Abnormal Cardiac and Skeletal Muscle Energy Metabolism in Patients With Type 2 Diabetes
Background— It is well known that patients with type 2 diabetes have increased risk of cardiovascular disease, but it is not known whether they have underlying abnormalities in cardiac or skeletal muscle high-energy phosphate metabolism.
Methods and Results— We studied 21 patients with type 2 diabetes with no evidence of coronary artery disease or impaired cardiac function, as determined by echocardiography, and 15 age-, sex-, and body mass index-matched control subjects. Cardiac high-energy phosphate metabolites were measured at rest using 31P nuclear magnetic resonance spectroscopy (MRS). Skeletal muscle high-energy phosphate metabolites, intracellular pH, and oxygenation were measured using 31P MRS and near infrared spectrophotometry, respectively, before, during, and after exercise. Although their cardiac morphology, mass, and function appeared to be normal, the patients with diabetes had significantly lower phosphocreatine (PCr)/ATP ratios, at 1.50±0.11, than the healthy volunteers, at 2.30±0.12. The cardiac PCr/ATP ratios correlated negatively with the fasting plasma free fatty acid concentrations. Although skeletal muscle energetics and pH were normal at rest, PCr loss and pH decrease were significantly faster during exercise in the patients with diabetes, who had lower exercise tolerance. After exercise, PCr recovery was slower in the patients with diabetes and correlated with tissue reoxygenation times. The exercise times correlated negatively with the deoxygenation rates and the hemoglobin (Hb)A1c levels and the reoxygenation times correlated positively with the HbA1c levels.
Conclusions— Type 2 diabetic patients with apparently normal cardiac function have impaired myocardial and skeletal muscle energy metabolism related to changes in circulating metabolic substrates.
Received January 24, 2003; revision received March 20, 2003; accepted March 24, 2003.
Cardiovascular disease is the leading cause of death in patients with type 2 diabetes,1 who have decreased survival after myocardial infarction and increased congestive heart failure and silent ischemia compared with nondiabetic control subjects.1,2 Type 2 diabetes mellitus is a chronic metabolic disorder characterized by insulin resistance, hyperglycemia, hyperinsulinemia, and elevated plasma free fatty acids, with poor glycemic control associated with an increased risk of heart failure.2,3 Therapeutic interventions that normalize glucose and lipid metabolism reduce the incidence of cardiovascular disease in patients with diabetes,4 with metabolic control of diabetes being the most important predictor of cardiovascular morbidity and mortality.1
In the normal adult heart, free fatty acids, glucose, and lactate are metabolized for ATP production in the mitochondria. However, in the diabetic heart, glucose and lactate oxidation are decreased5,6 and fatty acid oxidation is increased,7 increasing the oxygen requirement per ATP molecule produced.2,4,7,8 Positron emission tomographic studies have shown that patients with type 2 diabetes have decreased resting myocardial blood flow rates9 and decreased fluorodeoxyglucose uptake rates,10 yet little is known of cardiac high-energy phosphate metabolism in these patients. Similarly, skeletal muscle blood flow11 and glucose transmembrane transport and oxidation2 are decreased in diabetes. Patients with type 2 diabetes have limited exercise tolerance,2,12,13 which has been associated with decreased glycemic control12 and microvascular disease.13 It is not known whether decreased exercise tolerance in patients with type 2 diabetes is associated with abnormal skeletal muscle energy metabolism. Consequently, the purpose of this work was to determine whether heart and skeletal muscle energy metabolism was normal and whether any abnormalities were related to circulating glucose and free fatty acid concentrations in patients with type 2 diabetes.
Subjects and Protocol
Patients with type 2 diabetes (n=21) between 18 and 75 years of age with no evidence of cardiovascular disease or ECG-detectable evidence of ischemia were included in this study. Five patients were diet-controlled only, 6 patients each were treated with either a sulfonylurea drug or metformin, and 4 patients were treated with metformin and a sulfonylurea. Patients receiving insulin therapy were excluded. Patients were matched for age, sex, and body mass index with healthy control subjects (n=15).
All procedures were conducted on the same day, at the same time of day, for each subject. Subjects fasted overnight for 12 hours before blood sampling and echocardiography. After a small breakfast, the cardiac (rest) and skeletal muscle (exercise) magnetic resonance spectroscopy (MRS) protocols were performed and the subjects had lunch. For the near-infrared spectrophotometry (NIRS) measurements of muscle oxygenation, the MRS exercise protocol was repeated outside the magnet because the NIRS probe was magnetic. The MRS and NIRS exercise bouts were separated by 2 hours of ambulatory rest and 30 minutes of supine rest to ensure that all variables were stable. The local Oxford ethics committee approved all protocols, and subjects gave their informed consent.
