Effect of Sex and Impaired Glucose Tolerance on Organ-Specific Dietary Fatty Acid Metabolism in Humans

Sex, obesity, and insulin resistance appear to act independently but also in concert on myocardial metabolism, structure, and function. Studies conducted in the fasting state show that women have greater myocardial oxygen consumption (1), which increases with obesity (2). In addition, female sex appears to be an important predictor of higher diabetes-associated cardiac plasma nonesterified fatty acid (NEFA) utilization (3).

Excess exposure of lean tissues to fatty acids may stem from a disordered storage of dietary fatty acids (DFAs) in adipocytes (4–7). We recently developed a novel method for noninvasive measurement of organ-specific DFA uptake and partitioning using 14(R,S)-[¹⁸F]-fluoro-6-thia-heptadecanoic acid (¹⁸F-FTHA) and positron emission tomography (PET) (8). Using this method, we found that subjects with impaired glucose tolerance (IGT) displayed increased myocardial DFA partitioning and fractional uptake that correlated negatively with left ventricular ejection fraction (9). Subjects with IGT also displayed reduced DFA partitioning in abdominal adipose tissues (9), in accordance with findings using conventional tracer methods (10–12).

Due to the small sample size at the time of our previous report, it was not possible to assess whether any sex-related differences in organ-specific DFA metabolism were present and could modulate the effect of IGT on cardiac metabolism. Based on previous reports on cardiac NEFA metabolism (3), the hypothesis of the current study was that elevated DFA uptake and partitioning observed with IGT would be worsened in women versus men. A secondary exploratory objective of our study was to determine sex-related differences in liver, skeletal muscle, and adipose tissue DFA partitioning according to IGT status.

Forty-one Caucasian individuals (see Table 1) with normal (IGT⁻) and IGT (IGT⁺), defined as having a 2-h post 75-g oral glucose tolerance test between 7.8 and 11.1 mmol/L on two separate occasions, were enrolled. None of the participants had diabetes. Results of some of these subjects (17 IGT⁻ and 9 IGT⁺) were previously reported (9,13). Subjects with a history or clinical evidence of any cardiac disorder; any clinical or biochemical evidence for abnormal kidney, liver, or thyroid function; or any uncontrolled medical or surgical condition at the time of study enrollment were excluded. Subjects taking antidiabetic medication (except metformin), β-blocker, or fibrates; with a history of dietary or severe past allergic reaction; or who participated in any research trial involving radiation exposure within the past 12 months were also excluded. Those subjects who were on a statin or antihypertensive agents had to stop these medications 3 weeks and 7 days, respectively, prior to the metabolic assessments. One participant who was on metformin discontinued the medication a few weeks prior to study participation. Written informed consent was obtained from all participants in accordance with the Declaration of Helsinki, and all study protocols received approval from the Human Research Ethics Committee of the Centre Hospitalier Universitaire de Sherbrooke.

Subjects were instructed to follow an isocaloric diet (0% alcohol, 15–17% protein, 30–33% fat, and 50–55% carbohydrates) for 48 h before the experimental protocol. On arrival, body weight, height, and waist circumference were measured. Lean and fat body mass were determined with a body composition analyzer (TBF 300A; 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.

Each participant underwent a 6-h postprandial experimental protocol (Fig. 1), during which a standard liquid meal (14) was consumed. The oral intake of the meal corresponded to 400 mL over 20 min for a total of 906 kcal (approximately one-third of the median daily caloric intake of American adults [15]; 33 g or 33% as fat, 34 g or 17% as proteins, and 101 g or 50% as carbohydrates). The protocols started with 30 min bed rest, after which blood samples were taken at 30-min intervals. Blood was collected in tubes containing Na₂EDTA and Orlistat (30 μg/mL; Roche, Mississauga, Canada) to prevent in vitro triglyceride (TG) lipolysis.

