Increased postprandial nonesteri ﬁ ed fatty acid ef ﬂ ux from adipose tissue in prediabetes is offset by enhanced dietary fatty acid adipose trapping

The mechanism of increased postprandial nonesteri ﬁ ed fatty acid (NEFA) appearance in the circulation in impaired glucose tolerance (IGT) is due to increased adipose tissue lipolysis but could also be contributed to by reduced adipose tissue (AT) dietary fatty acid (DFA) trapping and increased “ spillover ” into the circulation. Thirty-one subjects with IGT (14 women, 17 men) and 29 with normal glucose tolerance (NGT, 15 women, 14 men) underwent a meal test with oral and intravenous palmitate tracers and the oral [ 18 F]- ﬂ uoro-thia-heptadecanoic acid positron emission tomography method. Postprandial palmitate appearance (Ra palmitate ) was higher in IGT versus NGT ( P < 0.001), driven exclusively by Ra palmitate from obesity-associated increase in intracellular lipolysis ( P = 0.01), as Ra palmitate from DFA spillover was not different between the groups ( P = 0.19) and visceral AT DFA trapping was even higher in IGT versus NGT ( P = 0.02). Plasma glycerol appearance was lower in IGT ( P = 0.01), driven down by insulin resistance and increased insulin secretion. Thus, we found higher AT DFA trapping, limiting spillover to lean organs and in part offsetting the increase in Ra palmitate from intracellular lipolysis. Whether similar ﬁ ndings occur in frank diabetes, a condition also charac-terized by insulin resistance but relative insulin de ﬁ ciency, requires further investigation (Clinicaltrials.gov: NCT04088344, NCT02808182). NEW & NOTEWORTHY We found higher adipose tissue dietary fatty acid trapping, limiting spillover to lean organs, that in part offsets the increase in appearance rate of palmitate from intracellular lipolysis in prediabetes. These results point to the adaptive nature of adipose tissue trapping and dietary fatty acid spillover as a protective mechanism against excess obesity-related palmitate appearance rate from intracellular adipose tissue lipolysis.


INTRODUCTION
Increasing plasma nonesterified fatty acid (NEFA) exposure leads to ectopic fat deposition, insulin resistance, and impaired b-cell function within several hours in animal models and in humans (1). The mechanisms by which ectopic fat accumulation affects insulin resistance are different between tissues, and previous studies have shown that disturbances in both triglycerides and NEFA supply and/or uptake in liver and skeletal muscle may contribute to the ectopic fat deposition and insulin resistance (2,3). Impaired adipose tissue (AT) storage of fatty acids as triglycerides has also been proposed as an important mechanism by which obesity can be linked to the development of insulin resistance and type 2 diabetes (T2D) (4,5). Genomic studies in humans support the concept that impaired storage capacity of peripheral body fat plays a causal role in the development of T2D and cardiovascular diseases (6,7). Reduced adipogenesis has been suggested as one mechanism underlying the relationship between peripheral AT expansion and metabolic health (8).
AT regulates postprandial NEFA appearance rate via two distinct but interrelated mechanisms (9): 1) intracellular lipolysis of stored triglycerides (TG), which is stimulated mostly by norepinephrine and inhibited by insulin and 2) trapping of dietary fatty acids (DFA) from intravascular TG-rich lipoprotein lipolysis that have escaped uptake, esterification, and storage in AT, spilling over into the circulation. Postprandial plasma NEFA appearance rate is increased in impaired glucose tolerance (IGT) (2) and in established T2D (10) and is partially corrected with interventions to improve insulin sensitivity (11,12). Impaired insulin-mediated suppression of NEFA release by lipolysis of AT stores is associated with ectopic liver fat content (13), with the risk of developing T2D (14) and with the increased cardiovascular risk score (15). Yet, it is unknown whether increased DFA spillover contributes to the increase in lipolysis of AT TG stores in dysmetabolic states. The largest study that examined DFA spillover in humans showed a lower NEFA appearance rate from spillover in obese versus lean individuals (16). However, the later study did not include subjects with IGT and did not assess relative contributions of AT lipolysis versus trapping of fatty acids released by intravascular lipolysis of circulating TG to postprandial NEFA appearance.
We determined whether the increase in postprandial NEFA appearance rate in IGT is due to enhanced mobilization of fatty acids by lipolysis of AT TG stores, reduced trapping, and/or increased DFA spillover. Because of the insulin resistance present in the majority of these individuals and in view of the important effect of insulin on suppression of TG lipolysis and stimulation of esterification of fatty acids taken up from the circulation by AT, we hypothesized that increased postprandial NEFA appearance rate occurs with IGT and results at least in part from impaired AT trapping and consequently increased DFA spillover from insulin resistance and/or impaired postprandial insulin secretion. Female sex, resting energy expenditure, and upper body fat distribution are additional factors that have previously been associated with increased plasma NEFA appearance from AT (17,18). Aging has also been associated with impaired adipose tissue catecholamine-stimulated lipolysis (19). Using a multivariate regression approach, we therefore also aim to answer those additional questions: 1) What is the role of age and sex on AT depots trapping, DFA spillover, and NEFA appearance from lipolysis of AT stores? 2) What is the role of adiposity parameters [body mass index (BMI), waist circumference, fat mass] on these latter variables? 3) Does energy expenditure influence AT trapping and DFA spillover?

