glucolipid toxicity (2–4). An important aspect of the regulation is properly cited, the use is educational and not of insulin secretion in healthy adults with normal glucose tolerance is that a for profit, and the work is not altered. Diabetes Care Publish Ahead of Print, published online August 19, 2014 2 Adipose Insulin Resistance and Secretion Diabetes Care

reduction in insulin sensitivity is gener- of chronic disease participated in the the duration of the study. During the ally accompanied by a compensatory in- study (14,15). Women were postmeno- last 30 min of tracer equilibration, plasma crease in insulin secretion (5,6). This pausal and were not receiving hor- samples were obtained at 230, 220, relationship typically is curvilinear and mone replacement therapy. Subjects 210, 25, and 0 min for determination has been referred to as the disposition who were being treated with insulin or of plasma [1-14C] palmitate radioactivity index (5). thiazolidinediones were excluded. Sub- and plasma FFA concentration. At 8 A.M. a The disposition index was originally jects were not taking any medications bolus of 3.5 uCi of sodium bicarbonate calculated from insulin sensitivity and known to influence glucose or insulin was given intravenously. Blood samples acute insulin response during the intra- metabolism. The diagnosis of type 2 di- were placed on ice and immediately venous (IVGTT) abetes was based on a standard 75-g centrifuged and stored at 2708C until (5,6). Insulin sensitivity measured with oral glucose tolerance test (OGTT): fasting subsequent analysis. Adipose insulin resis- the IVGTT is an aggregate of skeletal plasma glucose .126 mg/dL and/or 2-h tance index (Adipose-IR) was calculated as muscle, hepatic, and adipose insulin glucose .200 mg/dL. Subjects were in the product of the fasting plasma FFA 3 sensitivity. When the disposition index otherwise good health without evidence fasting insulin and compared with the is calculated using the hyperinsulinemic- of renal, hepatic, cardiovascular, or hema- product of basal rate palmitate turnover euglycemic clamp technique with supra- tological disease as determined by history, 3 fasting plasma insulin, as previously de- physiological insulin infusion rates (40–80 physical examination, screening blood scribed (12,13). mU z m22 z min21), hepatic glucose pro- chemistries, urinalysis, and electrocardiog- duction and lipolysis are almost com- raphy. The study protocol was approved Insulin Secretion At 8 A.M. on a different day, following a pletely suppressed, and 80–90% of by the institutional review board of the 10- to 12-h overnight fast, blood sam- glucose is disposed in skeletal muscle University of Texas Health Science Center ples were collected from a superficial (7). Therefore, when using the euglyce- at San Antonio and the Cleveland Clinic. antecubital vein to determine fasting mic insulin clamp to calculate the dispo- Subjects gave voluntary written informed plasma glucose, triglyceride, FFA, total sition index, two physiologically relevant consent before their participation. cholesterol, LDL, and HDL concentra- tissues and two physiologically relevant tions. HbA also was measured in pa- processes, that is, hepatic insulin sensitiv- Body Composition 1c tients with type 2 diabetes to further ity (2) and adipose insulin sensitivity (8), Height was measured without shoes characterize glycemic control. Following are not taken into account. Adipose tissue using a wall-mounted stadiometer, and collection of baseline samples, subjects should be considered in the interaction of weight was recorded on a digital platform orally consumed 75 g of glucose, and insulin secretion and insulin action be- scale with minimal clothing. Total body blood samples were obtained at 30, cause fat tissue secretes factors (e.g., fat and fat-free mass were measured by 3 60, 90, and 120 min to determine FFAs and