Abstract
Myocardial metabolic remodeling is the process in which the heart loses its ability to utilize different substrates, becoming dependent primarily on the metabolism of a single substrate such as glucose or fatty acids for energy production. Myocardial metabolic remodeling is central to the pathogenesis of a variety of cardiac disease processes such as left ventricular hypertrophy, myocardial ischemia, and diabetic cardiomyopathy. As a consequence, there is a growing demand for accurate noninvasive imaging approaches of various aspects of myocardial substrate metabolism that can be performed in both humans and small-animal models of disease, facilitating the crosstalk between the bedside and the bench and leading to improved patient management paradigms. SPECT, PET, and MR spectroscopy are the most commonly used imaging techniques. Discussed in this review are the strengths and weaknesses of these various imaging methods and how they are furthering our understanding of the role of myocardial remodeling in cardiovascular disease. In addition, the role of ultrasound to detect the inflammatory response to myocardial ischemia will be discussed.
- cardiology (basic/technical)
- molecular imaging
- PET
- myocardial imaging
- myocardial metabolic remodeling
- myocardial repair
As expertly highlighted in the previous review by Taegtmeyer (1), flexibility in myocardial substrate metabolism is fundamental to cardiac health. The ability of the heart to modify the type of substrate it metabolizes is critical to maintain a proper balance between myocardial energy production and function in response to various stimuli such as alterations in the plasma substrate or hormonal environment, myocardial blood flow, and cardiac work. The loss in flexibility leads to an overdependence on the metabolism of an individual category of substrates, such as fatty acids or carbohydrates, that can initiate a cascade of events detrimental to myocellular health, including impaired myocardial energetics, stimulation of cellular inflammation and oxidative stress, and enhancement of both cell growth and cell death pathways, resulting in reduced myocardial systolic and diastolic function. Indeed, this metabolic remodeling process is central to the pathogenesis of a variety of cardiac disease processes such as left ventricular hypertrophy, myocardial ischemia, and diabetic cardiomyopathy. However, important unresolved questions remain, including, what are the key determinants of these metabolic perturbations in relation to specific diseases, do they alter prognosis, and do they represent robust targets for novel therapeutics? As a consequence, the demand is growing for accurate noninvasive imaging approaches of various aspects of myocardial substrate metabolism that can be performed in both humans and small-animal models of disease, facilitating the crosstalk between the bedside and the bench and leading to improved patient management paradigms. This article discusses these various imaging methods and how they are furthering our understanding of the role of myocardial remodeling in cardiovascular disease. In addition, the role of ultrasound to detect the inflammatory response to myocardial ischemia is discussed.
METHODS TO IMAGE MYOCARDIAL METABOLISM
Several methods are available to image myocardial metabolism noninvasively; the most common are SPECT, PET, and MR spectroscopy (MRS).
SPECT
An advantage of SPECT is the inherent high sensitivity of the radionuclide method to measure metabolic processes. Moreover, the technology is widely available in both the clinical and research settings. Theoretically, assessing more than one metabolic process simultaneously is possible if the heart is imaged after the administration of radiopharmaceuticals labeled with radionuclides having different primary photon energies. Finally, small-animal SPECT and SPECT/CT systems are rapidly advancing, facilitating the performance of myocardial metabolic studies in rodent models of cardiac disease.
Several fatty acid imaging tracers have been developed for clinical imaging with SPECT. These include radioiodinated straight long-chain fatty acids, including 15-(p-iodophenyl) pentadecanoic acid and branched fatty acids, such as 123I-β-methyl-p-iodophenyl-pentadecanoic acid (BMIPP). Straight long-chain fatty acids enter the mitochondria and are metabolized by β-oxidation immediately, whereas BMIPP is not initially metabolized via β-oxidation because the methyl substitution precludes the formation of the ketoacyl coenzyme A intermediate. The prolonged retention of BMIPP in the cardiomyocyte is suitable for a longer acquisition time that is necessary for SPECT. Similar to 18F, 123I is cyclotron-produced but with a longer half-life of 13 h. The longer half-life of 123I facilitates centralized distribution of the radiotracer, as with a conventional radiopharmaceutical. No specific SPECT radiotracers are currently available to measure myocardial glucose metabolism. However, when combined with the appropriate detection scheme or collimator design, myocardial glucose metabolism can be measured with SPECT and 18F-FDG (2). The major disadvantage of SPECT is its inability to quantify cellular metabolic processes of interest. This limitation is primarily caused by the technical limitations of SPECT—relatively poor temporal and spatial resolution—and the limited metabolic information provided by the kinetics of the SPECT radiotracers.
PET
Noninvasive measurement of myocardial metabolism is most commonly performed with PET because of its intrinsic quantitative capability and the use of radiopharmaceuticals labeled with the positron-emitting radionuclides. The PET detection scheme permits both the accurate quantification of activity in the field of view and its temporal distribution. These 2 pieces of information are critical to quantifying metabolic processes of interest. The positron-emitting radionuclides of the biologically ubiquitous elements oxygen (15O), carbon (11C), and nitrogen (13N), as well as fluorine (18F) substituting for hydrogen, can be incorporated into a wide variety of metabolic radiotracers that participate in a variety of biochemical pathways without altering the biochemical properties of the substrate of interest (Table 1 in the preceding review by Taegtmeyer (1)). By analyzing different components of the tracer kinetic curves (e.g., uptake, fast washout, and slow washout) with appropriate and well-validated mathematic models, it is possible to quantify specific metabolic processes. Examples include the quantification of myocardial oxygen consumption, fatty acid uptake, oxidation and esterification, glucose uptake, glycolysis and oxidation, and lactate oxidation (3–5). As with SPECT, small-animal PET and PET/CT systems have been developed that permit measurements of myocardial metabolism in the rodent heart similar to those done in the human heart. Despite its strengths, PET has several limitations that hinder its broader applicability to measure myocardial metabolism. These include the high cost of the PET systems, the need for an on-site cyclotron and expertise in radiopharmaceutical synthesis (except for 18F-FDG), and the relative complexity of the image quantification schemes.