METHODS Study Participants
Healthy participants with normal glucose tolerance (NGT) and with IGT from previously published and ongoing studies (NCT04088344; NCT02808182) (20, 21) underwent identical postprandial protocols. IGT was defined as having a 2-h post-75 g oral glucose tolerance test, after a 12-h overnight fast, between 7.8 and 11.1 mmol·L À1 on two occasions. Subjects with NGT were selected based on a 2-h post-75 g glucose tolerance test below 7.8 mmol/L À1 and an HbA1c below 6.0% (http://guidelines.diabetes.ca/cpg/chapter3). Participants with a history or clinical evidence of any major medical or surgical condition, with a history of any dietary or severe past allergic reaction, or who participated in any research trial involving radiation exposure within the past 12 mo were excluded. Informed written consent was obtained from all participants in accordance with the Declaration of Helsinki, and the protocol received approval from the Human Ethics Committee of Centre de Recherche du CHUS.

Study Design and Experimental Procedures
The postprandial study was performed following an isocaloric diet with maintenance of their regular physical activities and avoidance of strenuous exercise during the 3-7 days before the metabolic protocol. After a 12-h overnight fast, body weight, height, and waist circumference were measured, and lean body mass was determined with a Body Composition Analyzer (TBF 300 A, Tanita Corporation of America Inc., Arlington Heights, IL). An intravenous catheter was placed in one forearm for infusions and another was placed in a distal vein in the contralateral arm maintained in a heating pad (55 C) for blood sampling.

