Elsevier

Metabolism

Volume 63, Issue 11, November 2014, Pages 1375-1384
Metabolism

Review
Using positron emission tomography to study human ketone body metabolism: A review

https://doi.org/10.1016/j.metabol.2014.08.001Get rights and content

Abstract

Ketone bodies – 3-hydroxybutyrate and acetoacetate – are important fuel substrates, which can be oxidized by most tissues in the body. They are synthesized in the liver and are derived from fatty acids released from adipose tissue. Intriguingly, under conditions of stress such as fasting, arterio-venous catheterization studies have shown that the brain switches from the use of almost 100% glucose to the use of > 50–60% ketone bodies. A similar adaptive mechanism is observed in the heart, where fasting induces a shift toward ketone body uptake that provides the myocardium with an alternate fuel source and also favorably affects myocardial contractility. Within the past years there has been a renewed interest in ketone bodies and the possible beneficial effects of fasting/semi-fasting/exercising and other “ketogenic” regimens have received much attention. In this perspective, it is promising that positron emission tomography (PET) techniques with isotopically labeled ketone bodies, fatty acids and glucose offer an opportunity to study interactions between ketone body, fatty acid and glucose metabolism in tissues such as the brain and heart. PET scans are non-invasive and thus eliminates the need to place catheters in vascular territories not easily accessible. The short half-life of e.g. 11C-labeled PET tracers even allows multiple scans on the same study day and reduces the total radiation burden associated with the procedure. This short review aims to give an overview of current knowledge on ketone body metabolism obtained by PET studies and discusses the methodological challenges and perspectives involved in PET ketone body research.

Introduction

Ketone bodies are produced by the liver when excess acetyl-CoA accumulates as a result of increased fatty acid oxidation and decreased glucose oxidation, and serve as important fuel sources during metabolic stress such as fasting and starvation. Ketogenesis is regulated by the supply of ketogenic precursors, primarily free fatty acids (FFAs), to the liver and by the ratio between insulin (inhibiting) and glucagon (stimulating) in the portal bed [1], [2]. Ketone bodies per se substantially reduce circulating levels of FFAs and glycerol, compatible with a regulated feed-back inhibition of lipolysis and decreased hepatic supply of ketogenic substrate [3]. Ketone bodies are exported to the bloodstream and metabolized by extrahepatic tissues to acetyl-CoA which enters the tricarboxylic acid cycle, enabling ATP production independently of glycolysis. This results in lower oxygen consumption per mole of ATP produced, compared to glucose [4]. The principal ketone bodies in humans are beta-hydroxybutyrate (OHB), acetoacetate and acetone (the latter being largely exhaled unused), although OHB and acetoacetate technically are ketoacids. Ketone bodies have often been thought of as merely the product of a so-called spillover pathway, produced as a result of accumulation of excess acetyl-CoA in the liver in times of decreased carbohydrate availability, and which other tissues then use as an alternative fuel source. This might suggest that ketone bodies are a dispensable part of human energy metabolism, rather than an essential component. However, the fact that ketones are metabolized through biochemical pathways that are evolutionarily conserved in most animals suggests a more important role in energy metabolism. The evolutionary role of ketones in brain is also supported by the fact that in many mammals the brains of young (e.g. suckling rats) have a largely ketone-dependent metabolism to support cerebral development [5]. It is now also thought that, in addition to liver hepatocytes, brain astrocytes can synthesize ketones to be used locally by neurones [6], [7].

In humans, circulating ketone bodies are always present in blood to some extent, albeit at low levels (< 0.1 mmol/L) under most conditions, and increase severalfold during fasting, starvation, carbohydrate restriction and some disease states. Ketones are now known to be not only an important alternative protein-sparing and glucose-sparing fuel source for the brain and other tissues, but have also been found to exert a number of beneficial effects. There is accumulating evidence for neuroprotective effects [8], [9], [10] including anticonvulsant activity, improving cognitive function and motor performance in people with cognitive impairment [11], [12], improving tolerance to insulin-induced hypoglycemia [13], protecting against oxidative stress [14] and decreasing the effects of acute brain injury and ischemic damage [15], [16], [17], as well as antitumoral effect in gliomas [18], [19]. All of which has led to the suggestion that ketones could be used therapeutically for a range of diseases [4], [20], [21], [22], [23], though currently the only widely recognized therapeutic use of ketones is in the form of ketogenic diets for the treatment of epilepsy [24], and the mechanism behind their anticonvulsive activity is incompletely understood [25].

