Effect of Insulin Infusion on Spillover of Meal-Derived Fatty Acids

Abstract

Context:

Spillover of chylomicron triglyceride fatty acids directly into the circulation as free fatty acids (FFAs) during lipoprotein lipase hydrolysis may contribute to the elevated total FFAs seen in insulin-resistant states.

Objective:

The objective of the study was to determine whether spillover is regulated by rates of intracellular lipolysis, we studied overweight and obese nondiabetic subjects (n = 7) on two occasions, during infusion of saline and insulin.

Design:

During insulin infusion (20 mU · m⁻² · min⁻¹), plasma glucose was clamped at the concentration achieved during saline infusion. On both study days, subjects sipped 1–2 oz of a liquid mixed meal every 15 min for 6.5 h to achieve steady-state chylomicron and FFA concentrations. Spillover was measured with infusions of [³H]triolein and [U-¹³C] oleate.

Results:

Glucose concentrations were similar during saline compared with insulin (113 ± 2 vs. 113 ± 1 mg/dl, P = NS). Insulin levels during saline and insulin infusion were 18 ± 3 and 44 ± 5 μU/ml, respectively. Glucose infusion rate during insulin infusion was 5.5 ± 1.0 mg · kg fat-free mass⁻¹ · min⁻¹. Plasma FFA concentrations were lower during insulin compared with saline (75 ± 8 vs. 124 ± 13 μmol/liter, P = 0.002). Oleate rate of appearance was lower during insulin vs. saline (27 ± 3 vs. 36 ± 5 μmol/min, P = 0.004). Spillover was similar during saline and insulin (26 ± 2 vs. 25 ± 2%, P = 0.60).

Conclusions:

These results indicate that suppression of intracellular lipolysis with insulin does not reduce lipoprotein lipase-mediated spillover in humans during meal absorption. It is possible that spillover did not decrease because of an impaired or absent antilipolytic effect of increased insulin concentrations in visceral fat.

Elevated concentrations of plasma free fatty acids (FFAs) are an important mediator of insulin resistance (1) and contribute to the dyslipidemia that often accompanies insulin resistance by stimulating secretion of very low-density lipoprotein triglycerides (2). Furthermore, acute elevation of FFA levels induce endothelial dysfunction (3) and raise blood pressure (4, 5). Although it is generally thought that the main source of FFAs is the adipocyte triglyceride depot, regulated by intracellular lipases (6), the action of lipoprotein lipase (LPL) on chylomicrons is an additional source of plasma FFAs, released into the circulation by a spillover mechanism (7). Fractional spillover in various studies has been estimated to be about 30–40% of LPL-generated fatty acids (7–9) and, depending on dietary fat intake, could account for 25% or more of total FFA flux (7). We previously reported that the majority of spillover in overweight and obese humans occurs in the splanchnic bed, probably in visceral adipose tissue (8, 10). Spillover correlates strongly with release of FFAs by intracellular lipases in visceral fat, suggesting that it is regulated by intracellular lipolysis (10). We undertook the present study to determine whether insulin infusion, via suppression of intracellular lipolysis, decreases spillover during meal absorption in overweight and obese subjects with dyslipidemia.

Materials and Methods

Subjects

Seven nondiabetic overweight and obese individuals with dyslipidemia were studied according to a protocol approved by the Mayo Institutional Review Board. Informed written consent was obtained before study participation. Subjects were healthy and weight stable and did not engage in regular vigorous exercise. Individuals with fasting hyperglycemia (>100 mg/dl), those taking lipid-lowering agents, and those with kidney, liver, or heart disease were excluded. A fasting plasma high-density lipoprotein cholesterol less than 40 mg/dl for men and less than 45 mg/dl for women, together with fasting triglycerides greater than 150 and less than 300 mg/dl was required for participation in the study. All subjects had body mass indices between 25 and 35 mg/dl. All meals were provided by the metabolic kitchen for 3 d before the study, were isocaloric, and contained 15% protein, 35% fat, and 50% carbohydrate. Subjects were admitted to the Clinical Research Unit the afternoon before the study; at that time, body composition was measured using dual-energy x-ray absorptiometry. The subjects consumed an evening meal at 1800 h. They had free access to water and were studied the next day after an overnight fast.

