Imaging Myocardial Metabolic Remodeling

Introduction

As expertly highlighted in the previous review by Dr. Taegtmeyer, 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/hormonal environment, myocardial blood flow and cardiac work. This loss in flexibility leads to an over dependence 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 that result 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 there is growing demand for accurate non-invasive 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 leading to improved patient management paradigms. Discussed below are 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.

Methods to Image Myocardial Metabolism

There are several methods to image myocardial metabolism noninvasively with the most common being single photon emission computed tomography (SPECT), positron emission tomography (PET) and magnetic resonance 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 setting. Theoretically, assessing more than one metabolic process simultaneously is possible if the heart is imaged after the administration of radiopharmaceuticals labeled with radionuclides with 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 mitcochondria 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 keto-acyl CoA intermediate. The prolonged retention of BMIPP in the cardiomyocyte is suitable for longer acquisition time that is necessary for SPECT imaging. Similar to fluorine-18, iodine-123 is cyclotron produced but with a longer half-life of 13 hrs. The longer half-life of iodine-123 facilitates centralized distribution of the radiotracer, like 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-fluorodeoxyglucose (FDG)¹. The major disadvantage of SPECT is the inability to quantify cellular metabolic processes of interest. This limitation is primarily due 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

Non-invasive 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 two 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 (see Table 1 in the preceding review by Dr. Taegtmeyer). By analyzing different components of the tracer kinetic curves (e.g., uptake, fast wash-out and slow wash-out) with appropriate and well-validated mathematical models it is possible to quantify specific metabolic processes. Examples include the quantification of myocardial oxygen consumption (MVO₂), fatty acid uptake, oxidation and esterification, glucose uptake, glycolysis and oxidation and lactate oxidation.^2–4 As with SPECT, small animal PET and PET/CT systems have been developed permitting similar measurements of myocardial metabolism in rodent heart as are done in the human heart. Despite its strengths, PET suffers from several limitations which limit its broader applicability to measure myocardial metabolism. These include the high cost of the PET imaging systems, the need for an on-site cyclotron and expertise in radiopharmaceutical synthesis (except for FDG), and the relative complexity of the image quantification schemes.

MRS

MRS of the heart has proven to be a powerful research tool. Its use offers great advantages for metabolic evaluation of whole organs, by providing unique, chemically specifically 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 whole hearts have proven valuable for evaluation of energetic state and energy substrate use by the heart, such as carbohydrate versus fatty acid use.^5–7 The dynamic processes described above, uniquely accessible by MRS, provide unique mechanistic insight linking metabolic derangements to the contractile dysfunction that are associated various cardiac disorders such as myocardial ischemia/reperfusion and hypertrophy.^{8, 9} Moreover, MRS enables detection of endogenous, nuclear magnetic resonance (NMR) sensitive metabolites (nuclei possessing net spin, I≥ ½), such as compounds containing phosphorus-31 (³¹P) and sodium-23.^{5, 6} Recent advances in the applications of localized detection of NMR signals, or MRS, from selected regions of interest within 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 due to low single to noise compared to 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 which 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 due to 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 complimentary 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 Imaging of Glucose Metabolism

The perfusion/metabolism mismatch pattern on FDG PET imaging is the gold standard for detecting reversible left ventricular dysfunction in the clinical setting. FDG PET imaging has shown proven benefit for recovery of function, improvement in heart failure symptoms, and in predicting survival after revascularization.^10–13 SPECT metabolic imaging with either Tl-201 or Tc-99m tracers is also reliable in detecting myocardial viability, particularly in patients with mild-to-moderately impaired LV dysfunction (LVEF 25%-50%).^14–16