MRS
MRS of the heart is a powerful research tool. Its use offers great advantages for metabolic evaluation of whole organs by providing unique, chemically specific information on true metabolic flux rates through pathways and specific enzymes, content and turnover of key metabolite pools, and transport kinetics across membranes. MRS of the whole heart has proven valuable for the evaluation of the energetic state and energy substrate use by the heart, such as carbohydrate versus fatty acid use (6–8). The dynamic processes described previously, uniquely accessible by MRS, provide unique mechanistic insight linking metabolic derangements to the contractile dysfunctions that are associated with various cardiac disorders, such as myocardial ischemia, reperfusion, and hypertrophy (9,10). Moreover, MRS enables detection of endogenous, nuclear MR (NMR)–sensitive metabolites (nuclei possessing net spin, I ≥ 1/2), such as compounds containing 31P and 23Na (6,7). Recent advances in the applications of localized detection of NMR signals, or MRS, from selected regions of interest in the in vivo heart of animal models and humans have also expanded the applicability of assessing cardiac metabolism for clinical evaluations and in vivo heart models of disease in animals.
However, MRS is not without limitations, the most notable being the relative insensitivity of the approach because of low signal-to-noise ratio (SNR) compared with radionuclide approaches for metabolic evaluation. The insensitivity of the technique requires the administration of exogenous metabolic precursors, labeled with NMR-detectable nuclei, in millimolar concentrations as opposed to radiotracer methods that typically use nanomolar and picomolar concentrations. In addition, the physics of NMR detection schemes for a deep structure such as the heart present unique challenges because of the potential for heating of tissue secondary to the application of the radiofrequency energy necessary to produce the MR signal (e.g., specific absorption rate). Nevertheless, the chemical specificity of MRS makes this approach highly complementary to radionuclide studies of spatial distribution, uptake, and clearance of radiolabeled metabolic precursors. A fundamental advantage of information from MRS of the heart is that the signals of interest emanate from intracellular metabolites, which can even be distinguished, in some circumstances, from extracellular pools.
APPLICATION OF SPECT
SPECT of Glucose Metabolism
The perfusion–metabolism mismatch pattern on 18F-FDG PET is the reference standard for detecting reversible left ventricular dysfunction in the clinical setting. 18F-FDG PET has proven benefit for the recovery of function, for improvement in detecting heart failure symptoms, and for predicting survival after revascularization (11–14). SPECT metabolic imaging with either 201Tl or 99mTc tracers is also reliable in detecting myocardial viability, particularly in patients with mild to moderately impaired left ventricular dysfunction (left ventricular ejection fraction of 25%−50%) (15–17).
Advances in high-energy collimation SPECT cameras have made imaging of positron emitting tracers, such as 18F, possible. Moreover, 18F-FDG metabolic imaging using SPECT equipped with a high-energy collimator simultaneously with a SPECT perfusion agent has the additional advantage of acquiring myocardial perfusion data. Studies have shown that when SPECT flow tracers such as 99mTc-sestamibi or 201Tl are used in combination with 18F-FDG SPECT, the detection of myocardial viability is similar to that with PET (18–20). However, the poorer spatial resolution of SPECT, its lower sensitivity, and the lack of attenuation correction may cause some discordance between images of the same tracer measured with SPECT and PET (17,21). Given these technical limitations of 18F-FDG SPECT, in addition to the limited number of publications on a small number of subjects, this technique has not received approval from the Food and Drug Administration or the Centers for Medicare and Medicaid Services for imaging the heart in the clinical setting.
Beyond its proven value for myocardial viability, the metabolic signature of reduced fatty acid metabolism and increased glucose use may serve as a sensitive marker of myocardial ischemia (22,23). Indeed, this metabolic signature may persist well after the resolution of the inciting ischemic event; this persistence has been termed ischemic memory (24,25). Applying a dual-isotope sestamibi and 18F-FDG simultaneous injection and acquisition protocol, this metabolic switch from fatty acid to glucose was recently shown to occur promptly when myocardial ischemia is induced during exercise and to persist for up to 24 h despite normal perfusion under resting conditions (Fig. 1) (2,26). Similarly, decreases in regional fatty acid metabolism and delayed recovery of fatty acid metabolism, long after regional blood flow has returned to baseline, have been shown when using BMIPP. Taken in sum, these studies show that ischemic memory occurs in humans and can be imaged with conventional radionuclide approaches, raising the potential for using these techniques in managing patients with ischemic heart disease.
Simultaneous myocardial perfusion and metabolism imaging after dual intravenous injection of 99mTc sestamibi and 18F-FDG at peak exercise (Ex). Dual-isotope simultaneous acquisition was performed 40–60 min after exercise study was completed. Rest 99mTc-sestamibi imaging (R) was performed separately. In this patient with angina and no prior myocardial infarction, note evidence of extensive reversible perfusion defect in anterior, septal, and apical regions. Coronary angiogram showed 90% stenosis of left anterior descending and 60% of left circumflex coronary arteries. Corresponding 18F-FDG image shows intense uptake in regions with reversible sestamibi defects reflecting metabolic correlate of exercise-induced myocardial ischemia. (Adapted from (2).)
Although the technical limitations of a modified SPECT camera to detect 18F signals over PET are well recognized, other challenges also exist for acquiring 18F-FDG studies with either SPECT or PET. For example, studies to detect myocardial ischemia are typically performed under fasting conditions to maximize the differences in tracer uptake between ischemic myocardium (increased uptake) and nonischemic tissue (reduced uptake). However, under fasting conditions, there is heterogeneous distribution of 18F-FDG in the myocardium even in healthy subjects, which creates another level of complexity in image interpretation (27). Thus, when it comes to the cardiac application of 18F-FDG to detect myocardial ischemia under fasting conditions, it may be difficult to resolve and conclusively localize the hot-spot 18F-FDG signal in the ischemic myocardium when the rest of the myocardial regions have no discernable signal to serve as a reference standard. Such studies will require either simultaneous acquisition of myocardial perfusion or hybrid SPECT/CT with anatomic coregistration of the heart.