Study protocol

Subjects were studied on two occasions, 2 wk apart. After an overnight fast, an infusion catheter was placed in a forearm vein. A catheter was placed in a retrograde fashion into a vein of the dorsum of the hand contralateral to the infusion catheter. The hand was placed in a heated Plexiglas box maintained at 55 C to allow sampling of arterialized blood. On the first day, saline was infused. A liquid meal [prepared from chocolate flavored Ensure Plus (Abbott Nutrition, Columbus, OH) with additional canola oil to achieve the macronutrient distribution described above] was given beginning at 0800 h (0 min) as a priming dose and continuing at 1000 h as aliquots given every 15 min until the end of the study (390 min) at a rate providing approximately 6% of estimated basal energy requirements (Harris-Benedict equation) per hour. At 1100 h (180 min), an infusion of [U-¹³C]oleate (0.5 nmol · kg⁻¹ · min⁻¹) was initiated and continued until the end of the study. At 270 min, an infusion of [9,10-³H (N)oleyl]triolein (1.2 μCi/min) was started and continued to 390 min. Blood samples were taken hourly between 0 and 300 min and every 15 min from 330 min to 390 min. FFA, triglycerides, glucose, and insulin levels were measured at all time points. [U-¹³C]oleate enrichment and ³H oleate specific activity were determined from 330 to 390 min. The second study day was identical to the first except that insulin was infused at a rate of 20 mU · kg⁻² · min⁻¹ together with 20% dextrose to maintain glucose concentrations at the same level as observed on the baseline day from 180 min to study completion (1430 h or 390 min). Glucose levels were determined every 10 min on both days during saline or insulin infusion to facilitate the glucose clamp.

Analyses

Blood samples were collected in chilled 10-mL EDTA tubes containing 0.5 mg of paraoxon to inhibit LPL and kept on ice until centrifugation at 4 C. Chylomicrons were isolated using a triple-spin ultracentrifugation method (11). The plasma FFA concentration and specific activity were determined with HPLC (12), using [²H₃₁]palmitate as an internal standard (13). Meal fatty acid content was also determined with HPLC. Plasma triglyceride concentrations were measured using a commercial enzymatic method (COBAS Integra autoanalyzer; Roche, Indianapolis, IN). [U-¹³C]oleate atoms percent enrichment (APE) was determined by liquid chromatography-mass spectrometry (14).

Calculations

Mean values from the 300- to 390-min sampling interval were used for calculation of kinetic data. Rate of appearance (R_a) of oleate was determined using the equations of Steele for steady-state conditions as previously described (12, 13). Fatty acid clearance was calculated as FFA tracer infusion rate/plasma tracer concentration (15). The rate of appearance of ³H oleate was calculated using the following formula:

$$\text{Ra}{}^3\text{H}\text{oleate}=\frac{\left[\text{U-}{}^{\text{13}}\text{C}\right]\text{oleate}\text{infusion}\text{rate}}{\left[\text{U-}{}^{\text{13}}\text{C}\right]\text{enrichment}/{}^3\text{H}\text{specific}\text{activity}}$$

Fractional spillover (percentage) is then derived using the following formula:

$$\text{Spillover}=\frac{\text{Ra}{}^3\text{H}\text{oleate}}{\text{Rd}{}^3\text{H}\text{triolein}}\times 100$$

where Rd ³H triolein indicates the infusion rate of the triglyceride tracer.

The contribution of systemic spillover to oleate Ra during continuous feeding was calculated as follows:

$$\frac{\text{Oleate}\text{ingestion}\text{rate}\left(\mu \text{mol}/\min \right)\times \text{fractional}\text{spillover}}{\text{Oleate}{\text{R}}_{\text{a}}}$$

Statistical methods

Data from the 2 study days were compared using paired t tests to determine significance (α < 0.05).

