The Circadian Biology of Heart Failure

Abstract

Driven by autonomous molecular clocks that are synchronized by a master pacemaker in the suprachiasmatic nucleus, cardiac physiology fluctuates in diurnal rhythms that can be partly or entirely circadian. Cardiac contractility, metabolism, and electrophysiology, all have diurnal rhythms, as does the neurohumoral control of cardiac and kidney function. In this review, we discuss the evidence that circadian biology regulates cardiac function, how molecular clocks may relate to the pathogenesis of heart failure, and how chronotherapeutics might be applied in heart failure. Disrupting molecular clocks can lead to heart failure in animal models, and the myocardial response to injury seems to be conditioned by the time of day. Human studies are consistent with these findings, and they implicate the clock and circadian rhythms in the pathogenesis of heart failure. Certain circadian rhythms are maintained in heart failure patients, a factor that can guide optimal timing of therapy. Pharmacologic and non-pharmacologic manipulation of circadian rhythms and molecular clocks show promise in the prevention and treatment of heart failure.

Introduction

Heart failure is prevalent, affecting around 6 million Americans over 20 years of age; a number projected to increase to 8 million by 2030.¹ It is a complex condition that incurs a high societal burden due to its debilitating effects on patients. The complexity of heart failure lies in its numerous etiologies, the variability in its symptomatology and the multiple systems involved both in causing or exacerbating the illness.

Adding to that complexity is the functional fluctuation of the heart and related systems throughout the 24-hour day. These diurnal rhythms may be partly or entirely circadian, the result of the synchronization of autonomous peripheral clocks by a master pacemaker located in the suprachiasmatic nucleus (SCN) of the hypothalamus. The core clock genes include the main helix-loop-helix transcription factors, circadian locomotor output cycles kaput (CLOCK) and brain and muscle ARNT-like protein-1 (BMAL1) driving transcription of target genes through E boxes located in their promoters. Included in these target genes are period 1-3 (PER1-3) and cryptochrome 1-2 (CRY1-2) whose respective proteins inhibit the expression of CLOCK and BMAL1. This leads to a cyclic oscillation of expression of all clock components in opposing phases, forming what is known as the transcription-translation feedback loops (TTFL). Other genes interacting with these feedback loops include REV-ERB-α and -β.²

Despite our increased understanding of molecular clocks, this knowledge has contributed little or nothing to the treatment of heart failure. In this review, we will consider the evidence that circadian biology regulates cardiac function and how dysregulation of molecular clocks may initiate or exacerbate cardiac decompensation. Furthermore, we will discuss advances in chronopharmacology and how they might influence the treatment of this condition.

The Cardiac Clock its Modulators, and its Outputs

We summarize the circadian control of cardiovascular function in figures 1 and 2. Similar to other peripheral tissues, core clock genes oscillate rhythmically over 24 hours in the murine and human heart.^2–6 This rhythm is seen both in heart tissue and in isolated and cultured cardiomyocytes, pointing to an innate cellular clock mechanism that can run independently.^7,8 Although the cardiomyocyte clock is cell autonomous, it can be modulated and synchronized by different cues or zeitgebers. The effect of light on the rhythmic expression of cardiac clock genes is well established.^9–12 Stimulated by light, the SCN acts as the central clock and synchronizer of all peripheral clocks. The SCN induces melatonin secretion by the pineal gland which influences clock gene expression in the heart.^13,14 It also connects to peripheral tissues via sympathetic nerve fibers² and through neurons secreting vasoactive intestinal peptide (VIP) which can alter heart rate.¹⁵ Food and metabolic cues are also important modulators of the cardiomyocyte clock as with other peripheral clocks.^11,16,17 Reversal of food availability from the active phase to the rest phase reversed the phasic rhythmicity of core clock genes in rat hearts both under regular and reverse light dark cycles.¹¹ Peroxisome proliferator activated receptor α (PPARα) has been implicated in food entrainment of the cardiac clock.¹⁸ While insulin dependent diabetes did lead to phase changes in clock genes of the rat heart,¹⁹ the hearts of insulin deficient rats exhibited phase changes in gene expression in response to restricted feeding.²⁰ Insulin thus can be one of multiple factors that mediate the regulation of the cardiac clock by feeding. Interestingly, redox states and the pentose phosphate pathway (PPP) seem to regulate molecular clocks. Inhibition of the PPP in human osteosarcoma cells, and mouse SCN and liver cells altered the circadian expression of clock genes.²¹ Whether this effect is seen in heart tissue or if it contributes to the cardiac clock regulation by feeding is yet to be determined.

Figure 1. Cardiac function under the regulation of the core molecular clock.

Food, light, and hormonal signals entrain the cardiomyocyte core molecular clock which can run autonomously without these cues. Of the elements that make up the core clock are BMAL1 and CLOCK which heterodimerize and initiate the expression of REV-ERB, PER1-3, and CRY1-2, which in turn suppress the expression of BMAL1 and CLOCK leading to a rhythmic oscillation in expression among these transcription factors. Each of these proteins then modulates the expression of various output genes which includes genes relating to cardiac contractility, electrophysiology, metabolism, and repair. SCN: Suprachiasmatic nucleus of the hypothalamus. SNS: Sympathetic nervous system. VIP: Vasoactive intestinal peptide. T3: Triiodothyronine. EP: Electrophysiology. Created with biorender.com. Illustration credit: Ben Smith.

