Introduction

Approximately 60% to 70% of energy is derived from oxidation of fatty acids in the adult heart, while the rest is being provided mostly by glucose and lactate. Insulin resistant states and type 2 diabetes are associated with increased circulatory free fatty acid (FFA) and triglyceride fluxes [1] and with a very important increase in myocardial fatty acid utilization at the expense of reduced glucose utilization [2]. More reliance on fatty acid oxidation with reduced oxidation of glucose is associated with reduced energetic efficiency, more oxygen being required from the same cardiac workload [3]. This may pose a problem in critical situations such as during ischemia.

Increased FFA delivery to lean tissues, including the heart, reduces glucose utilization [4] (see [1] for review). Another potentially very important source of fatty acids to the heart is circulating triglyceride-rich lipoproteins. Because triglycerides circulate in the millimolar range and contain three fatty acids per molecule and because the heart highly expresses lipoprotein lipase (LpL) [5], this has been proposed as the major source of fatty acids to the heart [6]. Experimental evidence in perfused heart models [7, 8] and from in vivo studies in rodents [9, 10] have suggested that the heart is a major user of fatty acids derived from circulating triglycerides. Several genetically modified mouse models with either reduced [11–13] or increased [14, 15] myocardial LpL activity have demonstrated the potential of chronic modulation of myocardial fatty acid delivery from circulating triglycerides to change substrate selection and induce cardiomyopathy. However, changes in circulating triglyceride levels occur acutely after dietary fat intake. The impact of very acute changes in fatty acid delivery from circulating triglycerides on myocardial energy substrate selection is currently unknown.

The aim of the present study was to study the impact of myocardial fatty acid delivery from circulating triglycerides on myocardial glucose utilization in the presence or absence of intravenous infusion of Intralipid, a triglyceride emulsion, to simulate a hypertriglyceridemic state. Our hypothesis was that inhibition of myocardial fatty acid delivery from circulating triglycerides would lead to reversal of Intralipid-induced reduction of myocardial glucose utilization if this is a major source of fatty acid delivery to the heart in vivo (Fig. 1).

Fig. 1
figure 1

Myocardial fatty acid uptake from circulating triglycerides (TG) vs. circulating free fatty acids (FFA). TG contained in lipoproteins are hydrolyzed by lipoprotein lipase (LpL), a process that may contribute to myocardial fatty acid delivery with consequent reduction in myocardial glucose metabolism during acute hypertriglyceridemia. Triton WR-1339 is a detergent that binds circulating TG and prevents LpL-mediated triglyceride hydrolysis (intravascular lipolysis). The physiological processes that are described by the multi-compartmental analysis of fluoro-deoxyglucose (FDG) (i.e., K 1, k 2, k 3, and myocardial metabolic rate of glucose—MMRG) are also depicted.

Full size image

Materials and Methods

Animals

Male Wistar rats (Charles River, Quebec, Canada) weighting 350–400 g were acclimatized to 12 h/day light cycle at constant temperature (22°C) for 7 days. A total of 63 rats were used for the experiments described herein. After a 12-h fasting period, animals were anesthetized with Isoflurane (Abbott laboratories, Montreal, Canada) delivered through a nose cone at a concentration of 2.0% (volume:volume). Anesthesia was maintained for the duration of the in vivo experiments. Catheters were placed into a carotid artery for blood sampling and into a jugular and two tail veins for intravenous infusions. The catheters were kept open with infusion of 0.9% saline. All animal protocols were approved by the Animal Ethics Committee of the Faculty of Medicine of the Université de Sherbrooke in accordance with the guidelines of the Canadian Council on Animal Care.