Ketones have long been studied in humans, mainly in connection with diabetic ketoacidosis and with starvation, where their primary physiological function appears to be as a protein-sparing source of energy for extrahepatic tissues. Surprisingly little is known about their role in less extreme physiological conditions, however. Increased ketogenesis resulting in moderate hyperketonemia occurs with short-term fasting regimens such as the 5:2 fast (5 days per week of ad lib eating and 2 days of consuming only 500–600 calories), low-carbohydrate diets, and, to a lesser extent, in controlled diabetes and heart failure. It is unknown to what extent states of sustained moderate hyperketonemia are metabolically equivalent to transitory or pronounced elevations in blood ketone levels, though the evidence so far suggests that the brain responds differently to hyperketonemia depending on duration and onset.

Ketones can be used by a number of organs including brain, heart and skeletal muscle, but not by the liver where they are produced. The brain in particular is a quantitatively important consumer of ketone bodies under hyperketonemia, and is the most extensively investigated organ in relation to ketone metabolism. Despite this, a great deal remains unknown about the cerebral effect of ketones, and about their interplay with other energy pathways, notably glucose and fatty acids. Due to the intrinsic methodological difficulties of studying the human brain at a molecular level, a large proportion of studies on brain ketone metabolism to date has been carried out on animals.

Early catheterization experiments on fasting humans (5–6 weeks) found that the fasting brain uses circulating ketone bodies as a major energy source [26]. Such invasive experiments are hardly reproducible nowadays, not least for ethical reasons, and less invasive methods are required to further study in vivo ketone metabolism in humans. Some evidence is derived from in vitro data on cell cultures, which may not sufficiently reflect how processes occur in the living brain, where complex intercellular trafficking and signaling pathways, particularly between neurons and astrocytes, play a central role in metabolism. It is therefore important to develop usable methods for studying such processes in vivo. Positron emission tomography (PET) has a number of advantages that make it promising as a tool to further our understanding of human ketone metabolism in a range of physiological conditions. A recent case report concerning an epileptic child treated with a ketogenic diet [27] highlights the profound shift in cerebral fuel preference away from glucose and presumably toward ketone bodies that is observed in humans with a sustained hyperketonemia. The case report also demonstrates the extent to which this reduction is apparent using 18F-FDG brain PET. The precise effect of ketones on the brain is however unclear, and experimental results are at times contradictory.

This article aims to review the contribution of PET to ketone metabolism research, first by providing a short overview of the general role of PET in human metabolic research, then discussing the experimental evidence obtained thus far in both humans and animals, with particular focus on the role of previous PET brain studies and how this modality can help to address remaining unsolved questions regarding human ketone metabolism.

In recent years PET has made a significant contribution to human metabolic research, enabled by the development of an increasingly large range of targeted positron emitting tracers for in vivo visualization and measurement of metabolic activity. The advantages of PET for metabolic research are numerous: it is non-invasive, quantitative, dynamic (providing kinetic data) and can be performed under close to physiological conditions with minimal patient discomfort. It is also becoming increasingly available and is relatively low risk. One of the clearest advantages of imaging-based approaches to metabolic research is that they allow regional non-invasive measurements of molecular processes which could otherwise only be studied through more invasive methods such as biopsy or arteriovenous difference measurements. This makes PET a particularly attractive option for organs inaccessible to biopsy such as the brain, heart and liver. Although e.g. 11C-palmitate PET has been used to study cardiac metabolism since the 80s [28], recent years have witnessed a great increase in metabolic PET studies as well as an increasing variety of tracers. For example, PET has recently been used to study brain fatty acid metabolism, using labeled heptadecanoate (18F-FTHA) [29] and docosahexanoic acid (11C-DHA) [30], providing interesting discoveries such as increased brain fatty acid uptake in humans with metabolic syndrome [29] and in chronic alcoholics [30].