adipokines) that contribute to underwater weighing or use of H-water plasma glucose and insulin concentra- multiorgan insulin resistance (8). These radioactivity, as previously described tions. Whole-body insulin action, which factors may contribute not only to skele- (14,15). reflects the aggregate of skeletal muscle, tal muscle and hepatic insulin resistance Control Period hepatic, and adipose insulin sensitivity, but also to impaired b-cell function in in- Before metabolic testing, subjects were was estimated using the Matsuda Index sulin-resistant states, including impaired instructed to avoid strenuous physical ac- (16). The HOMA of insulin resistance also fasting glucose, impaired glucose toler- tivity for 24 h. All subjects also were in- was calculated as a more reflective surro- ance, type 2 diabetes mellitus, and obe- structed to consume weight-maintenance gate of hepatic insulin resistance (17,18). sity (9,10). Because the fasting plasma meals (;55% carbohydrates, 30% fat, and Glucose and insulin total area under the FFA concentration is influenced by a feed- 15% protein). curve (AUC) during the OGTT were cal- back loop between adipose tissue and the culated using the trapezoidal method. (11), a valid estimate of adipose Palmitate Turnover First- and second-phase GSIS, that is, the insulin resistance can be obtained from Palmitate turnover, an index of lipolysis, insulinogenic index, was calculated during the product of the basal rate of FFA turn- wasassessedwithanisotopicallyla- the first 30 min and last 60 min of the over and the fasting insulin concentration beled palmitate infusion. After an over- OGTT as insulin AUC divided by plasma glu- (12,13). Relating adipose insulin re- night fast, a polyethylene catheter was cose AUC (i.e., I – /G – and I – – / sistance with glucose-stimulated insulin inserted into an antecubital vein for the 0 30 0 30 0 60 120 G – ). The first- and second-phase ad- secretion (GSIS) could yield a novel index infusion of [1-14C] palmitate. A second 60 120 ipose disposition indices were calculated of b-cell function. In this study we exam- catheter was inserted retrogradely into as follows: I – /G – 3 adipose plasma ine the physiologic relevance of the adi- the distal portion of a vein on the contra- 0 30 0 30 FFA IR and I – /G – 3 adipose pose disposition index as it relates to lateral hand for blood sampling. The hand 60 120 60 120 plasma FFA IR, respectively (19). blood glucose. was kept in a box warmed to ;608Cfor the duration of the study to collect arte- Biochemical Analysis RESEARCH DESIGN AND METHODS rialized blood. A prime (2.5 uCi) and con- Plasma glucose was determined using 14 Subjects tinuous (0.1 uCi/min) infusion of 1- C the glucose oxidase method (YSI 2300 Forty-four obese individuals (Table 1) palmitate was prepared as previously de- STAT Plus; YSI Life Sciences, Yellow who had a stable weight (,2-kg change scribed (14,15) and started at 8 A.M. Springs, OH). HbA1c was measured using over the previous 6 months), were sed- Briefly, [1-14C] palmitate was bound to affinity chromatography (Isolab, Akron, entary (,90 min/week), and were free human serum albumin and infused for OH). Plasma insulin was measured using a care.diabetesjournals.org Malin and Associates 3

Table 1—Demographics, cardiometabolic risk, and basal rate of palmitate insulin secretion and insulin sensitivity appearance, fasting plasma FFA concentration, and adipocyte insulin resistance would be represented by b = 21. Thus a disposition index details for nondiabetic and type 2 diabetic subjects hyperbolic relationship can be satisfied if b Nondiabetic subjects Type 2 diabetic is approximately equal to –1 and the 95% (n =11) subjects (n =33) confidence interval of b excludes 0. If all Demographics assumptions with linear regression model- Patients (n) ing, y = bx+E(whereEisthesampling Male 3 15 error associated with the dependent ran- Female 8 18 dom variable y) are met, such that the Age (years) 65.5 6 1.4 57.1 6 1.5* variance of sampling error associated 6 6 Weight (kg) 94.0 5.4 86.4 2.6 with the dependent random variable is BMI (kg/m2)33.86 2.0 30.5 6 0.7* Fat mass (kg) 36.3 (30.3, 44.5) 33.4 (25.9, 40.9) constant; if we assume that neither x nor Fat free mass (kg) 54.5 6 2.4 52.5 6 1.6 y has measurement error, then orthogonal Cardiometabolic outcomes least squares regression will give a more