We used the oral ¹⁸F-FTHA method (8) to determine whole-body DFA partitioning and cardiac DFA uptake. Each subject was positioned supine in a PET/computed tomography (CT) scanner (Philips Gemini TF, 16 slices CT; Philips, Eindhoven, NV) using a line of response row action maximum likelihood algorithm without sinogram rebinning for reconstructing images with isotropic voxel size of 4 mm, 60-cm diameter by 18-cm axial field of view. Whole-body CT scans (16 mAs) were performed to correct for attenuation and for definition of PET regions of interest (ROIs). Approximately 70 MBq ¹⁸F-FTHA (16) were mixed into polyunsaturated fatty acid lipid emulsion (200 µL), incorporated into gel capsules (T.U.B. Enterprises), and given orally at time 0 min with the liquid meal. After a segmental CT (40 mAs) to measure hepatic radiodensity and for definition of PET aortic and cardiac ROIs, dynamic list-mode PET scanning centered on the thoraco-abdominal segment was performed between time 90 and 120 min (15 × 120 s) after meal intake to determine oral ¹⁸F-FTHA uptake rate in the heart to determine myocardial fractional DFA uptake (K_i) early in the postprandial state when a large fraction of ¹⁸F-FTHA in circulation is still transported in chylomicron-TG. At time 360 min, a whole-body PET/CT acquisition was performed to determine ¹⁸F-FTHA tissue partitioning (i.e., biodistribution of DFA in all organs from direct uptake from chylomicrons and from recirculation into NEFA and VLDL-TG) in all organs of interest. Technical problems occurred during PET/CT acquisitions between time 90 and 120 min in three participants and during the whole-body PET/CT acquisition at time 360 min in one participant, leading to missing K_i and biodistribution data, respectively, for these subjects. Liver volume was extrapolated by drawing the liver surface on each 5-mm slice of the whole-body CT scan at time 360 min using OsiriX 5.8.2 software (Pixmeo SARL, Bernex, Switzerland). The maximal gastrointestinal tract radioactivity exposure for oral ¹⁸F-FTHA administration was estimated at 2.35 mSv at the stomach. Total radioactivity exposure to the participants was <20 mSv. ¹⁸F-FTHA was tested for sterility and nonpyrogenicity.

For dynamic PET acquisitions, a mean value of pixels (mean kBq/mL) for each frame was recorded. ROIs were drawn on the heart (three consecutive 5-mm slices starting from the middle of the long cardiac axis toward the cardiac valves), liver, and thoracic aorta to generate tissues and blood time-radioactivity curves. The ROIs were first defined from the transaxial CT slices and then copied to ¹⁸F-FTHA image sequences. Myocardial fractional DFA uptake (K_i) was determined using Patlak linearization (17). Input function was taken from the aorta and we showed that no correction is needed for spillover and partial volume effect in this instance because the diameter of the aorta is large enough and distant enough from the esophagus to rule out contamination of the input function from spill-in activity (18). For this analysis, it was assumed that all circulating DFAs and ¹⁸F-FTHA were circulating in chylomicron-TG between 90 and 120 min post–meal intake (8,19). Myocardial net DFA uptake (K_m) was then calculated by multiplying K_i by chylomicron-TG levels 120-min post–meal administration. As shown in Supplementary Fig. 1, most but not all plasma ¹⁸F-FTHA activity was retrieved in circulating TG and chylomicron-TG. This leads to underestimation of myocardial DFA uptake assessed by our method. For whole-body scans, mean value of pixels (mean standard uptake value [SUV]) for all tissues of interest was recorded. ROIs were drawn on the liver, heart, quadriceps femoris, subcutaneous thigh and anterior abdominal adipose tissues, and right peri-renal adipose tissue (a visceral adipose tissue depot; this depot was chosen to avoid gastric and intestinal spillover of ¹⁸F activity that occurred in most other splanchnic adipose tissue depots) (8).

Glucose, insulin, total NEFA, and TG were measured as previously described (9). Fasting glucose, insulin, NEFA, and TG levels were calculated as the mean levels at time −60 and 0 min (i.e., just prior to meal intake), whereas postprandial levels were calculated as the total area under the curve (AUC) between time 0 and 360 min. Chylomicrons and plasma lipids were separated by ultracentrifugation and counted for ¹⁸F activity and TG concentration (18). Postprandial chylomicron-TG levels were calculated as the total AUC between time 0 and 360 min. The homeostasis model assessment of insulin resistance (HOMA-IR) index was calculated as fasting glucose (in mmol/L) ⋅ fasting insulin (in mU/L)/22.5 (20).

