Fundamental Limitations of [18F]2-Deoxy-2-Fluoro-d-Glucose for Assessing Myocardial Glucose Uptake
Background The glucose tracer analog [18F]2-deoxy-2-fluoro-d-glucose (FDG) is widely used for assessing regional myocardial glucose metabolism in vivo. The reproducibility of this method has recently been questioned because of a discordant affinity of hexokinase for its substrates glucose and 2-deoxyglucose. We therefore compared rates of glucose utilization simultaneously with tissue time-activity curves of FDG uptake before and after changes in the physiological environment of the heart.
Methods and Results Isolated working rat hearts were perfused for 60 minutes with recirculating Krebs buffer containing glucose (10 mmol/L), FDG (1 μCi/mL), [2-3H]glucose (0.05 μCi/mL), and [U-14C]2-deoxyglucose (2-DG; 0.025 μCi/mL). Myocardial glucose uptake was measured by tracer ([2-3H]glucose) and tracer analog methods (FDG and 2-DG) before and after the addition of either insulin (1 mU/mL), epinephrine (1 μmol/L), lactate (40 mmol/L), or d,l-β-hydroxybutyrate (40 mmol/L) at 30 minutes of perfusion and after acute changes in cardiac workload. Under steady-state conditions, myocardial rates of glucose utilization as measured by tritiated water (3H2O) production from metabolism of [2-3H]glucose, FDG uptake, and 2-DG retention were linearly related. The addition of competing substrates decreased glucose utilization immediately. The addition of insulin increased the rate of glucose utilization as measured by the glucose tracer but not as measured by the tracer analogs. The ratio of 3H2O release/myocardial FDG uptake increased by 111% after the addition of insulin, by 428% after the addition of lactate, and by 232% after the addition of β-hydroxybutyrate. Epinephrine increased rates of glucose utilization and contractile performance, whereas there was no increase in glucose uptake with a comparable increase in workload alone. There was no change in the relation between the glucose tracer and the tracer analog either with epinephrine or with acute changes in workload.
Conclusions The uptake and retention of FDG in heart muscle is linearly related to glucose utilization only under steady-state conditions. Addition of insulin or of competing substrates changes the relation between uptake of the glucose tracer and FDG. These observations preclude the determination of absolute rates of myocardial glucose uptake by the tracer analog method under non–steady-state conditions.
During the past two decades [18F]2-deoxy-2-fluoro-d-glucose (FDG) has been widely used to assess glucose uptake in normal and diseased heart muscle.1 2 3 4 5 The rationale is relatively simple: like glucose and its analog 2-deoxyglucose, FDG is transported into the cell by specific glucose transporters and phosphorylated by the enzyme hexokinase. Once inside the cardiac myocyte, the phosphorylated tracer analog is not metabolized any further. Hence, the accumulation of FDG has the potential of tracing glucose uptake, although it cannot trace the metabolic fate of glucose inside the cell.
After the usefulness of FDG for metabolic imaging of the heart had been demonstrated,1 investigators applied the tracer kinetic model of Sokoloff et al6 to quantify the regional “myocardial rate of glucose metabolism in vivo.”7 8 A correction factor that equates 2-DG uptake to glucose uptake was experimentally derived and called the “lumped constant.” However, since the lumped constant used to determine myocardial rates of glucose metabolism was derived as a codependent variable, we have argued that the quantitative analysis of regional rates of myocardial glucose metabolism may be of limited value.9
Although the uptake of FDG is linear and constant10 and traces glucose uptake under steady-state conditions, the relation between tracer analog and a true glucose tracer is not always the same.11 At least two reports have shown that the lumped constant changes with insulin,11 12 and this change has been attributed to compartmentation of hexokinase II (EC 184.108.40.206) as well as to a differential affinity of hexokinase for glucose and 2-DG.12 These findings raise the question of whether other physiological perturbations change the relation between a glucose analog and glucose in a setting akin to the normal metabolic milieu of the heart.
To validate the tracer analog method, uptake of FDG must be measured simultaneously with a true glucose tracer under conditions similar to those in vivo in which FDG is used to assess myocardial glucose metabolism. We therefore examined the effects of a variety of experimental interventions, such as catecholamine stimulation, acute changes in workload, and the addition of competing substrates, on myocardial uptake of the glucose analogs FDG and 2-DG and a true glucose tracer, [2-3H]glucose. The isolated working rat heart model was used to measure substrate utilization, cardiac work, and uptake of glucose, FDG, and 2-DG precisely and under defined experimental conditions. We found that the addition of insulin or of a second substrate (particularly lactate) changed the relation between the glucose tracer and FDG. We believe that the disproportionate change in FDG uptake relative to glucose is important to the use of FDG in positron emission tomography (PET) of the heart.
Male Sprague-Dawley rats (weight, 275 to 300 g) obtained from Harlan were fed ad libitum on Purina laboratory chow or were fasted overnight (16 hours) with free access to water.
All chemicals were obtained from Fisher Scientific or Sigma Chemical Co. All enzymes and cofactors were obtained from Boehringer Mannheim or Sigma. Regular human insulin (Novolin R) was obtained from Novo-Nordisk.