The development of ketone PET tracers [31], [32] likewise makes PET an increasingly relevant tool for addressing human ketone metabolism (see Fig. 1). The possibility of making regional measurements is particularly useful in the brain, where function is highly regionalized. The size of the individual regions where tracer activity can be measured is limited chiefly by the spatial resolution of PET, although this can be expected to increase in the future. An 11C-acetoacetate biodistribution study performed in mice [33] shows relatively even distribution of the tracer in most organs, although the kidneys receive a somewhat higher concentration in the early phases after injection. The uptake is relatively rapid, with highest uptake of 11C-acetoacetate between 5 and 30 minutes after injection, and rapid elimination thereafter.

While there are published human studies on the effect of fasting on brain uptake of FDG using PET from as early as 1989 [34], there have subsequently only been few studies investigating the independent effect of ketosis on brain metabolism. In particular, only a handful of studies have specifically investigated the quantitative relationship between glucose and ketone metabolism. This may be due to the fact that an appropriate ketone tracer for use in humans has long been unavailable. There are to date two published human studies using a ketone tracer, 11C-betahydroxybutyrate [35], [36], but no dual-tracer studies. In animals, rodent studies have been performed using a ketone tracer (11C-acetoacetate) alone [33], [37] or as dual-tracer studies with 18F-FDG [38], [39], as well as some FDG-only studies to assess glucose metabolism under ketosis [40], [41]. Unsurprisingly, they are primarily focused on the brain, with a few studies using other tissues such as heart for comparison.

It is established that ketones, while able to diffuse across cell membranes, are transported into the brain across the blood–brain barrier by monocarboxylic acid transporters (MCTs), which also transport pyruvate, lactate and alpha-ketoacids (see Fig. 2). There are two main subtypes of MCT in the brain, MCT1 in endothelial cells and MCT2 in neurons and glial cells. Brain ketone uptake consistently shows a roughly linear relationship to blood ketone levels in both rats and humans [35], [36], [38], with low uptake at relatively low levels of ketonemia in postprandial states and severalfold increased uptake in hyperketonemic states. In rats, ketone metabolism accounts for up to a quarter of brain O2 consumption after 48 hours fasting [42]. Furthermore, brain ketone uptake increases with increasing duration of fasting [43], with a twofold increase in transport rate after 5 days of fasting, suggesting an induction mechanism in response to prolonged hyperketonemia.

Cerebral metabolic rate of ketones has been measured in a handful of studies on both humans and rats using dynamic PET with the ketone tracers 11C-betahydroxybutyrate and 11C-acetoacetate, respectively [35], [36], [37], [38], [39] (see Fig. 3). Additional studies using more traditional methods such as arteriovenous difference measurements provide similar kinetic results for ketones to those obtained from dynamic PET [44], confirming that PET is an adequate method to replace more invasive techniques for quantitation of cerebral ketone metabolism.

Various models and analysis methods have been used to characterize the kinetics of ketone metabolism, including Gjedde–Patlak graphical analysis and compartment models (Fig. 3). All methods of analysis yield similar results at both normoketonemia and acute hyperketonemia [35], [36], suggesting that the transport of ketones across the blood–brain barrier is largely irreversible and that they are quickly metabolized, with efflux (k2 constant) contributing relatively little to the equation. Both ketogenic diets and short-term fasting result in increased cerebral metabolic rate of ketones (CMRket), although the increase is greater with fasting, at least in rats [37]. In line with this, fasted rats also have higher k1 (influx) and k2 (efflux) constants for ketone bodies compared to rats on either a ketogenic or standard diet [37]. The increase in CMRket under hyperketonemia has been found by various methods and with various degree of ketonemia to be between 8-fold [37] and 13-fold [44]. At baseline, in healthy humans on a standard diet and after an overnight fast, with circulating beta-hydroxybutyrate levels of 0.04 mmol/L, the CMRket has been found to be approximately 0.5 nmol/ml/min in the brain as a whole, with cortex showing higher uptake than white matter [35]. Acute ketone infusion producing blood ketone levels of around 1 mmol/L also results in an increase in CMRket, with no significant CMRket differences between healthy and diabetic humans [36]. The relationship between circulating ketone levels and CMRket remains roughly linear both at baseline (low ketone levels) and at higher ketone levels. This suggests that the uptake of ketones does not saturate at physiological ranges of ketonemia (contrary to glucose uptake which tends to decrease with increased blood glucose levels), and that transport through MCTs across the blood–brain barrier is thereby the rate-limiting step in their uptake and metabolization.