Calculations
The homeostasis model assessment of insulin resistance (HOMA-IR), Matsuda index, and insulin secretion rate (ISR) from plasma C-peptide deconvolution were determined as previously described (23,25). The disposition index (DI), an index of b-cell function, was determined by the product of the postprandial area under the curve (AUC) ISR/AUC glucose and HOMA-IR (23,25).
All postprandial metabolic rates were calculated using Steele's nonsteady state equation (26)(27)(28)(29)(30)(31). Total postprandial plasma palmitate appearance rate (Ra palmitate ) was calculated as: where I M þ 4 palmitate is the infusion rate of [7,7,8, Ra palmitate spillover and Ra palmitate ICL were determined between times 120 and 360 min because enrichment of plasma palmitate with oral [U-13 C]-palmitate is not detected earlier after the meal in many subjects. Plasma NEFA appearance rate (Ra NEFA ) was calculated as: where C NEFA is the plasma NEFA concentration. Plasma glycerol appearance rate (Ra glycerol ) was calculated as: where I M þ 5 glycerol is the infusion rate of [1,1,2,3,3-2 H]-glycerol, V is the volume of distribution of plasma glycerol (i.e., 230 mL·kg À1 · the subject weight, Ref. 28), and TTR M þ 5 glycerol is the tracer-to-tracee ratio of plasma [1,1,2,3,3-2 H]-glycerol. C glycerol is the plasma concentration of glycerol adjusted for the presence of glycerol tracers (28): where C m glycerol is the measured plasma glycerol concentration. Meal palmitate fractional oxidation (Fox dietary palmitate ) was calculated as (20,34): where TTR CO 2 is the tracer-to-tracee ratio of breath CO 2 , Meal M þ 16 palmitate is the meal content of [U-13 C]-palmitate, and R acetate is the acetate recovery factor. It is preferable to determine R acetate in each individual from each experimental condition (35). R acetate was found to be 8% higher in healthy subjects versus those with IGT or T2D during physical exercise (36). It was, however, not possible to do so in the present study due to the complexity and high cost of the procedures. We previously established that the acetate recovery factor fitted a linear function of time during [1-13 C]-acetate administration (10): In support of this approach, another group demonstrated that application of average R acetate during the postprandial state leads to <10% over or underestimation of fatty acid oxidation (37). In the latter study, it was shown that no demographic or anthropometric characteristics significantly affected R acetate .
Dietary palmitate oxidation (Ox dietary palmitate ) was then calculated as: where Meal palmitate is the total meal fatty acid content. Total postprandial energy expenditure (TEE) was calculated using indirect calorimetry, corrected for protein oxidation, as previously published (38). AT DFA trapping and DFA distribution to lean organs were determined (20,21) as the product of [ 18 F]-FTHA SUV and volume of each specific organ divided by the sum of the product of SUV and volume. All AUC were calculated using GraphPad Prism version 7.0 (San Diego, CA).

Statistical Analyses
The Mann-Whitney test was used to compare the NGT versus IGT groups, and two-way ANOVA for repeated measures with NGT versus IGT groups, postprandial time, and interaction as the independent variables was used to analyze repeatedly measured postprandial variables. In addition, analysis of covariance (ANCOVA) with age, BMI, waist circumference, lean or fat mass, HOMA-IR, or TEE was also performed with the same dependent and independent variables. Linear regression analyses were performed after mathematical transformation of variables that failed the Shapiro-Wilk test to normalize their distribution before analysis. Univariate linear regression using Pearson's correlations between postprandial AUC of palmitate or glycerol fluxes and sex (male = 0, female = 1), age, fat mass, waist circumference, HOMA-IR, postprandial AUC ISR, and postprandial energy expenditure was performed. Multiple linear regression analyses were performed using the forwardbackward method using all variables showing reasonable trends for association in univariate analysis (r ! 0.20) and forcing the glucose tolerance group, sex, and age in the model to determine the model showing significant independent association, with maximum F value and R 2 of the model, and a Durbin-Watson test closest to 2. All independent variables retained in the models had a variance inflation factor (VIF) below 5. Because our data set includes many additional variables that were not a priori included in this plan and because many of the variable display intercorrelations, Principal Component Analysis (PCA) was performed following parallel analysis with Monte-Carlo simulation with 7,000 permutations in SPSS software to determine the most statically likely number of principal components (39). After direct Oblimin rotation with Kaiser normalization of extracted principal components, the pattern matrix of factor loadings was produced. A two-tailed P value <0.05 was considered statistically significant. All analyses were performed with the SPSS software for Mac OS (IBM, Armonk, NY) or GraphPad Prism version 7.0 (San Diego, CA).

Participant Characteristics
Thirty-one participants in the IGT (14 women, 17 men) and 29 participants (15 women, 14 men) in the NGT group underwent the postprandial study. The characteristics of IGT versus NGT participants are shown in Table 1. The IGT participants had significantly higher BMI, waist circumfer- ence, body fat mass, fasting plasma glucose, insulin, and TG levels and were more insulin resistant (i.e., higher HOMA-IR) than the NGT group.