The positron-emitting glucose analog FDG (specific activity >5000 Ci/mmol, ≈200 μCi per perfusion) was prepared at the University of Texas cyclotron facility at Houston by the method of Hamacher et al as described in detail previously (Nguyêñ et al10 ). High-performance liquid chromatography–purified [2-3H]glucose (10 μCi per perfusion) and [U-14C]2-deoxyglucose (5 μCi per perfusion) were obtained from Amersham Corp. The purity of [2-3H]glucose was ascertained by measuring the intrinsic 3H2O content of the tracer.
Working Heart Preparation
The preparation has been described in detail earlier.10 13 Briefly, rats were anesthetized by injection of sodium pentobarbital (100 mg/kg body wt IP). After injection of heparin (200 IU) into the inferior vena cava, the heart was rapidly removed and placed in ice-cold Krebs-Henseleit bicarbonate buffer. A brief period of retrograde perfusion (<5 minutes) with oxygenated buffer containing glucose (10 mmol/L) was necessary to wash out any blood from the heart and to perform the left atrial cannulation. Hearts were then switched to the working preparation. The perfusate (37°C) consisted of Krebs-Henseleit bicarbonate saline14 (200 mL) equilibrated with 95% O2/5% CO2 and contained glucose (10 mmol/L) as substrate, as well as other additions according to the individual protocols. Bovine serum albumin free of fatty acid (BSA; Cohn fraction V, fatty acid free, 1% wt/vol) was added to the perfusate, to which insulin or epinephrine was added. The perfusate was recirculated for the duration of the experiment. Aortic flow and coronary flow were measured every 5 minutes. Cardiac output was calculated as the sum of aortic and coronary flow. Heart rate and aortic pressures were continuously measured with a Hewlett-Packard transducer and recording system. The perfusion chamber was placed between a pair of coincidence detectors to monitor the tissue uptake of FDG. In addition, radioactivity in the perfusate (input function)10 was measured.
Six groups of rats were studied. The protocols are shown in Fig 1⇓. Perfusions were carried out at standard workload conditions (left atrial pressure, 10 cm H2O; aortic afterload, 100 cm H2O), with the following exceptions: (1) Perfusions to which insulin was added were performed at low workload (left atrial pressure, 7.5 cm H2O; aortic afterload, 70 cm H2O), since myocardial responsiveness to insulin is increased at low workload.15 (2) In the work jump experiments, hearts were initially perfused at low workload before it was changed to high workload (left atrial pressure, 15 cm H2O; aortic afterload, 140 cm H2O). In group 1 (control), hearts (n=4) were perfused at standard workload. This group served as control for hearts to which either lactate, β-hydroxybutyrate, or epinephrine was added. In a second control group, hearts from fasted animals (n=4) were perfused at low workload. This group served as control for hearts to which insulin was added. In group 2, insulin (1 mU/mL) was added at 30 minutes (n=8); hearts from fasted animals were perfused at low workload. In group 3, sodium l(+)lactate (40 mmol/L) was added at 30 minutes (n=9); hearts from fed animals were perfused at standard workload. In group 4, sodium d,l-β-hydroxybutyrate (40 mmol/L) was added at 30 minutes (n=7); hearts from fed animals were perfused at standard workload. In group 5, epinephrine (1 μmol/L) was added at 30 minutes (n=6); hearts from fasted animals were perfused at standard workload. Solutions of epinephrine were prepared immediately before their addition, and the perfusions continued in the dark so as to avoid degradation of epinephrine. In group 6, work jump (n=12), hearts from fasted animals were perfused at low workload for the first 20 minutes of perfusion. At 20 minutes, the workload was increased and maintained at high workload for the next 20 minutes before returning to low workload for the final 20 minutes of the experiment.
Radioactive Tracer Protocols
FDG and [2-3H]glucose were present in all experiments throughout. Activity of FDG in the perfusate and myocardial FDG accumulation were recorded continuously. 3H2O was measured in perfusate samples withdrawn at 5-minute intervals. [U-14C]2-DG was added at 30 minutes together with other additions made at the same time. In the work jump experiments, [U-14C]2-DG was added at 40 minutes along with a decrease in the workload on the heart. All hearts were subjected to a 5-minute “washout” period with nonradioactive perfusate before being freeze-clamped.
To assess the uptake of 2-DG before any intervention was made, hearts were perfused at low or at standard workload for 30 minutes in a separate series of experiments. 2-DG was added at the beginning of the experiment, and the hearts were subsequently freeze-clamped at 30 minutes of perfusion. In separate experiments, hearts were perfused for 20 minutes (n=4) at low workload and then for 40 minutes, 20 minutes at low workload followed by 20 minutes at high workload (n=4), and subsequently freeze-clamped. 2-DG was added either at the beginning of the perfusion or after 20 minutes of perfusion to measure the uptake of 2-DG at low and at high workloads, respectively. [2-3H]glucose was present throughout the perfusion period in both cases for comparison with the glucose analog 2-DG.
Assessment of Contractile Performance
Contractile performance, measured at 5-minute intervals, was assessed as cardiac power (in milliwatts) by the method of Kannengiesser et al16 : Power=aortic pressure (mm Hg)×cardiac output (mL/min)×(2.22×10−3).