Results

Baseline characteristics of the subjects are as shown in Table 1, including screening laboratory tests. Plasma triglyceride concentrations (not shown) were similar at baseline during saline and insulin infusion (181 ± 23 vs. 168 ± 29 mg/dl, P = 0.50) and increased during meal absorption to steady-state concentrations (257 ± 22 vs. 230 ± 35 mg/dl during saline and insulin infusion, respectively, P = 0.12). Glucose concentrations were 113 ± 2 and 113 ± 1 mg/dl during saline and insulin, respectively, P = NS (Fig. 1, upper panel). The glucose infusion rate during insulin was 5.5 ± 1.0 mg · kg fat-free mass⁻¹ · min⁻¹. Insulin concentrations during the baseline sampling period were 7 ± 1 μU/ml on both study days (not shown) and increased to 18 ± 3 and 44 ± 5 μU/ml during saline and insulin, respectively (Fig. 1, lower panel).

Table 1.

Subject characteristics

Sex	1 F, 6 M
Age (yr)	44 ± 3
Weight (kg)	94.8 ± 4.4
BMI (kg/m2)	30 ± 1
Body fat (%)	35 ± 3
Total cholesterol (mg/dl)	191 ± 8
Triglycerides (mg/dl)	201 ± 25
HDL cholesterol (mg/dl)	35 ± 2

F, Female; M, male; BMI, body mass index; HDL, high-density lipoprotein.

Fig. 1.

Plasma glucose (upper panel) and insulin (lower panel) concentrations during infusion of saline vs. insulin.

FFA concentrations at baseline were similar during saline and insulin (264 ± 36 vs. 284 ± 20 μmol/liter, P = NS, not shown). Oleate was 42 ± 1% of meal fat, and the oleate ingestion rate was 111 ± 7 μmol/min. Oleate concentrations during meal absorption were lower during insulin than saline (25 ± 3 vs. 42 ± 5 μmol/liter, P = 0.003, not shown). Plasma total FFA concentrations during the 330- to 390-min sampling interval (Fig. 2) were lower during insulin than saline (75 ± 8 vs. 124 ± 13 μmol/liter, P = 0.002).

Fig. 2.

Plasma FFA concentrations during infusion of saline vs. insulin.

Steady-state conditions were achieved with respect to [U-¹³C]oleate APE and [³H]oleate specific activity during the last hour of the study (Fig. 3). Oleate clearance was higher during insulin than during saline (1116 ± 61 vs. 888 ± 100 ml/min, P < 0.02, not shown). As shown in Fig. 4, oleate R_a was lower during insulin compared with saline (27 ± 3 vs. 36 ± 5 μmol/min, P = 0.004), but there was no effect on fractional spillover (25 ± 2 vs. 26 ± 2%, respectively, P = NS).

Fig. 3.

Plasma [U-¹³C]oleate enrichment and [³H]oleate specific activity during infusion of saline vs. insulin.

Fig. 4.

R_a and fractional spillover during infusion of saline vs. insulin. *, P = 0.004 vs. saline.

Discussion

The present study was undertaken to determine whether suppression of lipolysis with insulin infusion during continuous meal absorption reduces the spillover of fatty acids generated by the action of LPL on chylomicron triglyceride in overweight and obese individuals with dyslipidemia. In a previous study in overweight and obese subjects, we found that more than half of systemic spillover occurs in the splanchnic bed (8); studies in dogs showed a strong relationship between fractional spillover and intracellular lipolysis in visceral fat (10). In the present study, insulin was infused during continuous meal absorption to induce suppression of adipose tissue lipolysis additional to that achieved by the meal. Despite a 25% suppression of lipolysis and a 40% decrease in plasma FFA concentrations, there was no change in spillover.