HbA1c (%) d 8.5 6 0.2 accurate linear unbiased estimate of HbA1c (mmol/mol) d 70.1 6 3.2 b. However, if x or y is associated with Fasting plasma glucose (mg/dL) 107 6 31816 8* random error, larger variances in either Fasting plasma insulin (mU/mL) 17.3 (10.7, 24.4) 14.3 (10.4,19.7) parameter will bias the regression coeffi- 6 6 2-H glucose (mg/dL) 156 732112* cient and alter the slope (i.e., b). Further 2-H insulin (mU/mL) 105.3 (62.6, 119.3) 35.6 (24.3, 60.4)* Matsuda Index 1.9 (1.3, 3.2) 1.9 (1.5, 2.5) complicating these issues is the view that HOMA-IR (glucose 3 INS) 4.8 (2.8, 6.2) 6.6 (4.3, 9.3)* hyperglycemia amplifies measurement er- Triglyceride (mg/dL) 164 6 16 159 6 11 ror (20). Because the dependent variable Cholesterol (mg/dL) (insulin secretion) and independent vari- Total 192 6 81806 5 able (insulin sensitivity) are measured 6 6 LDL 129 71124 with error, use of ordinary least squares HDL 38 (33.0, 42.0) 34 (30.0, 43.0) regression would yield both a biased co- Adipose outcomes fi b 6 6 ef cient (underestimate of )andabiased Fasting FFA (mEq/L) 0.69 0.06 0.72 0.03 fi Fasting palmitate Ra (mg/kg FFM/min) 97.3 (51.4–133.9) 161.4 (116.1, 237.0)* standard error of coef cient estimate GSIS (overestimate of b) (5). Thus, as done by First phase (0–30 min) 0.27 (0.18, 0.46) 0.08 (0.06, 0.22)* Kahn et al. (5), we used perpendicular Second phase (60–120 min) 0.57 (0.31, 0.74) 0.09 (0.06, 0.23)* least squares properly weighted re- Adipose IR index gression, a method that accounts for the 3 FFA INS 11.4 (6.3, 17.3) 12.3 (6.3, 14.9) error in both independent and depen- 14C palmitate Ra 3 INS 1595 (650, 3243) 2725 (1322, 4324)* dent variables to accurately predict b First-phase DI adipose plasma† fi 0–30 min 0.03 (0.01, 0.03) 0.008 (0.007, 0.01)* (21). We also calculated 95% con dence 60–120 min 0.05 (0.04, 0.05) 0.01 (0.006, 0.01)* intervals to establish whether the hyper- Second-phase DI adipose 14C palmitate*† bolic criteria have been met. We used the 0–30 min 0.1 (0.1, 0.2) 0.03 (0.02, 0.04)* bootstrap method to calculate the stan- 60–120 min 0.2 (0.2, 0.4) 0.03 (0.02, 0.05)* dard error of the coefficient of estimate Normally distributed data are mean 6 SEM. Nonnormally distributed data are presented of b, as described previously (22,23). as median (interquartile range). INS, insulin; Ra, rate of appearance. *Different between nondiabetic vs. type 2 diabetic subjects. †Disposition index (DI) was calculated as: GSIS 3 1/adipose Statistical Analysis insulin resistance (plasma FFA and 14C palmitate Ra), where GSIS is INS/glucose area under the Data were assessed using R (version 14 curve. DI calculations for C palmitate Ra were multiplied here by 1,000 for data presentation of 2.4.0; The R Foundation, Vienna, Aus- first and second phase GSIS. tria). Nonnormally distributed data were log transformed for analysis. Inde- pendent two-tailed t tests were used to radioimmunoassay (Millipore, Billerica, Determining Hyperbolic Relationships compare nondiabetic and type 2 dia- MA). Plasma triglyceride and cholesterol The hyperbolic relationship between betic cohorts. Pearson product moment (total, HDL, and LDL) concentrations were insulin secretion and insulin sensitivity correlation was used to examine the re- analyzed using enzymatic methods with is characterized as insulin secretion = lationship between adipose disposition an automated platform (Roche Modular