Measurement of Radioactivity
18F radioactivity in tissue was externally recorded by a pair of coincidence γ-photon detectors placed on opposite sides of the heart, as previously described in detail.10 The processed signals were sent to their respective scalers (models 1774 and 1790C, Canberra) and then passed through a controller interface (model 188, Canberra). The input function was counted by β-counting of a portion of the recirculating perfusate as described earlier.10 All counts were decay-corrected from the time at which measured activity of FDG was added. Tissue time-activity of myocardial FDG uptake and perfusate activity were analyzed as described by Patlak et al.17 The slope of the Patlak plot represents the fractional rate of transport and phosphorylation of [18F]FDG from an extracellular compartment to an intracellular compartment under steady-state conditions. Tissue coincidence counts (cps) were not converted into microcuries per gram because such a conversion requires assumptions about the size and shape of the heart and incurs substantial potential inaccuracies. A well counter (Gamma Products) was used to determine the specific activity of FDG in the perfusate.
3H and 14C
Dual-label counting of these isotopes was performed on a Packard 1500 liquid scintillation counter by the method of spectral index analysis as described by the manufacturer (Packard Instruments). Glucose utilization was determined by the rate of 3H2O production from [2-3H]glucose.18 3H2O was separated from [2-3H]glucose by anion exchange chromatography19 on AG-1X8 resin (Bio-rad Laboratories, Inc). The method for measuring [U-14C]2-deoxyglucose has been described previously in detail12 and was adapted from the method of Hom et al.20 In the present work, the perfusate was changed to a nonrecirculating, tracer-free perfusate for the last 5 minutes of perfusion. The rationale for this “washout” technique is based on our earlier observation that the phosphorylation product of 2-DG is the only form of the tracer analog retained by the heart,10 which obviates the need for a separate isotopic extracellular fluid space marker. The efficacy of the washout was assessed by monitoring the radioactivity of the perfusate during the washout and by determining the specific activity of a sample of perfusate withdrawn at 65 minutes, immediately before the heart was freeze-clamped. The specific activity of the perfusate at the end of the washout was comparable to background activity, and the error introduced in assessment of myocardial 2-DG uptake was determined to be 1% to 2%.
Myocardial uptake of the glucose tracer and tracer analogs were simultaneously compared, as illustrated in Fig 2⇓. The rate of 3H2O release, measured at 5-minute intervals and expressed as the slope of the time-activity graph, was compared with Patlak graphical analysis of tissue time-activity curves of myocardial [18F]FDG uptake. Slopes of time-activity curves were determined by linear regression analysis. Patlak slopes (FDG uptake) and 3H2O slopes (glucose uptake) are expressed in units of μL perfusate · g tissue−1 · min−1. Rates of glucose uptake were also measured from an end-point analysis as determined by retention of [U-14C]2-deoxyglucose and by the release of 3H2O from [2-3H]glucose.
The accuracy of the coincidence counting system over the experimental range of counts was confirmed from a correlation study of counts as a function of activity decay with [18F]FDG. The AG-1X8 resin was checked by passing a known quantity of [2-3H]glucose and 3H2O through a column of resin.
Samples (1 mL) of the recirculating perfusate were withdrawn every 5 minutes. Glucose and lactate assays were performed immediately with a glucose/lactate analyzer (2300 STAT, YSI Inc). The glucose concentration (μmol/mL perfusate) was used to determine the specific activity of [2-3H]glucose and [U-14C]2-deoxyglucose. In experiments that included insulin, the concentration of insulin in perfusate samples determined by radioimmunoassay (Pharmacia Fine Chemicals) was 250±8 μU/mL and did not change significantly in the course of the perfusion.
Tissue Extraction and Metabolites
At the end of the perfusion, hearts were freeze-clamped between aluminum blocks cooled to the temperature of liquid nitrogen21 and stored at −70°C. The frozen tissue was ground under liquid nitrogen. Glycogen content of the tissue powder was determined as described by Walaas and Walaas.22 Total lipid extracts were prepared from a portion of the ground myocardial tissue.23 Triglyceride fractions were isolated by thin-layer chromatography on Whatman silica gel 60 A, with chloroform as mobile phase, and extracted from the silica gel.23 Triglycerides were determined from glycerol liberated upon alkaline hydrolysis.24 A small portion of the pulverized tissue was oven-dried for determination of the dry weight.
All data are presented as mean±SD. Statistical analysis was performed on a Macintosh computer with a statview se statistical package. Single-factor ANOVA and repeated-measures ANOVA were performed. Post hoc comparisons were performed with Scheffé F test if significant statistical differences were demonstrated.25
Fig 3⇓ shows a representative experiment with a heart perfused at standard workload. Rates of glucose uptake were linear under steady-state conditions when measured by either the rate of 3H2O production from [2-3H]glucose, slope of FDG accumulation (Patlak slope), or the retention of 2-DG. The relation between glucose analog and the glucose tracer did not change when measured by either the 3H2O/Patlak slope ratio or the ratio of uptake of glucose/2-DG. The first two tables show glucose uptake by hearts from fed animals perfused at standard workload (Table 1⇓) and by hearts from fasted animals perfused at low workload (Table 2⇓). Mean cardiac power was 8.56±1.22 mW and 5.78±0.28 mW (P<.05) for control hearts perfused at standard and low workloads, respectively. Comparison of Tables 1⇓ and 2⇓ shows that the uptake of glucose was higher in hearts from fed animals perfused at standard workload than in hearts from fasted animals perfused at low workload (P<.05). However, neither the nutritional state of the animal nor the workload affected the relation between glucose tracer and tracer analog.