All data are presented as means ± 95% CI unless otherwise stated. Data were tested for normal distribution and were mathematically transformed when appropriate to normalize their distribution prior to statistical testing. Two-way ANOVA was performed to examine the influence of IGT status and sex, as well as their interaction, on all dependent variables. Spearman ρ analyses were carried out to examine associations between continuous variables. Stepwise multivariate linear regression was performed to determine the best model to predict cardiac DFA partitioning (maximum R²) using the forward-backward method with all variables associated with the dependent variable (P value <0.10) and forcing IGT status and sex in the model (maximum of four variables in the model at all steps because n = 40 approximately). The best sex-specific models to predict cardiac DFA partitioning were determined using a similar approach with a maximum of two independent variables in the models at all steps (because n = 20 per group approximately). A P value <0.05 was considered statistically significant unless otherwise noted. All statistical analyses were performed using IBM SPSS Statistics version 20.0 for Windows (Armonk, NY) or GraphPad Prism version 6.0e for Mac OS X (San Diego, CA).

Data from 12 IGT⁻ men, 9 IGT⁻ women, 8 IGT⁺ men, and 12 IGT⁺ women were analyzed for this study. IGT⁺ subjects were significantly older and had more fat mass; higher percent fat mass; larger BMI; greater waist circumference; and higher fasting plasma glucose, insulin, NEFA, TG, HOMA-IR, and resting systolic and diastolic blood pressure than IGT⁻ subjects. Men were significantly heavier and had lower percent fat mass; greater waist circumference, fasting plasma glucose, TG, and HOMA-IR; and lower fasting plasma NEFA levels than women. There was no significant sex-IGT interaction for all these variables, although elevation of fasting plasma TG in IGT⁺ was mostly attributable to elevation in women (interaction P = 0.10) (Table 1).

IGT⁺ was associated with significantly higher postprandial glucose, insulin, NEFA, total plasma TG, and chylomicron-TG AUC and with larger liver volume and lower radiodensity versus IGT⁻. Men had significantly higher chylomicron-TG levels at 120 min postprandially with higher postprandial total plasma TG and chylomicron-TG AUC versus women. The IGT-associated increase in postprandial total plasma TG AUC and in liver volume, and reduction in liver radiodensity, were mostly attributable to differences in women (interaction P = 0.06, P = 0.04, and P = 0.04, respectively). IGT, sex, and sex-IGT interaction did not significantly affect ¹⁸F activity in total plasma or plasma NEFA fraction. There was higher ¹⁸F activity in total TG (P = 0.04) and chylomicron-TG (P = 0.12) in IGT⁺ versus IGT⁻ and men had higher ¹⁸F activity in chylomicron-TG versus women (P = 0.03). However, there was no sex-IGT interaction with regards to these variables (Table 2).

Myocardial DFA partitioning (Fig. 2A) was significantly higher in IGT⁺ versus IGT⁻ (P = 0.006) in both men (2.14 [95% CI 1.70–2.18] vs. 1.28 SUV units [0.80–1.76], respectively) and women (1.95 [1.57–2.33] vs. 1.64 SUV units [1.32–1.96], respectively) without difference between sexes (P = 0.67). IGT-associated increase in myocardial DFA partitioning tended to be higher in men (sex-IGT status interaction P = 0.18). Myocardial DFA fractional uptake (K_i) (Fig. 2B) was also significantly higher in IGT⁺ versus IGT⁻ (P < 0.001) in men (0.063 [0.032–0.095] vs. 0.016 min⁻¹ [0.007–0.025], respectively) and women (0.050 [0.024–0.077] vs. 0.030 min⁻¹ [0.013–0.047], respectively) without significant sex difference (P = 0.96). IGT-associated increase in myocardial DFA fractional uptake tended to be higher in men (sex-IGT status interaction P = 0.15). Myocardial DFA net uptake (K_m) (Fig. 2C) was also higher in IGT⁺ versus IGT⁻ (P = 0.01) in men (10.9 [0–24.1] vs. 2.5 nmol ⋅ g⁻¹ ⋅ min⁻¹ [0.4–4.6], respectively) and women (3.3 [95% CI 1.3–5.4] vs. 1.2 nmol ⋅ g⁻¹ ⋅ min⁻¹ [0.2–2.2], respectively), but levels were also significantly higher in men versus women (P = 0.03). IGT-associated increase in net myocardial DFA uptake also tended to be higher in men (sex-IGT status interaction P = 0.12).