Blood Tests and Echocardiography
Fasting blood was taken to determine glucose and glycosylated hemoglobin (HbA1c) levels, lipid profiles, and free fatty acid levels (Wako NEFA C enzyme assay, Wako Chemicals). Because our magnetic resonance (MR) scanner was capable of MR spectroscopy but not of precise left ventricular function analysis, we assessed cardiac function using a SONOS 5500 echocardiography machine (Hewlett Packard). Left ventricular dimensions and mass index were obtained with the use of M-mode echocardiography, and ejection fraction was calculated from left ventricular volumes, derived by using the modified Simpson’s rule. Diastolic function (early flow velocity [E] and late atrial contraction [A]; E:A) was evaluated by acquisition of a pulsed Doppler recording trace through the mitral valve, with the sample volume positioned just above the mitral valve leaflet tips.
Measurements of Cardiac Muscle Metabolism
Cardiac high-energy phosphate metabolism was measured using 31P MRS on a 2-Tesla whole-body magnet (Oxford Magnet Technology), which was interfaced to a Bruker Avance spectrometer (Bruker Medical GmbH). Cardiac 31P MRS was performed with the subject in the prone position, as previously described.14 Briefly, subjects were positioned with their heart at the isocenter of the magnet, which was confirmed by using standard multislice spin-echo proton (1H) images (Figure 1) acquired with a double-rectangular surface coil placed around the chest (relaxation time TR=heart rate, echo time TE=25 ms, slice =10 mm, 15 mm spacing). Once the position was verified, the coil was exchanged for a circular proton surface coil (diameter 15 cm), and shimming was performed to optimize the magnetic field homogeneity over the heart. Finally, a 31P surface coil (diameter 8 cm) was used to acquire cardiac spectra, using a slice-selective, 1-dimensional chemical shift imaging (1D-CSI) sequence, including spatial presaturation of lateral muscle (Figure 1). An 8-cm-thick transverse slice was then excited, followed by 1-dimensional phase encoding into the chest to subdivide signals into 64 coronal layers, each 1 cm thick (TR=heart rate, 16 averages). All 1H and 31P spectral acquisitions were cardiac-gated and saturation-corrected.14 Spectra were Fourier-transformed, and a 15-Hz line broadening was applied. Spectra were fitted using a purpose-designed interactive frequency domain-fitting program. After fitting, the ATP peak area was corrected for blood contamination according to the amplitude of the 2,3-diphosphoglycerate (2,3-DPG) peak, and the phosphocreatine (PCr)/ATP and phosphodiester (PDE)/ATP ratios were calculated and corrected for saturation as described earlier.14 The chemical shift differences between the α- and β-phosphate peaks of ATP were used as a measure of intracellular free magnesium concentrations.
Measurements of Skeletal Muscle Metabolism
31P MRS of the right gastrocnemius muscle was performed by using the 2-Tesla magnet (see above) with the subject in a supine position and a 6-cm-diameter surface coil under the muscle, as previously described.15 Spectra were acquired using a 2-second interpulse delay at rest (64 scans/spectrum) and during exercise and recovery (16 scans/spectrum).15 The muscle was exercised by plantar flexion against a standardized weight (10% lean body mass) at 0.5 Hz through a distance of 7 cm, with subsequent further increases of weight (2% of lean body mass every minute), and subjects were exercised until fatigued. Relative concentrations of inorganic phosphate, PCr, and ATP were obtained using a time-domain fitting routine (VARPRO, R. de Beer) and were corrected for partial saturation. Absolute concentrations were obtained assuming that the concentration of cytosolic ATP was 8.2 mmol/L intracellular water, and intracellular pH was calculated from the chemical shift of the Pi peak relative to PCr (δPi; measured in parts per million, ppm), using the equation
The chemical shift differences between the α- and β-phosphate peaks of ATP were used as a measure of intracellular free magnesium concentrations. Free cytosolic [ADP] was calculated from pH and [PCr] using a creatine kinase equilibrium constant16 (Kck) of 1.66×109/mol and assuming a normal total creatine content of 42.5 mmol/L, using the equation
At the end of exercise, because glycogenolysis had stopped and PCr resynthesis was purely oxidative, analysis of PCr recovery provided information about mitochondrial function. Recovery half-times for PCr and ADP and initial rates of PCr recovery were calculated as previously described.15
Measurements of Skeletal Muscle Oxygenation
Muscle oxygen saturation (Smo2) was measured using dual-wavelength NIRS (INVOS 4100 Oximeter, Somanetics), with the light emitter and two sensors placed over the medial head of the right gastrocnemius muscle.17 Smo2 was determined using the ratio of absorbance at the wavelengths of 733 nm and 809 nm, which estimated deoxygenated and the sum of deoxygenated and oxygenated hemoglobin, respectively. Smo2 was measured in deep tissue, predominantly at a depth of 2 cm, this being dependent on differentiating between absorption at the interoptode distances of 3 and 4 cm. As determined by such spatial resolution, the Smo2 was little, if at all, influenced by cutaneous and subcutaneous blood flow.17 In muscle, ≈75% of blood is in venules or veins, and the INVOS 4100 spectrophotometer has been calibrated against a tissue oxygen saturation in arterial (25% of the signal) and internal jugular vein (75% of the signal) blood. With spatially resolved dual-wavelength NIRS of skeletal muscle, 100% saturation refers to total oxygenation of hemoglobin and myoglobin, as myoglobin attenuates near-infrared light with an absorption spectrum comparable with that of hemoglobin. The muscle NIRS measurements were made in 12 control subjects and in 14 patients with type 2 diabetes.