The addition of insulin (Fig 4⇓) caused a significant increase in glucose uptake without a significant increase in the uptake of FDG. The ratio of 3H2O release to FDG uptake (3H2O/Patlak) increased by 111% (P<.004). Although glucose utilization and retention of 2-DG increased (Table 3⇓) after addition of insulin (255% and 57%, respectively), the ratio of glucose uptake to 2-DG retention (Glu/2-DG) increased by 63±25% (P<.03). Mean cardiac power was stable before and after the addition of insulin (5.63±0.82 versus 5.53±1.12 mW, P=NS).
The addition of lactate (40 mmol/L) or β-hydroxybutyrate (40 mmol/L) significantly decreased the uptake of the glucose tracer and the tracer analog (Figs 5⇓ and 6⇓). Lactate induced a decrease of 89% and β-hydroxybutyrate, a decrease of 82% in the Patlak slope (Table 4⇓). Glucose uptake decreased by 56% (6.54±1.85 versus 2.74±1.48 μmol · min−1 · g dry wt−1, P<.0001) after addition of lactate. The addition of β-hydroxybutyrate suppressed glucose uptake by 45% (6.90±1.52 versus 3.81±1.21 μmol · min−1 · g dry wt−1, P<.002). The 3H2O/Patlak slope ratio increased by 428% (P<.004) and by 232% (P<.02) after the addition of lactate and β-hydroxybutyrate, respectively. Cardiac power remained stable both before and after the addition of lactate (8.24±1.45 versus 8.49±1.12 mW, P=NS) and before and after the addition of β-hydroxybutyrate (8.32±1.3 versus 8.19±1.9 mW, P=NS).
The addition of epinephrine (1 μmol/L) (Fig 7⇓) caused a significant increase in the Patlak slope (+68%, P<.05), as well as a significant increase in the slope of 3H2O (+86%, P<.0003). Cardiac power increased by 25% (P<.05) after epinephrine was added (Table 5⇓). Glucose uptake increased by 75% and by 88% when traced by [2-3H]glucose and [U-14C]2-DG, respectively.
To test the hypothesis that the increase in glucose uptake observed with epinephrine was due to the inotropic effect of epinephrine on the heart, we raised both preload and afterload of hearts perfused with glucose alone (Table 6⇓). Although the work jump achieved by changing the workload was greater than that achieved by epinephrine stimulation, no change in glucose uptake was noted (Fig 8⇓). Hence, it is unlikely that the increase in cardiac power accounts for the increase in glucose uptake. The most likely source of energy was from the breakdown of endogenous substrates (see below).
Neither the addition of epinephrine nor acute changes in workload significantly changed the 3H2O/Patlak slope ratio or the Glu/2-DG ratio (Tables 5⇑ and 6⇑). Epinephrine-induced depletion of myocardial glycogen content was significantly greater than that induced by an acute increase in afterload (68.1% versus 35.6%, respectively; P<.0001). Epinephrine also caused a significant depletion of myocardial triglyceride content (98.3±7.84 versus 40.1±16.1 μmol · g dry wt−1 for control and epinephrine-treated hearts, respectively; P<.001). The results indicate that epinephrine mobilizes endogenous as well as exogenous fuel sources simultaneously with an increase in contractile performance.
The present study was undertaken to determine whether the glucose tracer analog method using FDG for the assessment of regional myocardial glucose metabolism yields consistent and reproducible results under controlled experimental conditions. We found a significant disparity between glucose and glucose analog methods in the assessment of glucose utilization not only after insulin but also after addition of either lactate or β-hydroxybutyrate. Comparable results were obtained when glucose utilization was traced by either method after stimulation with epinephrine or with acute changes in workload. The differential uptake of FDG relative to glucose in the presence of insulin or lactate may be clinically significant to the use of FDG in PET studies.
The results must be viewed in the context of the experimental preparation and the tissue under examination. Myocardial glucose metabolism has been studied in monkey,26 dog,1 rabbit,8 rat,10 and human heart.27 In the working rat heart preparation, workload and substrate concentration can be controlled precisely, and stable concentrations of glucose and insulin make it possible to apply a euglycemic hyperinsulinemic clamp. A new steady-state condition was reached in all groups in the present study either immediately (after addition of lactate or β-hydroxybutyrate) or after a delay of several minutes (with insulin or epinephrine). Since the uptake of tracer by the myocardium is dependent on the extracellular concentration (input function) of the tracer in the perfusate,10 it is essential to document that the input function is constant throughout the experiment (±1.5% in the present series). In most in vivo studies, FDG is presented as a bolus, and the rate of disappearance of the tracer from the blood pool is calculated with reference to activity measured in the blood pool of the left ventricle soon after injection.27 Clinically, after oral glucose loading, glucose levels are not constant over the observation period, insulin and lactate levels vary, and frequently the arterial input function is not measured, all of which are confounding variables in the interpretation or quantification of FDG uptake. In clinical PET, increased uptake of FDG is observed with an increase in insulin levels in response to oral glucose. Inconsistencies in FDG uptake have also been overcome by coadministration of glucose and insulin,28 by oral glucose loading after fasting,29 or by the euglycemic clamp technique.30 The clinical observations are consistent with our earlier findings that isolated hearts from fasted animals show greater insulin responsiveness when presented with glucose, lactate, and insulin.12 It is likely that the effect of an oral glucose load in clinical studies with FDG is related to an increased myocardial sensitivity to insulin in the presence of a variety of additional exogenous substrates. However, the present study failed to show a significant increase in the uptake of FDG after insulin when glucose was the only substrate. We chose not to include an additional substrate to avoid the confounding effect of competing substrates on the uptake of glucose. This experimental design most likely explains the lack of a significant effect of insulin on the uptake of FDG.