Experimental Protocols

All in vivo experiments described herein were initiated 60-min after insertion of catheters and performed under euglycemic–hyperinsulinemic conditions, as previously described [4]. At time 0 of the experiment, a primed (180 mU/kg) constant insulin infusion [12 mU/kg/min in 0.1% bovine serum albumin (BSA) in normal saline; Novolin®ge, Toronto, Canada] was started and continued for 2 h to standardize the metabolic condition in all in vivo protocols. Whole-blood glucose level was maintained using a variable 25% dextrose intravenous infusion according to the glucose level, determined every 10 min using a blood glucose monitor (Accusoft Advantage™, Roche, USA) [16]. Also at time 0, Triton WR-1339 (Tylaxapol, Sigma Chemicals, 400 mg/kg; referred to as the TRI group) vs. normal saline (referred to as the SAL group) was injected intravenously to inhibit intravascular triglyceride lipolysis during the experiment [17]. Our first set of in vivo experiments (Fig. 2a) was performed to validate the use of Triton WR-1339 to reduce fatty acid availability to tissues specifically derived from circulating triglycerides and not from plasma FFA. [1-14C]-2-bromopalmitate (a FFA tracer—injected in n = 18 animals) in 5% BSA (Moravek Biochemicals, California), 5 μCi, and 60 μCi of [1–3H]-triolein (a triglyceride tracer—injected in n = 15 animals) sonicated in 0.5 ml of Intralipid 20% were given intravenously at time 120 and 125 min, respectively, during the euglycemic–hyperinsulinemic clamp, as previously described [16], with and without intravenous administration of Triton WR-1339 as described above. Arterial blood samples were taken at time −0 and 0 min and at times 90, 100, 110, and 120 min to determine blood glucose and plasma insulin, FFA, and triglyceride levels. Blood samples were also taken at times 121, 123, and 125 to determine plasma [1-14C]-2-bromopalmitate and at times 126, 128, and 130 to determine plasma [1-3H]-triolein radioactivity, and organs were removed and processed to determine tissue tracer uptake as described below.

Fig. 2
figure 2

a Experimental protocol for the in vivo validation of Triton WR-1339 method. b Experimental protocol of the in vivo experiments using micro-positron emission tomography (μPET).

Full size image

Imaging Protocols

Figure 2b depicts the imaging protocol used in our second set of in vivo experiments. To determine whether our in vivo imaging methods could detect change in myocardial metabolism with change in circulating lipids, we performed experiments using intravenous infusion of Intralipid 20% (20 μl/min) (IL) vs. saline (SAL) from time 0 and continued for the 2 h of the experiment. Intralipid is a chylomicron-like emulsion that may be employed to increase plasma triglyceride levels in vivo [18] and has been shown by others to be metabolized similarly as endogenous lipoproteins by the heart in rodents [6, 15]. In a third group of rats, we administered intravenous injection of Triton WR-1339 together with IL infusion (IL + TRI group) to determine the role of intravascular triglyceride lipolysis on any effect of IL on myocardial metabolism. Imaging experiments were performed with the avalanche photodiode-based small animal PET scanner (μPET) of the Sherbrooke Molecular Imaging Centre [19]. Before imaging, the heart position was localized with a Doppler probe (0.64 cm [1/4 in.], 9 MHz; Parks Medical Electronics). During imaging, the animals rested supine on the scanner bed and were kept warm with a heating pad. At time 50-min, approximately 5 mCi [13N]-ammonia in 0.5 ml of saline was administered intravenously over a period of 30 s using an automated injection system to determine myocardial blood flow [20]. This method had been shown to give similar estimates of myocardial blood flow when compared to the [15O]-water technique [21]. Image acquisition was performed using a 10-min list-mode dynamic acquisition. At time 70 min, an intravenous injection of approximately 3 mCi [11C]-acetate in 0.5 ml of saline was administered over 30 s to determine myocardial oxidative metabolic rate [22]. Image acquisition was also performed using a 10-min list-mode dynamic acquisition. At time 90-min, an intravenous injection of approximately 1 mCi 2-deoxy-2-[F-18]fluoro-d-glucose (FDG) in 0.5 ml of saline was given over a period of 30 s to determine myocardial glucose utilization. Image acquisition was performed using a 30-min list-mode dynamic acquisition. The PET imaging protocols were previously validated in other studies [23, 24]. Blood samples were taken at time 90, 100, 110, and 120-min to determine blood glucose and plasma FFA and TG.