In-vivo magnetic resonance (MR) spectroscopy and polarization transfer spectroscopy have been used to non-invasively measure tissue concentration of ketones in human brain during both fasting and acute hyperketonemia [45], [46], [47], revealing some interesting differences in how the brain responds to acute hyperketonemia versus fasting. During fasting, with blood OHB levels climbing to 3–3.5 mmol/L after 3 days, brain tissue OHB concentration increases to 1.1 mmol/L after 3 days. With acute infusion of sodium OHB resulting in circulating levels of 2–3 mmol/L, intracerebral OHB concentration is only 0.24 mmol/L in the steady state (from 0.16 mmol/L at baseline) [45], [47]. That is, brain OHB concentration is considerably lower with acute ketone infusion compared to fasting, despite comparable ketonemia. While this shows that the brain is indeed able to acutely take up ketones, it also confirms that ketone uptake capacity is largely inducible, with full capacity being reached only after an adaptation period of sustained hyperketonemia, such as with a ketogenic diet or prolonged fasting. This is in line with rat studies showing upregulation of MCT-1 in rat brain after several weeks on a ketogenic diet [38], [48], though not after a 48-hour fast [38]. It is at present unclear how long the adaptation period is and whether there are individual or species-related differences, but the fact that there are changes in the brain uptake index of ketone bodies already during a 5-day fast in rats [43] suggests that the changes occur relatively early.

While MR spectroscopy does not, unlike PET, provide kinetic parameters, it has one clear advantage over PET – namely, that spectroscopy distinguishes chemical species, such that tissue concentrations of OHB and AcAc are measured separately. Using PET with a single ketone tracer, be it 11C-acetoacetate or 11C-betahydroxybutyrate, provides only a measure of total ketone body concentration and kinetics, since the two ketone bodies are interconvertible by the enzyme betahydroxybutyrate dehydrogenase and some of the injected tracer will thus be converted to the other ketone body [36]. Betahydroxybutyrate tends to predominate in hyperketonemic states, with concentrations severalfold higher than those of acetoacetate (the circulating OHB:AcAc ratio is approximately 1:1 postprandially, but increases up to 6:1 with fasting), regardless of which tracer is used. Should the two species have different uptake kinetics, however, this may affect results of PET studies depending on whether they use labeled AcAc or OHB as a tracer.

The predominance of animal studies on ketone metabolism should be taken into account when interpreting results, since there are potentially a number of caveats that limit their extrapolability to humans. Firstly, the rodent brain appears to be less efficient at utilizing ketones than the human brain [42], [49]. Some early data suggested that fasted rat brain can derive around 20% of energy requirements from ketones under fasting [34], [42], whereas the human brain may derive as much as 60% from ketones with prolonged fasting [26]. Secondly, it is unclear to what extent animal models of hyperketonemia, using fasting or ketogenic diets, simulate the human metabolic response to the same interventions. Thus dietary ketosis using either short-term fasting or ketogenic diet produces a 7–8 fold increase in brain ketone use in rats [37], whereas short-term fasting results in up to a thirteenfold increase in humans [44]. Thirdly, some animals lack key enzymes involved in ketogenesis and may thus not produce ketone bodies at all. This is most obviously illustrated in pigs, that lack the mitochondrial enzyme HMG-CoA lyase and thus do not produce ketone bodies. Finally, blood ketone concentrations in fasting humans increase more than in rats fasting for the same period of time.

In addition to these species inherent differences, experimental setups vary significantly between human and animal studies. The latter are generally performed under general anesthesia with isoflurane or halothane, which likely have an effect on brain energy metabolism. One study [50] assessed both anesthetized and awake rats before and after intravenous infusion of sodium betahydroxybutyrate, and found that while none of the groups showed increased CMRglc or CMRO2 at hyperketonemia, global cerebral blood flow (CBF) increased at hyperketonemia by different amounts depending on the anesthetic used (greater increase with halothane compared to pentobarbital), as well as lower CBF and CMRO2 at baseline in rats anesthetized with pentobarbital compared to non-anesthetized rats. The different effects of particular anesthetic compounds on brain metabolism, and potentially different effects on the cerebral response to hyperketonemia, should be taken into account when comparing rat studies with differing methodologies.