AT DFA Trapping and Distribution in Lean Organs
From whole body PET acquisition performed 6 h after meal intake and oral administration of [ 18 F]-FTHA (Fig. 4, A  and B), visceral adipose tissue (VAT) DFA trapping was higher (P = 0.02) and total AT DFA trapping tended to be higher (P = 0.09) in IGT versus NGT (Fig. 4C). However, abdominal (P = 0.30) and peripheral (P = 0.81) subcutaneous adipose tissue (SCAT) DFA trapping (Fig. 4C) and cardiac (P = 0.72) and hepatic (P = 0.73) DFA distribution (Fig. 4D) were not different between the two groups. Skeletal muscle DFA distribution was not significantly lower in IGT versus NGT (Fig. 4D, P = 0.11).
Factors Accounting for the Differences in Postprandial Palmitate and Glycerol Fluxes between IGT and NGT ANCOVA analyses were performed and showed that the IGT versus NGT difference in Ra palmitate was dependent on BMI, waist circumference, or fat mass, but not dependent on lean mass, total energy expenditure, or HOMA-IR (Table 2). Although this group difference in Ra palmitate was also dependent on Ra palmitate ICL , it remained significant when Ra palmitate spillover was entered as a covariate, demonstrating that the IGT status-related increase in Ra palmitate was a result of increased Ra palmitate ICL . The group difference in Ra palmitate ICL was also dependent on BMI, waist circumference, fat mass, and 6-h total energy expenditure (6-h TEE). The IGT status-associated decrease in Ra glycerol was not a consequence of increased obesity (BMI, waist, lean mass and fat mass).

Factors Independently Associated with Postprandial Palmitate and Glycerol Fluxes
In the best multivariate model, postprandial AUC of Ra palmitate and Ra palmitate ICL were positively predicted by 6-h TEE and total AT DFA trapping (Table 3). Ra glycerol was positively associated with younger age, fat mass, waist, and abdominal AT DFA trapping and negatively associated with insulin resistance, postprandial insulin secretion, and peripheral AT DFA trapping. In the best multivariate model, AUC Ra palmitate spillover was independently associated with 6-h TEE (positively) and lean mass (negatively).

Factors Independently Associated with AT DFA Trapping and Dietary Palmitate Oxidation Rate
The best multivariate model to predict VAT DFA trapping included IGT status, increased lean mass and   (Table 4). Abdominal SCAT trapping was predicted by female sex, fat mass, fasting TG, and lower postprandial insulin secretion rate. Peripheral SCAT trapping increased with waist circumference and decreased with age and lean mass. Total AT trapping was predicted by female sex, IGT status, younger age, increased fat mass, fasting TG, and lower postprandial insulin secretion rate. Low total AT DFA trapping and younger age were independently associated with dietary palmitate oxidation rate.

