Cardiac magnetic resonance spectroscopy : applications in a mouse model of fatty acid oxidation deficiency

Under normal, well-fed conditions, the primary source of energy for the healthy heart are long-chain fatty acids that fuel the mitochondrial fatty acid ß-oxidation (FAO) pathway. Patients with an inborn error in long-chain FAO may present with hypoketotic hypoglycemia and liver disease, and/or a life-threatening cardiac phenotype that includes conduction abnormalities, arrhythmias, and hypertrophic cardiomyopathy. Hypoketotic hypoglycemia can ultimately lead to coma or sudden death. Therefore, current treatment strategies for patients with a long-chain FAO defect aim at preventing hypoketotic hypoglycemia, particularly via avoidance of fasting. Unfortunately, no evidence-based treatment of cardiac disease in long-chain FAO disorders is currently available. Patients displaying a cardiac phenotype are therefore permanently exposed to the risk of unexpected death. The pathogenesis underlying the development of cardiac disease in long-chain FAO disorders is still unclear, explaining the lack of evidence-based treatment options. To facilitate the design of novel therapeutic strategies for the prevention and management of cardiomyopathy, a better understanding of the etiology of cardiac disease in long-chain FAO disorders is crucial. It has been suggested that both a chronic energy shortage and a potentially toxic accumulation of lipid metabolites can contribute to the cardiac phenotype. The objective of the work described in this thesis was to develop and apply magnetic resonance spectroscopy (MRS) methods for non-invasive in vivo investigations of myocardial lipid accumulation and energy metabolism in a mouse model of long-chain FAO deficiency, i.e. the long-chain acyl-CoA dehydrogenase (LCAD) knockout (KO) mouse. For MRS to be an accurate quantitative tool for the assessment of in vivo myocardial metabolite levels and metabolic processes, localization of the volume from which signal is acquired is essential to avoid contamination of the spectrum with signal originating from tissues other than the heart. The small heart size and the fast respiratory and cardiac cycle in the mouse pose major challenges for localized MRS of in the in vivo mouse heart. 1H-MRS is a widely applied method in clinical research as well as in animal studies to non-invasively quantify tissue lipid content. In Chapter 2, a comparison between two commonly used sequences (STEAM and PRESS) for localized 1H-MRS measurements is made for in vivo cardiac applications in the mouse. Using an acquisition timing strategy that employs dummy scans during respiratory gates, cardiac-triggered, respiratory-gated acquisitions were performed at an essentially constant repetition time. Based on a comparison of zero-order phase and amplitude stability for STEAM and PRESS acquisitions of the water signal, we concluded that PRESS acquisitions are superior to STEAM measurements for the detection of myocardial metabolites in the in vivo mouse heart. This was attributed to the inherently higher SNR obtainable with PRESS on the one hand, and the higher sensitivity to motion-induced signal loss for STEAM on the other. Using PRESS localization, we were able to perform saturation recovery 1H-MRS experiments, providing the first estimates of in vivo mouse myocardial water and metabolite T1 relaxation time constants at 9.4 T. The use of PRESS-localized 1H-MRS for the quantification of myocardial triglyceride (TG) levels was validated against conventional biochemical measurements of myocardial TG in Chapter 3. In addition, histology was performed to illustrate that higher levels of myocardial TG were associated with the presence of intracellular lipid droplets in cardiomyocytes. In Chapter 3, we applied 1H-MRS in combination with MRI measurements in the LCAD KO mouse in fed and fasted conditions to quantify myocardial lipid accumulation, and to determine cardiac function and morphology. Because fasting elevates the supply of fatty acids via the blood, and increases the heart’s reliance on FAO for ATP production, we hypothesized that in the FAO-deficient heart, fasting would lead to an elevation of myocardial lipid content and impaired cardiac function. We showed that the LCAD KO heart is hypertrophic, and that it harbors higher levels of myocardial TG in the fed state compared to wild-type (WT) mice. Cardiac function was normal in the fed LCAD KO mouse. We observed that myocardial TG levels decreased upon fasting in WT mice. In contrast, lipid levels further increased upon fasting in the LCAD KO myocardium, which was accompanied by a decrease in cardiac performance. Although elevation of myocardial TG may not be a pathologic mediator of cardiac dysfunction per se, it can be regarded as a marker for the accumulation of other metabolites that are lipotoxic. Indeed, the elevation of myocardial TG content observed in the fasted LCAD KO mice was accompanied by higher levels of myocardial ceramide, a known lipotoxic compound. Combined, these results may point to a role for lipotoxicity in the pathogenesis of cardiomyopathy in long-chain FAO disorders. To relieve the accumulation of long-chain fatty acyl compounds in the FAO-deficient myocardium, supplementation with L-carnitine has been suggested as a treatment option for patients. Furthermore, patients with long-chain FAO disorders may have secondary carnitine deficiency, which can be restored by carnitine supplementation. Ironically, carnitine supplementation could enhance the production of long-chain acylcarnitines, which are potentially lipotoxic