Advances in high energy collimation SPECT cameras have made imaging of positron emitting tracers, such as fluorine-18, possible. Moreover, FDG metabolic imaging with SPECT equipped with a high energy collimator has the additional advantage of acquiring myocardial perfusion data, simultaneously, with a SPECT perfusion agent. Studies have shown that when SPECT flow tracers such as ^99mTc-sestamibi or ²⁰¹thallium are used in combination with FDG SPECT, detection of myocardial viability is similar to PET.^17–19 However, the poorer spatial resolution of SPECT, the lower sensitivity, and lack of attenuation correction, may cause some discordance between images of the same tracer measured with SPECT and PET.^{16, 20} Given these technical limitations of FDG SPECT imaging, along with the limited number of publications, in small number of subjects, FDG SPECT has not received FDA and CMS approvals for imaging of 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.^{21, 22} Indeed, this metabolic signature well may persist well after the resolution of the inciting ischemic event and has been termed ischemic memory.^{23, 24} Applying a dual-isotope sestambi and 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 persisted up to 24 hrs despite normal perfusion under resting conditions (Figure 1).^{1, 25} Similarly, decreased in regional fatty acid metabolism and delayed recovery of fatty acid metabolism, long after regional blood flow has returned back to baseline, has been shown BMIPP. Taken is sum, these studies show that “ischemic memory” occurs in humans and can be imaged with conventional radionuclide approaches raising the potential of the use of these techniques in the management of patients with ischemic heart disease.

Figure 1.

Simultaneous myocardial perfusion and metabolism imaging after dual intravenous injection of Tc-99m sestamibi and FDG at peak exercise. Dual isotope simultaneous acquisition was carried out 40–60 minutes after the exercise study was completed. Rest Tc-99m sestamibi imaging was carried out separately. In this patient with angina and no prior myocardial infarction, there is evidence for extensive reversible perfusion defect in the anterior, septal, and apical regions. The coronary angiogram showed 90% stenosis of the left anterior descending and 60% of the left circumflex coronary arteries. The corresponding FDG image show intense uptake in the regions with reversible sestamibi defects reflecting the metabolic correlate of exercise-induced myocardial ischemia. (Adapted from He et al [1]).

While the technical limitations of modified SPECT camera to detect F-18 signals over PET are well-recognized, other challenges also exist for acquiring FDG studies either with SPECT or PET. For example, studies to detect myocardial ischemia are typically performed under fasting conditions to maximize differences in tracer uptake between ischemic myocardium (increased uptake) and non-ischemic tissue (reduced uptake). However under fasting conditions, there is heterogeneous distribution of FDG in the myocardium even in normal subjects creating another level of complexity in image interpretation.²⁶ Thus, when it comes to cardiac application of FDG to detect myocardial ischemia under fasting conditions, it may be difficult to resolve and conclusively localize the hot-spot 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 alternatively hybrid SPECT/CT with anatomic co-registration of the heart.

SPECT Imaging of Fatty Acid Metabolism

As mentioned above, measurement of myocardial fatty acid metabolism using BMIPP is the most commonly used metabolic application of cardiac SPECT. Similar to 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 (Figure 2), as well as in patients presenting with acute coronary syndromes.^{21, 22} This long-duration disturbance in metabolism is similar to that seen after FDG administration.

Figure 2.

Single-photon emission CT showing delayed recovery of regional fatty acid metabolism after transient exercise-induced ischemia, termed ischemic memory. Representative stress (left) and rest reinjection (middle) short-axis thallium tomograms demonstrate a reversible inferior defect consistent with exercise- induced myocardial ischemia. A BMIPP-labeled CT (center) injected and acquired at rest 22 h after exercise-induced ischemia shows persistent metabolic abnormality in the inferior region despite complete recovery of regional perfusion at rest, as evidenced by thallium reinjection image. The tomogram on the far right shows retention of BMIPP in the heart of a normal adult for comparison. (Adapted from [15]). Abbreviation: BMIPP, β-methylp-[¹²³I]-iodophenyl-pentadecanoic acid.