There were no IGT status or sex-related differences in liver (Fig. 3A) and skeletal muscle (quadriceps femoris) (Fig. 3B) DFA partitioning. However, visceral adipose tissue DFA partitioning (Fig. 3C) was lower in IGT⁺ versus IGT⁻ (P = 0.03) in men (0.26 [95% CI 0.12–0.39] vs. 0.40 SUV units [0.23–0.56]) and women (0.29 [0.16–0.42] vs. 0.46 SUV units [0.29–0.62]) without sex difference (P = 0.50) or sex-IGT status interaction (P = 0.88). Anterior abdominal subcutaneous adipose tissue DFA partitioning (Fig. 3D) was also lower in IGT⁺ versus IGT⁻ (P = 0.007) in men (0.14 [0.07–0.21] vs. 0.20 SUV units [0.12–0.29]) and women (0.18 [0.11–0.24] vs. 0.31 SUV units [0.25–0.38]) and was significantly lower in men versus women (P = 0.04) but without significant sex-IGT status interaction (P = 0.30). Femoral subcutaneous adipose tissue DFA partitioning (Fig. 3E) was not different in IGT⁺ versus IGT⁻ (P = 0.74) but was significantly lower (P = 0.04) in men (0.11 SUV units [0.02–0.20] in IGT⁺; 0.08 SUV units [0.03–0.14] in IGT⁻) versus in women (0.16 SUV units [0.10–0.22] in IGT⁺; 0.17 SUV units [0.07–0.26] in IGT⁻), without significant sex-IGT status interaction (P = 0.63).

In the entire group of participants, only IGT status, liver and skeletal muscle DFA partitioning, as well as myocardial DFA fractional uptake were significantly and positively associated with myocardial DFA partitioning (Table 3). When sexes were analyzed separately, however, the positive association of IGT status and liver or skeletal muscle DFA partitioning with myocardial DFA partitioning was observed only in men, whereas the positive association of variables associated with insulin resistance (fasting and postprandial insulin levels, HOMA-IR, and fasting and postprandial NEFA levels) was present only in women. In addition, markers of adiposity (BMI, fat mass, percent fat mass, and waist circumference), disordered DFA storage in central adipose tissues (i.e., visceral and subcutaneous abdominal adipose tissue DFA partitioning), and liver steatosis were significantly associated with increased myocardial DFA partitioning in women but not in men. Postprandial chylomicron-TG (at 120 min) and plasma TG levels, as well as resting cardiac rate-pressure product, were negatively associated with myocardial DFA partitioning in men, not in women. As expected, myocardial DFA fractional uptake between 90 and 120 min postprandially was associated with myocardial DFA partitioning over 6 h in the entire group, but this relationship, surprisingly, was only observed in men, not in women.

The best multivariate model to predict myocardial DFA partitioning included IGT status (P = 0.05), AUC of postprandial NEFA (P = 0.10), and liver DFA partitioning (P = 0.002) (R² = 0.39, P < 0.001) in the total group of participants; IGT status (P = 0.003) and resting rate-pressure product (P = 0.005) in men (R² = 0.56, P = 0.001); and BMI only in women (R² = 0.43, P = 0.002).

IGT, fasting insulin, HOMA-IR, and postprandial AUC of glucose and insulin were significantly associated with increased myocardial DFA fractional uptake between 90 and 120 min after meal intake (Supplementary Table 1). These associations remained in men but not in women. In women, no measured variable was significantly associated with myocardial DFA fractional uptake. The best multivariate model to predict myocardial DFA fractional uptake only included IGT status (R² = 0.28, P = 0.001) in the total group of participants. In men, the best multivariate model included IGT status (P = 0.03) and postprandial AUC of insulin (P = 0.02) (R² = 0.65, P < 0.001).

