Hyperpolarized [1,4-13C2]Fumarate Enables Magnetic Resonance-Based Imaging of Myocardial Necrosis

Objectives The aim of this study was to determine if hyperpolarized [1,4–13C2]malate imaging could measure cardiomyocyte necrosis after myocardial infarction (MI). Background MI is defined by an acute burst of cellular necrosis and the subsequent cascade of structural and functional adaptations. Quantifying necrosis in the clinic after MI remains challenging. Magnetic resonance-based detection of the conversion of hyperpolarized [1,4–13C2]fumarate to [1,4–13C2]malate, enabled by disrupted cell membrane integrity, has measured cellular necrosis in vivo in other tissue types. Our aim was to determine whether hyperpolarized [1,4–13C2]malate imaging could measure necrosis after MI. Methods Isolated perfused hearts were given hyperpolarized [1,4–13C2]fumarate at baseline, immediately after 20 min of ischemia, and after 45 min of reperfusion. Magnetic resonance spectroscopy measured conversion into [1,4–13C2]malate. Left ventricular function and energetics were monitored throughout the protocol, buffer samples were collected and hearts were preserved for further analyses. For in vivo studies, magnetic resonance spectroscopy and a novel spatial-spectral magnetic resonance imaging sequence were implemented to assess cardiomyocyte necrosis in rats, 1 day and 1 week after cryo-induced MI. Results In isolated hearts, [1,4–13C2]malate production became apparent after 45 min of reperfusion, and increased 2.7-fold compared with baseline. Expression of dicarboxylic acid transporter genes were negligible in healthy and reperfused hearts, and lactate dehydrogenase release and infarct size were significantly increased in reperfused hearts. Nonlinear regression revealed that [1,4–13C2]malate production was induced when adenosine triphosphate was depleted by >50%, below 5.3 mmol/l (R2 = 0.904). In vivo, the quantity of [1,4–13C2]malate visible increased 82-fold over controls 1 day after infarction, maintaining a 31-fold increase 7 days post-infarct. [1,4–13C2]Malate could be resolved using hyperpolarized magnetic resonance imaging in the infarct region one day after MI; [1,4–13C2]malate was not visible in control hearts. Conclusions Malate production in the infarcted heart appears to provide a specific probe of necrosis acutely after MI, and for at least 1 week afterward. This technique could offer an alternative noninvasive method to measure cellular necrosis in heart disease, and warrants further investigation in patients.

C ell death is a hallmark of cardiovascular disease (CVD), including heart failure and myocardial infarction (MI). Cardiomyocytes die via distinct mechanisms-necrosis, apoptosis, and autophagy-each of which occurs by different signaling events (1). Necrosis is marked by distinct morphologic changes, including cell and organelle swelling, plasma membrane rupture, and depletion of chemical energy in the form of adenosine triphosphate (ATP) (1,2). Emerging evidence suggests that instead of being a passive form of cell death, necrosis can be regulated by signaling cascades (1,2).
In vivo imaging of necrosis in the clinic remains challenging. Myocardial necrosis associated with ischemia and reperfusion is detected clinically by the presence of cardiomyocyte-specific proteins released into the circulation, ideally troponin. However, circulating troponin assays are not specific to MI, particularly in the settings of renal failure or sepsis (8). Furthermore, biomarkers found in the blood do not enable spatial localization, preventing tissue damage from being assigned to regions of the heart.
Recently, the metabolic production of [1,4-13 C 2 ] malate from hyperpolarized [1,4-13 C 2 ]fumarate was demonstrated to offer positive MR contrast to identify cellular necrosis in vivo in tumor cells and acute kidney injury (20,21). The fumarate-to-malate hydration reaction is catalyzed by the intracellular enzyme fumarase as part of the tricarboxylic acid cycle. Unlike many metabolic reactions requiring cofactors such as nicotinamide adenine dinucleotide to proceed, the fumarase reaction requires no cofactors and maintains activity during cell death (21). Imaging malate using 13 C MRI ensured specificity to necrosis by targeting loss of cell membrane integrity: [1,4-13 C 2 ]malate production was only observed when the cell membrane was disrupted, enabling the infused hyperpolarized [1,4-13 C 2 ]fumarate to access the fumarase enzyme (21).
Malate imaging may be valuable to detect necrosis in heart disease. A dicarboxylic acid transporter with a known capacity to import fumarate into cells has not been detected in the heart (22,23), implying that malate production may only be observed due to cell membrane rupture. Furthermore, the clinical translation of hyperpolarized [1,4-13 C 2 ]fumarate is actively underway and hyperpolarized [1-13 C] pyruvate is already being used in humans (24,25).