Exploratory Analyses Using PCA Identifying Association Patterns between the Outcomes of Interest and Clusters of Variables
As PCA reduces the number of intercorrelated variables in a large data set by clustering them into a limited number of principal components (PC) that describe  independent variation structures in the data, we used this method to explore, in an unbiased fashion, variables that may be best associated with the metabolic outcomes of interest (Tables 5 and 6).
A new set of five variables (PCs), accounting for 72% of variability in the data set, was generated by PCA ( Table 5). The first principal component (i.e., PC1 or "metabolic syndrome") had the highest eigenvalue of 10.4, accounted for 26.7% of  b-Coefficient represents the standardized b-coefficient. AT, adipose tissue; AUC, area under the curve; BMI, body mass index; DFA, dietary fatty acid; HOMA-IR, homeostatic assessment model of insulin resistance; ICL, intracellular lipolysis; IGT, impaired glucose tolerance; ISR, insulin secretion rate; NGT, normal glucose tolerance; Ox, oxidation rate; R 2 , R-squared; Ra, appearance rate; Sig., P value of the linear model; SCAT, subcutaneous adipose tissue; SQRT, square root; TG, triglyceride; VAT, visceral adipose tissue; 6-h TEE, 6-h total energy expenditure. For sex, female = 1, male = 0 (i.e., a positive b-coefficient indicate positive association with the female sex). the variability in the data set, and captured, after direct Oblimin rotation, variations in insulin resistance and variables associated with glucose homeostasis (with positive loading coefficients from log HOMA-IR, log AUC of insulin, log fasting insulin, and negative loading coefficients from log Matsuda index, SQRT DI, 1/fasting glucose, and 1/AUC glucose). This PC also included log fasting TG and SQRT AUC Ox dietary palmitate . PC2 (i.e., energy expenditure, body size, and insulin secretion) accounted for 15.5% of variance and captured variations in a wide range of variables, including net postprandial oxidative metabolism of fat and carbohydrates, 6-h TEE, and also anthropometric measures (lean mass, weight, waist, and BMI), measures of insulin secretion (ISI, ISR), and adiponectin. The third component, PC3 (i.e., "muscle and liver DFA distribution and fat mass," 14.1% of variance before rotation), mainly captured DFA distribution in muscles, lower body SCAT, and in the liver, as well as fat mass. DFA distribution values in SCAT were included in PC4 (9.4% of variance), along with leptin and the adiponectin/ leptin ratio. The last component, PC5 (i.e., "ectopic fat," 6.7% of variance), included DFA distribution in VAT and in the heart, along with liver radiodensity, a marker inversely related to liver TG content.
Multiple linear regressions were then performed using these five PCs together with sex and age to predict postprandial palmitate and glycerol appearance rates (Table 6). Log AUC Ra palmitate and log AUC Ra palmitate ICL were both independently and positively predicted by PC2, PC4, and PC5 Data are presented as loading coefficients, eigenvalues, and variance corresponding to each principal component (PC). All data distributions were normalized when necessary before analyses. Parallel analysis Monte-Carlo simulation with 7,000 permutations has been used to determine the most statically likely number of PC. Loading coefficients were determined using direct Oblimin rotation with Kaiser normalization. Variable names in bold fonts: complex variables (loaded on multiple PC). Bold: loading coefficient of 0.5-1.0 (absolute value). Italics: loading coefficient of 0.3-0.5 (absolute value). AUC, area under the curve; DFA, dietary fatty acid; DI, disposition index; EE, energy expenditure; HOMA-IR, homeostatic model assessment for insulin resistance; ICL, intracellular lipolysis; IGT, impaired glucose tolerance; ISI, insulin secretion index; ISR, insulin secretion rate; NGT, normal glucose tolerance; Ra, appearance rate; SCAT, subcutaneous adipose tissue; SQRT, square root; TG, triglycerides; VAT, visceral adipose tissue.