Uptake of BMIPP from the plasma into myocardial cells occurs via CD36 transporter protein present on the sarcolemmal membrane. Once BMIPP is taken up by the myocyte, it undergoes 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. An important clinical application of the latter include early assessment of chest pain among patients presenting to the emergency department with acute coronary syndromes as well as assessment of myocardium at risk among patients presenting with acute myocardial infarction and early reperfusion with either percutaneous intervention or thrombolysis. Preliminary 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 to the sensitivity of the initial clinical diagnosis alone, while maintaining specificity.²⁷

The deleterious effect of altered myocardial metabolism was recently demonstrated in patients with end-stage renal disease as well.^{28, 29} A shift in myocardial metabolism from fatty acid to glucose may be contributory to the cardiomyopathic process observed in late stage kidney disease.²⁹ The cardiomyopathy typical of chronic kidney disease and the associated uremia is thought to lead to a myocyte-capillary mismatch, with a diminished vascular supply relative to the number and volume of functioning myocytes.³⁰ The oxygen-poor milieu leads to diffuse myocardial ischemia with an anticipated decline in aerobic myocardial fatty acid utilization. Such altered cardiac metabolism (indicating silent myocardial ischemia) was recently shown to be highly prevalent among asymptomatic dialysis patients without history of prior myocardial infarction, and was able to identify the subgroup of patients who were at high risk for cardiac death.²⁹

Future Directions and Challenges

SPECT imaging 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 chronic and in the acute-care 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 overcome. From a technical perspective imaging protocols need to be optimized with respect to timing of imaging post-chest pain, acquisition schemes that account for iodine-123 need to be standardized and questions regarding the need to standardize the substrate environment resolved. From a clinical point of view, the diagnostic and prognostic accuracy and the added clinical information provided by these metabolic signatures needs to be determined.

Application of PET

FDG PET: Clinical Application – A Model of Molecular Imaging Probes

As mentioned above myocardial viability imaging using FDG has emerged as highly valuable clinical tool. As molecular imaging evolves seeking tools that could be translated to clinical practice, it is worthwhile to examine FDG and what has led to its success as a probe, namely: i) more than 30 years of experience ³¹, ii) straightforward production and imaging, iii) wide availability; iv) high accuracy; v) linkage to disease and outcomes; and vi) impact on patient management (Figure 3). FDG PET is the most sensitive viability method for predicting wall motion recovery ³² and identifies patients at increased risk for death if they do not undergo revascularization. The PARR-2 trial showed a trend for benefit when FDG PET was used to assist management.³³ In patients who adhered to recommendations, significant outcome benefits were attained with the degree of hibernation predicting the likelihood of response to revascularization. ³⁴ A post-hoc subgroup analysis demonstrated improved outcomes in an experienced center.³⁵ As a consequence, regional networks that work to ensure quality and clinical expertise at different FDG PET sites are now being established. ³⁶

Figure 3.

Example of hibernating myocardium. Images show a moderate perfusion defect involving the apex, mid to distal anterior and septal walls (superior row) but preserved FDG uptake; which allows delineation of hibernating myocardium in the distribution of the LAD territory (inferior row).

Phenotype Characterization and Translational Imaging

Quantitative measurements of myocardial metabolism are now possible in rodent heart using small animal PET. For example, measurements of myocardial glucose uptake using small animal PET have been shown to relate to protein expression of glucose transport-4 protein levels.³⁷ Moreover small animal PET imaging 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 alpha (PPARα), a key regulator of myocardial fatty acid uptake, oxidation and storage, demonstrate a metabolic phenotype that is similar to diabetic hearts.³⁸ Small animal PET studies with ¹¹C-palmitate and FDG in these 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 demonstrated an increase glucose uptake and reduced fatty acid uptake and oxidation (Figure 4).³⁹ Taken is sum these observations demonstrate that PPARα and PPARβ/δ drive different metabolic regulatory programs in the heart and that imaging can help characterize genetic manipulations in mouse heart. These observations highlight the importance of small animal PET to evaluate phenotype and transgene expression. However, small-animal PET has been met with several challenges in imaging the rodent heart including partial-volume effects, extraction of the input function for quantification, and other technical factors. Thus, reproducing the success observed in animals when applied in the human setting can represent a challenge.