Data analysis comparing patients with type 2 diabetes and control subjects was performed using the Student’s t test; correlations between data sets were determined using the Pearson correlation coefficient. Data are presented as mean±SEM. Statistical significance was taken at P<0.05.
Patient Characteristics and Echocardiography Results
There were no significant differences in sex, age, or body mass index between the patients with type 2 diabetes and the control subjects (Table 1). Mean duration of type 2 diabetes was 3.3±0.6 years from the time of diagnosis. Systolic and diastolic blood pressures and heart rates were similar in the two groups. Echocardiography showed normal left ventricular systolic and diastolic function in patients with no abnormalities in left ventricular chamber thickness or diameter or any other parameter (Table 1). The patients with diabetes had no history of cardiovascular disease and no clinical signs of impaired cardiac or skeletal muscle blood flow.
Fasting blood HbA1c and glucose levels were 1.5-fold and 1.9-fold higher, respectively, in patients with type 2 diabetes than in control subjects (Table 1). Plasma levels of free fatty acids were 1.4-fold higher in diabetic patients, as were lactate levels. Total cholesterol, triglycerides, and HDL cholesterol were normal in the patients with diabetes.
Cardiac High-Energy Phosphate Metabolism
Figure 2 shows typical examples of cardiac 31P MR spectra from a normal subject (PCr/ATP=2.35) and a patient with type 2 diabetes (PCr/ATP=1.35). The mean cardiac PCr/ATP ratio was 2.30±0.12 in control subjects but was decreased by 35%, to 1.50±0.11 (P<0.001), in patients with diabetes. The PCr/ATP ratios correlated negatively with the plasma free fatty acid concentrations in all subjects (r2=0.32; P<0.01; Figure 3) and positively with fasting plasma glucose concentrations in the diabetic patients (r2=0.55; P<0.05; Figure 3), but there were no correlations with plasma lactate or HbA1c levels. The PDE/ATP ratios were the same in the control subjects (0.51±0.06) and the diabetic patients (0.51±0.12), as were the chemical shift differences between the α- and β-phosphate peaks of ATP, being 8.3±0.4 and 8.5±0.1 ppm for control and diabetic patients, respectively.
Skeletal Muscle High-Energy Phosphate Metabolism
We found that the average exercise times for the patients with diabetes were 32% shorter, at 7 minutes, compared with the control subjects, at 11 minutes (Table 2 and Figure 4). Figure 5 shows examples of skeletal muscle spectra before and at the end of the standardized exercise protocol in a patient with type 2 diabetes and at the equivalent time (5.1 minutes) of exercise in a control subject. Under resting conditions, skeletal muscle pH and PCr, free ADP, and inorganic phosphate concentrations were the same in control subjects and patients with type 2 diabetes (Table 2). During exercise, PCr hydrolysis was 2-fold faster and the pH decrease was 3-fold faster in the patients with diabetes compared with the control subjects, but the free ADP production rates were not significantly different (Table 2). In all subjects, fatigue occurred when PCr depletion was ≈50% (50±4% in control subjects versus 51±4% in diabetics) and at the same pH and free ADP concentrations (Table 2). The free magnesium concentrations remained unaltered during exercise in all subjects (Table 2). After exercise, the initial rate of PCr recovery was 25% slower and the PCr recovery half-times were 1.6-fold longer in patients with type 2 diabetes than in control subjects, but the free ADP recovery half-times were the same (Table 2).