Figure 2: Distribution of peaks and troughs of gene expression, protein concentration and physiologic measurements in healthy human studies(A) and in mice and rats (B).

Values plotted are approximations of the timing of peaks and troughs extracted from the cited studies. A. BMAL 1 expression in healthy human hearts in close to opposite phase with REV-ERB and PER expression (also in the heart). ANP peaks in the early AM hours coinciding with a decrease in SBP. Cortisol, neurohormones, renin, and aldosterone peak in the morning in preparation for the rise in blood pressure at or before awakening. Blood pressure then peaks at noon to the afternoon. Heart rate (inversely proportional to RR interval) follows the peak in epinephrine and norepinephrine. QT interval and PR segment have a close pattern to that of the RR interval. SBP: Systolic Blood Pressure. HR: Heart Rate.^{3,39,109,110,126} B. X-axis depicts zeitgeber time (time from onset of light). All studies included housed mice or rats under 12:12 light: dark cycles. Like human hearts Bma1 expression is in close to opposite phase to Rev-erb and Per1-2. Also similarly to humans, cortisol, neurohormones, renin, and aldosterone cycle in a way to raise heart rate and blood pressure in the active phase.^{4,9,105,111,112,172} Created with biorender.com

The cardiomyocyte clock is also influenced by multiple hormonal factors such as angiotensin II, and aldosterone, two important therapeutic targets in heart failure.^22,23 The effect of aldosterone on clock gene expression in cultured rat cardiomyoblasts was attenuated, but not completely inhibited by spironolactone, an aldosterone receptor inhibitor.²³ Angiotensin II induces the expression of clock genes in vitro, and blocking its receptor abolishes this induction.²² Infusing rats with angiotensin II caused significant phase shifts in the expression of Per2 and Rev-erb-α.²² The thyroid hormone, triiodothyronine (T3) also regulates the cardiac molecular clock by inducing the expression of its components.²⁴

The cardiomyocyte clock itself influences expression of various genes leading to rhythmic oscillations in cardiomyocyte function. In one study, the expression of ~40% of 12,500 studied genes oscillated rhythmically in the murine heart.²⁵ Notably, clock genes can bind directly to enhancer sequences,² or can modulate expression by affecting histone modifications.⁵ Therefore, the influence of the clock on gene expression combined with its autonomous rhythmicity and the modulation of multiple inputs translates into variation in multiple cardiomyocyte functions across the 24-hour cycle. Sexually dimorphic outputs of clock function are recognized whereby females show higher amplitude oscillations in circadian rhythms. The consequences of clock disruptions are also different between males and females.²⁶ Sex differences are also recognized in heart failure, with the higher predisposition to heart failure with reduced ejection fraction (HfrEF) in males compared to a higher predisposition to heart failure with preserved ejection fraction (HfpEF) in females.²⁷

Clock Genes and the Development of Heart Failure

Disrupting the cardiac clock by gene knockouts leads to different cardiomyopathies. Knocking out Bmal1 specifically in mouse cardiomyocytes causes an age dependent hypertrophic cardiomyopathy with adverse remodeling of the myocardium characterized by increased cardiomyocyte size, a hypertrophic and enlarged left ventricle, increased serum BNP levels, and interstitial fibrosis with a later decrease in systolic function.^28–32 Global Bmal1 KO mice have a different phenotype with myocardial thinning and loss of systolic function in keeping with a dilated cardiomyopathy, likely due to circadian disruption in other tissues as well.³³ Mutations inactivating the Clock gene both in a cardio-specific manner or globally show a similar hypertrophic phenotype with a later decrease in systolic function that seems to be age dependent in both sexes but more delayed in females.^34–37. Knocking out Rev-erbα/β specifically in the mouse cardiomyocyte led to progressive worsening of systolic function and a dilated cardiomyopathy phenotype in both males and females. The same was seen when the KO was induced later in life. Interestingly, a high fat diet combined with insulin resistance alleviated systolic dysfunction in these KO mice by partly restoring rhythms in metabolic states of glucose and fatty acid utilization.⁶ Indeed Rev-erb seems to mediate this effect through E4BP4, through which it also proved essential for maintaining NAD+ production by the salvage pathway.^6,38 Thus, Clock and Bmal1 disruption cause cardiac hypertrophy, Rev-erb KO causes dilated cardiomyopathy, and cardiac hypertrophy is ameliorated when a REV-ERB agonist is administered to mice with specific Bmal1 KO in cardiomyocytes.²⁹ This indicates that the different clock components regulate both distinct and common pathways involved in cardiac decompensation.

In a study of human failing hearts with dilated cardiomyopathy removed for cardiac transplantation, Song et al.⁶ assigned a “molecular chronotype” based on the REV-ERB to BMAL1 expression ratio. Those with a ratio below that of normal hearts had a decreased left ventricular end diastolic diameter and a higher prevalence of mitral regurgitation seen on both echocardiography and cardiac MRI. This indicates that a change in REV-ERB/BMAL1 rhythms is correlated with the severity of heart failure in dilated cardiomyopathy.