Our hypothesis was that IGT-associated increase in myocardial DFA uptake and partitioning would be worsened in women compared with men. We found that women had lower levels of myocardial DFA net uptake without a sex effect in cardiac DFA fractional uptake or in cardiac DFA partitioning. This sex effect was driven by increased postprandial chylomicron-TG levels seen in some male participants. There was no significant sex-IGT interaction with regards to DFA metabolism in the myocardium, although IGT-associated increase in fractional and net cardiac uptake of DFA tended to be more important in men. We also found no significant sex-IGT interaction on the IGT- or obesity-associated reduction in DFA partitioning in visceral and abdominal subcutaneous adipose tissues that we (9) and others (10–12) previously observed. Other sex differences included significant increases in femoral and abdominal subcutaneous adipose tissue partitioning in women, findings that confirm those of previous studies (21–23).

Peterson and colleagues (1,3) found that women had significantly higher myocardial oxidative metabolism compared with men, which may have explained increased myocardial NEFA uptake and oxidation. The latter observations were performed during fasting and may not necessarily apply to the postprandial state, however. The subset of women from the current study (n = 11) in whom we also measured cardiac oxidative metabolism (using the ¹¹C-acetate PET method) also displayed higher myocardial oxidative metabolism than men (data not shown), consistent with the findings of Peterson and colleagues. We, and others, also previously reported enhanced myocardial oxidative metabolism in obesity and/or IGT and/or type 2 diabetes versus control subjects (2,9). We also observed, in a subset of participants (14 men and 11 women) that increased myocardial blood flow in women (also using the ¹¹C-acetate PET method) did not significantly correlate with fractional DFA uptake (data not shown).

We found a significant association between myocardial DFA partitioning and markers of adiposity (BMI, waist circumference, fat mass, and percent fat mass), impaired adipose tissue metabolic function (reduced central adipose tissues DFA partitioning and increased fasting and postprandial NEFA levels), insulin resistance (increased fasting and postprandial insulin levels and HOMA-IR), and hepatic steatosis (reduced liver radiodensity) in women but not in men. Our multivariate analyses demonstrated that BMI was the only factor entering the prediction model of cardiac DFA partitioning in women, whereas IGT status and cardiac rate-pressure product entered the model in men. The reason for these sex-related differences is unclear. It is possible that increased partitioning of DFAs associated with IGT may happen primarily through fatty acid recirculation into VLDL-TG and/or plasma NEFA in women, whereas it is driven by direct LPL-mediated uptake of chylomicron-TG in men. This interpretation is further suggested by increased postprandial chylomicron-TG and ¹⁸F activity in chylomicron-TG in men and the close relationship between fractional uptake of DFAs between 90 and 120 min (mostly driven by chylomicron-TG uptake) and myocardial DFA partitioning over 6 h in men but not in women. Peterson and colleagues (2,24) have previously reported a significant association between myocardial oxidative rate or NEFA uptake or oxidation with BMI in women but not in men, further suggesting that the IGT-mediated increase in cardiac DFA partitioning in women could be attributable to recirculation and uptake through the NEFA pool. More studies are needed to directly address these issues.

Our data show that during the postprandial period, men and women, regardless of whether they have normal or IGT, do not differ in the sum of DFA that has been ultimately oxidized plus stored in their liver over 6 h post–meal intake. Liver uptake of DFA occurs via direct uptake of chylomicron-remnant particles, hepatic lipase-mediated hydrolysis of chylomicron-remnant TG, and recirculation into plasma NEFA after adipose tissue lipolysis. The absence of sex effect on liver DFA partitioning in the current study was observed despite the known increase in hepatic lipase activity in men versus women (25). The lack of IGT, sex, or sex-IGT interaction effect on liver DFA partitioning also occurred despite a significant increase in liver steatosis in IGT⁺ attributable mostly to women (i.e., significant sex-IGT interaction). However, the liver secretes some ¹⁸F-FTHA from chylomicron-TG into VLDL-TG over 6 h (8). Thus, liver DFA partitioning over 6 h as determined by the oral ¹⁸F-FTHA method likely underestimates liver DFA uptake, limiting our conclusion regarding this organ. It is possible that VLDL-TG secretion from DFAs may have been increased in women in the current study. However, no sex difference in DFA incorporation into VLDL-TG was observed in a previous report using ¹⁴C-labeled oleate tracer (26).