CARDIOMYOCYTE FUMARATE UPTAKE IS NEGLIGIBLE.
To examine bulk fumarate uptake into the healthy myocardium we infused hyperpolarized [1,4-13 C 2 ] fumarate into the perfused heart and used magnetic resonance spectroscopy (MRS) to search for evidence of its enzymatic conversion into other species.
Extremely low levels of metabolism were detected (Online Figure S1A). As an additional test, we measured mRNA expression of putative fumarate transporters in control heart tissue and tissue collected from hearts 45 min after reperfusion. Online Figure S1B shows that the genes encoding each transporter were detected robustly in kidney tissue (Slc3a2 and Slc3a3, relative to internal control 18s).
The mRNA from Slc3a2 could not be detected in heart tissue. Expression of Slc3a3 was approximately 4 orders of magnitude lower than in the kidney, for both healthy and reperfused tissue (kidney, 4.1 Â 10 À6 AE 1.5 Â 10 À6 ; healthy heart, 4.2 Â 10 À10 AE 2.7 Â 10 À10 ; reperfused heart, 1.4 Â 10 À9 AE 3.7 Â 10 À10 ).     triphhenyltetrazoliumchloride-staining can first reveal the necrotic region reliably [27,28]). As with malate MRS and LDH release, a significant increase in necrosis was observed after late reperfusion (Ei). The degree of increase was larger (w12Â) due to ongoing necrosis during the reperfusion period (27). Representative triphhenyltetrazoliumchloride-stained heart slices are shown to the right (Eii). *p < 0.05 compared with healthy group. Abbreviations as in Figure 1.   Our data agree with numerous studies demonstrating a permissive role for ATP levels and irreversible myocardial damage (16,17,19,35). However, the permissive value identified by our data was twice as high as that from previous work (19,35). Additionally, our results diverge from previous work by suggesting a weak correlation between necrosis and ATP, and a stronger relationship between necrosis and lactate (19). The differences may be explained by differences in ATP demand, and the difference in parameters representing "necrosis" that were measured. Our study targeted a fundamental event of necrosis-cell membrane rupture-as an output, whereas previous work relied on scar size (19,36).

DISCUSSION
Scar may encompass other mechanisms of cell damage beyond necrosis that cloud the strength of the ATP-necrosis relationship (apoptosis, for example, requires ATP to proceed) (1). In cerebral infarction, experimental work has implicated tissue lactate as a predictor of destruction of the tissue at risk (37), whereas regions of ATP depletion pinpoint necrotic core (37). Future work with hyperpolarized [1- 13   during reperfusion, by supporting its production or delaying the restoration of its use (39,40 (2), but could also indicate accumulation and persistence of released fumarase enzyme in the extracellular space.
Understanding how long fumarase persists after MI will be essential in quantifying 13 C-malate MRI signal as a marker of necrosis. Furthermore, correlating perfusion imaging techniques, such as first pass gadolinium or co-infusion with hyperpolarized 13 C-urea (30), and 13 C-fumarate imaging may enable the 13 C-fumarate images themselves to facilitate interpretation of 13 C-malate images in the context of altered perfusion. Ultimately, the application of 13 C-fumarate imaging in patients with prior MI or coronary artery disease and comparison with current well-established imaging techniques will determine the value of this promising method.

CONCLUSIONS
Malate production in the infarcted heart appears to provide a specific probe of necrosis acutely following MI, and for at least 1 week afterwards. This technique could offer a new noninvasive method to measure cellular necrosis in heart disease, and warrants further investigation in patients. Hyperpolarized MR Assessment of Necrosis