Triple Tracer Technique
FDG has the advantage of a high time resolution (500 milliseconds) for the assessment of glucose uptake. Patlak transformation of decay-corrected tissue time-activity curves normalizes the slope of FDG uptake for the measured input function. A disadvantage is that quantification of FDG uptake from Patlak slopes involves assumptions regarding the shape and size of the heart. The release of 3H2O from the detritiation of [2-3H]glucose occurs in a reversible reaction catalyzed by glucose 6-phosphate isomerase (EC 220.127.116.11) and can be used to measure glucose transport and phosphorylation. Although the question could be raised of whether insulin or competing substrates affect glucose 6-phosphate isomerase activity, this is unlikely, for two reasons. First, the turnover rate of the cycle between glucose 6-phosphate and fructose 6-phosphate is significantly greater than flux through the glycolytic pathway,18 and second, we have already shown that insulin changes the kinetic properties of hexokinase II.12 The validity of this method in skeletal muscle tissue has been established by Katz and Dunn,18 and the method has subsequently been used in heart muscle by Ng et al11 and Russell et al.12 Although the time resolution with 3H2O production is not as good as it is with FDG, it is possible to obtain slopes (rates) of glucose utilization over defined periods of time and compare those with Patlak slopes of FDG. We used the rate of 3H2O production from [2-3H]glucose as our gold standard because it is a direct measure of glucose utilization. The [U-14C]2-deoxyglucose method was used to provide an independent control glucose analog for the uptake of FDG. A disadvantage is that determination of the uptake and retention of [U-14C]2-deoxyglucose requires freeze-clamping of the heart. Since 2-DG provides a single end-point measurement, the uptake of 2-DG must be extrapolated from multiple experiments with different hearts. Hence, although comparisons can be made between hearts perfused under experimental conditions and control hearts, it is not possible to compare the rates of uptake of 2-DG under baseline and new steady-state conditions in the same heart.
The present results have direct bearing on the correction factor used to convert rates of tracer uptake to uptake of tracee (lumped constant, LC). Phelps et al31 reported a value of 0.38 for the LC in intact canine myocardium. Subsequently, the same investigators reported an experimentally derived value of 0.67 for the LC3 that has since been widely used. It is possible that the different values reported by the same laboratory are due to the acquisition of data under different steady-state conditions. Although it has previously been stressed that the lumped constant does not change in heart muscle,32 33 it must be noted that these results were obtained in an isolated interventricular septum preparation that was externally paced and perfused at a low coronary flow (≈1 mL saline perfusate · min−1 · g wet wt−1). The isolated working rat heart used in our study is of comparable mass, but it beats spontaneously and exhibits coronary flow rates in excess of 20 mL perfusate · min−1 · g wet wt−1. Ng et al11 34 used the same preparation and demonstrated a greater utilization of glucose relative to FDG after insulin at high workload. These investigators suggested that the LC is not constant but is independently sensitive to the concentration of glucose in the perfusate and that the LC rises as glycolysis becomes rate limiting.11
We have previously shown that insulin increases hexokinase activity associated with the mitochondrial fraction of tissue extracts.12 Although the redistribution of hexokinase to the mitochondria did not affect the apparent affinity constant for glucose, hexokinase bound to mitochondria exhibited an 8.5-fold increase in the Km for 2-DG compared with hexokinase in the cytosol. Our findings support and extend the observations of Ng et al and Russell et al.12 The difference in values for the LC in the report by Russell et al and in the present study is most likely due to the different experimental design. Whereas Russell et al used an extracellular fluid space marker12 20 to correct for the uptake of 2-DG, we used a washout technique but came to the same conclusions. The washout technique allows accurate determination of the error introduced from the radioactivity of the perfusate in each experiment. We determined that the optimal duration of the washout period causing the least error was 5 minutes; however, the error with even a 60-second washout was less than 5%. It is also likely that a 5-minute washout period will decrease the error from an intracellular accumulation of nonphosphorylated 2-DG. This experimental detail appears to account for the higher rate of uptake of 2-DG in the study by Russell et al compared with that in the present study.