constant/insulin sensitivity. Using log index and metabolic characteristics. Sig- Diagnostics, Indianapolis, IN). Plasma FFA transformation, this equation becomes: nificance was accepted as P # 0.05. Data concentration was measured with a color- log(insulin secretion) = constant – log are expressed as mean 6 SEM or median imetric assay (Wako Chemicals, Richmond, (insulin sensitivity). When viewed in (interquartile range), as appropriate. VA). Basal FFA turnover was determined by this context, regression analysis can be dividing the [1-14C] palmitate infusate rate applied to OGTT-derived indices of insu- RESULTS 14 by the plasma [1- C] palmitate–specific lin secretion and insulin sensitivity to Metabolic Characteristics activity. Plasma FFA and specific activity determine the regression coefficient b. Individuals with type 2 diabetes were were constant during the final 30 min of As indicated by Kahn et al. (5), a hyper- slightly younger and had a lower BMI the tracer infusion. bolic relationship between a measure of compared with nondiabetic subjects 4 Adipose Insulin Resistance and Secretion Diabetes Care

(P , 0.05; Table 1). Although whole- disposition index between nondiabetic plasma FFA concentrations were closely body insulin sensitivity was comparable and type 2 diabetic subjects for plasma correlated with fasting (r = 20.87, P , between groups, type 2 diabetic sub- FFA (first phase: P = 0.73; second phase: 0.00001 and r = 20.89, P , 0.00001, re- jects had higher hepatic and adipose P =0.88)and[1-14C] palmitate (first spectively) and 2-h plasma glucose insulin resistance compared with nondi- phase: P = 0.93; second phase: P = (OGTT) (r = 20.86, P , 0.00001 and r = abetic subjects (P , 0.05; Table 1). Ad- 0.98), suggesting that both groups 20.90, P , 0.00001; Fig. 2). The first- and ipose tissue insulin resistance using the exhibited a similar curvilinear relation- second-phase adipose tissue disposition plasma FFA concentration was strongly ship. Importantly, in nondiabetic sub- indices using plasma FFA concentration correlated with the [1-14C] palmitate– jects the curve was shifted up and also was significantly correlated with derived measures (r =0.80;P , to the right compared with patients age (r = 0.35, P , 0.03 and r = 0.45, 0.00001), although this relationship was with type 2 diabetes for both first- and P , 0.004, respectively) and BMI (r = attenuated when plasma FFA and [1-14C] second-phase plasma FFA and [1-14C] 0.42, P , 0.007 and r =0.39,P , palmitate were analyzed without adjust- palmitate calculations (P , 0.001; data 0.01, respectively). ing for insulin (r =0.45;P , 0.003). not shown). First- and second-phase ad- ipose tissue disposition index using CONCLUSIONS Insulin Secretion plasma FFA concentration was strongly Elevated fasting plasma FFA concentra- As expected, GSIS and the adipose dis- correlated with the first- and second- tions and impaired suppression of position index were lower in type 2 di- phase adipose tissue disposition index de- plasma FFA by insulin plays an impor- abetic compared with nondiabetic rived from [1-14C] palmitate (r = 0.86, P , tant role in not only the development subjects (P , 0.05; Table 1). Both 0.000001 and r = 0.89, P , 0.000001, re- of insulin resistance (8,14,24) but also [1-14C] palmitate– and plasma FFA– spectively). A similar correlation between b-cell dysfunction (9,10). The major derived adipose disposition indices the plasma FFA–derived and [1-14C] finding from this study is that the (i.e., GSIS 3 adipose tissue insulin re- palmitate–derived adipose disposition plasma FFA–derived adipose tissue dis- sistance) satisfied the hyperbolic rela- indices was observed in men (first position index, whether calculated using tionship whether calculated using the phase: r = 0.80, P , 0.00001; second first- or second-phase GSIS, approxi- first phase (0–30 min) or second phase phase: r =0.89,P , 0.000001) and mates the hyperbolic relationship (Fig. (60–120 min) of GSIS in nondiabetic and women (first phase: r =0.88,P , 1) that is purported to maintain normal type 2 diabetic subjects (Fig. 1 and Table 0.00001; second phase: r =0.90,P , glucose concentrations and correlates 2). Following log transformation there 0.000001). Both first- and second-phase strongly with [1-14C] palmitate kinetic was no difference in the slope of the adipose disposition indices derived from assessments. The variability in slope

