Circadian Rhythms in Cardiac Metabolic Flexibility

Abstract

Numerous aspects of cardiovascular physiology (e.g., heart rate, blood pressure) and pathology (e.g., myocardial infarction and sudden cardiac death) exhibit time-of-day-dependency. In association with day-night differences in energetic demand and substrate availability, the healthy heart displays remarkable metabolic flexibility, through temporal partitioning the metabolic fate of common substrates (glucose, lipid, amino acids). The purpose of this review is to highlight the contribution that circadian clocks provide towards 24-hr fluctuations in cardiac metabolism, and to discuss whether attenuation and/or augmentation of these metabolic rhythms through adjustment of nutrient intake timing impacts cardiovascular disease development.

Introduction

Metabolic flexibility (defined as the ability to change flux through metabolic pathways in response to physiologic stimuli and pathologic stresses) is critical for homeostatic processes governing normal cellular/organ function (Taegtmeyer, Golfman et al., 2004; Riera & Dillin, 2015). Loss of metabolic flexibility is invariably associated with pathology, through impaired energetics and/or accumulation of metabolites (ultimately perturbing processes involved in transcription, protein turnover, signaling, and even cell survival/death). As an organism ages, metabolic flexibility declines; restoration of metabolic flexibility in aged animals (through, for example, caloric restriction) increases lifespan, underscoring the importance of this attribute (Riera & Dillin, 2015). It is noteworthy that mammals exhibit appreciable metabolic flexibility over a normal 24hr period, at cellular, organ, and whole-body levels. Classically, daily rhythms in metabolism have been ascribed to fluctuations in substrate availability and/or energetic demands that are secondary to fasting/feeding and sleep/wake cycles. In simplistic terms, this adaptation dogma is founded on established signal-transduction and enzyme kinetic mechanisms, wherein alterations in neurohumoral stimuli and carbon for enzymatic reactions, drive changes in metabolic fluxes over the course of the day. More recently, an anticipation model has emerged as a significant contributor to daily perturbations in metabolism. This model is based on an intrinsic timekeeping mechanism (termed the circadian clock; see below) modulating capacity and/or flux through metabolic pathways, in anticipation of predicted daily fluctuations in extrinsic factors (such as substrate availability). In a healthy organism, adaptation and anticipation mechanisms operate in synchrony, thereby optimizing metabolic flexibility. Conversely, impairment or temporal misalignment of either mechanism precipitates metabolic inflexibility (and ultimately pathology). In this article, we review current knowledge regarding 24hr fluctuations in cardiac carbohydrate (e.g., glucose), lipid (e.g., oleate/palmitate), and amino acid (e.g., branched chain amino acid; BCAA) metabolism, the mechanisms responsible for these metabolic rhythms, and their intersection with behaviors (such as timing of nutrient intake).

Circadian Clocks

Prior to discussing 24hr rhythms in cardiac metabolism, it is important to briefly introduce the mammalian circadian clock. Circadian clocks are cell autonomous molecular mechanisms that are found within virtually all mammalian cells, including cardiomyocytes, fibroblasts, vascular smooth muscle cells, and endothelial cells (Balsalobre, Damiola et al., 1998; McNamara, Seo et al., 2001; Durgan, Hotze et al., 2005; Takeda, Maemura et al., 2007; Takahashi, Hong et al., 2008). They are composed of a series of transcriptional-translational feedback loops, with a periodicity of approximately 24hr (Takahashi, Hong et al., 2008). At the core of the mechanism are circadian locomotor output cycles kaput (CLOCK) and brain and muscle ARNT-like 1 (BMAL1) that, upon heterodimerization, have the ability to bind to E-boxes within target genes leading to their transcription (Gekakis, Staknis et al., 1998; Hogenesch, Gu et al., 1998). Direct target genes encoding for multiple negative feedback components, include period (PER1/2/3), cryptochrome (CRY1/2) and REV-ERB (REV-ERBα/β) isoforms; the PER/CRY heterodimer physically interacts with and repress CLOCK/BMAL1, while REV-ERBα/β represses transcription of the Bmal1 gene (Kume, Zylka et al., 1999; Shearman, Sriram et al., 2000; Preitner, Damiola et al., 2002). This cyclical transcription and repression leads to rhythms in clock-controlled / output genes (Durgan, Hotze et al., 2005; Takahashi, Hong et al., 2008). It has been estimated that 3% to 16% of an organ’s transcriptome is circadian clock controlled (Zhang, Lahens et al., 2014). Rodent-based studies reveal that ~10% of the cardiac transcriptome is regulated by the cardiomyocyte circadian clock (CCC); these genes have established functions in a diverse number of processes, including cellular transport, signal transduction, and metabolism (Bray, Shaw et al., 2008; Young, Brewer et al., 2014). The latter is the primary focus of the current review.