SPECT of Fatty Acid Metabolism
As mentioned previously, measurement of myocardial fatty acid metabolism using BMIPP is the most commonly used metabolic application of cardiac SPECT. Similar to 18F-FDG, BMIPP is taken up by the myocyte but is not further metabolized after the first step in this metabolic pathway. After a transient ischemic event, defects in fatty acid metabolism, as assessed by BMIPP distribution, are present in patients undergoing exercise myocardial perfusion SPECT studies (Fig. 2), as well as in patients presenting with acute coronary syndromes (22,23). This long-duration disturbance in metabolism is similar to that seen after 18F-FDG administration.
SPECT image shows delayed recovery of regional fatty acid metabolism after transient exercise-induced ischemia, termed ischemic memory. Representative stress (left) and rest reinjection (second from left) short-axis thallium tomograms demonstrate reversible inferior defect consistent with exercise-induced myocardial ischemia. After injection of BMIPP, labeled tomogram (second from right) acquired at rest 22 h after exercise-induced ischemia shows persistent metabolic abnormality in inferior region despite complete recovery of regional perfusion at rest, as evidenced by thallium reinjection image. Tomogram on far right shows retention of BMIPP in heart of healthy adult for comparison. (Adapted from (25).)
Uptake of BMIPP from the plasma into myocardial cells occurs via the CD36 transporter protein present on the sarcolemmal membrane. Once BMIPP is taken up by the myocyte, it undergoes adenosine triphosphate (ATP)–dependent thioesterification but does not undergo significant mitochondrial β-oxidation. Retention of BMIPP in the intracellular lipid pool of the myocardium most likely reflects activation of fatty acids by coenzyme A, and indirectly, of cellular ATP production resulting from fatty acid metabolism. Thus, in the setting of myocardial ischemia, reduction in ATP production secondary to diminished fatty acid metabolism is mirrored by decreased myocardial BMIPP uptake. Important clinical applications of the latter are the early assessment of chest pain among patients presenting to the emergency department with acute coronary syndromes and the assessment of myocardium at risk among patients presenting with acute myocardial infarction and early reperfusion with either percutaneous intervention or thrombolysis. Results from a recent multicenter clinical trial suggest that the combination of BMIPP SPECT with initial clinical information results in improved sensitivity for identifying patients with acute coronary syndrome compared with the sensitivity of the initial clinical diagnosis alone, while at the same time maintaining specificity (28).
The deleterious effect of altered myocardial metabolism was recently also demonstrated in patients with end-stage renal disease (29,30). A shift in myocardial metabolism from fatty acid to glucose may contribute to the cardiomyopathic process observed in late-stage kidney disease (30). The cardiomyopathy typical of chronic kidney disease and the associated uremia are thought to lead to a myocytes–capillary mismatch, with a diminished vascular supply relative to the number and volume of functioning myocytes (31). The oxygen-poor milieu leads to diffuse myocardial ischemia with an anticipated decline in aerobic myocardial fatty acid use. Such altered cardiac metabolism, indicating silent myocardial ischemia, was recently shown to be highly prevalent among asymptomatic dialysis patients with no history of myocardial infarction, and was able to identify the subgroup of patients who were at high risk of cardiac death (30).
Future Directions and Challenges
SPECT of myocardial metabolism has focused primarily on the concept of ischemic memory. Delayed recovery of fatty acid metabolism detected with SPECT up to 30 h after the resolution of transient myocardial ischemia provides the potential for diagnosing antecedent myocardial ischemia both in the chronic and in the acute setting. As a consequence, targeting intracellular fatty acid metabolism with SPECT may expand our ability to diagnose and treat subclinical myocardial ischemia or progressive cardiomyopathy (e.g., chronic kidney disease) that often remains elusive with traditional imaging approaches. However, to achieve these goals, several hurdles must be overcome. From a technical perspective, imaging protocols need to be optimized with respect to timing of imaging after chest pain, acquisition schemes that account for 123I need to be standardized, and questions regarding the need to standardize the substrate environment must be resolved. From a clinical point of view, the diagnostic and prognostic accuracy and the added clinical information provided by these metabolic signatures need to be determined.
APPLICATION OF PET
18F-FDG PET
As mentioned previously, myocardial viability imaging using 18F-FDG has emerged as a highly valuable clinical tool. As molecular imaging evolves, seeking tools that could be translated to clinical practice, it is worthwhile to examine 18F-FDG and what has led to its success as a probe—namely, more than 30 y of experience (32), straightforward production and imaging, wide availability, high accuracy, linkage to disease and outcomes, and impact on patient management (Fig. 3). 18F-FDG PET is the most sensitive viability method for predicting wall motion recovery (33) and identifies patients at increased risk for death if they do not undergo revascularization. The PET and Recovery Following Revascularization (PARR-2) trial showed a trend for benefit when 18F-FDG PET was used to assist management (34). In patients who adhered to recommendations, significant outcome benefits were attained, with the degree of hibernation predicting the likelihood of response to revascularization (35). A post hoc subgroup analysis demonstrated improved outcomes in an experienced center (36). As a consequence, regional networks that work to ensure quality and clinical expertise at different 18F-FDG PET sites are now being established (37).
Example of hibernating myocardium. Images show moderate perfusion defect involving apex and mid to distal anterior and septal walls (superior row) but preserved 18F-FDG uptake, which allows delineation of hibernating myocardium in distribution of territory of left anterior descending artery (inferior row). SA = short axis, HLA = horizontal long axis, VLA = vertical long axis.