Clock function is altered in established heart failure. Core clock genes keep their circadian rhythms in a rat model of cardiac hypertrophy, but clock output genes, such as PAR transcription factors, have an attenuated induction.⁴ This was also seen in the hearts of end-stage heart failure patients at the time of transplant where core clock genes had rhythmic oscillations with robust rhythms that were phase shifted when compared to brain dead donor hearts.³⁹ However a comparison with brain dead donors should be interpreted with caution as circadian rhythms are affected by the underlying condition and the treatments given.³⁹ The attenuation of clock output gene induction is propagated to downstream targets such as metabolic genes.⁴⁰

Cardiac Circadian Rhythms in Health and in the evolution of Heart Failure

Contractility

Cardiac function oscillates rhythmically over the 24-hour day. One such function is contractility that is impaired in heart failure with reduced ejection fraction HFrEF. Contractile function in the mouse heart fluctuates rhythmically through the light-dark cycle.^35,40,41 For example, in ex vivo perfused hearts, contractility (assessed by measuring cardiac power) increased by 30% compared to the rest phase.³⁵ Indeed echocardiographic parameters of ejection fraction and fractional shortening are higher in mice in the active phase compared to the rest phase under both sedentary and forced exercise conditions.⁴²

Rhythms in cardiac contractility were abolished in mice with a cardiomyocyte Clock mutation (CCM). CCM mice also have suppressed rhythmicity in the activity of proteins involved in contractile function, an effect that resembled that seen with disruption of the regular light dark cycle. ^35,41,43 CCM hearts have altered phosphorylation patterns of proteins related to contractility (Erka, p38, Akt, Gsk-3β, myosin binding protein C, desmin, troponins T and I, and tropomyosin) along with alterations in the rhythmic expression of some of these genes, including L-type voltage gated calcium channel (VGCCα1D) which also had diminished mRNA and protein levels.^41,43 These L-type voltage-gated channels are responsible for the excitation contraction coupling leading to the contractile response to an action potential. Transmembrane voltage changes also showed rhythmic oscillations in expression.^44,45

Bmal1 deletion in mouse hearts did not cause any changes in locomotor activity. However, Bmal1 deletion blunted the rhythmic expression of Pik3r1, Akt, and Gsk-3β genes that are important for contractile function, in a similar fashion to the pattern seen in CCM mice.³⁰ The majority of the studies on circadian control of contractile function have been done in CCM mice rather than in Bmal1 cardiomyocyte specific KO mice. However, Clock has functional redundancy with another gene (Npas2) which may have attenuated the rhythmic perturbations in CCM mice.^46,47

In wildtype non-mutant mice, differences between active and rest phases were detected in intra-sarcomeric calcium concentrations along with the functions of sarco/endoplasmic reticulum Ca²⁺-ATPase (SERCA), an ATP coupled transporter important for muscle relaxation and maintenance of sarcomeric calcium stores.⁴⁸ The sensitivity of calcium homeostasis to sympathetic stimulation also differed between active and rest phases.⁴⁸ These experiments investigating excitation contraction coupling only measured outcomes at two opposing times which is insufficient to establish the presence of a circadian rhythm. Measurements of basal calcium levels in mouse ventricular explants over multiple time points over multiple days did not show robust circadian rhythms.⁴⁹ However, Sachan et al. studied calcineurin activity and phospholamban phosphorylation, two important proteins in excitation-contraction coupling, across multiple time points. Calcineurin activity was higher in the rest phase while phospholamban phosphorylation was highest in the active phase. Both are influenced by β-adrenergic signaling. Thus, high β-adrenergic signaling during the active phase leads to more phospholamban phosphorylation, resulting in enhanced contractility.⁵⁰ Tcap is another sarcomeric protein under circadian control and mutations have been linked to dilated cardiomyopathy.⁵¹ However, the presence of rhythmic oscillations does not always implicate the molecular clock. While some studies have elegantly demonstrated the core clock dependency of many of the rhythms discussed above, others have only reported diurnal rhythmicity which may be environmentally rather than endogenously driven.

Rhythms in cardiac contractility are altered in heart failure rodent models. In hypertrophied hearts of spontaneously hypertensive rats, left ventricular developed pressure lost its variation between daytime and nighttime. The sensitivity of left ventricular pressure to the non-selective β-agonist isoproterenol also lost its day-night variation in spontaneously hypertensive rats compared to control rats. Cytosolic calcium concentrations, calcium currents, and SERCA activity all had abolished diurnal variations in pressure overloaded hearts.⁵² In another mouse model of pressure overload, systolic function was worse in the morning while diastolic function was the same throughout the day. In pressure overloaded hearts the pattern of phospholamban phosphorylation was lost whereby phosphorylated phospholamban levels remained elevated throughout the day.⁵⁰ Due to its involvement in lusitropy, the absence of rhythms in phosphorylation/de-phosphorylation of phospholamban can be a reason behind the loss of rhythms in diastolic function.

Metabolism

Metabolic states are known to vary rhythmically over the day. The heart responds differently to fatty acids and other nutrients throughout the day and genes related to the fasting response depend on molecular clock to fluctuate in expression and ensure a ready adaptation to the availability or lack of food.^53,54 Ex vivo perfused rat hearts responded in a time-specific manner to nutrient-rich perfusions in regards to cardiac power, myocardial oxygen consumption, and carbohydrate oxidation.⁴⁰ Triglyceride levels in the mouse heart are also subject to circadian variation. The cardiomyocyte clock inactivates hormone sensitive lipase in the active phase at both transcriptional and post-translational levels, thus increasing triglyceride synthesis at the end of the active phase.³⁴ Many genes involved in metabolism, particularly fatty acid metabolism, have diurnal oscillations peaking at the light-dark transition.⁶ Ith et al used magnetic resonance spectroscopy to demonstrate in nine young healthy men a decrease in intracardiac myocellular lipids (ectopic lipids in the myocardium possibly linked to adverse cardiovascular states) in the evening compared to morning.⁵⁵ Indeed, a higher lipid synthesis in the heart towards the end of the active phase is observed in both mice and humans. Glucose utilization is higher during the active phase while reliance on fatty acids for energy is higher in the rest phase.^6,56–58 By contrast, protein synthesis, growth and repair occur mainly in the rest phase, particularly at the transition from wakefulness to sleep, possibly contributing to a time-dependency in the response to injury.^56,58–60 In fact, transverse aortic constriction led to significantly less cardiac hypertrophy in mice fed a high branched chain amino acid diet at the end of the active phase compared to at its beginning.⁶⁰