Women had lower chylomicron-TG levels and lower ¹⁸F activity in chylomicron-TG postprandially when compared with men in both IGT⁻ and IGT⁺ groups, consistent with findings by others that could be explained by faster chylomicron-TG clearance in women (26). In rats, testosterone reduces chylomicron clearance and postheparin LPL activity (27). LPL expression is higher in subcutaneous adipose tissues in women (28), but there has been debate as to what organs are responsible for the apparent faster chylomicron-TG clearance in women. Our findings are more in favor of subcutaneous adipose tissues (23) than skeletal muscles (29) as the site explaining higher chylomicron-TG clearance in women. Delayed gastric emptying may also have contributed to the reduction in postprandial chylomicron-TG level and ¹⁸F activity we observed in women (30).

The strengths of the current study included the ability to simultaneously measure DFA partitioning in most organs of the body and the inclusion of a similar number of men versus women without and with IGT and of similar age and BMI. The oral ¹⁸F-FTHA method, however, cannot assess intracellular fate (i.e., oxidation vs. esterification) of the DFA tracer (8). It is important to consider that DFA partitioning as assessed by the oral ¹⁸F-FTHA method integrates organ-specific distribution of DFA from digestion and absorption of DFA, chylomicron transport, and recirculation into plasma NEFA and VLDL-TG. This method also integrates any variation in the rate of DFA digestion and/or absorption. Although the majority of plasma ¹⁸F is recovered into chylomicron-TG at 120 min postprandially, our measure of net DFA uptake through chylomicron-TG is likely an underestimation because some tissue uptake may also occur through the NEFA pool even that early after meal intake (9). However, we reached the same conclusion when we calculated net cardiac DFA uptake using chylomicron-TG + NEFA (Supplementary Table 2). Because of the short radioactive half-life of ¹⁸F and the delay required to isolate VLDLs, it was not possible to determine ¹⁸F-FTHA recycling in VLDL-TG in the current study. The oral ¹⁸F-FTHA method is further limited by the relatively large radioactivity exposition and by the incapacity to assess tissue DFA uptake before 90 min, which would more closely reflect DFA uptake through chylomicron-TG specifically. Because of the relatively small number of participants, the current study also lacked power to detect a small difference between groups or a modest degree of association between variables. Other limitations include lack of matching between sexes for percent fat mass, waist circumference, and degree of insulin resistance. Our patient selection, however, included a wide range of adiposity and insulin sensitivity in both sexes.

In conclusion, contrary to our initial hypothesis based on a previously reported increase in plasma NEFA cardiac uptake and utilization in women, we found that men have higher net early cardiac uptake of DFA driven by a higher chylomicron-TG level. Increased cardiac DFA partitioning 6 h after a meal is nevertheless seen in both sexes with IGT. In men, cardiac partitioning of DFAs is independently associated with IGT per se, whereas it is directly related to obesity in women. DFA partitioning is increased in subcutaneous adipose tissues in women.

Acknowledgments. The authors acknowledge the collaboration of Diane Lessard, Caroll-Lynn Thibodeau, Maude Gérard, Éric Lavallée, Frédérique Frisch, and Lucie Bouffard (all from Centre de Recherche du Centre Hospitalier Universitaire de Sherbrooke) for their technical assistance. The authors also thank all study participants who dedicated their time for all protocol-related activities.

Funding. This work was supported by a grant from the Canadian Institutes of Health Research (CIHR-MOP 53094 to A.C.C.). M.K. is the recipient of a Université de Sherbrooke scholarship and a Diabète Québec studentship award. C.N. is the recipient of a CIHR postdoctoral fellowship award. J.-P.B. is the recipient of a Senior Clinician Investigator Award of the Fonds de Recherche du Québec - Santé. A.C.C. holds the CIHR-GSK Chair in Diabetes. The Centre de Recherche du Centre Hospitalier Universitaire de Sherbrooke is funded by the Fonds de Recherche du Québec – Santé.

Duality of Interest. No potential conflicts of interest relevant to this article were reported.

Author Contributions. M.K., C.N., S.P., B.G., J.-P.B., and E.E.T. collected, analyzed, and interpreted data and drafted the manuscript or revised it critically for important intellectual content. A.C.C. conceived and designed the experiments; collected, analyzed, and interpreted data; and drafted the manuscript or revised it critically for important intellectual content. A.C.C. is the guarantor of this work and, as such, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.