Phenotype Characterization and Translational Imaging
Quantitative measurements of myocardial metabolism are now possible in the rodent heart using small-animal PET. For example, measurements of myocardial glucose uptake using small-animal PET have been shown to relate to the protein expression of the glucose transporter-4 protein levels (38). Moreover, small-animal PET has helped clarify the mechanisms responsible for the metabolic alterations that occur in various diseases. For example, in mice with cardiac-restricted overexpression of the nuclear receptor, peroxisome proliferator-activated receptor α-(PPARα), a key regulator of myocardial fatty acid uptake, oxidation, and storage, demonstrates a metabolic phenotype that is similar to that in diabetic hearts (39). Small-animal PET studies with 11C-palmitate and 18F-FDG in mice demonstrate an increase in fatty acid uptake and oxidation and an abnormal suppression of glucose uptake. In contrast, in mice with cardiac-restricted overexpression of PPARβ/δ, small-animal PET measurements showed an increase in glucose uptake and reduced fatty acid uptake and oxidation (Fig. 4) (40). Taken in sum, these observations show that PPARα and PPARβ/δ drive different metabolic regulatory programs in the heart and that imaging can help characterize genetic manipulations in the mouse heart. These observations highlight the importance of small-animal PET to evaluate phenotype and transgene expression. However, small-animal PET has to overcome several challenges in imaging the rodent heart, including partial-volume effects, extraction of the input function for quantification, and other technical factors. Thus, reproducing in the rodent setting the success observed in humans can be a challenge.
Increased myocardial glucose use in myosin heavy chain (MHC)-PPARβ/δ mice. (A) On left, standardized uptake value time–activity curves for 11C-palmitate and 11C-glucose into female MHC-PPARβ/δ–high-expression (HE) and nontransgenic (NTG) hearts as determined by small-animal PET. On right, representative small-animal PET images at 20 s after tracer injection. Images are normalized to total amount of radioactivity injected and body weight. Relative amounts of tracer uptake are indicated by color scale. (B) Oxidation of palmitate and glucose was assessed in isolated working hearts of 12-wk-old male MHC-PPARβ/δ-HE and NTG control mice. Bars represent mean oxidation rates expressed as nanomoles of substrate oxidized per gram of dry mass per minute. (C) Glycogen levels were assessed in mouse hearts from male MHC-PPARα–low-expression and MHCPPARβ/δ-HE mice and NTG controls. Results are presented as glucose released from glycogen and normalized to tissue weight. *P < 0.05 vs. NTG. (Reprinted with permission of (40).)
To further translate metabolic imaging in the clinical field, the myocardial metabolic phenotype in experimental models of disease should be also applicable in humans. One such example is the PRKAG2 cardiac syndrome, which is characterized by arrhythmias and hypertrophy caused by a mutation in the PRKAG2 gene that codes for adenosine monophosphate–activated protein kinase (AMPK), a key metabolic regulator in the myocyte (41). The central role of AMPK in regulating glucose metabolism had led to the hypothesis that the pathologic basis of the PRKAG2 cardiac syndrome was caused by impaired glucose metabolism and excessive glycogen storage (42). Indeed, patients and the corresponding transgenic mouse model are characterized by increased glycogen storage. Preliminary data in the transgenic model show reduced 18F-FDG uptake that can be quantified (Fig. 5) (43). Recently, human studies demonstrated reduced 18F-FDG uptake (44). At first blush, the lower level of glucose uptake appears paradoxical because cardiac hypertrophy caused by excessive glycogen storage is a hallmark of this condition. This apparent paradox may be explained by the developing nature of the disease process and highlights the potential importance of serial imaging to better characterize the time course of the disease. In another example, altered fatty acid and oxidative metabolism were shown in a group of patients with mutation in the α-tropomyosin gene and associated left ventricular hypertrophy (45). Taken together, these data highlight the advances in small-animal PET technology to quantify metabolic parameters and exemplify the translational capabilities of metabolic PET.
On left, cardiac imaging in mouse model. On right, short-axis display of 18F-FDG myocardial uptake (60 min) in NTG and transgenic-mutant (TG-mut) of PRKAG2 gene. Note that 18F-FDG uptake in NTG is normal, whereas in transgenic-mutant it is significantly reduced. Below, graphs show arterial blood (red) and myocardial (blue) time–activity curves. In transgenic-mutant, myocardial curve shows lower rate of 18F-FDG uptake compared with that in NTG. (Courtesy of M.H. Gollob and S. Thorn.)
Future Directions and Challenges
Several promising developments augur well for the continued need for metabolic imaging with PET. Clinical research studies are exploring new roles for metabolic imaging in the diabetic heart, myocardial hypertrophy, and ischemic memory. The emergence of new drugs that target specific metabolic processes, such as fatty acid oxidation (e.g., PPARα and PPARβ/δ agonists) or insulin sensitivity (glucagon-like peptide-1), is creating a need to direct and monitor metabolic response, paving a path for personalized therapy directed by metabolic imaging.
Metabolic imaging works well as a clinical tool in the context of viability imaging, but metabolic alterations in nonischemic disorders are common and not well understood. Imaging tools are available to enhance metabolic phenotype characterization, and translation to humans is possible. However, comparative effectiveness research is needed to understand the clinical value of metabolic imaging and how it compares with other emerging methodologies.
APPLICATION OF MRS
Myocardial Energetics
31P NMR has long been used to assess high-energy phosphate content in the intact beating heart. Indeed, the earliest clinical studies with cardiac MRS provided evaluations of the myocardial energetic state from the relative contents of high-energy phosphates via 31P detection (46–48). Cardiac 31P MRS in patients provides information on the relative contents of phosphocreatine (PCr) and ATP (46–50). A widely used parameter for assessing the relative energy state of the heart has been the ratio of PCr to ATP content, as detected by 31P MRS, and this approach was instrumental in showing the now well-recognized impaired energetic state of the failing heart.