Knocking out Bmal1 in mouse cardiomyocytes changed the cardiomyocyte response to fasting, whereby fatty acid oxidation remained at high levels and glucose oxidation, glycolysis and glycogen synthesis remained at low levels, independent of feeding status.³⁰ The absence of Bmal1 also seems to increase protein synthesis and reduce autophagy resulting in cardiac hypertrophy.⁵⁸ Furthermore, heart specific knockout of Bmal1 altered mitochondrial function and increased oxidative stress leading to the activation of apoptosis, autophagy, inflammation, and remodeling contributing to the decompensation seen in this mouse model. Light-dark cycle disruption produced similar metabolic changes.⁶¹ Other molecular clock components also influence rhythms in metabolic activity. In the mouse heart, REV-ERBα/β enhances fatty acid oxidation in the light phase and counteracts glucose metabolism in the dark phase.⁶ Xu et al describe how the cardiac clock is run by Kruppel like factor 15 (KLF15) that oscillates under the control of Bmal1 and coordinates catabolism and reactive oxygen species formation in the active phase while being subject to negative feedback from metabolic cues.^58,62 In already hypertrophied rat hearts, the induction of glucose metabolism (including GLUT1 and GLUT 4 transporters) and fatty acid metabolism genes along with mitochondrial function genes was attenuated. Rhythmic diurnal variation occurs in the expression of these genes in normal rat hearts.⁴⁰

The effect of sodium glucose transporter 2 (SGLT2) inhibitors on cardiac metabolism is of recent interest. Their role in decreasing heart failure hospitalizations and death, regardless of diabetic status, is supported by randomized controlled trials.⁶³ The SGLT2 channel is expressed in the kidney, but not in the heart. However, SGLT2 inhibitors bind glucose transporters (GLUT 1 and GLUT4) in the mouse heart, affecting glucose metabolism in cardiomyocytes and attenuating adverse cardiac remodeling in a pressure overload heart failure model.⁶⁴ GLUT1 and GLUT4 expression fluctuate rhythmically in rat cardiomyocytes. This rhythmicity is blunted in pressure overload induced cardiac hypertrophy.⁴⁰ Whether or not SGLT2 inhibitors have a time dependent effect is yet to be studied.

Remodeling and Response to Injury

Remodeling and repair seem to occur mostly during the rest phase.^56,59,60 The response to injury is also time of day dependent. Both are orchestrated by several cells and processes. Cardiac fibroblasts have a prominent role in remodeling, the response to injury, and interstitial fibrosis is associated with heart failure and cardiomyopathies.^65,66 Fibroblasts in culture show a circadian expression pattern of clock components,⁶⁷ and seem to function in a rhythmic manner to maintain collagen homeostasis in surrounding tissues.⁶⁸ Lung fibroblasts also showed robust circadian rhythms that are accentuated in fibrotic areas. ⁶⁹ However, the circadian biology of cardiac fibroblast function remains to be elucidated. Angiotensin II, which is strongly implicated in adverse cardiac remodeling, was shown to stimulate cardiac fibroblast proliferation and the fibrotic process. As we discuss below, angiotensin II levels fluctuate rhythmically throughout the day further pointing to possible diurnal variation in cardiac fibroblast function.⁶⁶ Consistent with this expectation are circadian oscillations in fibrotic mouse lungs that are dampened by lung fibroblast specific deletion of Bmal1.⁶⁹ Circadian regulation of fibroblast function is seen in other tissues and the importance of fibroblasts in cardiac remodeling and fibrosis deserves further study. This is necessary to understand the extent and mechanisms of circadian regulation of cardiac remodeling, fibrosis, and injury response.

Another player in cardiac remodeling is the immune system. Myocardial ischemia, hemodynamic overload, and pathogens activate innate and adaptive immune responses in the heart leading to repair and adaptive remodeling. This response can become maladaptive, leading to decompensation and heart failure.⁷⁰ Molecular clocks are present in immune cells, and circadian clocks regulate immune function from cytokine production, to immune cell trafficking, maturation, proliferation, and differentiation.^71–73 The molecular clock regulates the inflammatory response and inflammation can feed back, in turn, on the molecular clock.⁷¹ However, studies investigating the circadian regulation of immune responses in the heart are lacking, even though the response to ischemia-reperfusion injury in mice is time-of-day dependent⁷⁴. There is some evidence that inflammasome activity and immune cell recruitment to the myocardium are regulated by core clock genes such as Rev-erb.⁷⁵ In a myeloid specific Bmal1 KO mouse model, vascular macrophages were implicated in enhanced collagen deposition and thickness of the vascular media, resulting in increased endothelial dysfunction, and higher blood pressure levels in response to angiotensin II infusion compared to wild type mice.⁷⁶ Whether a similar effect is observed on the myocardium remains to be elucidated.