Figure 4.

Increased myocardial glucose utilization in MHC-PPARβ/δ mice. A) Left: Standardized uptake value time-activity curves for ¹¹C-palmitate and ¹¹C-glucose into female MHC-PPARβ/δ–HE and NTG (non-transgenic) hearts as determined by micro-PET. Right: Representative micro-PET images at 20 sec after tracer injection. Images are normalized to total amount of radioactivity injected and body weight. The relative amounts of tracer uptake are indicated by the color scale. B) Oxidation of Palmitate and Glucose was assessed in isolated working hearts of 12-week-old male MHC-PPARβ/δ-HE and NTG control mice. Bars represent mean oxidation rates expressed as nanomoles substrate oxidized per gram dry mass per minute. C) Glycogen levels were assessed in mouse hearts from male MHC-PPARα-LE and MHCPPARβ/δ-HE mice and NTG controls. Results are presented as glucose released from glycogen and normalized to tissue weight. *P < 0.05 versus NTG [39] (permission pending).

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 PRKAG2 gene which codes for adenosine monophosphate-activated activated protein kinase (AMPK), a key metabolic regulator in the myocyte.⁴⁰ The central role of AMPK in regulatingglucose metabolism had led to the hypothesis that the pathological basis of the PRKAG2 Cardiac Syndrome was due to impaired glucose metabolism and excessive glycogen storage.⁴¹ Indeed, patients and the corresponding transgenic mouse model are characterized by increased glycogen storage. Preliminary data in the transgenic model demonstrate reduced FDG uptake that can be quantified (Figure 5).⁴² Recently, human studies demonstrated reduced FDG uptake.⁴³ At first blush the lower level of glucose uptake appears paradoxical because cardiac hypertrophy caused by excessiveglycogen 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 was demonstrated in a group of patients with mutation in the alpha-tropomyosin gene and associated left ventricular hypertrophy.⁴⁴ Taken in sum, these data highlight the advances in small animal PET technology to quantify metabolic parameters and exemplifies the translational capabilities with metabolic PET imaging.

Figure 5.

Left: Cardiac Imaging in the mouse model. Right: Short Axis display of FDG myocardial uptake (60 minutes) in NTG (non-transgenic) and TG-mut (transgenic-mutant) of the PRKAG2 gene. It can be appreciated that FDG uptake in NTG is normal whereas in the TG-mut is significantly reduced. Below: Arterial blood (red) and myocardial (blue) time-activity curves. In the TG-mut, the myocardial curve shows lower rate of FDG uptake compared to that in NTG. With permission from MH Gollob and S Thorn.

Future Directions and Challenges

There are several promising developments that augur well for the continued need 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) are creating a need to direct and monitor metabolic response, paving a path for metabolic imaging directed personalized therapy.

It is clear that metabolic imaging works well as a clinical tool in the context of viability imaging but also metabolic alterations in non-ischemic disorders are common but 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 to other emerging methodologies.

Application of MRS

Myocardial Energetics

³¹P NMR has been a long used detection scheme for assessing high energy phosphate content in the intact, beating heart. Indeed, the earliest clinical studies with cardiac MRS provided evaluations of myocardial energetic state from the relative contents of high energy phosphates via ³¹P detection.^45–47 Cardiac ³¹P MRS in patients provides information on the relative contents of phosphocreatine (PCr) and adenosine triphosphate (ATP)^45–49. 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 demonstrating the now well recognized impaired energetic state of the failing heart.

However, the diseased myocardium demonstrates a generally reduced PCr:ATP ratio, and one limitation of the parameter for clinical use is the lack of specificity. Reduced PCr:ATP is generally evident evidence in patients with either hypertrophic or dilated cardiomyopathies, allograft rejection following cardiac transplantation, and coronary artery disease with stenosis.^47–50 Fortunately, important new advances in ³¹P MRS detection extend well beyond the original uncertainties of the pathophysiological specificity of the PCr: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 utilization. Indeed, application of this approach in clinical subjects suggests better delineation of failing from non-failing, hypertrophic hearts when compared to using the PCr:ATP ratio (Figures 6 and 7).⁵¹

Figure 6.