Incorporation of Tracers into Plasma, Intracellular Lipids, and Mitochondria

Determination of uptake of plasma [1-3H]-triolein and [1-14C]-2-bromopalmitate into tissue lipids was performed as described previously [16]. Briefly, the heart was removed immediately and snap-frozen after thorough washing with 0.9% NaCl and blotting to remove excess water. Samples of heart (100 mg) were homogenized in a 0.9% NaCl solution. Plasma and tissue lipids were extracted according to the method described by Folch et al. [25] and were applied onto thin-layer chromatography plates (Silica Gel 60, F-254, Selecto Scientific, Suwanee, GA, USA) with standards for free fatty acids (FFA), triglycerides (TG), diglycerides (DG), monoglycerides (MG), cholesterol esters, and phospholipids. The plates were eluted in a hexane/diethyl ether/acetic acid (80:20:2) solution and stained with dichlorofluorescein (1 mg/mL ethanol). Relative uptake (%) of [14C] and [3H] in different tissue lipids was calculated by dividing the activity in each fraction by the total recovered in all fractions multiplied by 100.

Mitochondria from the heart were extracted by using the method previously described in [16, 25]. Briefly, 500 mg of the heart were homogenized on ice (PowerGen 125, Fisher Scientific International Inc., USA) in an ice-cold homogenization solution (0.075 M sucrose, 0.225 M sorbitol, 1 mM EGTA, 0.1% fatty-acid-free BSA, and 10 mM Tris–HCl, pH 7.4). The homogenates were centrifuged at 1,000 × g for 10 min at 4°C, and the supernatants were centrifuged again at 12,000 × g for 10 min at 4°C. The supernatants were then removed, and 1,000 μL of an incubation solution (10 mM Tris–HCl, 25 mM sucrose, 75 mM sorbitol, 100 mM KCl, 10 mM K2HPO4, 0.05 mM EDTA, 5 mM MgCl2, and 1 mg/mL FFA-free BSA) was added to the pellet containing mitochondria. Fifty microliters of the mitochondrial samples was kept frozen at −80°C for the measurement of glutamate dehydrogenase (GDH) activity to correct for the fractional recovery of mitochondria extracts from tissues [25]. The GDH activity was measured according to the Sigma quality control test procedure (EC 1.4.1.3, Enzymatic Assay of L-GDH) modified as previously described [16]. The fractional recovery of mitochondria was then calculated by dividing mitochondrial GDH activity (units per gram of tissue) by whole-tissue GDH activity (units per gram of tissue). Mitochondrial [14C] and [3H] activity were corrected by dividing by the fractional recovery to determine mitochondrial uptake of [1-14C]-2-bromopalmitate and [1-3H]-oleate from labeled triolein, respectively.

Imaging Data Analysis

For [13N]-ammonia and [11C]-acetate, dynamic series of 23 frames each were sorted out from the list mode data, using the following sequence: 1 × 30 s, 12 × 5 s, 8 × 30 s, 2 × 150 s. For FDG, a dynamic sequence series of 29 frames was sorted out using the following sequence: 1 × 30 s, 12 × 5 s, 8 × 30 s, 6 × 150 s, 2 × 300 s. Image planes were reconstructed on a 128 × 128 matrix with a 0.475 × 0.475 mm pixel size using the maximum likelihood expectation maximization algorithm with 15 iterations [26, 27]. Regions of interest (ROI) were drawn manually over the myocardium in the left ventricle and on the left ventricular cavity from the best frame image acquired with [13N]-ammonia acquisition and then used on all frames of the three tracers. Tissue time–activity curves were generated from the ROI on the myocardium, and the arterial tracer activity curve was obtained from the ROI drawn over the left ventricular cavity. The PET kinetic parameter values were assessed with the input function derived from the images. We used a three-compartment kinetic model for [13N]-ammonia that provides an estimate of blood flow from k 1 value [27]. For [11C]-acetate, we used a three-compartment kinetic model that estimates the generation of CO2 from the citric cycle in the myocardium using k 2 value [26]. For the modeling of myocardial uptake of FDG, we used the classical three-compartment model of Sokoloff [28]. According to this model, K 1 is an index of membrane glucose transport, k 2 is an index of FDG escape, k 3 is an index of glucose phosphorylation. The myocardial metabolic rate of glucose (MMRG) can be derived from these constants, the arterial blood glucose concentration, and the lumped constant (LC):