The diseased myocardium shows a generally reduced PCr-to-ATP ratio, but one limitation of the parameter for clinical use is the lack of specificity. Reduced PCr-to-ATP is generally evident in patients with either hypertrophic or dilated cardiomyopathies, allograft rejection after cardiac transplantation, or coronary artery disease with stenosis (48–51). Fortunately, important new advances in 31P MRS detection extend well beyond the original uncertainties of the pathophysiologic specificity of the PCr-to-ATP ratio of the heart. One such example is the use of MRS magnetic transfer techniques to measure the activity of the creatine kinase system, which provides a balance between energy production, transport, and use. Indeed, application of this approach in clinical subjects suggests better delineation of failing from nonfailing, hypertrophic hearts when compared with the PCr-to-ATP ratio (Figs. 6 and 7) (52).
(A) Axial MR image of patient with left ventricular hypertrophy and congestive heart failure in which region of localized 31P NMR spectra from chest and left ventricle are identified (rectangular outline). (B) 31P NMR spectra from chest muscle (bottom) and left ventricle (top) with control saturating radiofrequency irradiation (arrow). (C) 31P NMR spectra from chest muscle (bottom) and left ventricle (top) with selective, saturating radiofrequency irradiation at γ-phosphate resonance (arrow). Note decreased magnitude of PCr signal in panel C due to chemical exchange with saturated 31P nuclei of γ-phosphate of ATP. Decreased PCr signal depends on rate of ATP synthesis through creatine kinase reaction (52).
(A) Pseudo first-order rate constant (Kfor) for creatine kinase (CK) in hearts of healthy subjects (Normal), patients with left ventricular hypertrophy (LVH), and patients with LVH and congestive heart failure (LVH + CHF). (B) ATP flux through CK in each group. Note depressed flux rate in LVH + CHF. (Reprinted with permission of (52).) ns = not significant.
Although 31P MRS provides unique information on the energetic state of the myocardium, the underlying mechanisms of energy production pathways via the intermediary metabolism can be detected by exploiting stable isotope kinetics. Naturally abundant at only 1.1%, the nonemitting 13C is well suited for enrichment studies of metabolism and is commonly used for in vitro analysis of the fractional contributions of carbon-based substrates for oxidative production of ATP, as determined by 13C isotopomer distributions in the glutamate pool of tissue extracts (53,54). 13C detection of enrichment rates in metabolites provides the basis for understanding metabolic flux regulation and the metabolic support of contractile function through shifts in enzyme isoform expression in normal and diseased hearts. After the introduction of 13C-enriched precursors to the heart, such as glucose or fatty acids, the enrichment rates of targeted metabolites can be monitored sequentially for kinetic analysis either in isolated perfused hearts or in vivo (8,10,55). For example, NMR spectra can be obtained from the beating heart to monitor rates of palmitate oxidation or storage into triacylgyceride (Fig. 8). Recent applications for 13C MRS in experimental animal models enable measurement of metabolic enzyme activity caused by altered gene expression in the intact functioning heart (10,55).
(A) Sequential 13C NMR spectra (2 min each) of isolated rat heart oxidizing 13C-enriched palmitate. Relative rates of isotope enrichment of glutamate 2-, 3-, and 4-carbons (GLU C-2, GLU C-3, GLU C-4, respectively) provide measures of oxidative rates. Progressive enrichment of triacylglyceride (TAG) provides TAG turnover. (B) Selected 13C NMR spectra (2 min each) of isolated mouse heart, perfused with 13C palmitate, show progressive enrichment of TAG for measures of TAG turnover in transgenic models.
Dynamic-mode 13C MRS has provided a mechanistic understanding of enzyme isoform changes, and even mitochondrial transport function, in isolated organs and animal models of heart disease (10,55,56). The use of 13C-enriched long-chain fatty acids in isolated rat and mouse heart perfusion studies has enabled comparison of the rates of fatty acid oxidation by mitochondria and storage in the triacylglyceride pool, providing a comprehensive view of lipid dynamics in normal and diseased hearts (57–59). For example, 13C MRS of hypertrophied rat hearts, isolated and perfused with the long-chain fatty acid, palmitate, that is 13C-enriched at novel carbon-chain sites, has shown reduced rates of long-chain fatty acid oxidation and impaired turnover in the triacylglyceride pool (10,59). The reduced fat oxidation coincides with the apparent compensatory recruitment of an alternate mode of carbohydrate entry into oxidative pathways, via anaplerosis, through upregulation of malic enzyme expression in hypertrophied hearts (10,60). Pharmacologic reversal of this up-regulated anaplerotic activity in cardiac hypertrophy produced a surprising increase in contractility of the hypertrophied rat heart (60). These findings show how such basic evaluations of experimental models with MRS contribute to elucidating links between metabolic flux in the intact beating heart and contractile performance. These links hold obvious relevance for the development of potential therapeutic, metabolic protocols that combat contractile dysfunction in the diseased heart.
Just as important, whereas 13C MRS has enabled detection of increased anaplerosis in hypertrophied hearts, experimental evaluation of metabolism with radioactive 14C and release rates of 14CO2 from oxidation of 14C-labeled precursors are insensitive to the detection of anaplerotic mechanisms. 14CO2 release methods, when used alone, do not detect changes in CO2 fixation into the citric acid cycle intermediate pools that are associated with anaplerotic reactions, and consequently may lead to an overestimation of mismatches between the rates of glycolysis and glucose oxidation. Thus, despite the limitation of nontracer levels of exogenous materials, 13C MRS studies have already offered unique perspectives not previously accessible by other labeling methods.