It is noteworthy that circadian clocks in multiple cell types have the potential to modulate cardiac metabolism over a normal 24-hr period. One of the most well characterized mammalian circadian clocks is located within the suprachiasmatic nucleus (SCN; a collection of approximately 20,000 neurons within the hypothalamus) (Antle & Silver, 2005). The SCN (so called central/master) clock perceives light via the retinohypothalamic track, and propagates this signal to clocks within other cells/organs (termed peripheral/slave clocks) via neurohumoral signals, thereby facilitating whole-body temporal synchronization with the environment (Hirota & Fukada, 2004). Circadian clocks within both the central nervous system and periphery in turn modulate cardiovascular-relevant stimuli/stresses over the course of the day, by regulating the timing of behaviors (sleep/wake and fasting/feeding cycles), sympathetic and autonomic tone, endocrine factor secretion, digestion and absorption, as well as metabolic homeostasis and circulating substrate levels (Takahashi, Hong et al., 2008; Gamble, Berry et al., 2014; Rana, Prabhu et al., 2020). The latter is illustrated by the hepatocyte circadian clock, which affects circulating glucose, lipid, and amino acid levels, via temporal regulation of gluconeogenesis, and VLDL production/secretion respectively (Lamia, Storch et al., 2008; Jeyaraj, Scheer et al., 2012; Pan, Bradfield et al., 2016). For example, the hepatocyte circadian clock modulates gluconeogenesis through CRY dependent regulation of cAMP response element binding protein (Zhang, Liu et al., 2010). The skeletal myocyte clock also influences glucose homeostasis through temporal governance of muscle insulin sensitivity and glucose utilization (Harfmann, Schroder et al., 2016). It is therefore apparent that an interplay between numerous extra-cardiac clocks (in liver, skeletal muscle, adipose tissue, pancreatic β-cells, intestinal epithelial cells, etc) collectively modulate nutrient availability for utilization by the myocardium, as well as neurohumoral signals (e.g., insulin) that differentially influence anabolic versus catabolic pathways. The interaction between nutrient (glucose, lipid, and amino acid) availability and the CCC will be a primary focus of the following sections.

Glucose Metabolism

The heart must contend with two primary behavioral cycles over the course of the day; namely sleep/wake and fasting/feeding cycles. Both impact myocardial metabolism. In simplistic terms, increased physical activity and increased food intake during the awake period will challenge the heart with increased workload and postprandial hormones (e.g., insulin). Both are known to promote myocardial glucose utilization (Goodwin, Taylor et al., 1998; Lloyd, Brocks et al., 2003). It is therefore not surprising that multiple studies report augmentation of cardiac glucose oxidative and non-oxidative metabolism in rodent hearts during the awake period (approximately 2-fold higher, relative to the sleep period) (Young, Razeghi et al., 2001; Bray, Shaw et al., 2008; Durgan, Pat et al., 2011; Brewer, Collins et al., 2018). Here, we address the relative contribution of extrinsic (e.g., workload, neurohumoral factors, nutrients) versus intrinsic (i.e., CCC) factors as mediators of daily rhythms in myocardial glucose utilization.

In humans, blood glucose levels transiently rise following consumption of carbohydrate containing meals(Wolever & Bolognesi, 1996). However, the magnitude of this meal response is time-of-day-dependent. More specifically, blood glucose levels rise to a greater extent in response to a carbohydrate meal in the evening (i.e., dinner) relative to the same meal in the afternoon (i.e., lunch)(Pearce, Noakes et al., 2008). Continuous infusion of glucose in healthy volunteers, thus bypassing digestion/absorption, results in hyperglycemia during the night, indicative of decreased glucose tolerance at this time(Van Cauter, Desir et al., 1989). Despite augmented skeletal muscle insulin sensitivity immediately prior to waking, blood glucose levels rise immediately prior to waking (termed the dawn phenomenon), secondary to increased hepatic glucose output (i.e., gluconeogenesis) (Bolli, De Feo et al., 1984; Bolli & Gerich, 1984). Twenty four hour rhythms in glucose homeostatic mechanisms persist when individuals are subjected to constant conditions or cyclic conditions outside the normal entrainment range (e.g., 28hr day; termed a desynchrony model) (Scheer, Hilton et al., 2009). Light dark cycle manipulations known to disrupt circadian clocks, as well as shift work and BMAL1 polymorphisms, impair glucose tolerance and increase diabetes mellitus risk in humans (Marcheva, Ramsey et al., 2010; Lee, Kim et al., 2011; Lee, Moulik et al., 2013; Gan, Yang et al., 2015). Glucose tolerance also exhibits diurnal variations in rodent models. Rodents eat smaller, more frequent meals (as opposed to 3 meals per day in humans), with approximately two thirds daily food intake during the more active (dark) phase(Stephan, 2002; Bray, Ratcliffe et al., 2013). However, blood glucose levels increase during the less active (light) period (when glucose tolerance is low) (la Fleur, Kalsbeek et al., 2001; Ando, Ushijima et al., 2016). These rhythms persist during fasting and are therefore independent of daily feeding patterns (la Fleur, Kalsbeek et al., 2001). Importantly, disruption of circadian clocks through surgical (e.g., SCN ablation), environmental (e.g., light/dark cycle manipulation), and/or genetic (e.g., BMAL1/CLOCK manipulation) means impairs various glucose homeostasis parameters in mice (la Fleur, Kalsbeek et al., 2001; Barclay, Husse et al., 2012; Ando, Ushijima et al., 2016; Kolbe, Leinweber et al., 2019). Collectively, these observations suggest that circadian clocks decrease glucose tolerance during the sleep period, potentially to prevent hypoglycemia during this period of reduced food intake and to anticipate increased glucose demand (from, for example, contracting skeletal and cardiac muscle) immediately upon awaking.