Endothelial cells form the largest non-myocyte cell population in the heart. They line the vessels that distribute blood from the epicardial coronary arteries to the subendocardial capillaries that can supply blood to individual muscle fibers. Endothelial cells also secrete factors that modulate cardiac contractility, lusitropy, and remodeling.^{77 67,78} Endothelial cells are important in angiogenesis, maintaining the needed capillary density during physiologic myocardial hypertrophy. This is impaired in heart failure.^79,80 Endothelial cell function exhibits circadian variation as seen in the fluctuations in endothelial nitric oxide synthase activity (eNOS), vascular endothelial growth factor (VEGF), plasminogen activator inhibitor (PAI) and tissue plasminogen activator (tPA) production.^67,81 Indeed, cultured endothelial cells display circadian expression of clock components⁶⁷ Global deletion of Bmal1 affected vascular endothelial functions such as the secretion of Von-Willebrand factor, vasorelaxation, and the activity of eNOS while endothelial cell specific deletion abolished diurnal variation in thrombogenesis.^78,82,83 The core clock gene Period 2 is essential for restoring endothelial cell function and angiogenesis after induction of myocardial infarction in mice.⁸⁴ Clock disruption also undermines reparative angiogenesis following mouse hind limb ischemia⁸⁵.

Despite the gaps linking each of these cell types to the circadian regulation of injury response and remodeling, there is substantial evidence that the clock influences remodeling and outcomes post injury. Rats injected with doxorubicin at the end of the rest phase had higher markers of cardiotoxicity compared to those injected at the end of the active phase. Doxorubicin plasma levels were also higher in the active phase injection group suggesting that pharmacokinetics might contribute to the differential response.⁸⁶ However exposing synchronized cultured cardiomyocytes to doxorubicin also showed a circadian pattern in the damage response and apoptosis.^87,88 Oscillations in the global transcriptome were affected in mice by injections of doxorubicin. Apart from a global decrease in Rev-erb expression, the expression of core clock genes was largely unaffected. However, circadian Bmal1 deacetylation was altered by doxorubicin treatment.⁸⁹ Cardiotoxicity due to doxorubicin thus seems to affect and to be affected by circadian rhythms. The response to ischemic insults also involves the clock. Wild type mice with a left anterior descending artery (LAD) occlusion imposed at the sleep to wake transition had larger infarct sizes than those in which the surgery was performed at the wake to sleep transition phase.⁷⁴ Cardiac function was also worse in this group combined with higher levels of fibrosis and adverse remodeling. CCM mice did not show these differences.⁷⁴ One possible mechanism behind these findings is the role of Clock in regulating mitochondrial fission, fusion, and autophagy during myocardial ischemia, clearing disrupted mitochondria, and preventing the accumulation of reactive oxygen species. Mice with a mutated Clock gene had worse systolic function after ischemia re-perfusion injury than wild type mice. They had a higher number of fragmented myocardial mitochondria and a downregulation of genes involved in mitochondrial metabolism and mitophagy.⁹⁰ The expression of clock components was also disrupted in ischemia.⁹¹

Misalignment of internal clocks with external cues such as light has been associated with cardiac decompensation in mice. Transverse aortic constriction in mice causes pressure overload on the left ventricle leading to hypertrophy, fibrosis, and perivascular remodeling. Switching these mice to a 20-hour rhythm consisting of 10 hours of light and 10 hours of darkness causes myocardial thinning instead of hypertrophy, along with decreased systolic function and increased systemic blood pressure. Returning these mice to 24 hour rhythms converted this myocardial pathology into a hypertrophic cardiomyopathy.⁹² Disruption in the light-dark schedule after myocardial infarction also led to a deterioration in cardiac structure and function in a mouse cardiac ischemia model; an altered immune response to injury may underlie this functional deterioration.⁹³

After a myocardial infarction, it is common to admit patients to a coronary care unit where noise and light are constant. Considering the evidence presented, efforts should be made to decrease the disruption of a patient’s circadian rhythms as much as possible, particularly those who are critically ill and known to experience circadian light-dark exposure misalignment.⁹⁴

Studies of clinical outcomes are also consistent with a role for circadian rhythms in the response to injury and remodeling. Among a cohort of 6710 consecutive ST-elevation myocardial infarction (STEMI) patients, those with symptom onset between midnight to 6 am had a 30% higher risk of developing acute heart failure. Those with symptom onset in the morning had the smallest mean infarct size. It is important to note that in this cohort, the mean times between symptom onset and revascularization did not differ significantly between the different time intervals.⁹⁵ This was also confirmed by another study showing a higher risk for congestive heart failure with infarctions at night.⁹⁶ As in the animal models discussed above, it seems that ischemic insults at the wake sleep transition or through the rest phase have more detrimental effects and likely due to the disruption of the repair process that usually occurs during that time. Misaligning intrinsic rhythms with sleep rhythms also increases the risk of developing heart failure. A recent secondary analysis of the Sleep Heart Health Study found a significant association between a late bedtime (after midnight), and heart failure incidence. Also, later wake times had a similar association.⁹⁷ Long naps and sleep durations of less than 6 hours were also seen to be associated with a higher incidence of heart failure,⁹⁸ although, a recent analysis of the UK Biobank showed no association between night shift work and heart failure incidence. This was unlike atrial fibrillation and coronary heart disease, both of which were significantly associated with long term night shift work.⁹⁹