The exercise times correlated negatively with the HbA1c levels (r2=0.32; P<0.01; Figure 6) and the plasma glucose levels (r2=0.23; P<0.01; correlation not shown), but there were no correlations with the plasma free fatty acid or lactate levels. The rates of PCr hydrolysis and pH decrease during exercise did not correlate with any of the fasting metabolite concentrations. However, the PCr recovery half-times correlated positively with the HbA1c levels (r2=0.40; P<0.001; correlation not shown) and the plasma glucose concentrations (r2=0.16; P<0.05; correlation not shown) for all subjects, but there were no correlations with the plasma free fatty acid or lactate concentrations.
Skeletal Muscle Oxygenation
At rest, gastrocnemius muscle oxygen saturation was stable and the same for both groups, 68% in control subjects and 71% in diabetics, and all subjects stopped exercising after an 11% decrease in tissue oxygenation measured with NIRS (Table 2). The first diabetic patient stopped exercising after 3 minutes (Figure 4), therefore, during the first 3 minutes of exercise, the rate of deoxygenation was 3.1-fold faster in the type 2 diabetic patients than in the control subjects (Table 2) and correlated with exercise time (r2=0.29, P<0.01, Figure 6). Similarly, the reoxygenation times during recovery after exercise were 2.5 times longer in patients with diabetes compared with control subjects (Table 2), correlating with the HbA1c levels (r2=0.35; P<0.01; Figure 6) and with PCr recovery half-times (r2=0.25; P<0.01; Figure 6) in all subjects but not with the plasma free fatty acid or lactate levels.
In this study, we have shown that high-energy phosphate metabolism was significantly impaired in cardiac and skeletal muscle in patients with type 2 diabetes who had apparently normal cardiac morphology and function. The PCr/ATP ratios correlated negatively with the circulating free fatty acids in all subjects and positively with the plasma glucose in the patients with diabetes. Furthermore, we found faster skeletal muscle PCr loss, pH decline, and deoxygenation during exercise in the patients with diabetes and slower PCr recovery after exercise, with the PCr recovery half-times correlating with the reoxygenation times for all subjects.
Hyperinsulinemia, hyperglycemia, and increased lipid and lipoprotein abnormalities associated with type 2 diabetes may negatively influence myocardial performance,4 but, in the early stages of diabetes mellitus, systolic function is often preserved despite changes in cardiac substrate metabolism.5,6 It is unknown whether substrate changes in diabetes mellitus2,4–7 alter myocardial high-energy phosphate metabolism. 31P MRS is the only noninvasive tool for measurement of high-energy phosphate metabolism in the human heart, although limited to measurement of the PCr/ATP ratio in routine patient studies. Despite the limited information obtainable from human heart, compared with 31P MRS of isolated heart6 and skeletal muscle,15 a 31P MRS study of patients with dilated cardiomyopathy has shown a low cardiac PCr/ATP ratio to be a strong predictor of total and cardiovascular mortality, superior to the measurement of ejection fraction.18 Here, we found that the myocardial PCr/ATP ratio was 35% lower in type 2 diabetic patients, who had normal cardiac function, than in healthy control subjects. The PCr/ATP ratio correlated negatively with the plasma free fatty acid concentrations in all subjects because free fatty acid concentrations are not under tight metabolic control (Figure 3). Increased fatty acid availability results in increased free fatty acid uptake and oxidation in the mitochondria2,7 and increased expression of mitochondrial uncoupling proteins,19 both of which decrease the amount of ATP produced per molecule of oxygen consumed in the mitochondrial electron transport chain.2,19 Therefore, the diabetic heart has an increased requirement for oxygen.8
The hyperglycemia that occurs with diabetes is known to compensate for the impaired capacity for myocardial glucose transport.7 Glucose uptake is important for glycolytic ATP production during ischemia, low glucose uptake increasing ischemic injury in the heart.6 Patients with type 2 diabetes have decreased fluorodeoxyglucose uptake rates,10 decreased resting myocardial blood flows,9 and an increased incidence of silent ischemia.1 In our study, the lower cardiac PCr/ATP ratios in the patients who had lower plasma glucose concentrations suggested that decreased glucose availability may have limited glucose uptake. Although plasma lactate levels were 40% higher in the diabetic patients, and lactate is a metabolic substrate for the heart, the lack of a correlation between lactate levels and the cardiac PCr/ATP ratio was possibly because lactate oxidation is inhibited more than glucose oxidation in the diabetic heart.20