Although extracardiac influences (i.e., plasma glucose and insulin levels, as well as workload) likely contribute to 24hr rhythms in myocardial glucose utilization, evidence has accumulated indicating mediation by an intracardiac mechanism. When examining glucose utilization in the ex vivo perfused rodent heart (during which glucose, insulin, and workload are constant), day-night differences in glucose utilization persist (Young, Razeghi et al., 2001; Bray, Shaw et al., 2008; Durgan, Pat et al., 2011; Brewer, Collins et al., 2018). Moreover, glucose utilization rhythms are abolished when the circadian clock is genetically ablated specifically in cardiomyocytes (during which behaviors and the neurohumoral milieu are unaffected); this timekeeping mechanism appears to augment myocardial glucose uptake, glycolysis, pyruvate oxidation, glycogen synthesis, and protein O-GlcNAcylation during the active period (Durgan, Pat et al., 2011). Although the precise mechanisms by which the CCC temporally governs cardiac glucose utilization remain undefined, multiple candidates have been proposed. Given the relatively close temporal alignment of oxidative and non-oxidative glucose metabolic pathways in the heart, clock control of glucose uptake has been a major focal point. The uptake of glucose into cardiomyocytes is primarily dependent on two glucose transports; GLUT1 and GLUT4 (Aerni-Flessner, Abi-Jaoude et al., 2012). At the mRNA level, both isoforms exhibit 24hr oscillations in the heart (peaking at the beginning of the active period), which are dependent on the CCC (although GLUT1 and GLUT4 protein level assessments have not revealed consistent day-night differences) (Durgan, Pat et al., 2011). Glucose transport is also regulated through posttranslational events, including translocation of GLUT4 from intracellular vesicles to the sarcolemmel membrane; this process is stimulated by both contraction- and insulin- mediated pathways (Shao & Tian, 2015). AMP-activated protein kinase (AMPK) plays an important regulatory role in the contraction-mediate pathway; both the phosphorylation status and activity of AMPK oscillates in the heart in a CCC-dependent manner, mirroring the rhythm in glucose utilization (peaking in the middle of the active period) (Tsai, Kienesberger et al., 2010). It is noteworthy that AMPK also stimulates phosphofructokinase (Marsin, Bertrand et al., 2000), and may therefore simultaneously augment myocardial glycolytic flux during the active period. Although insulin signaling fluctuates in the heart over the course of the day, peak phosphorylation status of key insulin signaling components (e.g., Akt, AS160) is approximately 6hr phase advanced relative to myocardial glucose uptake (McGinnis, Tang et al., 2017). Interestingly, intrinsic cardiac insulin sensitivity does not appear to fluctuate over the course of the day; instead phosphorylation status of Akt and AS160 in hearts of ad libitum fed mice correlate closely with circulating insulin levels (McGinnis, Tang et al., 2017). An important crossroads between glycolysis and the TCA cycle is undoubtedly oxidative decarboxylation of pyruvate, via the pyruvate dehydrogenase complex (PDC); PDC activity is augmented in the heart during the active period, in synchrony with other glucose utilization parameters (Young, 2006). PDC activity is dependent on phosphorylation (leading to inhibition) secondary to a balance between pyruvate dehydrogenase kinases (PDKs) and a Ca²⁺-activated phosphatase; interestingly, PDK4 levels and Ca²⁺ homeostatic mechanisms both fluctuate in the heart with a periodicity of 24hr (Stavinoha, RaySpellicy et al., 2004; Collins & Rodrigo, 2010; Sachan, Dey et al., 2011; Wang, Tapa et al., 2020). With regards to non-oxidative metabolism of the glucose, myocardial glycogen levels oscillate, even during fasting (albeit at a lower amplitude) (Brewer, Collins et al., 2018). Glycogen synthesis rates are antiphase (i.e., 12hr phase shift) relative to oscillations of in inhibitor-1 (an established glycogen synthase inhibitor); 24hr rhythms in both glycogen synthesis and inhibitor-1 are abolished when the CCC is disrupted (Durgan, Pat et al., 2011; McGinnis, Tang et al., 2017). It is noteworthy that depressed cardiac glycogen synthesis following genetic CCC disruption is normalized by pharmacologic activation of REV-ERBα/β (Mia, Kane et al., 2020). Carbon from the glycolytic pathway can be shunted into the hexosamine biosynthetic pathway (HBP) via glutamine:fructose 6-phosphate amidotransferase (GFAT); UDP-glucosamine (end product of the HBP) is subsequently utilized by O-GlcNAc transferase (OGT) for protein O-GlcNAcylation (a post-translational modification known to affect stability, activity, and/or cellular localization of numerous proteins) (Wright, Collins et al., 2017). Both GFAT2 and OGT are regulated by the CCC; increased cardiac OGT protein levels during the active period align with increased protein O-GlcNAcylation at this time (Durgan, Pat et al., 2011). Interestingly, glutamine synthetase also oscillates in the heart in a CCC-dependent manner; increased myocardial glutamine synthesis during the active period may potentially provide substrate for the GFAT reaction (Durgan, Pat et al., 2011).