Although the examples discussed previously demonstrate a unique role for 13C studies of the intact heart, the requirements for a second resonance frequency for proton decoupling of these 13C NMR spectra and the limited sensitivity of 13C MRS detection present significant physical limitations for noninvasive in vivo applications. Therefore, the recent use of hyperpolarization of 13C in specific compounds to provide 10,000- to 20,000-fold increases in SNR holds great potential. The use of dynamic nuclear polarization to hyperpolarize 13C has already enabled high SNR detection for animal studies on isolated hearts and rapid temporal resolution for in vivo rat heart studies (Fig. 9) (61–63). With dynamic nuclear polarization, the hyperpolarized state of the 13C nuclei lasts for approximately 1 min, which is sufficient time for injection and collection of in vivo MRS data in serial fashion, with acquisition intervals of sequential spectra on the order of seconds (Fig. 9) (62,63). Although current in vivo studies with hyperpolarized 13C rely on detection after bolus intravenous injections of 13C-enriched precursors at relatively large, millimolar concentrations, the increase in SNR holds the potential for the eventual development of studies using tracer levels that would not influence the metabolic pathways of interest. At present, chemistry of the dynamic nuclear polarization excitation limits the availability of relevant precursor molecules that enable practical durations of the nuclear hyperpolarization for delivery and detection by MRS. For this reason, 13C hyperpolarization studies in hearts have been limited to the fate of pyruvate. However, increased development in the chemistry of these hyperpolarized nuclei in other precursors offers the promise for a more comprehensive application of hyperpolarized 13C NMR and MRS for in vivo cardiac metabolism.
In vivo 13C NMR spectra from heart of anesthetized rat after bolus tail vein injection of 1 mL of 80 mM sodium [1-13C] pyruvate. (A) Spectrum displaying 13C-enriched metabolites of pyruvate. (B) Sequential spectra acquired every second for 1 min after injection. (C) Time course of signals from pyruvate (solid circle), lactate (square), alanine (triangle), and bicarbonate (X). a.u. = arbitrary units. (Reprinted with permission of (62).)
Proton (1H) MRS is more accessible to current in vivo and clinical studies. Localized 1H MRS, of both animal models and human subjects, offers unique evaluation of lipid infiltration of internal organs (64,65). In the heart, 1H detection provides quantitative analysis of the intramyocardial mobile lipid from methylene group protons (CH2) in acyl chains generally associated with triacylglyceride. Notably, such evaluation relies on 1H signal, which is localized from the left ventricular septum, to eliminate contaminating signal from the pericardial fat that is adjacent to the free wall. Increased intramyocardial lipid, as identified by 1H MRS, may be linked to lipotoxic effects that impair cardiomyocyte function and has been associated with the development of diabetic cardiomyopathy. For example, the presence of increased myocardial lipid detected by 1H MRS has been shown in diabetics and appears to be associated with a decline in diastolic function (64,66). Moreover, a prolonged caloric restriction in obese diabetic patients and its attendant salutary effect of glucoregulation decreases myocardial lipid and improves diastolic function (67). Thus, it appears that myocardial lipid levels may have a detrimental effect on diastolic function in diabetes and that lipid levels are responsive to changes in plasma fatty acid delivery. As interest mounts in monitoring lipid accumulation in the heart, the advantages of high- and ultrahigh-field magnet systems become increasingly apparent. The advantageous improvements in SNR with high field strength are apparent in the comparison of in vivo lipid signals from 1H MRS of in vivo hearts, at the clinical field of 1.5 T and from a research magnet at 14.1 T, as shown in Figures 10 and 11, respectively.
Localized 1H MRS signal from myocardial triacylglyceride in human. Left panel shows MR image of heart displaying localized volume in left ventricular (LV) septum for MRS (yellow rectangle). Right panel displays 1H NMR spectrum with inset of expanded region of triacylglyceride signal (in red circle). (Adapted with permission of (64).) RV = right ventricle; RA = right atrium; LA = left atrium.
Localized 1H MRS signals from myocardial triacylglyceride (TAG) of in vivo anesthetized mouse heart at 14 T with localized volume of left ventricular septum for MRS indicated at right in axial MR image of heart (yellow box). Top panel displays signals from triacylglyceride within a 1 × 1 × 1 mm voxel. Bottom panel displays enhanced signal from increased voxel size of 1 × 2 × 1 mm. (Courtesy of E. Douglas Lewandowski.)
Preliminary results suggest that MRS of myocardial lipid content can be combined with high-resolution cardiac tagging MRI at 14.1 T for a comprehensive study combining metabolic evaluation with transmural measures of compression and stretch, via 2-dimensional principle strains (E1 and E2) across the left ventricular wall (68). This work indicates that intramyocardial lipid infiltration increases myocardial stiffness, as evidenced by reduced endocardial compression and stretch in the left ventricular wall.
Future Directions and Challenges
The specific examples highlighted here show an array of cardiac applications for NMR spectroscopy and MRS. Although some schemes hold diagnostic value for direct clinical evaluation, others provide basic information to elucidate mechanisms of contractile dysfunction and contribute to the development of therapeutic strategies for heart disease. Although the technical challenges presented by some of the more sophisticated methods for kinetic analysis limit current applications to experimental heart models, the chemically specific nature of MRS provides a valuable correlate to clinical evaluations with nuclear modalities such as PET and SPECT. Combining metabolic evaluation by MRS with functional assessment by MRI is emerging as a powerful combination.
APPLICATION OF ULTRASOUND
Ultrasound, Targeted Contrast Agents, and Ischemic Memory
Although ultrasound has not been used to measure specific myocardial metabolic processes, it can be used to detect cellular signatures of antecedent myocardial ischemia. As mentioned previously, transient myocardial ischemia results in reduced fatty acid metabolism and increased glucose metabolism that persist after the resolution of the ischemia, leading to the concept of ischemic memory. These metabolic signatures can be detected with either SPECT or PET as a means to indicate an antecedent ischemic event. Another approach to imaging recent myocardial ischemia is to target the stereotypical inflammatory response to an ischemic insult. This is characterized by sequential microvascular endothelial overexpression of leukocyte adhesion molecules that mediate leukocyte slowing, rolling, capture, and firm adhesion to the endothelium (69–71). On ischemic insult, the leukocyte adhesion molecule P-selectin is rapidly mobilized to the endothelial surface from preformed cytoplasmic stores within minutes of ischemia and reperfusion, where it mediates transient tethering interactions that result in leukocytes rolling along the endothelium (71). Leukocyte rolling enables leukocyte interaction with adhesion molecules such as intercellular adhesion molecule-1 (ICAM1) or vascular adhesion molecule (VCAM), which mediate firm attachment (69). Unlike preformed cytoplasmic P-selectin, which mobilizes quickly to the surface of endothelium on ischemia and reperfusion, postischemic expression of ICAM and VCAM is transcriptionally dependent, requiring several hours to manifest after ischemic insult. Because adhesion molecule expression and leukocyte adhesion are endothelial events, an imaging probe that remains in the intravascular space, such as an ultrasound contrast agent, could be adapted to detect these hallmarks of inflammation.