DISCUSSION
We found that higher postprandial NEFA flux in IGT results essentially from the larger fat mass in these patients, not from dysmetabolic features of this condition per SE. Increased mobilization of NEFA from intracellular lipolysis was the mechanism of this increase in NEFA flux, as Ra palmitate spillover was the same in both groups. Ra glycerol , a marker of total AT lipolytic rate, was in contrast lower in IGT, in association with the increase in insulin resistance and postprandial insulin secretion in this group of participants. In contrast to our initial hypothesis, VAT DFA trapping was higher and total adipose tissue DFA trapping tended to be higher in IGT. Furthermore, we found adipose tissue DFA trapping to be positively and independently associated with postprandial plasma NEFA flux from intracellular lipolysis.
Our finding of higher postprandial NEFA flux explained by obesity in IGT is consistent with the results from previous cross-sectional (2,10,18) and nutritional intervention studies (11,40). Here, we demonstrate a direct association between total AT DFA trapping and postprandial NEFA appearance rate, and also a strong relationship between total AT DFA trapping and systemic NEFA flux from AT intracellular lipolysis. Furthermore, VAT DFA trapping, which was increased in IGT, was independently associated with increased postprandial palmitate appearance rate. This supports the concept of enhanced VAT DFA trapping that limits total fatty acid exposure of lean organs in the face of obesity-induced increase in postprandial NEFA flux in IGT. We recently showed increased AT DFA trapping occurring in SCAT 7 days after overfeeding in healthy subjects, reducing DFA distribution to skeletal muscles and the heart during the development of insulin resistance (21). This suggests rapid functional expansion of DFA storage capacity in AT, which limits the exposure of lean organs to DFA early during positive caloric balance. We also recently reported increased VAT DFA trapping within 2 wk after bariatric surgery in patients with T2D, associated with the rapid reduction in HOMA-IR in these patients (41). Altogether, these pieces of evidence suggest that, during obesity-induced impairment of insulin sensitivity, VAT DFA trapping may act as an ultimate protection against total fatty acid exposure of lean organs.
Percent DFA spillover rate was previously found to be higher in lean versus otherwise healthy overweight or obese subjects (16). However, there was no change in postprandial total or spillover-mediated Ra NEFA between groups in the latter study, in accordance with the findings of others (42). To our knowledge, no previous study has assessed postprandial NEFA appearance with respect to intracellular lipolysis versus DFA trapping and spillover in NGT versus IGT subjects. Almost two decades ago, we postulated, without hard experimental evidence, that AT fatty acid spillover would be increased in insulin-resistant conditions (5). In the present study, Ra palmitate spillover was positively associated with 6-h TEE and negatively associated with lean mass, independently of insulin resistance, adipose tissue DFA trapping, or other characteristics associated with IGT. Our PCA also confirmed that Ra palmitate spillover is associated with energy expenditure and body composition, but not with impaired glucose metabolism or insulin resistance. In our previous studies reporting the effect of 7-day caloric restriction or overfeeding, with significant change in insulin sensitivity but minimal change in weight, we found no significant change in postprandial plasma Ra NEFA or in DFA spillover. In contrast, Alligier et al. (40) reported increased DFA spillover after significant weight gain induced by prolonged overfeeding in healthy men, associated with significant gain of visceral fat mass. Almandoz et al. (11) showed reduced postprandial Ra NEFA , but no change in % DFA spillover after continuous feeding in patients with T2D after weight loss and metabolic improvement induced by lifestyle changes. We recently reported lower DFA spillover without change in postprandial total plasma Ra NEFA within 12 days after bariatric surgery in morbidly obese patients with T2D (41). However, due to oral fat intolerance after bariatric surgery, we markedly reduced the lipid content of our liquid meal in the latter study, making comparison with the present study difficult. As DFA spillover has been shown to be reduced with continuous versus regular meal intake (26), with smaller versus larger meal fat content (43), in spread versus emulsified meal fat (42), and with previous meal intake (44), differences between studies may also be explained by differences in methods and experimental conditions. Furthermore, none of these studies took into account energy expenditure as a potential confounder of change in postprandial NEFA spillover. In the light of our findings, this variable should now be added to the list of factors that can affect NEFA spillover rate.
We found independent associations between increased HOMA-IR or insulin secretion rate and reduced rate of glycerol appearance. Previous studies also showed reduced fasting and postprandial NEFA output and AT lipolytic rate per abdominal fat mass associated with fasting hyperinsulinemia in abdominally obese men (45,46). The observed reduction in AT lipolytic rate may also be related at least in part to resistance to catecholamine-induced stimulation of AT lipolysis (47). Although these associations do not prove causality, in view of insulin's critical regulatory role in suppressing AT lipolysis and possibly enhancing AT fatty acid esterification and TG storage, they do suggest that postprandial hyperinsulinemia may play a key regulatory role in limiting total AT lipolytic rate with the development of insulin resistance, contributing to limit systemic availability of fatty acids to lean organs.