In addition to sleep, dysregulation of circadian blood pressure rhythms might be associated with heart failure. 24-hour blood pressure rhythms have been categorized based on the percentage decrease of mean nighttime/sleep blood pressure from mean daytime/awake blood pressure into a dipping (10% or greater decrease), non-dipping (less than 10% decrease), and riser patterns (any increase in mean nighttime blood pressure compared to mean daytime blood pressure). In a cohort of 951 elderly men with no heart failure or ventricular hypertrophy, a riser pattern of blood pressure was seen to be an independent predictor of heart failure. In patients with established systolic or diastolic heart failure, the riser pattern has been associated with increased morbidity and mortality.^100–102 Non-dipping and particularly reverse dipping was associated with the presence of left ventricular hypertrophy and higher late gadolinium enhancement seen on cardiac MRI indicating higher levels of interstitial fibrosis.^103,104 This is confirmed mechanistically by the fact that myocardial growth and repair happen at the wake-sleep transition and the rest phase which makes the myocardium more susceptible to hypertrophic stimuli during this time.⁵⁹ Despite these findings, since most studies are of a cross sectional design, more longitudinal studies that relate circadian blood pressure patterns to the development of heart failure are needed. Most longitudinal studies reporting a risk of cardiovascular disease describe composite outcomes rather than specifically heart failure. We summarize clock related factors impacting on the pathogenesis of heart failure in figure 3.

Figure 3. Circadian involvement of heart failure pathogenesis.

Disruption of core clock function is implicated directly in the pathogenesis of heart failure and indirectly in worsening decompensation after injury. A similar relationship can be seen when external cues are misaligned with already established rhythms which can also lead to myocardial injury particularly ischemic insults. Both myocardial injury and heart failure can lead to a disruption of core clock function which can further aggravate the deleterious effects of core clock disruption of cardiac function. Created with biorender.com. Illustration credit: Ben Smith.

Electrophysiology

Electrophysiologic studies in humans demonstrate robust diurnal rhythms in PR, RR, and QT intervals (Figure 2) that are all sensitive to shift change except for the PR interval indicating that the sinoatrial (SA) and atrioventricular (AV) nodes might be differentially sensitive to circadian disruption.¹⁰⁵ The cardiomyocyte clock seems to prevent abnormal QT-interval prolongation.¹⁰⁶ In hearts from patients with advanced heart failure, some genes known to be involved in electrical properties and arrhythmias showed strong circadian expression patterns.³⁹ There also seems to be a diurnal variation in the incidence of arrhythmias requiring implantable cardioverter defibrillator (ICD) therapy in heart failure patients.¹⁰⁷

Neurohumoral Factors and the Kidney

Volume status is a key factor behind heart failure symptoms and is a target of diuretic therapies. Secretion of natriuretic peptides by the heart oscillates rhythmically throughout the day. In rats, brain natriuretic peptide (BNP) has a rhythmic expression that is diminished when light is phase advanced or delayed. In human studies, blood pressure was not affected by the fluctuations of BNP across the circadian cycle.^9,22 Atrial natriuretic peptide (ANP) gene expression in rats was seen to be arrhythmic throughout the day.⁹ However, ANP protein quantification in perfused rat hearts demonstrated rhythmic levels of expression.¹⁰⁸ ANP concentrations oscillate rhythmically in humans, increasing at night and peaking before morning in an opposite trend to blood pressure.^109,110 (Figure 2).

The renin-angiotensin-aldosterone system (RAAS) is a key regulator of volume status and is maladaptively upregulated in heart failure. Under physiologic conditions, the RAAS oscillates rhythmically across the 24-hour day which is seen in both rats and humans^{109,111–120} (Figure 2). Plasma renin activity, angiotensin converting enzyme (ACE), angiotensin II, and aldosterone levels show circadian variation in humans, whereby the RAAS has a peak activity in the morning and a trough in the evening (Figure 2).^{109,113–120} The expression of angiotensin II receptors is rhythmic in rats; this rhythmicity was abolished when angiotensin II was constantly infused indicating the modulation of the receptor expression by its ligand which itself fluctuates in concentration across the day.²² The sympathetic nervous system can modulate the circadian rhythmicity of the RAAS via β-adrenergic receptors; blocking β-adrenergic receptors blunted this rhythm in human volunteers.¹²¹ Given the benefit of RAAS modulators in the treatment of heart failure, elucidating the molecular mechanisms behind such rhythmicity is important.

As a key regulator of fluid balance, the kidney is of importance in the pathogenesis and treatment of heart failure. Glomerular filtration rate (GFR), renal blood flow, and electrolyte excretion by the kidney fluctuate in a circadian pattern.¹¹⁸ Knocking out Bmal1 in mouse podocytes blunted this circadian fluctuation in GFR, ¹²² and knocking out Bma1 in renal tubular cells altered the renal transcriptome and metabolome.¹²³ Per1 seems to decrease sodium reabsorption by the epithelial sodium channel (ENaC), one of the genes induced by aldosterone, in the distal tubule.¹²⁴ The Na⁺/H⁺ channel in the mouse renal medulla was found to have a circadian expression pattern.¹²⁵ Although little is known about other ion channels, knowledge of their expression and functional patterns across the day might bring novel insights to understanding and treating and heart failure.