$${\text{MMRG}} = \left( {\frac{{{\text{glycemia}}_{\left( {{{{\text{mmol}}} \mathord{\left/ {\vphantom {{{\text{mmol}}} {\text{l}}}} \right. \kern-\nulldelimiterspace} {\text{l}}}} \right)} }}{{{\text{LC}}}}} \right) \times K_{\text{m}}$$
(Eq.1)
$$K_{\text{m}} = \frac{{\left( {K_1 \times k_3 } \right)}}{{\left( {k_2 + k_3 } \right)}}$$
(Eq.2)

where LC is a term made up of six constants that relate the transport and phosphorylation kinetics of FDG to that for glucose [29].

Statistical Analysis

All data are reported as mean ± SE. Plasma glucose, insulin, FFA, and TG levels at steady state (baseline and between 90 to 120 min during the euglycemic clamp) were averaged for statistical analysis. Unpaired Student t test or one-way ANOVA with Tukey’s post hoc test were performed using GraphPad Prism version 5.00 for Windows (San Diego California USA). A P value of less than 0.05 was considered significant.

Results

Triton WR-1339 Selectively Reduces Myocardial Uptake of Circulating Triglycerides but not Circulating FFA

Blood glucose and plasma insulin, FFA, and triglyceride levels at baseline and during euglycemic–hyperinsulinemic clamp in rats injected with SAL vs. TRI are shown in Table 1. There was no significant difference in these parameters at baseline between the two groups. During the clamp, blood glucose and plasma FFA levels were similar between the two groups. As expected, plasma triglyceride levels were significantly increased approximately fourfold in TRI (P < 0.001). However, plasma insulin levels were lower in TRI vs. SAL (P = 0.04). This effect was explained by an in vitro interference of Triton WR1339 with the insulin immunoassay. This was further suggested by glucose infusion rates (GINF) that were virtually identical with and without TRI injection (Table 1).

Table 1 Plasma glucose, insulin, FFA and TG levels at baseline and during the normoglycemic–hyperinsulinemic clamp
Full size table

Table 2 depicts the effect of injection of Triton WR-1339 on area under curve (AUC) of plasma 3H-triolein activity and on myocardial uptake and incorporation of 3H-triolein-derived 3H in cellular lipids and in mitochondria. TRI was associated with an ~168% increase of plasma 3H-triolein activity AUC (P = 0.02) and with ~48% (P = NS) and ~44% (P = 0.03) reductions in myocardial 3H-triolein-derived label uptake in cellular lipids and in mitochondria, respectively. However, TRI had no effect on relative uptake of 3H-triolein in the various cellular lipid fractions. Thus, Triton WR-1339 reduced myocardial fatty acid uptake from circulating triglycerides without change in partitioning into the different intracellular lipids.

Table 2 Effect of Triton WR-1339 on plasma AUC of 3H-triolein activity and on myocardial uptake of 3H-triolein
Full size table

In contrast, TRI had no significant effect on plasma [1-14C]-2-bromopalmitate AUC activity and on myocardial uptake of this tracer in cellular lipids and in mitochondria (Table 3). Furthermore, TRI did not affect relative uptake of [1-14C]-2-bromopalmitate into the various cellular lipid factions.