A) Axial MR image of a patient with left ventricular hypertrophy and congestive heart failure with the region of localized ³¹P NMR spectra from chest and left ventricle identified (rectangular outline). B) ³¹P NMR spectra from chest muscle (bottom) and left ventricle (top) with control saturating RF irradiation (arrow). C) ³¹P NMR spectra from chest muscle (bottom) and left ventricle (top) with selective, saturating RF irradiation at the gamma-phosphate resonance (arrow). Note decreased magnitude of PCr signal in panel C due to chemical exchange with saturated ³¹P nuclei of the gamma phosphate of ATP. Decreased PCr signal depends on rate of ATP synthesis through the creatine kinase reaction. [51]

Figure 7.

A) pseudo first-order rate constant (K_for) 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. [51]

While ³¹P MRS provides unique information on the energetic state of the myocardium, the underlying mechanisms of energy production pathways via intermediary metabolism can detected by exploiting stable isotope kinetics. Naturally abundant at only 1.1%, the non-emitting 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 within the glutamate pool of tissue extracts.^{52, 53} Carbon-13 (¹³C) detection of enrichment rates within 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 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.^{7, 9, 54} For example, NMR spectra can be obtained from the beating heart to monitor rates of palmitate oxidation or storage into triacylgyceride (Figure 8). Recent applications for ¹³C MRS in experimental animal models enable measurement of metabolic enzyme activity due to altered gene expression in the intact, functioning heart.^{9, 54}

Figure 8.

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) 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, showing progressive enrichment of the TAG for measures of TAG turnover in transgenic models

Dynamic- mode ¹³C MRS has provided mechanistic understanding of enzyme isoform changes, and even mitochondrial transport function, in isolated organs and animals models of heart disease.^{9, 54, 55} 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 into the triacylglyceride pool, providing a comprehensive view of lipid dynamics in normal and diseased hearts.^56–58 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 demonstrated reduced rates of long chain fatty acid oxidation and impaired turnover within the triacylglyceride pool.^{9, 58} 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.^{9, 59} Pharmacological reversal of this up-regulated anaplerotic activity in cardiac hypertrophy produced a surprising increase in contractility of the hypertrophied rat heart.⁵⁹ These findings demonstrate 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 to the development of potential therapeutic, metabolic protocols that combat contractile dysfunction in the diseased heart.

Just as important, while ¹³C MRS 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, is insensitive to 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 consequentially may lead to an overestimation of mismatches between the rates of glycolysis and glucose oxidation. Thus, despite the limitation of non-tracer levels of exogenous materials, ¹³C MRS studies have already offered unique perspectives not previously accessible by other labeling methods.

While the examples discussed above 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 within specific compounds to provide 10–20,000 fold increases in signal to noise (SNR), holds great potential. The use of dynamic nuclear polarization (DNP) 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 (Figure 9).^60–62 With DNP, the hyperpolarized state of the ¹³C nuclei lasts for approximately one minute, 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 (Figure 9).^{61, 62} While current in vivo studies with hyperpolarized ¹³C rely on detection following bolus, intravenous injections of ¹³C enriched precursors at relatively large, millimolar concentrations, the increase in SNR holds potential for the eventual development of studies using tracer levels that would not influence the metabolic pathways of interest. At present, chemistry of the DNP 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/MRS for in vivo cardiac metabolism.

Figure 9.

In vivo ¹³C NMR spectra from heart of an anesthetized rat following 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 minute, post-injection. C) Time course of signals from pyruvate (solid circle), lactate (square), alanine (triangle) and bicarbonate (X). [62] Copyright 2007 by the National Academy of Sciences of the USA.

