High resolution respirometry is considered the gold standard to measure mitochondrial function ex vivo. However, as respirometry requires invasive tissue sampling, its application in human research is limited, especially in tissues that can not be easily sampled. Hence, there is need for a non-invasive method for the measurement of cardiac mitochondrial function. In skeletal muscle, it has been shown that non-invasive PCr resynthesis rates by MRS can be used to measure in vivo oxidative capacity and correlates well with ex vivo mitochondrial respiration (27). However, PCr resysnthesis rate can not be determined in the heart. Alternatively, the PCr/ATP ratio, as determined by non-invasive MRS might be used as a marker of mitochondrial function in the heart. Since PCr buffers ATP concentration when ATP demand is increased, the PCr/ATP ratio reflects the equilibrium between ATP synthesis and utilization, the latter depending on the external myocardial work. Therefore, PCr/ATP indicates myocardial energy status and, as energy supply in the heart is mainly covered by mitochondrial oxidative metabolism, PCr/ATP may be an indicator of mitochondrial function. Indeed, PCr/ATP ratio has been shown to be diminished in T2DM (12–15) and heart failure (16) and to have predictive value for cardiovascular morbidity and mortality (16), while in parallel mitochondrial function is suggested to be hampered in T2DM and heart failure (1, 2, 8). Therefore, decreased mitochondrial function may underlie the decreased PCr/ATP ratio that is found in these patients. Hence, the PCr/ATP ratio may be a good in vivo marker of mitochondrial function. However, the direct relationship with cardiac mitochondrial function has never been established in humans. Here we determined PCr/ATP ratio and ex vivo mitochondrial function in a population with a wide range of cardiac metabolic health. We chose to encompass this population, as cardiac mitochondrial function is hampered in metabolic disease, preceding the onset of cardiac dysfunction (28, 29), allowing us also to investigate the associations with cardiac function/geometry.
In contrast to our expectations, we did not find any correlations between PCr/ATP ratio in vivo and any of the mitochondrial respiration states measured ex vivo. Apparently, the PCr/ATP ratio is not solely influenced by mitochondrial function. While it is very likely that a severely diminished mitochondrial function impairs ATP generation, and hence will lead to a low PCr/ATP ratio by relying more heavily on the buffering by PCr to keep ATP concentrations relatively stable, the PCr/ATP ratio may also be influenced by additional factors that do not affect mitochondrial function. For example, a decrease in PCr/ATP ratio may partly be explained by variation in creatine availability. If creatine availability is low, this may limit PCr formation and therefore PCr/ATP ratio, while mitochondrial function may remain unaffected. Furthermore, creatine content decreases in heart failure and its level reflects the severity of heart failure (30, 31). Therefore, the time course of diminishing creatine availability runs in parallel with lowering PCr/ATP ratio and it can be suggested that not ATP but creatine availability is limited in heart failure (31–34). Variations in creatine availability in the current population may have weakened any potential relationship between PCr/ATP ratio and mitochondrial function. However, as we did not measure cardiac creatine content this remains speculative.
The rationale of PCr/ATP potentially being a marker of mitochondrial function originates from ATP being largely produced by mitochondrial metabolism. However, one should keep in mind that in the resting state, ATP demand is low and mitochondrial metabolism is not maximally stimulated. Therefore, it is conceivable that PCr/ATP can be normal in the resting state, even if mitochondrial metabolism is hampered. It is therefore possible that mild mitochondrial dysfunction only becomes apparent during physiological stress. Indeed, it was shown that normal physiological adaptations to stress were hampered in the obese heart (18). This may have troubled the correlation of the in vivo PCr/ATP measurements in resting conditions with the ex vivo measurements wherein the tissue is provoked to its maximal respiratory capacity. Also, it is likely that the cardiac PCr/ATP ratios are influenced by substrate availability rather than by mitochondrial oxidative capacity (14). The in vivo measurement is performed after an overnight fast, and since it is known that there are differences in plasma glucose and whole-body metabolism between individuals with T2DM and lean non-diabetic patients, this overnight fast before the test measurements will lead to a different substrate supply for the heart during the in and ex vivo measurements. This might bring some heterogeneity. Ex vivo mitochondrial function is a measure of mitochondrial capacity under optimal experimental conditions and does not reflect mitochondrial respiration in situ. Therefore, next to variations in creatine, variations in stress and substrate availability may be a modulator of PCr/ATP ratio that is lost in the ex vivo situation.
Interestingly, we did find correlations between the PCr/ATP ratio and cardiac function parameters, whereas the relationship between these parameters and ex vivo mitochondrial function were lacking. This suggests that PCr/ATP ratio may be a relevant parameter, though not necessarily reflecting mitochondrial capacity. Indeed, the relationship between cardiac energy status measured with PCr/ATP ratio and cardiac function parameters has been found previously both in diabetes (12, 29) and heart failure (16, 35, 36) and seems consistent throughout studies.
The average value of the PCr/ATP ratio, was low in our study (average PCr/ATP ratio was 1.0 ± 0.3) compared to earlier studies (PCr/ATP ratio 1.7–2.3 (12, 14, 37) in healthy controls; and 1.5 (12, 14) or even 1.9 (38) in T2DM, and around 1.6 (16) in patients with dilated cardiomyopathy). Differences in correction for T1 relaxation and fitting routines may partially explain such differences. Furthermore, the scan technique we used, 3D ISIS, is known to produce lower PCr/ATP ratios compared to 1D CSI and 1D ISIS (39), which is probably due to less contamination in the 3D technique from non-cardiac muscle tissue such as diaphragm or chest wall muscle with high PCr content. However, these methodological issues would not hamper the conclusion regarding the lack of correlation between PCr/ATP ratio and mitochondrial function. In addition, we here study a population with known cardiac disease, which may also partly explain a lower cardiac energy status.
The main limitation of the current study is that we measured PCr/ATP ratio in the LV of the heart, whereas we measured respiratory capacity in tissue of the right atrial appendage. This was the case because respiration measurements in ventricular tissue specimens would require a LV biopsy, which is a very invasive procedure, while the right atrial appendage can be obtained during open heart surgery without any extra risk for the patient. Importantly, Lemieux et al. (1) previously showed very good agreement between respiratory capacity of the LV and the right atrial appendage. In addition, our measurements of mitochondrial oxidative capacity were within normal range compared to literature (1, 40). Therefore, we are confident that the respiratory rates measured in tissue from the right atrial appendage are characteristic for the whole myocardium and therefore reflect respiratory rates in the LV.