Table 3 Effect of Triton WR-1339 on plasma AUC of [1-14C]-2-bromopalmitate activity and on myocardial uptake of [1-14C]-2-bromopalmitate
Full size table

μPET Studies

Weight of the animals was similar in SAL, IL, and IL + TRI experiments (387 ± 12, 387 ± 11, and 418 ± 14 g, respectively, P = 0.14). After surgery, heart rate and body temperature were also similar in all three groups (heart rate: 354 ± 14, 350 ± 21, and 303 ± 14 beats/min, respectively, P = 0.09; body temperature: 33.9 ± 0.7, 34.5 ± 0.4, and 35.1 ± 0.4°C, respectively, P = 0.30). Baseline blood glucose and plasma FFA levels were not significantly different between SAL, IL, and IL + TRI (blood glucose: 6.5 ± 0.7, 6.4 ± 0.7, and 5.9 ± 0.4 mmol/l, respectively, P = 0.81; baseline plasma FFA: 0.68 ± 0.12, 0.78 ± 0.06, and 0.72 ± 0.10 mmol/l, respectively, P = 0.81). Baseline plasma triglyceride levels were also not significantly different between the three groups (0.21 ± 0.02, 0.37 ± 0.09, and 0.25 ± 0.04 mmol/l, respectively, P = 0.17).

Table 4 shows the metabolic parameters of interest during the euglycemic–hyperinsulinemic clamp and μPET study. By design, plasma glucose levels were similar between the three experimental groups. Plasma FFA levels were significantly higher in IL and IL + TRI vs. SAL. Plasma triglyceride levels in IL and IL + TRI experiments were also significantly higher than SAL (P < 0.001). FDG K 1 and k 2 were similar in all three groups, whereas k 3 was significantly reduced in IL and IL + TRI compared with SAL (P = 0.02). MMRG (Fig. 3a) was significantly lower in IL and IL + TRI vs. SAL (P < 0.05). IL also significantly reduced myocardial blood flow ([13N]-ammonia k 1—Fig. 3b) vs. SAL (P < 0.05), whereas TRI did not significantly improve this IL-mediated reduction in blood flow. Total myocardial oxidative rate (11C-acetate k 2—Fig. 3c) was not significantly different between the three groups (P = 0.06), although it tended to be lower in the IL + TRI group vs. the two others. Figure 4 shows representative μPET scans after FDG injection in the three groups of rats.

Fig. 3
figure 3

a Myocardial metabolic rate of glucose (MMRG), b myocardial blood flow, and c total myocardial oxidative rate during euglycemic clamp in rats treated with saline (SAL, open bars), Intralipid 20% (IL, closed bars), and Intralipid 20% + Triton WR-1339 (IL + TRI, grey bars). a P < 0.05 vs. SAL group by ANOVA with Tukey’s post hoc test. Data are mean ± SE.

Full size image
Fig. 4
figure 4

Representative myocardial transaxial μPET scan images after injection of [18F]FDG in the saline (SAL - upper panel), Intralipid 20% (IL - middle panel), and Intralipid 20% + Triton WR-1339 (IL + TRI - lower panel) groups.

Full size image
Table 4 Metabolic parameters during the μPET imaging protocol with euglycemic–hyperinsulinemic clamp
Full size table

Discussion

In the present study, IL resulted in increased plasma triglyceride and FFA levels together with significant reduction in MMRG and reduction in myocardial blood flow but no change in myocardial oxidative capacity, heart rate, or body temperature. In contrast to our initial hypothesis, we found no significant reversal of IL-induced reduction in MMRG with Triton WR-1339. Triton WR-1339 also did not reverse IL-induced reduction in myocardial blood flow. We demonstrated that Triton WR-1339 reduced clearance of circulating triglycerides and reduced myocardial uptake of triglycerides without significantly changing plasma FFA clearance and myocardial uptake. Thus, our results suggest that IL-induced reduction in blood flow and MMRG is not primarily caused by increased fatty acid delivery through increased intravascular triglyceride lipolysis during euglycemic–hyperinsulinemia in rats. Triton WR-1339 was not associated with change in insulin sensitivity, as assessed by lack of change in glucose infusion rates during euglycemic–hyperinsulinemic clamps. Plasma insulin levels were lower after injection of Triton WR-1339, but this effect was caused by interference of Triton WR-1339 with the insulin immunoassay used in the present study.