In summary, myocardial glucose utilization is augmented during the active period, a time at which both physical activity and food intake typically increase (in association with workload and neurohumoral alterations). Twenty four hour rhythms in myocardial glucose utilization appear to be the consequence of both extracardiac factors (e.g., circulating insulin levels) and the CCC. The latter likely synchronizes myocardial glucose metabolism with predicted daily fluctuations in energetic demand and nutrient availability, thereby optimizing cardiac functional responses. Although several candidate mediators between the CCC and myocardial glucose utilization have been proposed (e.g., AMPK), additional studies are required to establish direct mechanistic links.

Lipid Metabolism

During baseline conditions, the healthy myocardium relies on mitochondrial fatty acid β-oxidation for up to 70% of the ATP required for maintenance of contractile function; during acute periods of increased fatty acid availability (e.g., fasting), myocardial fatty acid utilization increases further (i.e., metabolic flexibility) (Lopaschuk, Ussher et al., 2010; Glatz, Nabben et al., 2020). Fatty acid sources for the heart include plasma lipoproteins and non-esterified fatty acids (NEFA), as well as intramyocardial triglyceride (Park, Yamashita et al., 2007). The importance of myocardial lipid metabolism is underscored at multiple levels. First, the heart is the initial metabolically active recipient of dietary lipids (following packaging as chylomicrones by intestinal epithelial cells) (Labbé, Grenier-Larouche et al., 2011). Second, the heart has extremely high lipoprotein lipase (LPL) activity relative to other mammalian tissues, aiding in efficient utilization of plasma lipoproteins; cardiomyocyte-specific genetic deletion of LPL results in dyslipidemia (Benavides, Siches et al., 1998; Noh, Okajima et al., 2006). Third, mutation/deletion of genes encoding for proteins involved in fatty acid metabolism often precipitates cardiomyopathy (in both animal models and humans) (Abdurrachim, Luiken et al., 2015). Fourth, chronic elevation of lipid availability (e.g., during dyslipidemic states, such as obesity and diabetes mellitus) results in cardiac fatty acid uptake in excess of β-oxidation capacity, leading to accumulation of lipid species in the heart; the latter impairs contractile function through numerous ‘lipotoxic’ mechanisms (Young, McNulty et al., 2002). These observations highlight a need for metabolic flexibility, such that fatty acid availability is balanced with cardiac utilization, for maintenance of both systemic homeostasis and cardiac contractility. This is evident across a normal 24hr day.

Circulating lipoprotein, triglyceride, and NEFA levels fluctuate with respect to time of day, in both humans and animal models. In humans, plasma triglyceride and NEFA peak towards the end of the sleep phase, which is prevented by pharmacologic inhibition of lipolysis; these observations suggest that lipolysis during the sleep phase likely provides fatty acids for hepatic triglyceride synthesis (Schlierf & Dorow, 1973). Consistent with this concept, proteins involved in hepatic triglyceride synthesis/secretion (e.g., microsomal triglyceride transfer protein) exhibit peak activity at the end of the active period (in both the liver and intestine) (Pan, Zhang et al., 2010). Moreover, lipid enriched meals consumed during the sleep phase result in greater postprandial elevations in circulating triglyceride levels (Hampton, Morgan et al., 1996). Time-of-day-dependent fluctuations in circulating lipids appear to be governed by endogenous circadian clocks, as they persist under constant environmental/behavior conditions, and are lost in rodent models of clock disruption (Morgan, Arendt et al., 1998; Pan, Zhang et al., 2010). Interestingly, inter-individual variability has been reported for circadian rhythmicity of numerous circulating lipid species in humans (probably secondary to polymorphisms in circadian clock components, leading to morningness versus eveningness chronotypes) (Chua, Shui et al., 2013). With regards to fatty acid utilization, whole body reliance on fatty acid oxidation is elevated during the sleep period (consistent with increased substrate availability at this time) (Bray, Ratcliffe et al., 2013). Collectively, these observations highlight striking time-of-day-dependent fluctuations in lipid metabolism.