A further factor which can complicate comparability between studies using ketogenic diets is the exact composition of the diets used. There is no universally accepted definition of the ketogenic diet, other than one which results in some degree of hyperketonemia, and thus the exact macronutrient ratios (fat:protein:carbohydrate, as a proportion of total calories) can vary considerably, even between various commercially available ketogenic diet preparations used in animal studies. It is unclear to what extent such differences affect the variables measured, but it is not unlikely that the amount of carbohydrate in the diets can considerably affect the degree of ketonemia and/or glycemia and thereby the experimental results. A strict hyperketogenic diet as prescribed for control of refractory epilepsy in humans can result in blood ketone levels of up to 12 mmol/L [51], whereas more moderate ketogenic diets with higher protein and/or carbohydrate content typically result in ketonemia around 4–6 mmol/L (reviewed in [52]).

Thus the evidence collectively indicates that a) uptake of ketones by the brain is directly related to blood ketone levels, b) can be considered irreversible, c) is non-saturable at physiological ranges of ketonemia and d) it is this initial uptake, i.e. transport across the capillary blood–brain barrier, that is the limiting factor in cerebral ketone metabolism.

Less explored is the relationship between glucose and ketone metabolism. Using 18F-FDG PET, some studies with diet-induced hyperketonemia (fasting or ketogenic diet) report decreased glucose uptake [40], [41], [44], whereas others on the contrary report that such dietary regimens stimulate glucose uptake [38], [39]. Acute and experimental hyperketonemia in humans has also been found to result in reduced cerebral glucose metabolism assessed by 18F-FDG PET [53]. Rats under sustained hyperketonemia appear to develop both increased brain capillary density [54], [55] and upregulation of GLUT-1 transporters in brain endothelial cells [48], which may serve to explain the counterintuitive concomitant increase in brain glucose uptake despite the expected shift from glucose-based to ketone-based oxidative brain metabolism. However, both rat studies reporting increased 18F-FDG uptake were hampered by methodological flaws: one was performed on aged rats [39], which may very well have altered cerebral metabolism, and the other reported only semi-quantitative standard uptake values (SUV) values rather than true kinetic data [38]. Human studies using arteriovenous difference measurements on fasted subjects [34], [44] have shown significantly decreased cerebral metabolic rate of glucose (CMRglc) after both 3 weeks and 3.5 days fasting. However, in the case of a 3.5-days fast, the k1 constant of glucose (reflecting relative uptake into the brain) was increased [44], whereas it was found unchanged after a 3-week fast [34]. In both cases, significantly decreased plasma glucose levels rather than decreased relative uptake (k1) served to produce an overall decrease in cerebral glucose metabolism (see Fig. 3).

Interestingly, another study on humans fasting for 3.5 days found increased blood–brain barrier (BBB) permeability to glucose after fasting [56]. Blood glucose concentrations decreased, and the permeability surface-area product for glucose transport into the brain across the BBB increased by 55%, which was 33% more than expected from calculations based on decreased glycemia alone. These data suggest that glucose transport across the BBB is upregulated during fasting, and are in agreement with the previous finding of increased relative glucose uptake (k1). There was no significant change in the permeability surface-area product for transport in the other direction, i.e. out of the brain into the blood. It is unclear how increased BBB transport and increased glucose uptake can be reconciled with a decreased CMRglc as shown by other human studies, but it may suggest decreased glycolysis and accumulation of unmetabolized glucose, an idea also supported by the finding of markedly decreased k3 constant after prolonged fasting [34]. On the other hand, BBB permeability to betahydroxybutyrate did not decrease as might be expected from the increased blood ketone levels, again suggesting that ketone transport across the BBB is non-saturable at physiological ranges of ketonemia, and that blood ketone concentration is the main determinant of ketone transport rate into the brain, with ketones being rapidly metabolized thereafter.

A recent meta-analysis of data from 11 previous rat and human studies reporting CMRglc under hyperketonemia [41] found a linear relationship between CMRglc and total blood ketone body concentration, with an approximately 9% decrease in CMRglc for each mmol/L increase in blood ketone concentration, though the analysis did not include data from the two rat studies reporting increased CMRglc. Thus, most evidence points toward decreased CMRglc under sustained ketosis, with some remaining discrepancy. There is considerable methodological variability between studies, with varying degrees of ketonemia, and the difficulty in accurately detecting small changes in CMRglc at lower ranges of ketosis may explain some of the discrepancies. Markedly decreased cerebral glucose uptake is supported by a recent case study [27] showing decreased brain 18F-FDG uptake with sparing of the basal ganglia in an 11-year old child following a ketogenic diet to prevent epilepsy (see Fig. 4). However, age-related differences in ketone metabolism and ketonemia exist between children and adults [57], and such a dramatic effect may not be observed in adults.