Skeletal Muscle Metabolism
Although cardiac high-energy phosphate metabolism was abnormal in the patients with diabetes, we found that skeletal muscle energetics, pH, and oxygenation were normal at rest. All subjects fatigued after the same tissue deoxygenation and with the same loss of PCr, increase in free ADP, and at the same acidic pH. This suggests that substrate availability or metabolism and glycogen levels were not limiting the skeletal muscle energetic changes. Additionally, we found faster loss of PCr and decrease in pH during exercise with slower PCr recovery after exercise in the patients with diabetes. The PCr recovery half-times correlated with the plasma HbA1c and glucose but not with fatty acid or lactate concentrations. However, deoxygenation was faster during exercise in the patients with diabetes, and reoxygenation was slower after exercise and correlated with the PCr recovery half-times, suggesting that tissue oxygen availability was limiting ATP production. Elevated levels of HbA1c have been associated with microvascular complications3,21,22 and reduced exercise capacity.12 In our study, the exercise times and the reoxygenation times correlated with the HbA1c levels, indicating that abnormal skeletal muscle oxygenation in the patients with diabetes may have been related to microvascular disease. In diabetic patients with intermittent claudication, skeletal muscle reoxygenation took ≈4 times longer than in normal subjects and provided a more sensitive measure of lower leg claudication than ankle pressure measurements.23 Consequently, microvascular disease may explain most if not all of the abnormalities in skeletal muscle high-energy phosphate metabolism that we observed in patients with diabetes.
In summary, we have shown significantly altered cardiac high-energy phosphate metabolism despite apparently normal cardiac morphology and function in patients with type 2 diabetes, which correlated with circulating free fatty acid and glucose concentrations. In contrast, we found that skeletal muscle energetics and oxygenation were normal at rest, but deoxygenation and loss of PCr were faster during exercise, and reoxygenation and PCr recovery were slower after exercise. Thus, in the patients with diabetes, cardiac and skeletal muscle differed, in that energetics and the availability of metabolic substrate appeared to be closely linked in the heart, whereas energetics were limited by the availability of oxygen in exercising skeletal muscle. These findings suggest that alterations in cardiac and skeletal muscle energetics occur early in the pathophysiology of type 2 diabetes and are associated with alterations in metabolic substrates.
The British Heart Foundation and the Medical Research Council supported this work.
Taegtmeyer H, McNulty P, Young ME. Adaptation and maladaptation of the heart in diabetes, I: general concepts. Circulation. 2002; 105: 1727–1733.
Iribarren C, Karter AJ, Go AS, et al. Glycemic control and heart failure among adult patients with diabetes. Circulation. 2001; 103: 2668–2673.
Chatham JC, Seymour A-ML. Cardiac carbohydrate metabolism in Zucker diabetic fatty rats. Cardiovasc Res. 2002; 55: 104–112.
Sidell RJ, Cole MA, Draper NJ, et al. Thiazolidinedione treatment normalizes insulin resistance and ischemic injury in the Zucker fatty rat heart. Diabetes. 2002; 51: 1110–1117.
Stanley WC, Lopaschuk GD, McCormack JG. Regulation of energy substrate metabolism in the diabetic heart. Cardiovasc Res. 1997; 34: 25–33.
Estacio RO, Regensteiner JG, Wolfel EE, et al. The association between diabetic complications and exercise capacity in NIDDM patients. Diabetes Care. 1998; 21: 291–295.
Lodi R, Kemp GJ, Muntoni F, et al. Reduced cytosolic acidification during exercise suggests defective glycolytic activity in skeletal muscle of patients with Becker muscular dystrophy. An in vivo 31P magnetic resonance spectroscopy study. Brain. 1999; 122: 121–130.
Veech RL, Lawson JWR, Cornell NW, et al. Cytosolic phosphorylation potential. J Biol Chem. 1979; 254: 6538–6547.
Neubauer S, Horn M, Cramer M, et al. Myocardial phosphocreatine-to-ATP ratio is a predictor of mortality in patients with dilated cardiomyopathy. Circulation. 1997; 96: 2190–2196.
Boehm EA, Jones BE, Radda GK, et al. Increased uncoupling proteins and decreased efficiency in the palmitate-perfused hyperthyroid rat heart. Am J Physiol. 2001; 280: H977–H983.
Chatham JC, Gao ZP, Bonen A, et al. Preferential inhibition of lactate oxidation relative to glucose oxidation in the rat heart following diabetes. Cardiovasc Res. 1999; 43: 96–106.
Libby P, Plutzky J. Diabetic macrovascular disease: the glucose paradox? Circulation. 2002; 106: 2760–2763.