Results of the present study in IGT cannot be extrapolated to subjects with frank T2D, who have not yet been studied in a similar fashion. It is possible that the hyperinsulinemia of those with IGT compensates for the associated insulin resistance, overcoming insulin resistance-induced impairments in adipocyte TG esterification and mobilization by intracellular lipolysis, analogous to the compensated glucose uptake in insulin-resistant, hyperinsulinemic individuals. This concept is supported by the demonstration that reduction of "adipose tissue disposition index" (i.e., fasting NEFA Â insulin levels-a marker of adipose tissue insulin resistance multiplied by insulin secretion) is associated with impaired glucose tolerance (48). As insulin secretion declines with progressive pancreatic b-cell failure in longstanding T2D, there may be an associated progressive increase in adipocyte TG mobilization and impairment in esterification, with decompensation of adipocyte metabolism manifesting in reduced trapping and increased fatty acid spillover. Future studies in those with relatively insulin-deficient T2D are required to resolve this important issue.
The average meal [U 13 C]-palmitate TTR enrichment from our studies was 0.071, whereas the peak of the [U 13 C]-palmitate TTR in the chylomicron fraction was 0.046. This difference was not due to important contamination of the chylomicron fraction by VLDL-TG since the average molar ratio of apolipoprotein B48:(apolipoprotein B48 þ B100) in a subset of 80 chylomicron fractions from 20 of the participants was 89.1%. It, however, might be explained by the dilution due to stored intestinal lipids, which are used to support chylomicron-TG secretion early in the postprandial period (49). In our subjects, plasma TG reach their peak by 300 min on average, which is concordant with the peak of TTR in the chylomicron faction at 300 min. Pich e et al. (16) found TTR in the chylomicron faction to peak at 240 min postprandially in their obese subjects. The delay in peak TTR in the chylomicron faction in the present study may be explained by various factors, including the large liquid meal (168 g of nutrients vs. 80 g in Pich e et al.), the larger protein content of the meal, and intake over 20 min, not all at once, in the present study.
We previously observed lower DFA uptake per subcutaneous AT volume in obese IGT versus healthy lean subjects with the oral 18 FTHA PET method (50), analogous to what was shown with the AV balance method across abdominal subcutaneous AT circulation (51). In the present study, the difference in DFA uptake per volume of subcutaneous AT (expressed as SUV, see Table 1) tended to be lower in IGT but did not reach statistical significance. However, this trend for lower DFA uptake per AT volume in obese IGT individuals was compensated by a larger total adipose depot volume, leading to similar total subcutaneous AT DFA uptake between IGT and NGT subjects.
Measurement of AT DFA trapping and distribution in lean organs using the oral [ 18 F]-FTHA PET method has been well validated and accepted (20,21,52), but some limitations need to be mentioned. DFA distribution to lean organs measured by this method depends not only on AT-derived DFA spillover as circulating NEFA, but also on direct chylomicron-mediated DFA delivery (or chylomicron remnantsmediated delivery in the liver). The latter mechanism may assume greater importance for DFA distribution in some organs, such as the heart. Because the liver reassembles and secretes [ 18 F]-FTHA into VLDL-TG over the 6-h postprandial period, interpretation of hepatic DFA distribution can be confounded by change in VLDL-TG secretion rate. Oral and intravenous palmitate and glycerol tracers are indirect methods to assess AT metabolism. However, DFA spillover stems mostly from chylomicron lipolysis in AT with a minor contribution from skeletal muscles (53,54) and none from the heart and liver (55)(56)(57), and with very little fatty acid spillover from VLDL (58). Only AT are known to produce net release of glycerol in circulation (59), but this release can occur from intracellular and intravascular lipolysis in vivo. Therefore, glycerol appearance rate is a marker of total lipolytic rate. Finally, we did not measure splanchnic DFA spillover, which may exert a very important impact on the liver with abdominal obesity (55).
In conclusion, enhanced postprandial NEFA flux in IGT occurs as a result of increased intracellular AT lipolysis from the larger body fat mass. Contrary to our initial hypothesis, DFA spillover was not increased and VAT DFA trapping was even increased in IGT. Increased AT DFA trapping was in fact associated with increased postprandial palmitate flux, independently of all other dysmetabolic features, except postprandial energy expenditure. Increased AT DFA trapping therefore limits DFA distribution and the total fatty acid exposure to lean organs in subjects with IGT. The negative association of AT lipolytic rate with insulin secretion leads us to speculate that the compensatory hyperinsulinemia may play an important role in overcoming enhanced postprandial NEFA flux associated with obesity. Whether this metabolic adaptation becomes impaired with the progressive pancreatic b-cell failure that characterizes T2D remains to be determined.

DATA AND RESOURCE AVAILABILITY
The data sets generated during and/or analyzed during the current study are available from the corresponding author upon reasonable request. No applicable resources were generated or analyzed during the current study.