Figure 1—GSIS adjusted for adipose insulin resistance (IR) (adipose tissue disposition index). A: 14C palmitate–calculated adipose disposition index: first phase. B: Plasma FFA–calculated adipose disposition index: first phase. C: 14C palmitate–calculated adipose disposition index: second phase. D: Plasma FFA–calculated adipose disposition index: second phase. Open circles represent nondiabetic subjects. Closed circles represent diabetic subjects. care.diabetesjournals.org Malin and Associates 5

Table 2—Estimated regression coefficients (b) and 95% CIs using orthogonal slope curve (20) (see Fig. 1). However, it for b based on first- and second-phase GSIS and adipose insulin sensitivity should be acknowledged that, in vivo, Insulin secretion Insulin sensitivity b 95% CI the compensatory increase in pancre- atic insulin secretion in response to in- Nondiabetic subjects First-phase GSIS (0–30 min) 1/Plasma FFA 3 INS 20.81 (21.42, 20.20) sulin resistance may be unable to 1/Ra palmitate 3 INS 20.88 (212.81, 11.06) completely normalize circulating glu- Second-phase GSIS (60–120 min) 1/Plasma FFA 3 INS 20.85 (21.19, 20.50) cose concentrations. In Pima Indians 1/Ra palmitate 3 INS 20.87 (21.24, 20.50) and Caucasians, Stumvoll and col- Type 2 diabetic subjects leagues (26) reported that the feedback First-phase GSIS (0–30 min) 1/Plasma FFA 3 INS 21.47 (22.02, 20.92) between insulin resistance on pancre- 1/Ra palmitate 3 INS 20.96 (21.27, 20.65) atic b-cell demand involves a signal – 3 2 2 2 Second-phase GSIS (60 120 min) 1/Plasma FFA INS 1.93 ( 3.13, 0.73) from circulating glucose, such that 1/Ra palmitate 3 INS 21.21 (22.63, 0.21) when GSIS is divided by insulin resis- No between-group difference in slope parameters was detected, suggesting diabetic and tance to depict the b-cell demand for nondiabetic groups had similar curvilinear relationships between insulin secretion and sensitivity. INS, insulin; Ra, rate of palmitate appearance. maintaining blood glucose concentra- tions (as opposed to multiplying GSIS by insulin resistance), glycemia in- parameters detected in our study (Table metabolic abnormalities associated creases for a given decrease in insulin 2) are comparable with those reported with b-cell dysfunction in obesity- and sensitivity despite normal compensa- using whole-body estimates of insulin diabetes-related glucose intolerance. tion of insulin secretion (termed glucose resistance (23) and is consistent with To maintain normoglycemia the pan- allostasis). Thus GSIS compensation for work by Ferrannini and Mari (25) de- creatic b-cell must compensate for insulin resistance is likely more complex scribing the utility of the hyperbolic re- multiorgan insulin resistance with an ap- than that provided by a single calcula- lationship when scaling GSIS to insulin propriate increase in insulin secretion. tion of a whole-body insulin sensitivity– resistance across the glucose tolerance Our data are consistent with the view derived disposition index, and addi- spectrum. Therefore, based on our mod- that the capacity of pancreatic b-cells tional models of b-cell response are est sample size of obese nondiabetic to secrete insulin for the prevailing de- needed (25). Recent work suggested and diabetic subjects, adjusting GSIS gree of insulin resistance declines as that alterations in hepatic disposition in- for adipose insulin resistance seems to glucose intolerance worsens in obese dex may precede changes in peripheral be a reasonable approach to under- subjects, that is, diabetic subjects disposition index in response to over- standing the integrated multitissue have a downward, left-shifted insulin feeding and/or physical inactivity,