Similar to the whole body and extra-cardiac rhythms highlighted above, myocardial lipid metabolism fluctuates over the course of the day, at both oxidative and non-oxidative levels. However, unlike glucose metabolism, 24-hr rhythms in distinct lipid metabolism pathways are temporally misaligned with one another, suggesting that an upstream process (such as substrate availability and/or uptake) is not a common mediator. Metabolic tracer studies in rodent hearts reveal that triglyceride synthesis exhibits an approximate 2.5-fold oscillation, peaking towards the end of the active period; consistent with this, myocardial triglyceride levels peak at the active-to-sleep phase transition (Tsai, Kienesberger et al., 2010). In contrast to lipid storage, fatty acid β-oxidation exhibits a relatively modest diurnal variation (1.3-fold), peaking towards the end of the sleep phase (Brewer, Collins et al., 2018). Moreover, synthesis of lipid-derived signaling molecules (e.g., diacylglycerol, phospholipids, and cholesterol esters) is elevated during the sleep phase (Durgan, Moore et al., 2007). Questions therefore arise regarding the mechanisms involved in temporal partitioning of distinct branches of lipid metabolism. Circadian clocks have emerged are strong candidates. For example, germline disruption of PER2 and pharmacologic activation of REV-ERBα/β alters myocardial expression of genes involved in β-oxidation, consistent with perturbations in fatty acid oxidation capacity and triglyceride levels (Bonney, Kominsky et al., 2013; Zhang, Zhang et al., 2017a). Importantly, genetic disruption of the CCC abolished 24-hr rhythms in triglyceride synthesis, which are chronically depressed (concomitant with a time-of-day-independent elevation of cardiac fatty acid oxidation) (Tsai, Kienesberger et al., 2010).

Although the precise molecular links between the CCC and 24hr rhythms in lipid metabolism have not been elucidated fully, some progress has been made. LPL activity fluctuates over the course of the day, peaking towards the sleep-to-active phase transition (in parallel with plasma triglyceride levels) (Benavides, Siches et al., 1998). LPL appears to be regulated directly by circadian clocks, leading to speculation that diurnal variations in cardiac LPL activity optimizes lipoprotein utilization in temporal synchronization with daily oscillations in substrate availability (Gimble & Floyd, 2009). Regarding myocardial fatty acid uptake, CD36 and FATP1 (primary fatty acid transports in the heart) both exhibit 24hr rhythms at the mRNA level, peaking towards the end of the sleep phase (Mia, Sonkar et al., 2021). Once in the cardiomyocyte, metabolic fate of the fatty acid species will be determined in part by the activity of enzymes/proteins within distinct pathways. With regards to triglyceride turnover, enzymes within both synthesis and degradation pathways fluctuate in the heart at mRNA, protein, and posttranslational modification levels (Tsai, Kienesberger et al., 2010). These include DGAT2 (diacylglycerol acyltransferase 2, involved in triglyceride synthesis) and HSL (hormone sensitive lipase, involved in lipolysis). In the case of DGAT2, mRNA levels peak in the heart towards to beginning of the active period (approximately 6 hours prior to the peak in triglyceride synthesis, thus allowing for a potential delay between mRNA and protein) (Tsai, Kienesberger et al., 2010). DGAT2 is a direct clock-controlled gene, possessing E-box consensus sequences in the promoter region known to be recognized by the BMAL1/CLOCK heterodimer; genetic disruption of either CLOCK or BMAL1 results in chronic repression of DGAT2 expression (consistent with lower triglyceride synthesis) (Young, Brewer et al., 2014). With regards myocardial fatty acid oxidation, time-of-day-dependent oscillations have been reported in expression levels of distinct canonical β-oxidation enzymes (e.g., MCAD; medium chain acyl-CoA dehydrogenase) and regulators (e.g., MCD; malonyl-CoA dehydrogenase); however, daily oscillations in these transcripts tend to be relatively small in the heart, consistent with modest fluctuations in fatty acid oxidation rates (Bray, Shaw et al., 2008; Brewer, Collins et al., 2018; Mia, Sonkar et al., 2021). It is noteworthy that a key end product of fatty acid oxidation is acetyl-CoA, which can ultimately be utilized for acetylation of numerous proteins in the cell, including various metabolic enzymes (Fukushima & Lopaschuk, 2016). Conversely, deacetylases are required to remove this posttranslational modification; one family of deacetylases includes NAD-dependent sirtuins (Lavu, Boss et al., 2008). NAD levels fluctuate in the heart over the course of the day, as does NAMPT (nicotinamide phosphoribosyltransferase; a key enzyme in the NAD salvage pathway) (Powanda & Wannemacher, 1970; Durgan, Pat et al., 2011). Interestingly, NAMPT is regulated directly by the CCC, leading to suggestions that the clock may modulate cardiac processes (including fatty acid metabolism) through protein acetylation (Durgan, Pat et al., 2011; Young, Brewer et al., 2014; Peliciari-Garcia, Goel et al., 2016).

In summary, extensive temporal partitioning of cardiac lipid metabolism exists, with augmented fatty acid oxidation at the end of the sleep period, channeling of excess fatty acids into triglyceride at the end of the active period, and enhanced synthesis of lipid-derived signaling molecules during the sleep period. Elevated circulating lipid species at the end of the sleep period is therefore synchronized with enhanced oxidative capacity at this time; mediation by circadian clocks may be in anticipation of prolongation of the sleep phase fast, in the event that the animal in the wild is initially unsuccessful in its forage for food upon awakening.