Glucose is an essential fuel for energy production in the heart.35 36 Control of the metabolic flux of glucose metabolism in the heart is not exerted by a single enzyme but rather is variably distributed among enzymes depending on substrate availability, hormonal stimulation, and other conditions such as the dietary state of the animal, coronary flow, cardiac output, and the workload on the heart.37 In heart muscle and in other tissues able to meet energy needs through oxidation of a variety of substrates besides glucose, it is therefore reasonable to assume that regulatory mechanisms affect glucose transport and phosphorylation differently from transport and phosphorylation of the glucose tracer analog. In contrast, tissues such as the brain or tumor cells normally meet their energy requirements exclusively through glucose. The uptake and retention of the glucose tracer analog in these “glucose-dependent” tissues is a function of the metabolic activity of the tissue and of the input function of FDG. Hence, the quantitative relation between glucose and the glucose analog is unlikely to be affected, and FDG would provide a more accurate assessment of glucose utilization in those tissues than in heart muscle.
In heart and in skeletal muscle, glucose uptake has been shown to be suppressed with the addition of competing substrates such as lactate, fatty acids, or their breakdown products, ketone bodies. Preliminary studies with FDG from our laboratory have estimated that the uptake of glucose is suppressed by 90% with 40 mmol/L lactate and by 64% with 40 mmol/L β-hydroxybutyrate.38 In the present study, the uptake of 2-DG and [2-3H]glucose was also suppressed with the addition of either lactate or β-hydroxybutyrate. However, the addition of competing substrates induced a significantly greater suppression of FDG uptake compared with that of glucose. The mechanism of a disparate suppression of FDG uptake compared with that of glucose is not known. The effect of competing substrates on the affinity of hexokinase for its substrates has not yet been studied. It is also possible that both lactate and β-hydroxybutyrate effect an increase in the rate of transport of FDG from intracellular back to the extracellular space (k2 in the two-compartment model of Sokoloff et al6 ). This mechanism may also be responsible for the immediate suppression of glucose uptake noted on addition of either substrate.
Schneider et al39 observed that the enhanced lactate extraction in the reperfusion period after ischemia was associated with suppression of glucose utilization. In the present study, a supraphysiological concentration of lactate was used and induced a greater suppression of FDG uptake than the uptake of glucose. This observation may have clinical implications, although the studies were performed in a nonischemic preparation. In clinical PET studies on patients with ischemic heart disease, increased uptake of FDG and decreased coronary flow (flow-metabolism mismatch) are equated with reversibly ischemic (“viable”) heart muscle. Conversely, the lack of uptake of FDG is equated with “nonviable” myocardium. It is conceivable that excess lactate in an ischemic territory could disproportionately suppress FDG uptake relative to that of glucose, thereby obscuring true glucose uptake by the myocardium, and render invalid assumptions on the viability of the myocardium in the ischemic territory.
It is of interest that acute changes in workload were not associated with a change in rates of glucose utilization in the present study. Opie40 reported that an acute increase in workload increases myocardial glucose uptake in the isolated retrogradely perfused rat heart. Gertz et al41 reported that in humans, myocardial glucose utilization increases with exercise. Conversely, Nguyêñ et al10 showed that the uptake of FDG decreases with a decrease in the workload. The seemingly conflicting results of the present study point to the important contributions of endogenous substrates to myocardial fuel metabolism. Competition between exogenous and endogenous substrates for the fuel of respiration in the heart in response to changes in the workload has been recognized since the beginning of this century.42 Evans43 already suggested half a century ago that endogenous glycogen stores are used to meet the energy requirements associated with an increase in the workload on the heart.
Catecholamines have also long been known to increase glucose utilization.44 45 This increase is associated with an increase in cardiac performance and thus oxygen demand of the heart. However, increased cardiac work is an unlikely mechanism for an increase in glucose uptake. In the present study, we observed no increase in glucose uptake after an acute increase in the workload was the only intervention made. Furthermore, the increase in glucose utilization after epinephrine stimulation is observed after a delay of 5 to 7 minutes after administration of epinephrine, whereas cardiac performance increases immediately. The dichotomy between glucose uptake and an increase in workload suggests that other mechanisms may be responsible for the increase in glucose uptake noted after epinephrine stimulation. It has been suggested that an increase in glycogen turnover enhances glucose utilization.41 Nolte et al46 recently showed that epinephrine induced glycogen depletion and increased glucose transport in skeletal muscle in the absence of an increase in contractile performance. Like insulin, epinephrine has also been shown to recruit the glucose transporter GLUT4 to the sarcolemma,47 although it is not known whether an increase in glycogen turnover in heart muscle can cause this recruitment.
In contrast to the present in vitro experiments, Merhige et al28 observed a decrease in the uptake of FDG after an infusion of dopamine in an intact dog model. The depression in FDG uptake was attributed to an increase in circulating free fatty acids. The effect was reversed with insulin. Since an increase in fatty acid utilization for energy production results in a decrease in glucose oxidation,48 an increase in myocardial triglyceride breakdown could also have accounted for the decrease in uptake of FDG observed. However, in the present study, epinephrine caused an increase in glucose utilization despite a significant decrease in myocardial triglyceride content. Hence, it is most likely that the suppression of FDG uptake observed in vivo with catecholamine stimulation is not due to myocardial triglyceride breakdown but rather to an associated increase in circulating fatty acids that competes with glucose to mask the “insulin-like” action of epinephrine on myocardial glucose utilization. Thus, in the intact animal model of Merhige et al, catecholamine stimulation caused a decrease in FDG uptake due to an associated increase in fatty acids. However, in the more controlled isolated heart model used in this study, catecholamine stimulation had the direct effect of increased FDG uptake in the absence of changes in fatty acids or other competing substrates. These observations emphasize the importance of competing substrates on the uptake of glucose analogs under various physiological conditions.