Ultrasound contrast agents are gas-filled microspheres (microbubbles) used in ultrasound imaging for acoustically opacifying the blood pool. The microbubbles, ranging from 1 to 4 μm in diameter, are typically perfluorocarbon- or nitrogen-gas microspheres encapsulated by shells of varying composition, such as phospholipids, albumin, or biodegradable polymers (72). Because of their size and shell composition, the microbubbles do not interact with the endothelium and transit in unimpeded fashion through the microcirculation (73). In the presence of an ultrasound field tuned to the appropriate frequency and acoustic pressure, the microbubbles expand and contract (oscillate) in a nonlinear fashion, themselves becoming emitters of ultrasound that can be detected and displayed as a transient signal on a 2-dimensional ultrasound image (74). This acoustic property has been the basis for the use of microbubbles as red cell tracers in applications such as enhancement of the blood pool for endocardial border detection in technically difficult echocardiograms and myocardial perfusion imaging (75–78).
For ultrasound molecular imaging applications, targeting ligands are conjugated to the microbubble shell, conferring the ability of a microbubble to attach to a specific endothelial epitope, resulting in persistent ultrasound contrast enhancement and hence the capability for investigating the endothelial phenotype (79). The surface of microbubbles has been conjugated to a variety of targeting ligands, including monoclonal antibodies, peptides, and naturally occurring protein and carbohydrate ligands, for the in vivo acoustic detection of inflammation and angiogenesis in animal models of cardiac and noncardiac disease (Fig. 12A) (79–83). Alternatively, a nonligand-based approach for imaging inflammation has also been described, in which activated leukocytes adhering to the endothelium are rendered acoustically active by lipid microbubbles that attach to the leukocyte surface through a process that is thought to be complement-mediated (Fig. 12B) (84–86).
Schematic of vascular endothelium and approaches for ultrasound molecular imaging of inflammation during ischemia and reperfusion using microbubble attachment to endothelial cells. (A) Microbubbles bearing targeting ligand on surface can bind to specific endothelial target, such as leukocyte adhesion molecule. (B) Activated leukocytes may bind or phagocytose microbubbles and become acoustically active. Figure not drawn to scale.
Based on these considerations, it has been possible to acoustically detect antecedent ischemia using ultrasound contrast agents designed to bind to inflammatory endothelium. The rapidity with which P-selectin is expressed after ischemia and reperfusion makes it an attractive marker when the clinical dilemma is to determine whether myocardial ischemia has recently occurred. Furthermore, its location on the endothelial cell surface renders it accessible to an intravascular probe such as a microbubble.
The concept of P-selectin targeting by a microbubble was first proven in a murine model of renal ischemia and reperfusion. Monoclonal antibody to P-selectin was conjugated to a lipid microbubble and intravenously injected into mice undergoing transient renal artery occlusion followed by reperfusion (81). Persistent contrast enhancement of the postischemic kidney was noted on ultrasound imaging, which was not seen after the injection of nontargeted bubbles, or after the injection of targeted and nontargeted bubbles into normal control mice. These studies established the concept that P-selectin targeting with an ultrasound contrast agent is a feasible approach for detecting postischemic tissue during ultrasound imaging.
P-selectin targeting for the detection of antecedent myocardial ischemia has been recently described using microbubbles bearing the naturally occurring ligand for selectins. The major ligands for all 3 selectins are cell surface glycans possessing a specific sialyl Lewisx (sLex) epitope, meaning that a single moiety could theoretically be used to target both P- and E-selectin, which are expressed early and late, respectively, after reperfusion (87). Sialyl-Lewisx was conjugated to the surface of lipid microbubbles and initially tested for adhesive properties under direct visualization during intravital microscopy of rat cremaster microcirculation rendered inflammatory by intrascrotal tumor necrosis factor-α administration (82). Sialyl-Lewisx microbubbles adhered to activated endothelium, whereas control Lex-conjugated microbubbles minimally adhered, with neither bubble species adhering significantly to normal, noninflamed endothelium. A rat model of myocardial ischemia and reperfusion was then used to echocardiographically detect adhesion events (82). Both early (30 min) and late (up to 90 min) after reflow, video intensity in the postischemic bed after the intravenous injection of sLex microbubbles was significantly higher than that in the nonischemic bed. Furthermore, the region of persistent contrast enhancement colocalized with, and correlated in size to, the risk area (Fig. 13). There was also a trend toward similar findings after milder (10 min) ischemia. These data indicate that the selectin-targeted microbubbles are capable of recapitulating the presence, location, and spatial extent of previously ischemic myocardium.
Ultrasound ischemic memory imaging of myocardium using microbubbles targeted to bind to P-selectin via tetrasaccharide sialyl Lewisx in rat model of 15-min coronary occlusion followed by reperfusion. Short-axis nonlinear ultrasound images of heart are background-subtracted, and degree of contrast enhancement is color-coded. (A) Injection of nontargeted microbubbles during coronary occlusion shows risk area (region between arrows). (B) After reperfusion, nontargeted microbubble injection confirms restoration of myocardial perfusion. (C) Postmortem staining with triphenyl tetrazolium chloride (TTC) indicates no infarction. (D) Delayed imaging after injection of control microbubbles during reperfusion demonstrates no persistent contrast enhancement. (E) Delayed imaging after injection of P-selectin–targeted microbubbles during reperfusion shows persistent contrast enhancement in region that was previously ischemic (risk area, A). (Adapted with permission of (82).)