Figure 2—Correlation between plasma FFA–derived calculation of the adipose tissue disposition index versus fasting and 2-h plasma glucose concentration (OGTT). Open circles represent nondiabetic subjects. Closed circles represent diabetic subjects. 6 Adipose Insulin Resistance and Secretion Diabetes Care

suggesting that multiple organs influ- disposition index, and part of this finding 3.CobelliC,ToffoloGM,DallaManC,etal. ence compensatory insulin secretion could be related to estimation of the dis- Assessment of beta-cell function in humans, si- (2,27,28). The novel calculation presented position index including plasma glucose. multaneously with insulin sensitivity and he- fi patic extraction, from intravenous and oral herein justi es the use of the adipose dis- However, the possibility of autocorrelation glucose tests. Am J Physiol Endocrinol Metab position index to characterize the contri- is lessened by the fact that first-phase GSIS 2007;293:E1–E15 bution of adipose insulin resistance to GSIS (0–30 min of insulin divided by glucose) 4. Malin SK, Finnegan S, Fealy CE, Filion J, Rocco in obese nondiabetic and obese diabetic adjusted for adipose insulin resistance MB, Kirwan JP. b-Cell dysfunction is associated subjects. Future work is needed to exam- (fasting FFA 3 fasting insulin) correlates with metabolic syndrome severity in adults. Metab Syndr Relat Disord 2014;12:79–85 ine the role of skeletal muscle, hepatic, strongly with 2-h plasma glucose. This 5. Kahn SE, Prigeon RL, McCulloch DK, et al. and adipose disposition indices following time course difference in glucose and in- Quantification of the relationship between in- lifestyle modification and/or bariatric sur- sulin pattern suggests that the adipose dis- sulin sensitivity and beta-cell function in human gery to better understand the complexity position index calculation has physiologic subjects. Evidence for a hyperbolic function. Di- – of human glucose regulation. Ultimately, relevance in characterizing b-cell function abetes 1993;42:1663 1672 6. Bergman RN, Phillips LS, Cobelli C. Physio- fi these ndings may provide greater clinical across the glucose tolerance continuum logic evaluation of factors controlling glucose insight for targeted treatment strategies (Fig. 2). The index also correlated with tolerance in man: measurement of insulin sen- that prevent/reverse type 2 diabetes. age and body mass, which further con- sitivity and beta-cell glucose sensitivity from the There are important considerations firms that it has physiological relevance. response to intravenous glucose. J Clin Invest – of the current work that should be ac- In conclusion, GSIS is essential for 1981;68:1456 1467 7. Abdul-Ghani MA, Matsuda M, Balas B, knowledged. We recognize that using preventing the progression from normal DeFronzo RA. Muscle and liver insulin resistance the plasma C-peptide concentration glucose tolerance to type 2 diabetes indexes derived from the oral glucose tolerance as a measure of prehepatic insulin secre- (1,2,30). Our results suggest that adjusting test. Diabetes Care 2007;30:89–94 tion provides a more direct measure of GSIS for adipose insulin resistance 8. Boden G. Effects of free fatty acids (FFA) on fi GSIS and b-cell function, although work provides a reasonable index of b-cell func- glucose metabolism: signi cance for insulin re- sistance and type 2 diabetes. Exp Clin Endocri- by our laboratory suggested that he- tion in obese nondiabetic and obese dia- nol Diabetes 2003;111:121–124 patic insulin extraction does not affect betic subjects. Because a