More accessible to current in vivo and clinical studies is proton (¹H) MRS. Localized ¹H MRS, of both animal models and human subjects, offers unique evaluation of lipid infiltration of internal organs.^{63, 64} In heart, ¹H detection provides quantitative analysis of the intramyocardial mobile lipid, from methylene group protons (CH₂-) within acyl chains, generally associated with triacylglyceride. Notably, such evaluation relies on ¹H signal that 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 demonstrated in diabetics and appears to be associated with a decline in diastolic function.^{63, 65} Moreover, a prolonged caloric restriction in obese diabetic patients and its attendant salutary effects of glucoregulation decreases myocardial lipid and improves diastolic function.⁶⁶ Thus, it appears myocardial lipid levels may have a detrimental effect on diastolic function in both diabetic and non-diabetic men and that lipid levels are responsive to changes in plasma FA delivery. As interest mounts in monitoring lipid accumulation within the heart, the advantages of high and ultra-high 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.

Figure 10.

Localized ¹H MRS signal from myocardial triacylglyceride in human. Left panel shows MR image of heart displaying localized volume within LV septum for MRS (yellow rectangle). Right panel displays ¹H NMR spectrum with inset of expanded region of triacylglyceride signal (in red circle). [63]

Figure 11.

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 the axial MR image of the heart (yellow box). Top panel displays signals from triacylglyceride within a 1 × 1 × 1 mm voxel. Bottom panel displayed enhanced signal from increased voxel size of 1 × 2 × 1 mm. From Dr. E. Douglas Lewandowski, Program in Integrative Cardiac Metabolism, UIC College of Medicine, Chicago, IL.

Preliminary results suggest 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 2D principle strains (E1 and E2) across the LV wall.⁶⁷ 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 above, demonstrate an array of cardiac applications for NMR spectroscopy/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. While 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 to cellular signatures of antecedent myocardial ischemia. As mentioned previously, transient myocardial ischemia results 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 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.^68–70 Upon ischemic insult, the leukocyte adhesion molecule P-selectin is rapidly mobilized to the endothelial surface from pre-formed cytoplasmic stores within minutes of ischemia-reperfusion, where it mediates transient tethering interactions that result in leukocytes rolling along the endothelium.⁷⁰ Leukocyte rolling enables leukocyte interaction with adhesion molecules such as intercellular adhesion molecule-1 (ICAM1) or vascular adhesion molecule (VCAM), which mediate firm attachment.⁶⁸ Unlike pre-formed cytoplasmic P-selectin, which mobilizes quickly to the surface of endothelium upon ischemia/reperfusion, post-ischemic 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 within 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 microns in diameter, are typically perfluorocarbon- or nitrogen-gas microspheres encapsulated by shells of varying composition, such as phospholipids, albumin, or biodegradable polymers.⁷¹ Because of their size and shell composition, the microbubbles do not interact with the endothelium and transit in unimpeded fashion through the microcirculation.⁷² In the presence of an ultrasound field tuned to the appropriate frequency and acoustic pressure, the microbubbles expand and contract (oscillate) in a non-linear fashion, themselves becoming emitters of ultrasound that can be detected and displayed as a transient signal on a two dimensional ultrasound image.⁷³ 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.^74–77

For ultrasound molecular imaging applications, targeting ligands are conjugated to the microbubble shell, conferring the ability for a microbubble to attach to a specific endothelial epitope, resulting in persistent ultrasound contrast enhancement, and hence the capability for interrogating endothelial phenotype.⁷⁸ 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 non-cardiac disease (Figure 12A) (26–40).^78–82 Alternatively, a non-ligand-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 (Figure 12B).^83–85

Figure 12.

Schematic of vascular endothelium and approaches for ultrasound molecular imaging of inflammation during ischemia/reperfusion using microbubble attachment to endothelial cells. (A). Microbubbles bearing a targeting ligand on the surface can bind to a specific endothelial target, such as a leukocyte adhesion molecule. (B) Activated leukocytes may bind or phagocytose microbubbles and become acoustically active. Figure not drawn to scale.