As mentioned previously, another approach to inflammatory imaging for the detection of prior ischemia is the use of microbubbles that are avid for activated leukocytes. Lipid microbubbles augmented with phosphatidylserine in the shell adhere to activated leukocytes (41) and remain acoustically active even after leukocyte phagocytosis (86). Phosphatidylserine microbubbles have been intravenously injected into canines with reperfused myocardial infarction, and ultrasound imaging demonstrated persistent contrast enhancement in the region of prior ischemia and infarction (Fig. 14) (88). Note that this approach appears feasible in a model in which there is sizeable infarction; it is unclear whether a leukocyte-avid microbubble would be effective in delineating antecedent ischemia in which there is minimal or only mild necrosis, the clinical scenarios in which the emergency department evaluation of chest pain is particularly problematic.
Inflammatory imaging of reperfused infarcted canine myocardium using phosphatidylserine-augmented lipid microbubbles that attach to activated leukocytes. Short-axis nonlinear ultrasound images are background-subtracted and color-coded. (A) After injection of leukocyte-avid microbubbles during reperfusion, note persistent contrast enhancement of previously ischemic area. (B) Confirmation of leukocyte accumulation in postischemic zone on autoradiography of isotope-labeled leukocytes. (C) TTC-stained myocardial specimen demonstrates nontransmural infarction (88).
Future Directions and Challenges
Ultrasonic identification of acute adhesion molecule expression or leukocyte activation on postischemic endothelium could enable not only the identification of recent myocardial ischemia but also the mapping of its location and spatial extent. The ability to causally link a discrete episode of chest pain to true myocardial ischemia would be a powerful clinical tool for the triage and subsequent testing of patients presenting with symptoms and clinical signs suggestive, but not diagnostic, of coronary ischemia. Further studies will be required to establish whether this approach is incremental to existing diagnostic clinical tools. Furthermore, the time window for useful imaging, and sensitivity to varying degrees and durations of ischemia, need to be defined.
The clinical translation of the concepts described here has challenges that are inherent to ultrasound molecular imaging in general. Targeting ligands need to be nonimmunogenic (e.g., not monoclonal antibodies), with adhesive kinetics that are optimized for attachment and persistence on the endothelium. Maximal sensitivity for detection will require both optimized adhesive dynamics, which may include multitargeting and selection of ligands with better on–off rates to permit maximal microbubble adhesion, and imaging systems that can detect acoustic signals unique to adhered microbubbles, as opposed to those that are freely circulating. Such requirements will demand that multidisciplinary efforts be used to bring these promising developments to clinical fruition.
CHALLENGES AND FUTURE NEEDS FOR IMAGING MYOCARDIAL METABOLIC REMODELING
As seen in the preceding discussion, metabolic imaging using a variety of technologies is playing a key role in delineating the relationship between myocardial metabolic remodeling and cardiovascular disease. However, for metabolic imaging to achieve its full potential, advances in several areas must occur. First, there must be continued improvement in instrumentation design, acquisition schemes, and image analysis methods that permit quantitative measurements of metabolic processes of interest that are accurate, robust, and reproducible, and that are capable of being performed in both rodents and humans. There is a compelling need for the development of new radiopharmaceuticals and contrast agents that permit characterization of key metabolic pathways such as uptake, storage, or oxidation that are linked to disease manifestations. Moreover, new imaging agents are needed to provide insights into the detrimental effects on myocellular health of perturbations in myocardial metabolism, such as activation of inflammatory, oxidative stress, cell growth, and cell survival pathways. Ultimately, appropriately powered clinical trials will need to be performed to determine whether the metabolic signatures identified by imaging in various diseases provide unique information that will alter the care of the cardiac patient.
Acknowledgments
This work was supported in part by grants P01-HL13851 and HL69100 (RJG), R37HL49244, R01HL62702, R01HL56178 (EDL), and R01HL077434 (FSV) from the National Institutes of Health.
Footnotes
-
COPYRIGHT © 2010 by the Society of Nuclear Medicine, Inc.
References
- 1.↵
- 2.↵
- 3.↵
- 4.
- 5.↵
- 6.↵
- 7.↵
- 8.↵
- 9.↵
- 10.↵
- 11.↵
- 12.
- 13.
- 14.↵
- 15.↵
- 16.
- 17.↵
- 18.↵
- 19.
- 20.↵
- 21.↵
- 22.↵
- 23.↵
- 24.↵
- 25.↵
- 26.↵
- 27.↵
- 28.↵
- 29.↵
- 30.↵
- 31.↵
- 32.↵
- 33.↵
- 34.↵
- 35.↵
- 36.↵
- 37.↵
- 38.↵
- 39.↵
- 40.↵
- 41.↵
- 42.↵
- 43.↵
- 44.↵
- 45.↵
- 46.↵
- 47.
- 48.↵
- 49.
- 50.↵
- 51.↵
- 52.↵
- 53.↵
- 54.↵
- 55.↵
- 56.↵
- 57.↵
- 58.
- 59.↵
- 60.↵
- 61.↵
- 62.↵
- 63.↵
- 64.↵
- 65.↵
- 66.↵
- 67.↵
- 68.↵
- 69.↵
- 70.
- 71.↵
- 72.↵
- 73.↵
- 74.↵
- 75.↵
- 76.
- 77.
- 78.↵
- 79.↵
- 80.
- 81.↵
- 82.↵
- 83.↵
- 84.↵
- 85.
- 86.↵
- 87.↵
- 88.↵
- Received for publication February 26, 2010.
- Accepted for publication April 12, 2010.