high-fat diet and 9. McGarry JD. Banting lecture 2001: dysregu- insulin-derived calculations of b-cell func- lipid infusion in humans impair GSIS lation of fatty acid metabolism in the etiology of tion before or after lifestyle modification (27,31), characterization of in vivo pancre- type 2 diabetes. Diabetes 2002;51:7–18 b 10. Cusi K, Kashyap S, Gastaldelli A, Bajaj M, in a cohort of obese insulin-resistant atic -cell insulin secretion in relation to Cersosimo E. Effects on insulin secretion and adults similar to that used in the current adipose tissue insulin resistance may pro- insulin action of a 48-h reduction of plasma study (29). In addition, use of the IVGTT vide additional mechanistic insight into the free fatty acids with acipimox in nondiabetic or hyperglycemic clamp may provide a regulation of glucose homeostasis. subjects genetically predisposed to type 2 dia- more accurate assessment of insulin se- betes. Am J Physiol Endocrinol Metab 2007;292: E1775–E1781 cretion compared with OGTT-derived 11. Jensen MD, Nielsen S. Insulin dose response measures (5,6). However, exclusion of Acknowledgments. The authors thank the analysis of free fatty acid kinetics. Metabolism the gastrointestinal tract limits the phys- nursing staff for technical assistance and the 2007;56:68–76 iologic understanding of in vivo b-cell dedicated research assistants and subjects for 12. Abdul-Ghani MA, Molina-Carrion M, Jani R, their efforts. function. Further, circulating glycerol Jenkinson C, Defronzo RA. Adipocytes in sub- Funding. This research was supported by Na- jects with impaired fasting glucose and impaired may represent a better biomarker for tional Institutes of Health grant RO1-AG-12834 glucose tolerance are resistant to the anti-lipolytic lipolysis because plasma FFA can be re- (J.P.K.). S.K.M. was supported by National Insti- effect of insulin. Acta Diabetol 2008;45:147–150 esterified and/or taken up for storage, tutes of Health grant T32-DK-007319. The major- 13. Groop LC, Bonadonna RC, DelPrato S, et al. ity of R.A.D.’s salary is supported by the Veterans thereby limiting FFA as a lipolytic indica- Glucose and free fatty acid metabolism in non- Administration Health Care Service. insulin-dependent diabetes mellitus. Evidence tor. Consistent with this notion, basal fl Duality of Interest. No potential con icts of for multiple sites of insulin resistance. J Clin In- 14 interest relevant to this article were reported. [1- C] palmitate turnover was only vest 1989;84:205–213 S.K.M. developed the modestly associated with plasma FFA in Author Contributions. 14. Solomon TPJ, Haus JM, Marchetti CM, hypothesis for this study and performed statis- our study (r = 0.45; P , 0.003). There- Stanley WC, Kirwan JP. Effects of exercise train- tical analysis. J.P.K., S.R.K., and Y.M. collected ing and diet on lipid kinetics during free fatty fore, despite the observation that groups and organized data, performed analysis, and acid-induced insulin resistance in older obese in our study exhibited similar fasting edited the manuscript. J.H. performed statisti- humans with impaired glucose tolerance. Am J cal analysis. R.A.D. and J.P.K. wrote and revised plasma FFA concentrations, the higher Physiol Endocrinol Metab 2009;297:E552–E559 14 the manuscript. R.A.D. and J.P.K. are the basal [1- C] palmitate turnover in 15. Miyazaki Y, Glass L, Triplitt C, et al. Effect of guarantors of this work and, as such, had full diabetic compared with nondiabetic rosiglitazone on glucose and non-esterified fatty access to all the data in the study and take acid metabolism in Type II diabetic patients. Dia- subjects suggest differences in FFA me- responsibility for the integrity of the data and betologia 2001;44:2210–2219 tabolism (i.e., production vs. storage) the accuracy of the data analysis. 