The present studies in the isolated working rat heart demonstrate the limitations of FDG as an analog for measuring rates of myocardial glucose metabolism. Although the mechanisms by which insulin, lactate, or β-hydroxybutyrate decreases the uptake of FDG relative to glucose are not completely known, they are likely to involve the affinity of hexokinase for the tracer analog. The effect of competing substrates and insulin on changing the relation between the uptake of glucose and glucose analog must be considered when FDG is used to assess regional myocardial rates of glucose metabolism. The change in the relation between glucose tracer and tracer analogs renders FDG a qualitative, rather than a quantitative, tracer for myocardial glucose utilization. Although these limitations do not invalidate FDG as a tracer for the rapid kinetic analysis of glucose uptake, they call attention to the influence of the metabolic environment on the uptake of FDG relative to glucose.
This work was supported in part by the National Heart, Lung, and Blood Institute (RO1 HL-43133) and a Grant-in-Aid from the American Heart Association (National Center). Torsten Doenst was the recipient of a student fellowship from the German Academic Exchange Service (DAAD) and was on leave from the University of Göttingen, Germany. We acknowledge the technical assistance of Patrick Guthrie, Jonathan Lee, Sonya Carmical, and Alexis Woods and thank the cyclotron staff of the Positron Diagnostic and Research Center at the University of Texas Health Science Center, Houston, for preparation of FDG. We are also grateful to Drs K. Lance Gould and James E. Holden for their critical comments on the manuscript.
Guest editor for this article was Howard E. Morgan, MD, Sigfried and Janet Weis Center, Danville, Pennsylvania.
- Received October 27, 1994.
- Revision received December 28, 1994.
- Accepted January 10, 1995.
- Copyright © 1995 by American Heart Association
Phelps ME, Hoffman EJ, Selin C, Huang SC, Robinson G, Macdonald N, Schelbert H, Kuhl DE. Investigation of [18F] 2-fluoro-2-deoxyglucose for the measure of myocardial glucose metabolism. J Nucl Med. 1978;19:1311-1319.
Schelbert HR, Schwaiger M. Positron emission tomography studies of the heart. In: Positron Emission Tomography and Autoradiography: Principles and Applications for the Brain and the Heart. Phelps ME, Mazziota JC, Schelbert HR, eds. New York, NY: Raven Press; 1986:581-661.
Schwaiger M, Hicks R. The clinical role of metabolic imaging of the heart by positron emission tomography. J Nucl Med. 1991;32:565-578.
Sokoloff L, Reivich M, Kennedy C, Des Rosiers MH, Patlak CS, Pettigrew KD, Sakurada O, Shinohara M. The [14C] deoxyglucose method for the measurement of local cerebral glucose utilization: theory, procedure, and normal values in the conscious and anesthetized albino rat. J Neurochem. 1977;28:897-916.
Ratib O, Phelps ME, Huang SC, Henze E, Selin CE, Schelbert HR. Positron tomography with deoxyglucose for estimating local myocardial glucose metabolism. J Nucl Med. 1982;23:577-586.
Krivokapich J, Huang S-C, Phelps ME, Barrio JR, Watanabe CR, Selin CE, Shine KI. Estimation of rabbit myocardial metabolic rate for glucose using fluoro-deoxyglucose. Am J Physiol. 1982;243:H884-H895.
Taegtmeyer H. Utility and limitations of [18F]2-deoxy-2-fluoro-D-glucose for the assessment of flux through metabolic pathways in heart muscle: a critical appraisal. In: Schwaiger M, ed. Cardiac Positron Emission Tomography. Norwell, Mass: Kluwer Academic Publishers; 1995.
Nguyêñ VTB, Mossberg KA, Tewson T, Wong WH, Rowe RW, Coleman GC, Taegtmeyer H. Temporal analysis of myocardial glucose metabolism by 18F-2-deoxy-2-fluoro-D-glucose. Am J Physiol. 1990;259:H1022-H1031.
Ng CK, Holden JE, DeGrado TR, Raffel DM, Kornguth ML, Gatley SJ. Sensitivity of myocardial fluorodeoxyglucose lumped constant to glucose and insulin. Am J Physiol. 1991;260:H593-H603.
Russell RR, Mrus JM, Mommesin JI, Taegtmeyer H. Compartmentation of hexokinase in rat heart: a critical factor for tracer kinetic analysis of myocardial glucose metabolism. J Clin Invest. 1992;90:1972-1977.
Taegtmeyer H, Hems R, Krebs HA. Utilization of energy providing substrates in the isolated working rat heart. Biochem J. 1980;186:701-711.
Russell RR, Nguyêñ VTB, Mrus JM, Taegtmeyer H. Fasting and lactate unmask insulin responsiveness in the isolated working rat heart. Am J Physiol. 1992;263:E556-E561.
Marshall RC, Nash WW, Shine KI, Phelps ME, Ricchiuti N. Glucose metabolism during ischemia due to excessive oxygen demand or altered coronary flow in the isolated arterially perfused rabbit septum. Circ Res. 1981;49:640-648.