Importantly, there are to date only few studies investigating the effect of acute ketone infusion on CMRglc. Rats given an acute infusion of sodium OHB resulting in circulating ketone levels of 6 mmol/L have unchanged CMRglc [50], suggesting that the effect on brain glucose metabolism is indeed different with sudden versus sustained hyperketonemia. Whether this is also the case in humans subjected to a ketone infusion under euglycemia remains to be investigated. Most acute ketone infusion studies performed in humans (e.g. [45], [47]) have focused on brain ketone metabolism and omitted measurements of concomitant changes in glucose consumption. One human study [53] using intravenous administration of ketone bodies resulting in mean blood plasma betahydroxybutyrate levels of 2.16 mmol/L reported a 33% decrease in cerebral glucose metabolism as determined by Kety–Schmidt technique, but no subsequent studies have been performed to confirm this.

Considering that the balance between glucose and ketone metabolism is likely to be of evolutionary significance in the setting of human metabolic adaptations, and that most natural, physiological hyperketonemia occurs gradually and is sustained over time (fasting, starvation, carbohydrate deprivation), it can be expected that the brain and other organs have evolved adaptive mechanisms for dealing with sustained hyperketonemia but not necessarily with acute hyperketonemia. The latter may have its uses in modern medicine, however, if acute ketone administration is used therapeutically (e.g. for preventing protein catabolism in acutely ill patients or limiting damage in acute brain injury). Understanding the different metabolic effects of acute versus sustained hyperketonemia is therefore of clear importance.

There is some evidence that fasting and ketogenic diets can have beneficial effects on cognitive function [11] but the mechanism is unknown, and it is unclear whether the improvement can be directly related to increased neuronal oxidation of ketone bodies, or if it is due to a more indirect mechanism. The fact that brain glucose metabolism and mitochondrial function seems to be decreased in humans with cognitive impairment [58] suggests that the cognitively impaired brain could potentially benefit from an alternative energy source such as ketones.

PET may prove useful in elucidating this, by providing a means to image and measure brain metabolism in real time while subjects perform cognitive tasks or tests. No such studies in humans or animals have been published. A ketogenic diet has been shown to improve motor performance in a rat model of Alzheimer's disease, but not cognitive performance [12]. While PET studies on glucose and ketone metabolism have been performed on rats comparing aged and younger subjects, they have not investigated functional correlation. A dual-tracer study [39] comparing brain ketone (11C-acetoacetate) and glucose (18F-FDG) uptake in young 4-month and older 21-month rats found higher FDG uptake in the hippocampus of the aged rats compared to younger, and a lower percentage contribution of acetoacetate to total brain energy substrate uptake (11C-Acetoacetate + 18F-FDG), albeit ignoring the contribution of other possible energy substrates such as fatty acids and lactate. Aged 22-month rats do show improved cognitive performance after a 3-week ketogenic diet [55] together with increased capillary density, but this has not been correlated with changes in brain ketone or glucose metabolism.

A great deal is known about cardiac metabolism of fatty acids, lactate and glucose, but much less is known about the effect of ketones on the heart and their interplay with other myocardial fuel substrates. It is well documented that fatty acids are the preferred myocardial fuel substrate, with a shift toward increased use of glucose, lactate and ketones during acutely increased demand [59], [60]. This can be altered in various pathologies – the diabetic heart, for instance, relies more on fatty acid than glucose as oxidative fuel source[61], and in vitro studies suggest that ketone bodies inhibit fatty acid oxidation in diabetic cardiomyocytes [62], [63], which also appear to have impaired ketone oxidation when compared with healthy cardiomyocytes [62]. Interestingly, acute ketone infusion in non-diabetic pigs inhibits myocardial fatty acid oxidation, but not glucose oxidation, when administered together with a fat emulsion [64]. In vitro studies further suggest that prolonged exposure to ketones decreases myocardial glucose uptake in a concentration-dependent manner through inhibition of various signaling pathways, in particular insulin-dependent uptake, thus possibly promoting myocardial insulin resistance [65], [66], [67]. At the same time, however, an opposing effect has been shown with shorter exposure to ketone bodies (betahydroxybutyrate and acetoacetate in a 4:1 ratio) together with insulin, a combination that appears to potentiate the metabolic effects of insulin on cardiomyocytes reflected by decreased O2 use and increased mechanic efficiency [68]. Ketone bodies also decrease lactate oxidation in a concentration-dependent manner and decrease palmitate oxidation but not oxidation of octanoate, a medium chain fatty acid [62]. Controversy surrounds the effect of ketone bodies on myocardial contractility, since some studies report increased contractility [69] whereas others in contrast indicate that ketone bodies as the sole energy source result in decreased myocardial contractility [70]. These contradicting findings have not been further investigated in vivo in humans. It is therefore unclear to what extent ketones may contribute to myocardial metabolism in conditions of hyperketonemia, how this affects contractility, and how the effect varies with disease states such as congestive heart failure and diabetes. PET may be of help to answer these questions, by allowing non-invasive in vivo quantification of cardiac substrate metabolism in hyperketonemic subjects with simultaneous measurement of cardiac contractility.