Amino Acid Metabolism

Amino acid catabolism accounts for approximately 3–5% of the energetic demands of the heart (Ichihara, Neely et al., 1980) although it has not been extensively well studied. Despite this apparent low reliance on amino acids, in terms of ATP turnover, perturbations in amino acid metabolism can severely impair cardiac function. This likely stems from diverse functions of amino acids and their derivatives, ranging from generation of intra- and extra- cellular signaling molecules, to serving as the backbone of proteins and cofactors. Surprisingly little is known regarding circadian governance of amino acid metabolic fates in the heart. Indirect evidence includes description of the steady state levels of amino acids in the heart, as well as expression of genes encoding for enzymes in amino acid catabolic pathways, and how these indices fluctuate over a 24hr day in the heart. Given central roles of amino acids in cardiac processes, here we outline how daily rhythms in amino acid metabolism may impact function of the heart.

Circulating amino acid levels reflect a balance between ingestion/absorption, de-novo synthesis, proteolysis, and utilization. In humans, plasma amino acid levels trough in the early morning, and a peak in the late afternoon (Feigin, Klainer et al., 1968). Equivalent oscillations are observed in mice, with a peak in the middle of the active (dark) phase (Minami, Kasukawa et al., 2009). Protein ingestion is a major source of amino acids, with breakdown initiated by digestive enzymes (i.e., proteases) (Snook & Meyer, 1964). The latter includes chymotrypsin, which exhibits increased secretion/activity at the beginning of the active period; this daily oscillation persists during a prolonged fast, suggesting anticipation of food intake (ensuring efficient food breakdown once ingested) (Maouyo, Sarfati et al., 1993). Similar to digestive enzymes, expression of multiple amino acid transporters display diurnal rhythms in the gut (Pan, Terada et al., 2004; Zhang, Zhang et al., 2017b; Yi-lin, Ke et al., 2018; Oparija-Rogenmozere, Rajendran et al., 2020). In piglet intestinal mucosa, a host of solute carrier (SLC) amino acid transporters are elevated at the beginning of the active period (Zhang, Zhang et al., 2017b); similar observations have been reported for the amino acid uniporter LAT4 in mice (Oparija-Rogenmozere, Rajendran et al., 2020)and the small peptide transporter PEPT1 in rats (Pan, Terada et al., 2004). Diurnal variations in amino acid uptake mechanisms parallel food intake patterns and likely contribute towards daily rhythms in circulating amino acids levels. When mice are placed in constant conditions (e.g., continuous darkness) for 38hrs, circadian rhythms in plasma amino acids persist (Jeyaraj, Scheer et al., 2012). Similarly, when humans are subjected to a 40hr constant routine and environmental conditions, amino acids in the saliva exhibit persistent 24hr oscillations (Feigin, Klainer et al., 1968; Dallmann, Viola et al., 2012). These observations suggest that endogenous circadian clocks also contribute towards 24hr rhythms in circulating amino acids. Consistent with this concept, genetic disruption of the circadian regulated transcription factor KLF15 abolishes diurnal variations in plasma amino acid levels, potentially through regulation of peripheral tissue amino acid metabolism (Jeyaraj, Scheer et al., 2012).

The total pool of ‘free’ amino acids exhibits a modest diurnal variation in the rat heart, with increased levels during the sleep (light) period (Lyons, Squibb et al., 1967). Similar diurnal variations are observed in the murine heart, with approximately 8% higher total ‘free’ amino acids at the beginning of the sleep (light) period (relative to the beginning of the active period) (Latimer, Sonkar et al., 2021). A disconnect therefore exists between circulating amino acids (which peak in the middle of the active period) and cardiac ‘free’ amino acids (which peak at the beginning of the sleep phase). Potential explanations for this disconnect include the possibility that processes involved in cardiac amino acid import/export, utilization (e.g., catabolism and/or protein synthesis), and/or mobilization (e.g., proteolysis) are circadian regulated. Direct and indirect evidence exist in support of this concept. Although 24hr rhythms in cardiac amino acid import/export have not been assessed to date, transcriptomic analyses suggest that multiple mRNA species encoding for amino acid transporters exhibit diurnal variations in the heart; these include sodium-coupled neutral amino acid transporter 9 (SLC38A9) and sodium-dependent multivitamin transporter (SLC5A6)(He, Hamm et al., 2016). Similarly, expression of genes involved in branched chain amino acid catabolism (e.g. isovaleryl-CoA dehydrogenase and dihydrolipoamide branched chain transaceylase E2) have been reported to oscillate in the heart over the course of the day(Zhang, Prosdocimo et al., 2015); whether these transcriptomic fluctuations translate to the metabolic flux level currently remains unknown. Interestingly, these mRNA fluctuations appear to be dependent on the CCC, as they are abolished following genetic disruption of circadian clock components (e.g., BMAL1) or putative clock output genes (e.g., KLF15) (Young, Brewer et al., 2014; Zhang, Prosdocimo et al., 2015). Consistent with this, decreased expression of branched chain amino acid catabolic pathway genes following genetic disruption of the CCC is associated with attenuated cardiac leucine oxidation (He, Hamm et al., 2016). Unlike catabolic fates of amino acids, 24hr rhythms in cardiac protein synthesis have been reported through use of metabolic flux studies, which peaks at the beginning of the sleep phase (McGinnis, Tang et al., 2017). This is in alignment with the pool of ‘free’ amino acids in the heart, as well as the activation status of mTOR signaling components (an established pro-growth pathway) and ribosomal RNA levels; genetic disruption of the CCC attenuates 24hr oscillations in these parameters (suggesting governance by this intrinsic mechanism) (McGinnis, Tang et al., 2017; Latimer, Sonkar et al., 2021). Work from extra-cardiac tissues suggest that an important subset of proteins that are synthesized and/or secreted in a circadian fashion include collagen; although diurnal variations in collagen synthesis has not been investigated in the heart, disruption of the CCC leads to profound cardiac fibrosis (concomitant with chronic activation of protein synthesis) (Chang, Garva et al., 2019). Similar to protein synthesis, proteolysis exhibits 24hr oscillations in the heart, peaking early in the sleep phase. For example, macroautophagy is elevated during the first 4 hours of the sleep phase; this appears to be secondary to both behavioral (e.g., fasting) and circadian influences (Brewer, Collins et al., 2018).