The autonomic nervous system also undergoes circadian fluctuations. The sympathetic nervous system is dominant through the day while the parasympathetic is dominant through the night. This is consistent with the peaking of catecholamine levels in the morning, and the drop in heart rate and cardiac output during the night (Figure 2).^118,126 Also, β-adrenergic receptors in the heart do show rhythmic changes in their expression.¹²⁷ The extent to which the sympathetic nervous system controls circadian rhythms in tissues is unclear. Even though catecholamines regulate cardiac core clock genes in vitro, they do not seem to influence these rhythms to the same degree in vivo.^17,49 However, the autonomic nervous system modulates rhythms of blood pressure and heart rate,^105,128 and it is a possible mechanism behind the circadian regulation of voltage gated channels and cardiac remodeling after injury hinting at an important role for the autonomic nervous system in the cardiac circadian clock.^129,130 It is important to note also circadian rhythmicity in the baroreflex.^128,131,132 Takotsubo syndrome is an intriguing pathology leading to reversible cardiomyopathy. It is mediated mainly by a sympathetic storm followed by microvascular dysfunction, metabolic derangements, and the activation of inflammation.¹³³ Interestingly, symptom onset in most Takotsubo syndrome patients in one study commences in the morning likely reflecting the sympathetic surge at that time of day.¹³⁴

The autonomic nervous system is also heavily involved in electrical properties of the heart whereby ECG parameters can be an indicator of sympathetic to parasympathetic balance. In a rat model of ischemic cardiomyopathy, heart rate differences between night and day were significantly reduced compared to control rats.¹³⁵ In humans, heart rate differences between night and day were not seen in heart failure patients unlike in their healthy counterparts.¹³⁶ This may be putting them at a higher risk for sudden cardiac death and all-cause mortality.^137,138

Blood pressure differences between night and day were significantly reduced in an ischemic cardiomyopathy rat model compared to control rats.¹³⁵ However, a nocturnal dip in blood pressure is still present in heart failure patients.^139–141 A longer duration of HFpEF was associated with a higher prevalence of the non-dipping pattern, ^141,142 and the non-dipping or riser patterns in ambulatory blood pressure were associated with adverse events in heart failure patients.^101,139 The day and night differences in diastolic blood pressure had a negative correlation with ejection fraction in 50 patients with stable systolic and diastolic heart failure.¹⁴¹ In patients with left ventricular hypertrophy due to primary hypertension, a higher left ventricular mass index was associated with a non-dipping or riser pattern in ambulatory blood pressure.¹⁴³ In an interesting study of familial amyloid polyneuropathy, cardiac thickening was associated with dipping level.¹⁴⁴ Patients with left ventricular assist devices (LVAD) also have circadian variation in motor currents with higher levels during the day. These rhythms seem to be more established at day 30 post implantation and are due to pressure gradients between the aorta and left ventricles which the device algorithm uses to determine motor currents.^145,146 Large scale longitudinal studies that control for the multiple factors that can influence nocturnal blood pressure dipping are needed to understand the utility of the nocturnal dip in risk stratification of heart failure patients.

Multiple studies have investigated the presence of diurnal rhythms in clinical events in heart failure patients. Episodes of acute cardiogenic pulmonary edema had a significant diurnal distribution peaking at 1 am.¹⁴⁷ This can be corroborated by the higher prevalence of pulmonary manifestations of heart failure in night time hospital admissions compared to day time admissions,¹⁴⁸ and to the higher rates of intubation at night in heart failure patients compared to the day time.¹⁴⁹ These rhythms can be explained by the increase in venous return associated with recumbency at night compared to a more upright posture during the day.¹⁴⁷ Because of the association of sudden cardiac death with the morning surge in sympathetic tone, any diurnal patterns of sudden cardiac death or the absence of such patterns can point to autonomic rhythms in heart failure. Studies are contradictory, however, with one small study with 158 deaths showing a uniform distribution of sudden cardiac death with a peak between 4 and 8 pm and another small study showing a bimodal peak in the am and pm, and a larger study of 517 deaths showing a peak between 7 and 8 am.^150–152 All cause deaths in heart failure patients do not seem to have any diurnal distribution.¹⁵³ Many studies are now using remote monitoring technologies in the management of heart failure patients.¹⁵⁴ These are of particular potential value in determining circadian rhythms in heart failure in the home setting.

We see a large involvement of the clock in physiologic cardiac function and in heart failure. Disrupting clock genes produces loss of cardiac contractility and myocardial mass or hypertrophy and diastolic dysfunction thus mimicking both HFrEF and HFpEF. The clock controls rhythms in cardiac function whereby contractile function is increased concomitant with the most suitable metabolic and growth and repair states. Clinical data is concordant with these observations in rodents. Thus, the deleterious response to injurious stimuli and the risk of heart failure are both exacerbated when external and internal cues are misaligned with intrinsic rhythms.

Chronotherapy

Chronotherapy, or circadian medicine, can include targeting the molecular clock itself or basing conventional therapies on its rhythmic outputs.¹⁵⁵ Many drug targets oscillate but so do metabolizing enzymes and transporters.¹⁵⁶ Because of the involvement of the clock in the pathogenesis of heart failure, clock targeting therapeutics seem like a rational prospect to evaluate. Recent studies in model systems have shown that the clock mechanism can be manipulated for cardioprotection.