when accumulating in cardiomyocytes. It is unclear how carnitine supplementation affects the balance between enhanced acylcarnitine export on the one hand and potentially increased production of longchain acylcarnitines on the other hand. This makes carnitine supplementation in patients with long-chain FAO disorders controversial. In Chapter 4, we performed a longitudinal MR study in LCAD KO mice and WT controls with and without carnitine supplementation, starting at 5 weeks of age. We found that, in adolescent mice, LCAD deficiency induced hypertrophic growth of the heart, which was accompanied by elevated levels of myocardial TG compared to WT mice. After four weeks of carnitine supplementation at a clinically relevant dose, myocardial TG levels were lower in carnitine-supplemented animals. In accordance with this in vivo observation, the myocardial total fatty acid content in LCAD KO mice supplemented with carnitine was lower than in LCAD KO mice without carnitine supplementation. Importantly, circulating as well as myocardial levels of free carnitine were normalized by carnitine supplementation in LCAD KO mice, without inducing myocardial accumulation of potentially lipotoxic long-chain acylcarnitines. In addition, no effect of carnitine supplementation on cardiac performance was observed. As such, this study in mice lends support to the proposed beneficial effect of carnitine supplementation in patients, by alleviating lipid overload in the FAO-deficient myocardium. No evidence was found to substantiate the concern about potentially detrimental effects of supplementation-induced production of lipotoxic long-chain acylcarnitines. Based on these results, carnitine supplementation should therefore be considered as a candidate strategy for treatment of cardiomyopathy in patients with inborn errors of long-chain FAO. To be able to assess the in vivo myocardial energy status in mice, we implemented a cardiac-triggered, respiratory-gated 3D ISIS sequence for single-voxel localized 31PMRS of the in vivo mouse heart. Using an acquisition timing strategy similar to the approach described in Chapter 2 and Chapter 3, we demonstrated in Chapter 5 that differences in the myocardial energy status between healthy mice and a wellcharacterized mouse model of heart failure can be detected with localized 31P-MRS, evidenced by a lower myocardial PCr/ATP ratio in mice with a thoracic aortic constriction. To investigate the consequences of a disorder in long-chain FAO on the myocardial energy status, we applied the 31P-MRS method outlined in Chapter 5 in LCAD KO mice in fed and fasted conditions in the study described in Chapter 6. With this approach, we were able to establish that the cardiac dysfunction observed in our first study of the fasted LCAD KO mouse was not only accompanied by an increase of myocardial lipid content (Chapter 3), but also by a lower PCr/ATP ratio compared to fasted WT mice. Consequently, in addition to lipotoxicity, myocardial energy shortage may play a role in the development of cardiomyopathy in long-chain FAO disorders. Because fasting was needed to elicit a cardiac phenotype in the LCAD KO mouse, we hypothesized that enhanced glucose metabolism may act as a compensatory mechanism for the FAO defect in the fed state, which becomes inadequate when the LCAD KO mouse becomes hypoglycemic during fasting. To test this hypothesis, we used 13C-MRS to measure the 13C label incorporation from hyperpolarized [1- 13C]pyruvate through the pyruvate dehydrogenase (PDH) complex into bicarbonate in fed and fasted LCAD KO mice and WT mice (Chapter 6). The PDH flux is a direct measure of in vivo PDH activity, and reflects the contribution of glucose metabolism to myocardial energy homeostasis. We found that the PDH flux was normal in fed LCAD KO mice, but that in fasted conditions, the PDH flux was higher in LCAD KO mice than in WT mice. Additionally, we observed that myocardial deoxyglucose uptake was similar for fed LCAD KO and WT mice. After fasting, the myocardial deoxyglucose uptake reduced in WT mice, whereas the uptake was sustained in fasted LCAD KO mice. Combined, these results suggest that the FAO-deficient heart has an increased reliance on glucose metabolism during fasting. Interestingly, we noted incorporation of the 13C label from hyperpolarized [1-13C]pyruvate into the myocardial aspartate and malate pools in the fasted LCAD KO mouse. This may point to an increased need for anaplerosis in the LCAD KO heart. Indeed, steady-state levels of citric acid cycle intermediates were found to be lower in the fasted LCAD KO myocardium. We concluded that due to hypoglycemia, the sustained myocardial glucose uptake and PDH flux in LCAD KO mice are ineffective to maintain metabolic homeostasis during fasting, rendering the myocardial metabolic flexibility inadequate. This is reflected by a low myocardial energy status and impaired cardiac performance in fasted LCAD KO mice. Given the lower levels of citric acid cycle intermediates, and the enhanced flux of 13C label through anaplerotic pathways, therapeutic strategies that aim to provide anaplerotic substrates to the FAO-deficient heart may be effective in reducing or reversing the cardiac phenotype in patients with long-chain FAO disorders. To conclude, this work describes the development of methods for localized MRS in the in vivo mouse heart that were successfully applied in a mouse model of longchain FAO deficiency. With these non-invasive methods, we obtained evidence that both lipotoxicity and energy shortage may play a role in the development of cardiomyopathy in long-chain FAO disorders.