In summary, emerging evidence supports the concept that cardiac amino acid metabolism fluctuates over the 24hr day at multiple levels. The concomitant increase in cardiac amino acid availability, proteolysis, and protein synthesis during the early portion of the sleep period is consistent with this being a period of growth and repair (i.e., replacement of damaged cellular constituents) for the heart. In contrast, less is known regarding daily fluctuations in cardiac amino acid catabolism (beyond the transcriptional level). Given that amino acids are critical for the biosynthesis of numerous signaling molecules and cofactors, future studies are needed to gain a better understanding of circadian governance of cardiac amino acid metabolism.

Timing of Nutrient Intake: Implications for the Heart

Circadian influences over cardiovascular processes have far reaching translational implications, ranging from the timing of adverse events and severity of the insult, to biomarker efficiency and chronotherapeutics (Van Laake, Lüscher et al., 2017). In keeping with the latter, the American Heart Association published a Scientific Statement that recommended “eating a greater share of the total calorie intake earlier in the day to have positive effects on risk factors for heart disease and diabetes mellitus” (St-Onge, Ard et al., 2017). This stems in large part from epidemiologic/observational studies reporting that individuals who skip breakfast and/or consume snacks at night, have an increased incidence of heart disease. Similar to the heart, circadian regulation of metabolism has been reported for extra-cardiac tissues and at the whole body level. The possibility therefore exists that the timing at which distinct nutrients are consumed could impact the myocardium either directly (i.e., metabolic fate of nutrients taken up by the heart) and/or indirectly (i.e., alterations in the neurohumoral milieu secondary to differential responses of extracardiac tissues). Here, we consider evidence from studies that help provide insight regarding whether the time-of-day at which carbohydrate, lipid, and/or protein is consumed influences cardiac function and/or cardiometabolic disease risk.

To date, the majority of studies investigating whether the timing of nutrient intake influences cardiometabolic/cardiovascular disease related parameters have focused primarily on calorically-dense lipid consumption. Rodent-based studies revealed that consumption of high fat diets specifically during the sleep (light) period augmented weight gain, adiposity, dyslipidemia, glucose intolerance, insulin resistance, and hyperinsulinemia (relative to consumption of the same diet during the active period) (Arble, Bass et al., 2009; Bray, Tsai et al., 2010; Hatori, Vollmers et al., 2012). Moreover, several of these parameters can be improved in obese mice (and humans) by restricting lipid intake only to the active period (Garaulet, Gomez-Abellan et al., 2013; Chung, Chou et al., 2016; Olsen, Choi et al., 2017). Studies designed to address whether lipids have differential effects when consumed for breakfast versus dinner, the majority of studies report greatest cardiometabolic benefit when lipids are consumed for breakfast; this is associated with a greater activation of fatty acid oxidation following lipid consumption for breakfast, in both rodents and humans (Bray, Tsai et al., 2010; Bush, Resuehr et al., 2018). Importantly, similar outcomes have been reported for cardiac parameters. More specifically, restricting dietary lipid intake to the active (dark) period, and in particular, the beginning of the active (dark) period, increased cardiac contractility, associated with higher rates of fatty acid oxidation (Tsai, Villegas-Montoya et al., 2013; Peliciari-Garcia, Goel et al., 2016). In addition, restricting food intake to the active period improved cardiac contractility and partially reversed adverse cardiac remodeling in both obese flies and mice; the latter includes normalization of cardiac steatosis (Gill, Le et al., 2015; Mia, Sonkar et al., 2021). Collectively, these studies suggest that dietary lipid intake at the end of the active period and/or during the sleep period has adverse cardiometabolic and cardiovascular outcomes. With regards to circadian regulation of cardiac metabolism, the possibility exists that consumption of dietary lipid at the end of the active period aligns substrate availability with enhanced capacity of the heart to synthesize triglyceride and lipid signaling molecules (e.g., phospholipids, ceramides), thus augmenting cardiac steatosis and lipotoxicity.