In normal subjects, intracellular lipolysis is exquisitely sensitive to the suppressive effects of insulin, with half-maximal suppression occurring at plasma insulin concentrations less than 2 μU/ml (16). Considering the apparent strong linkage between spillover and intracellular lipolysis in the splanchnic bed (10), where the majority of spillover occurs (8, 10), our failure to observe a decrease in spillover in the present studies is somewhat surprising. However, visceral fat is significantly resistant to the antilipolytic effects of insulin compared with sc fat (17), even in normal-weight individuals (18). It is thus possible that the increase in plasma insulin concentration that resulted when insulin was infused during meal absorption in our studies had a minimal additional antilipolytic effect in visceral fat compared with that achieved during meal absorption alone. It is possible that higher insulin infusion rates and the higher insulin concentrations they would achieve would result in a significant decrease in spillover.

As expected, plasma insulin concentrations were higher during insulin infusion than during saline infusion in our study. It is possible that this resulted in greater activation of LPL on the insulin infusion day because insulin is known to activate LPL (19). If this occurred, it would be expected to increase the clearance and decrease the concentration of chylomicron triglycerides. However, once steady-state rates of meal absorption and chylomicron secretion are achieved, the flux of chylomicron triglyceride would be expected to be the same on the 2 study days. Whether activation of LPL would alter the trafficking of LPL-generated fatty acids is not known.

We have recently found that iv niacin produces an approximate one third decrease in spillover with comparable reductions in systemic lipolysis to those seen here (20). A possible explanation for differences between the effects of insulin- and niacin-suppressed lipolysis on the spillover phenomenon is that niacin may be more effective than insulin in suppressing visceral lipolysis. It would be necessary to conduct splanchnic balance studies during infusion of insulin and niacin to determine whether this is the case.

The implications of our findings are of considerable potential importance. We found that nearly three quarters of oleate appearance during continuous feeding and saline infusion derived from spillover. This is somewhat greater than previously reported by Nguyen et al. (21), who found that 45% of oleate R_a was from spillover during continuous feeding in normal men. This may be due to the fact that oleate R_a during continuous feeding in the study by Nguyen et al. was approximately 50% higher than in our subjects.

Considering that spillover of LPL-generated fatty acids reflects inefficiency in the storage of dietary fat and that meal absorption (especially that of dietary fat) is occurring for most of the 24-h day, differences in spillover could have an important impact on the magnitude of ectopic fat deposition in nonadipose tissue (22) by variably augmenting FFA availability beyond that due to intracellular lipolysis. Thus, spillover could contribute to insulin resistance in addition to the effect of increased FFA availability due to impaired insulin suppression of intracellular lipolysis.

Dietary fat may be an important source of hepatic fatty acid uptake. The theoretical contribution of spillover in visceral adipose tissue to portal venous plasma FFA concentration can be calculated for our continuous feeding model using assumptions derived from the present study concerning oleate R_a, concentration, and ingestion rate together with data on fractional extraction of FFAs by splanchnic tissues, blood flow, uptake of and spillover from chylomicrons in the splanchnic bed, and the contribution of the splanchnic bed to systemic lipolysis derived from other studies of splanchnic fat metabolism in animals (10) and humans (8, 18). These calculations indicate that spillover in visceral adipose tissue would account for approximately 20% of the portal venous plasma FFA concentration. The contribution would be proportionally higher on a higher fat diet and would be expected to vary over time with noncontinuous food intake (23). Of course, this estimate would require experimental verification with portal venous sampling in an animal model. The online Supplemental Data, published on The Endocrine Society's Journals Online web site at http://jcem.endojournals.org, provides a detailed discussion of our estimate of spillover to systemic and hepatic fatty acid uptake and addresses the separate contribution from direct hepatic uptake of chylomicron fatty acids (10). Those calculations indicate that dietary fat may be a substantial contributor to total hepatic fatty acid availability, accounting for as much as half of the total.

In summary, an increase in insulin concentrations during continuous feeding does not decrease fractional spillover of LPL-generated fatty acids in overweight and obese subjects with dyslipidemia. Whether this is due to impaired insulin antilipolysis in visceral adipose tissue or other factors will require further study.