16. Matsuda M, DeFronzo RA. Insulin sensitiv- were present. It is worth recognizing, ity indices obtained from oral glucose tolerance however, that FFAs are implicated in References testing: comparison with the euglycemic insulin the development of b-cell dysfunction, 1. DeFronzo RA, Abdul-Ghani MA. Preservation clamp. Diabetes Care 1999;22:1462–1470 and FFAs are released from adipocytes, of b-cell function: the key to diabetes prevention. 17. Matthews DR, Hosker JP, Rudenski AS, which make them a reasonable marker J Clin Endocrinol Metab 2011;96:2354–2366 Naylor BA, Treacher DF, Turner RC. Homeostasis 2. Faerch K, Brøns C, Alibegovic AC, Vaag A. The model assessment: insulin resistance and beta- of lipolysis (9,10,27,30). Both fasting disposition index: adjustment for peripheral vs. cell function from fasting plasma glucose and and 2-h plasma glucose concentrations hepatic insulin sensitivity? J Physiol 2010;588: insulin concentrations in man. Diabetologia were closely associated with adipose 759–764 1985;28:412–419 care.diabetesjournals.org Malin and Associates 7

18. Abdul-Ghani MA, Jenkinson CP, Richardson 23. Retnakaran R, Shen S, Hanley AJ, Vuksan V, peripheral insulin action, insulin secretion, and DK, Tripathy D, DeFronzo RA. Insulin secretion Hamilton JK, Zinman B. Hyperbolic relationship whole-body lipolysis in healthy young male off- and action in subjects with impaired fasting glu- between insulin secretion and sensitivity on oral spring of patients with type 2 diabetes. Diabetes cose and impaired glucose tolerance: results from glucose tolerance test. Obesity (Silver Spring) 2009;58:2749–2756 the Veterans Administration Genetic Epidemiol- 2008;16:1901–1907 29. Malin SK, Solomon TPJ, Blaszczak A, Finnegan ogy Study. Diabetes 2006;55:1430–1435 24. Haus JM, Solomon TPJ, Marchetti CM, S, Filion J, Kirwan JP. Pancreatic b-cell function 19. Phillips DI, Clark PM, Hales CN, Osmond C. Edmison JM, Gonzalez´ F, Kirwan JP. Free fatty increases in a linear dose-response manner fol- Understanding oral glucose tolerance: compar- acid-induced hepatic insulin resistance is atten- lowing exercise training in adults with prediabe- ison of glucose or insulin measurements during uated following lifestyle intervention in obese tes. Am J Physiol Endocrinol Metab 2013;305: the oral glucose tolerance test with specific individuals with impaired glucose tolerance. J E1248–E1254 measurements of insulin resistance and insulin Clin Endocrinol Metab 2010;95:323–327 30. Solomon TPJ, Haus JM, Kelly KR, Rocco M, secretion. Diabet Med 1994;11:286–292 25. Ferrannini E, Mari A. function and KashyapSR,KirwanJP.Improvedpancre- 20. Krudys KM, Kahn SE, Vicini P. Population its relation to insulin action in humans: a critical atic beta-cell function in type 2 diabetic pa- approaches to estimate minimal model indexes appraisal. Diabetologia 2004;47:943–956 tients after lifestyle-induced weight loss is of insulin sensitivity and glucose effective- 26. Stumvoll M, Tataranni PA, Stefan N, related to glucose-dependent insulinotropic ness using full and reduced sampling schedules. Vozarova B, Bogardus C. Glucose allostasis. Di- polypeptide. Diabetes Care 2010;33:1561– Am J Physiol Endocrinol Metab 2006;291:E716– abetes 2003;52:903–909 1566 E723 27. Brøns C, Jensen CB, Storgaard H, et al. Im- 31. Kashyap S, Belfort R, Gastaldelli A, et al. 21. Riggs DS, Guarnieri JA, Addelman S. Fitting pact of short-term high-fat feeding on glucose A sustained increase in plasma free fatty straight lines when both variables are subject to and insulin metabolism in young healthy men. J acids impairs insulin secretion in nondiabet- error. Life Sci 1978;22:1305–1360 Physiol 2009;587:2387–2397 ic subjects genetically predisposed to devel- 22. Efron B, Tibshirani R. Statistical data analysis 28. Alibegovic AC, Højbjerre L, Sonne MP, et al. op type 2 diabetes. Diabetes 2003;52:2461– in the computer age. Science 1991;253:390–395 Impact of 9 days of bed rest on hepatic and 2474