Reitz et al⁷⁵ treated wild type mice with a REV-ERB agonist, SR9009, after ischemia-reperfusion. This improved cardiac function and minimized cardiac hypertrophy. SR9009 decreased REV-ERBα and REV-ERBβ mRNA and protein concentrations which, in turn, decreased immune cell infiltration of the myocardium post injury and reduced NLRP3 (NOD-, LRR- and pyrin domain-containing protein 3) inflammasome function.⁷⁵ These findings are consistent with another study in which SR9009 was administered before surgery.¹⁵⁷ These beneficial effects of SR9009 agonism despite downregulation of REV-ERB are unexpected as cardiomyocyte specific deletion of Rev-erb leads to cardiac decompensation and global deletion leads to peritoneal macrophage dependent inflammation.^6,158 One explanation could be that the drug triggers a negative feedback loop of Rev-erb expression.¹⁵⁹ Another is that improved function is caused by a yet to be identified off target effect of the drug. SR9009 also reduced heart weight and left ventricular internal dimensions in transverse aortic constriction wild type mice, thus decreasing the expected hypertrophy in response to pressure overload.³⁶ This may be due to blocking cellular remodeling in response to neurohormonal stress through transcriptional repression.¹⁵⁹ Furthermore, SR9009 administration after hypertrophy was established in mice subjected to transverse aortic constriction halted the decline in cardiac function. This hints at the potential of an agonist in established heart failure.¹⁵⁹

Another REV-ERB agonist, GSK 4112, also restrained hypertrophy and stress marker expression in cultured murine cardiomyocytes. It also prevented transforming growth factor β (TGF-β) induced collagen expression in ex-vivo culture of human lung fibroblasts obtained from idiopathic pulmonary fibrosis patients.^69,159 Given the diversity of phenotypes seen due to disrupting different clock genes, drugs targeting specific elements of the clock might have differing effects. These early experiments highlight the likely complexity of directly manipulating the clock in organ systems. Given the diversity of tissues expressing molecular clocks, potentially different tissue access of systemically administered drugs, the high levels of redundancy amongst core clock genes and the diverse functionality of clocks in regulating immune and metabolic processes, targeting the clock itself may be no small challenge.

A more immediate opportunity is to widen the therapeutic index of existing drugs. An example of this approach is the labelled instruction to take statins in the evening, based on the oscillation of their target, HMG CoA Synthase. However, this is likely just the beginning of timed of cardiovascular therapeutics.¹⁵⁵ Spontaneously hypertensive rats given valsartan before the rest period had a significantly lower cardiac mass index than rats given valsartan at the beginning of the wake period.¹⁶⁰ In humans, the bioavailability of the orally administered cardiac glycoside, digoxin, is known to be influenced by time of day.^161,162 Up to 50% of the digoxin administered to patients can be inactivated by the gut microbiome which itself oscillates and is increasingly recognized as a modulator of drug metabolism.^163,164

The MAPEC and the HYGIA clinical trials,^165,166 report a lower risk of heart failure development in patients taking their prescribed anti-hypertensives at bedtime compared to ingesting them in the morning. However, these trials have both been controversial. There was no standardized treatment plan for either trial, which meant that the actual drugs prescribed differed between participants ^167,168

Intense light has been proposed as a cardioprotective strategy involving the clock. Intense light enhances the amplitude of oscillations of PER2, hypoxia inducible factor (HIF1A) and angiopoietin like 4. This seems to protect against oxidative stress and ischemic injury.¹⁶⁹

Finally, guidelines recommend exercise inclusion in heart failure treatment plans.¹⁷⁰ Exercise has been shown to entrain the mammalian circadian clock which points to the clock being one potential mechanism behind the benefits of exercise in heart failure.¹⁷¹ Perhaps the timing of exercise will prove to have a differential impact on cardiac outcomes.

Conclusion

Although our understanding of circadian physiology has advanced, this has yet to be applied to the treatment of heart failure. Cardiologists today have few therapeutic options to delay the progression of heart failure and to alleviate its symptoms. Given the evident circadian regulation of cardiac function in both healthy and failing hearts, the time-of-day dependency of these therapeutics needs to be better understood and recorded. Although the safety of directly targeting the clock may prove to be a challenge, we can likely exploit our knowledge of the clock to refine the administration of existing therapies, either to improve efficacy or to minimize risk.

Non-Standard Abbreviations and Acronyms

ACE

angiotensin converting enzyme

ANP

atrial natriuretic peptide

AV

atrioventricular

BMAL1

brain and muscle ARNT-like protein-1

BNP

brain natriuretic peptide

CCM

cardiomyocyte Clock mutant

CLOCK

circadian locomotor output cycles kaput

CRY

cryptochrome

ENaC

epithelial sodium channel

eNOS

endothelial nitric oxide synthase

GFR

glomerular filtration rate

GLUT

glucose transporter

HFpEF

heart failure with preserved ejection fraction

HFrEF

heart failure with reduced ejection fraction

ICD

implantable cardioverter defibrillator

KLF15

kruppel like factor 15

KO

knock out

LAD

left anterior descending artery

LVAD

left ventricular assist device

MRI

magnetic resonance imaging

NLRP3

NOD-, LRR- and pyrin domain-containing protein 3

PER

period

PPP

Pentose Phosphate Pathway

RAAS

renin-angiotensin-aldosterone system

SA

sinoatrial

SCN

suprachiasmatic nucleus

SERCA

sarco/endoplasmic reticulum Ca²⁺-ATPase

SGLT2

sodium glucose transporter 2

STEMI

ST-elevation myocardial infarction

T3

triiodothyronine

TGF- β

transforming growth factor-β

VGCC

voltage gated calcium channel

VIP

vasoactive intestinal peptide