Our results are consistent with those of other investigators showing reduced myocardial glucose utilization with intravenous infusion of Intralipid in vivo [4, 30]. Intravenous infusion of lipids has also been previously shown to alter myocardial blood flow in other animal models [31, 32]. This lipid-induced reduction in myocardial blood flow could also be mediated by increased blood viscosity [33]. Interestingly, human LpL overexpression in aorta vascular smooth muscles in mice alters vascular reactivity [34], suggesting that lipolysis of triglycerides leads to endothelial dysfunction in vivo. However, injection of Triton WR-1339 did not blunt IL-induced reduction in myocardial blood flow. The lack of Intralipid-induced change in total myocardial oxidative metabolism that we observed in the present study is also consistent with results from other investigators [35].

One possible explanation for this finding may be compensatory increase in myocardial plasma FFA utilization in the presence of inhibition of myocardial utilization of fatty acids derived from circulating triglycerides. Indeed, this phenomenon has been shown to occur in mice using Langendorff ex vivo myocardial perfusion protocols in some [6, 15] but not all studies [7]. We found that blocking myocardial availability of fatty acids from circulating triglycerides did not affect myocardial 14C-bromopalmitate cellular and mitochondrial uptake. Thus, elevation of FFA levels seen with IL infusion may have been the primary cause of IL-mediated reduction in myocardial blood flow and MMRG. It is also possible that the hyperinsulinemic condition in our experiments reduced myocardial LpL activity [36], thus, minimizing the relative contribution of fatty acid uptake from circulating triglycerides. Finally, it is also possible that fatty acids delivered from circulating triglycerides do not acutely mix into the same intracellular fatty acid pool than that from plasma FFA [10], with different acute metabolic consequences. Intriguingly, total myocardial oxidative rate tended to be lower in the IL + TRI vs. IL group, an effect that may have been expected on the basis of reduced myocardial fatty acid delivery from circulating triglycerides in the face of no significant change in myocardial FFA uptake and similar MMRG.

Cardiac-specific LpL knockout and cardiac-specific LpL overexpression mouse models have illustrated the potential importance of prolonged lack or excess of fatty acid delivery from circulating triglycerides for the development of cardiomyopathy associated with change in myocardial glucose utilization [11–15]. However, the relative contribution of fatty acids derived from circulating triglycerides vs. plasma FFA pool as a substrate for the heart is controversial in physiological conditions. Results from some investigators have suggested that utilization rate of circulating triglyceride-derived fatty acids in the mouse heart is quantitatively lower [37], similar [7], or greater [6] to plasma FFA. These disparities are likely explained by different experimental and physiological conditions since myocardial fatty acid delivery from circulating triglycerides may depend on its lipoprotein source, and cardiac LpL function is modified in various conditions [38]. Interestingly, a recent human study has shown that plasma FFA contribute to more than 80% of fatty acids utilized by the heart in vivo in humans with coronary artery disease during the fasting state, whereas circulating triglycerides contribute to less than 20% [39].

Conclusion

In conclusion, the results from the present study show that intravenous infusion of fat emulsion acutely reduces myocardial glucose utilization and blood flow in vivo during euglycemic–hyperinsulinemia in rats. However, these effects do not result from increased delivery of fatty acids from intravascular triglyceride lipolysis. Future studies will be needed to determine whether myocardial fatty acid delivery from circulating triglycerides may play a role in reduced myocardial glucose utilization in pathophysiological states such as type 2 diabetes.