There are to date no published human studies on myocardial metabolism during hyperketonemia. Some animal studies, while focusing on the brain, have however included heart scans for comparison [37]. Quantified by dynamic PET with an 11C-acetoacetate tracer, rat brain has a significantly higher ketone body metabolic rate (CMRket) than rat heart, regardless of whether hyperketonemia is achieved by fasting or a ketogenic diet. On the other hand, this higher cerebral/cardiac ratio in ketone body metabolic rate is not observed in animals fed a standard diet. This may be explained by higher k1 (influx) and k2 (efflux) rate constants for ketone bodies in the heart compared to the brain in animals on a standard diet, a difference not observed in hyperketonemic states. The myocardium thus appears to be more permeable to ketone bodies at physiological conditions, whereas the brain shows greater adaptability to ketone uptake and metabolism when conditions become hyperketonemic. Interestingly, both cardiac and cerebral k1 and k2 were lowest during fasting compared to ketogenic diet and controls, whereas both brain and heart MRket were highest in fasting rats. Given that blood glucose levels were lowest and ketone levels highest during fasting compared to ketogenic and standard diet, it may suggest that MRket is most dependent on blood ketone levels (indeed showing a linear relationship to ketonemia), whereas absolute influx and efflux rates are to a greater extent dependent on glucose availability [37]. Such findings provide a foundation for further studies on human myocardial substrate kinetics under ketosis with the use of glucose, ketone and fatty acid PET tracers – the latter being particularly relevant to the heart given its highly fatty acid-dependent metabolism. Labeled fatty acids such as 11C-palmitate have previously been used in humans to measure myocardial fatty acid metabolism [71], [72].

Section snippets

Conclusion

Ketone bodies are increasingly recognized as relevant metabolites in a range of physiological and pathological processes. Far from being the byproduct of an accessory pathway, ketone bodies are an integral part of human energy metabolism, and are also important for the synthesis of non-energetic biomolecules. Despite decades of research, however, many details of human ketone metabolism and especially its relationship with glucose metabolism remain unclear. There have been few dedicated studies

References (72)

  • R.L. Leino et al.

    Diet-induced ketosis increases monocarboxylate transporter (MCT1) levels in rat brain

    Neurochem Int

    (2001)
  • S. Cunnane et al.

    Brain fuel metabolism, aging, and Alzheimer's disease

    Nutrition

    (2011)
  • P. Herrero et al.

    PET detection of the impact of dobutamine on myocardial glucose metabolism in women with type 1 diabetes mellitus

    J Nucl Cardiol

    (2008)
  • Y. Kashiwaya et al.

    Substrate signaling by insulin: a ketone bodies ratio mimics insulin action in heart

    Am J Cardiol

    (1997)
  • Y. Kashiwaya et al.

    Control of glucose utilization in working perfused rat heart

    J Biol Chem

    (1994)
  • J.D. McGarry et al.

    Regulation of ketogenesis and the renaissance of carnitine palmitoyltransferase

    Diabetes Metab Rev

    (1989)
  • M. Beylot

    Regulation of in vivo ketogenesis: role of free fatty acids and control by epinephrine, thyroid hormones, insulin and glucagon

    Diabetes Metab

    (1996)
  • R.L. Veech et al.