Based on the above 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/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-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.⁸⁰ Persistent contrast enhancement of the post-ischemic kidney was noted on ultrasound imaging, which was not seen after injection of non-targeted bubbles, or after injection of targeted and non-targeted 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 post-ischemic 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 Lewis^x (sLe^x) 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.⁸⁶ sLe^x 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 TNF-α administration.⁸¹ sLe^x-microbubbles adhered to activated endothelium, whereas control Le^x-conjugated microbubbles minimally adhered, with neither bubble species adhering significantly to normal, noninflamed endothelium. A rat model of myocardial ischemia-reperfusion was then used to echocardiographically detect adhesion events.⁸¹ Both early (30 min) and late (up to 90 min) after reflow, videointensity in the post-ischemic bed after intravenous injection of sLe^x microbubbles was significantly higher than that in the non-ischemic bed. Furthermore, the region of persistent contrast enhancement colocalized with, and correlated in size to, the risk area (Figure 13). There was a trend towards similar findings after milder (10min) of ischemia, as well. These data indicate that the selectin-targeted microbubbles were capable of recapitulating the presence, location, and spatial extent of previously ischemic myocardium.

Figure 13.

Ultrasound ischemic memory imaging of myocardium using microbubbles targeted to bind to P-selectin via the tetrasaccharide sialyl Lewis^x in a rat model of 15 minute coronary occlusion followed by reperfusion. Short axis non-linear ultrasound images of the heart are background subtracted, and degree of contrast enhancement is color-coded. (A) Injection of non-targeted microbubbles during coronary occlusion shows the risk area (region between arrows). (B) After reperfusion, non-targeted microbubble injection confirms restoration of myocardial perfusion. (C) Post mortem staining with triphenyl tetrazolium chloride 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 the region that was previously ischemic (risk area, Panel A). [81]

As mentioned above, another approach to inflammatory imaging for detection of prior ischemia is to use microbubbles that are avid for activated leukocytes. Lipid microbubbles augmented with phosphatidylserine in the shell adhere to activated leukocytes (40), and remain acoustically active even after leukocyte phagocytosis.⁸⁵ Phosphatidaylserine 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 (Figure 14).⁸⁷ Note that this approach appears feasible in a model in which there is sizeable infarction, and it is unclear whether a leukocyte-avid microbubble would be effective in delineating antecedent ischemia where there is minimal or only mild necrosis, the clinical scenarios in which the ED evaluation of chest pain is particularly problematic.

Figure 14.

Inflammatory imaging of reperfused infarcted canine myocardium using phosphatidylserine-augmented lipid microbubbles which attach to activated leukocytes. Short axis non-linear ultrasound images are background subtracted and color-coded. (A) After injection of leukocyte-avid microbubbles during reperfusion, there is persistent contrast enhancement of the previously ischemic area. (B) Confirmation of leukocyte accumulation in the post-ischemic zone on autoradiography of isotope-labeled leukocytes. (C) TTC-stained myocardial specimen demonstrates non-transmural infarction. [87].

Future Directions and Challenges

Ultrasonic identification of acute adhesion molecule expression or leukocyte activation on post ischemic 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 above has challenges that are inherent to ultrasound molecular imaging in general. Targeting ligands need to be non-immunogenic (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 multi-targeting and selection of ligands with better on/off rates (to permit maximal microbubble adhesion), as well as imaging systems that are capable of detecting acoustic signals unique to adhered, as compared to freely circulating, microbubbles. Such requirements will require that multi-disciplinary efforts be entrained to bring these promising developments to clinical fruition.

Challenges and Future Needs for Imaging Myocardial Metabolic Remodeling

By the dint of the preceding discussion, metabolic imaging using a variety of imaging 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 will be the 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 man. 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 if the metabolic signatures identified by imaging in various diseases provide unique information that alter the care of the cardiac patient.