Hom FG, Goodner CJ, Berrie MA. A (3H) 2-deoxyglucose method for comparing rates of glucose metabolism and insulin responses among rat tissues in vivo. Diabetes. 1984;33:141-152.
Walaas O, Walaas E. Effect of epinephrine on rat diaphragm. J Biol Chem. 1950;187:769-776.
Bligh EG, Dyer WJ. A rapid method of total lipid extraction and purification. Can J Biochem Physiol. 1959;378:911-917.
Eggstein M, Kuhlmann E. Triglycerides and glycerol: determination after alkaline hydrolysis. Methods Enzymat Anal. 1974;4:1825-1831.
Zar JH. Biostatistical Analysis. 2nd ed. Englewood Cliffs, NJ: Prentice-Hall Inc; 1984.
Gallagher BM, Ansari A, Atkins H, Casella V, Christman DE, Fowler JS, Ido T, MacGregor RR, Sam P, Wan C-N, Wolf AP, Kuhl DE, Reivich M. Radiopharmaceuticals, XXVII: 18-F labelled 2-deoxy-2-fluoro-D-glucose as a radiopharmaceutical for measuring regional myocardial glucose metabolism in vivo: tissue distribution and imaging studies in animals. J Nucl Med. 1977;18:990-996.
Gambhir SS, Schwaiger M, Huang SC, Krivokapich J, Schelbert HR, Nienaber CA, Phelps ME. Simple noninvasive quantification method for measuring myocardial glucose utilization in humans employing positron emission tomography and fluorine-18 deoxyglucose. J Nucl Med. 1989;30:359-366.
Merhige ME, Ekas R, Mossberg K, Taegtmeyer H, Gould KL. Catecholamine stimulation, substrate competition, and myocardial glucose uptake in conscious dogs assessed with positron emission tomography. Circ Res. 1987;61(suppl II):II-124-II-129.
Knuuti MJ, Nuutila P, Ruotsalainen U, Saraste M, Harkonen R, Ahonen A, Teras M, Haaparant M, Wegelius U, Haapanen A, Hartiala J, Voipio-Pulkki LM. Euglycemic hyperinsulinemic clamp and oral glucose load in stimulating myocardial glucose utilization during positron-emission tomography. J Nucl Med. 1992;33:1255-1262.
Krivokapich J, Huang S-C, Selin CE, Phelps ME. Fluorodeoxyglucose rate constants, lumped constant, and glucose metabolic rate in rabbit heart. Am J Physiol. 1987;252:H777-H787.
Ng CK, Holden JE, DeGrado TR, Kornguth ML, Raffel DM, Gatley SJ. Variation of 2-FDG/glucose lumped constant with substrate and hormone concentration in isolated rat heart. J Nucl Med. 1986;27:966. Abstract.
Taegtmeyer H. Energy metabolism of the heart: from basic concepts to clinical applications. Curr Probl Cardiol. 1994;19:57-116.
Randle PJ, Priestman DA, Mistry SC, Halsall A. Glucose fatty acid interactions and the regulation of glucose disposal. J Cell Biochem. 1994;55S:1-11.
Kashiwaya Y, Sato K, Tsuchiya N, Thomas S, Fell DA, Veech RL, Passonneau JV. Control of glucose utilization in working perfused rat heart. J Biol Chem. 1994;269:25502-25514.
Taegtmeyer H, Doenst T, Mommessin JI, Guthrie PH, Williams CM. Further evidence for the importance of anaplerosis in the isolated working rat heart: a tracer kinetic study with [18F] fluoro-2-deoxyglucose (FDG). Circulation. 1993;88(suppl I):I-284. Abstract.
Schneider CA, Nguyêñ VTB, Taegtmeyer H. Feeding and fasting determine postischemic glucose utilization in isolated working rat hearts. Am J Physiol. 1991;260:H542-H548.
Gertz EW, Wisneski JA, Stanley WC, Neese RA. Myocardial substrate utilization during exercise in humans. J Clin Invest. 1988;82:2017-2025.
Rohde E. Stoffwechseluntersuchungen am ueberlebenden Warmblüterherzen, I. Mitteilung zur Physiologie des Herzstoffwechsels. Z Physiol Chem. 1910;68:181-235.
Evans G. The glycogen content of the rat heart. J Physiol (Lond). 1934;82:468-480.
Evans CL. The effect of glucose on the gaseous metabolism of the isolated mammalian heart. J Physiol (Lond). 1914;47:407-418.
Evans CL. The metabolism of cardiac muscle. In: Evans’ Recent Advances in Physiology. 6th ed. Newton WH, ed. Philadelphia, Pa: Blakiston’s Sons & Co; 1939;7:157-182.
Nolte LA, Gulve EA, Holloszy JO. Epinephrine induced in vivo muscle glycogen depletion enhances insulin sensitivity of glucose transport. J Appl Physiol. 1994;765:2054-2058.
Randle PJ, Tubbs PK. Carbohydrate and fatty acid metabolism. In: Handbook of Physiology: The Cardiovascular System. Berne RE, Sperelakis V, Geiger SR, eds. Bethesda, Md: American Physiological Society; 1979:805-844.