    Ketone bodies, potential therapeutic uses

    IUBMB Life

    (2001)
  • S. Rheims et al.

    GABA action in immature neocortical neurons directly depends on the availability of ketone bodies

    J Neurochem

    (2009)
  • N. Auestad et al.

    Fatty acid oxidation and ketogenesis by astrocytes in primary culture

    J Neurochem

    (1991)
  • Y. Kashiwaya et al.

    D-beta-hydroxybutyrate protects neurons in models of Alzheimer's and Parkinson's disease

    Proc Natl Acad Sci U S A

    (2000)
  • M.L. Prins

    Cerebral metabolic adaptation and ketone metabolism after brain injury

    J Cereb Blood Flow Metab

    (2008)
  • M.L. Brownlow et al.

    Ketogenic diet improves motor performance but not cognition in two mouse models of Alzheimer's pathology

    PLoS One

    (2013)
  • T. Veneman et al.

    Effect of hyperketonemia and hyperlacticacidemia on symptoms, cognitive dysfunction, and counterregulatory hormone responses during hypoglycemia in normal humans

    Diabetes

    (1994)
  • Y. Kim do et al.

    Ketone bodies are protective against oxidative stress in neocortical neurons

    J Neurochem

    (2007)
  • M.A. Puchowicz et al.

    Neuroprotection in diet-induced ketotic rat brain after focal ischemia

    J Cereb Blood Flow Metab

    (2008)
  • H. White et al.

    Clinical review: ketones and brain injury

    Crit Care

    (2011)
  • G. Zuccoli et al.

    Metabolic management of glioblastoma multiforme using standard therapy together with a restricted ketogenic diet: case report

    Nutr Metab

    (2010)
  • G.D. Maurer et al.

    Differential utilization of ketone bodies by neurons and glioma cell lines: a rationale for ketogenic diet as experimental glioma therapy

    BMC Cancer

    (2011)
  • Z. Zhao et al.

    A ketogenic diet as a potential novel therapeutic intervention in amyotrophic lateral sclerosis

    BMC Neurosci

    (2006)
  • C.E. Stafstrom et al.

    The ketogenic diet as a treatment paradigm for diverse neurological disorders

    Front Pharmacol

    (2012)
  • A. Paoli et al.

    Beyond weight loss: a review of the therapeutic uses of very-low-carbohydrate (ketogenic) diets

    Eur J Clin Nutr

    (2013)
  • E.H. Kossoff et al.

    Worldwide use of the ketogenic diet

    Epilepsia

    (2005)
  • O.E. Owen et al.

    Brain metabolism during fasting

    J Clin Invest

    (1967)
  • K. Korsholm et al.

    Effects of a ketogenic diet on brain metabolism in epilepsy

    Clin Nucl Med

    (2013)
  • A. Karmi et al.

    Increased brain fatty acid uptake in metabolic syndrome

    Diabetes

    (2010)
  • Cited by (16)

    • Efficacy and safety of D,L-3-hydroxybutyrate (D,L-3-HB) treatment in multiple acyl-CoA dehydrogenase deficiency

      2020, Genetics in Medicine
      Citation Excerpt :

      Despite early diagnosis and treatment, morbidity and mortality remain high in neonatal onset patients.1 Upon prolonged fasting, hepatic mitochondrial FAO fuels synthesis of ketone bodies (KB) acetoacetate and 3-hydroxybutyrate, as important alternative energy sources for the brain, skeletal muscle, and heart.2-4 Patients with mitochondrial FAO disorders, such as MADD, demonstrate multiorgan dysfunction especially during catabolism.2

    • Sleep profile during fasting in PPAR-alpha knockout mice

      2020, Physiology and Behavior
      Citation Excerpt :

      During fasting, the brain switches from the use of glucose to the use of ketone bodies [28]. A similar mechanism has been observed in the heart, in which fasting induces a transition to ketone body utilization from FFA to provide an alternative fuel source for the myocardium [29,30]. In the present study, PPARα KO mice showed fewer ketone bodies (AcAc and BHB) than WT mice during FD because of decreased mRNA expression of Hmgcs2, a target gene of PPARα and a rate-limiting enzyme of ketogenesis, in the liver and the hypothalamus (Fig. 5D, E, G and H).

    View all citing articles on Scopus
    1

    N.B., L.T.A., N.M. and L.C.G. have no disclosures.

    View full text