In comparison to lipid intake, far less is known regarding whether the timing of dietary carbohydrate consumption affects cardiovascular parameters. However, indirect evidence exists suggesting that this could be an important factor. As highlighted above, consumption of a carbohydrate-rich meal leads to a greater postprandial rise in blood glucose levels (compared to the same meal in morning) (Scheer, Hilton et al., 2009). The latter is associated with higher post-prandial insulin levels (Scheer, Hilton et al., 2009). Moreover, simulated shift work led to increased carbohydrate storage with decreased carbohydrate utilization (positive carbohydrate balance) when eaten during ‘scheduled sleep’ periods(McHill, Melanson et al., 2014). Collectively, these studies suggest that consumption of dietary carbohydrate at the end of the active period may exert detrimental effects on cardiometabolic parameters. What is less clear is whether circadian governance of cardiac glucose metabolism directly impacts the response of the heart to dietary carbohydrate intake at distinct times of the day. Future studies are required to address this question.

Evidence is beginning to emerge that the time of day at which protein is consumed influences cardiometabolic parameters. A recent NHANES-based study reported that protein intake at breakfast was associated with lower blood pressure and higher HDL levels; conversely, protein intake at dinner was associated with higher insulin levels and insulin resistance (Berryman, Lieberman et al., 2021). Such observations suggest cardiometabolic benefit when dietary protein is consumed at the beginning of the day. Conversely, kinesiology-based studies indicate that consumption of protein prior to sleep promotes skeletal muscle hypertrophy following resistance exercise; given that retention of skeletal muscle mass is critical for maintenance of whole-body metabolic health, protein intake at dinner following exercise may exert some benefit. With regards to the interaction between dietary amino acid intake and circadian governance of cardiac protein metabolism, Latimer et al recently reported that consumption of dietary branched chain amino acids at the end of the active period (when protein synthesis rates are high) leads to cardiac hypertrophy (Latimer, Sonkar et al., 2021). Moreover, prolonged consumption of branched chain amino acids at this time accelerated adverse remodeling of the hearts during cardiac disease (Latimer, Sonkar et al., 2021). Collectively, these studies are consistent with the concept that cardiometabolic/cardiovascular benefit is observed when dietary protein is consumed at the beginning of the day (although acute intake of protein at the end of the active period following exercise may aid in maintenance of lean body mass.

Summary and Future Directions

Coincident with dramatic daily fluctuations in workload, substrate availability, and the neurohumoral milieu, the healthy myocardium temporally partitions metabolic processes across a 24-hr period. A truly remarkable example of metabolic flexibility. This is observed at the levels of myocardial glucose, fatty acid, and amino acid metabolism (both oxidative and non-oxidative). In general terms, the beginning of the active period is a time of increased oxidative metabolism (presumably to meet increased energetic demand), the end of the active period is a time of increased storage of excess nutrients (presumably to prepare for the upcoming sleep phase fast), while the beginning of the sleep period is a time of increased cellular consistent turnover (presumably in anticipation of increased contractility upon awakening). Temporal partitioning, which is mediated in large part by the CCC, therefore allows the heart to synchronize metabolism with predicted 24hr fluctuations in the environment and behaviors. However, both attenuation (e.g., disruption of the normal 24hr environment during shift work) and augmentation (e.g., macronutrient consumption at the ‘wrong’ time of the day) of these circadian rhythms outside of a physiologic range precipitates cardiometabolic disease. A better understanding of the molecular links between circadian clocks and cardiac metabolism, as well as how lifestyle interventions impact these parameters, has the potential to identify novel therapeutics for the prevention and/or treatment of cardiac disease.

Figure 1. Temporal partitioning of myocardial glucose metabolism.

At the beginning of the active period (i.e., breakfast time), AMP-activated protein kinase (AMPK) promotes glucose uptake, glycolysis, and the hexosamine biosynthetic pathway, as well as full oxidation via pyruvate dehydrogenase (PDH). At the end of the active period (i.e., dinner time), decreased levels of inhibitor-1 promotes protein phosphatase 1 (PP1) activity, and concomitant storage of excess glucose as glycogen (which can serve as a fuel source during the upcoming sleep period).

Figure 2. The metabolic fate of lipids in the heart is time-of-day-dependent.

At the beginning of the active period (i.e., breakfast time), increased lipoprotein lipase (LPL) activity promotes lipo-protein-derived fatty acid uptake and oxidation; resultant acetyl-CoA can serve as a substrate in protein acetylation (Ac) reactions. At the end of the active period (i.e., dinner time), fatty acids are channeled into triglyceride synthesis (which can serve as a fuel and/or source for lipid signaling molecules during the upcoming sleep period).

Figure 3. Amino acids selectively augment cellular consistent turnover at the beginning of the sleep phase.

At the beginning of the active period (i.e., breakfast time), the heart utilizes amino acids for basal processes. In contrast, at the end of the active period (i.e., dinner time), a concomitant increase in mTOR signaling activation, ribosomal RNA, and amino acid content, augments protein synthesis (and growth/repair).