Sleep calibrates atherosclerotic cardiovascular disease

Abstract

Sleep modulates cardiovascular health, and recent studies have begun to uncover underlying mechanistic links. An integrated translational approach that combines animal models and human trials will enrich scientific discovery, improve therapy, and help to alleviate the global burden of insufficient sleep and cardiovascular disease.

Atherosclerotic cardiovascular disease (ASCVD) and its associated consequences — myocardial infarction, stroke and peripheral artery disease — remain major contributors to mortality worldwide. Over the past few decades, new therapeutic treatments have emerged that control hypercholesterolemia, hypertension and diabetes, while public health initiatives have substantially decreased smoking and continue to address obesity. Despite these achievements, the rates of ASCVD continue to rise. Moreover, global success at combating communicable diseases has extended lifespan and expanded the pool of people at risk for ASCVD to include populations in developing countries that often live in bright, noisy and overcrowded cities. Together, these factors have contributed to a persistent residual ASCVD risk that remains high.

The influence of lifestyle factors, including sleep and its disruption, on ASCVD risk and pathology remains poorly defined. Sleep is integral to health and adults should spend one-third of their life asleep. Despite its importance, an estimated 35–40% of adults do not get the recommended 7–8 hours of sleep per night. This may not capture the full extent of the sleep-deficiency crisis, as self-reported sleep is often overestimated and objective measures of sleep remain insufficiently used. Contemporary culture and lifestyles disincentivize sleep, and modern technology, instantaneous communication and global travel have contributed to an epidemic of insufficient sleep. Furthermore, sleep disorders, such as obstructive sleep apnea (OSA), are common and under-recognized, with global estimates reaching almost one billion people. We know that sleep has a profound effect on cardiovascular health. Short or disrupted sleep increases the risk of ASCVD events¹, subclinical atherosclerosis², vascular inflammation³ and incidence of myocardial infarction⁴. Importantly, insufficient sleep increases the prevalence of ASCVD despite other risk factors or underlying genetic predispositions^4,5, which suggests that sleep is an important and understudied variable that influences ASCVD incidence and pathology. Sleep disorders including OSA, insomnia, narcolepsy and, paradoxically, excessive sleep, also increase ASCVD risk¹. Studies using animal models have further cemented the association and begun to uncover underlying mechanistic links. In mice, sleep fragmentation or disruption increases hematopoietic stem cell (HSC) proliferation and myelopoiesis, inflames the vasculature and enlarges atherosclerotic lesions^6–8. Despite these advances, the fundamental biological, cellular and molecular processes that link sleep to cardiovascular health remain ill-defined.

Animal models are indispensable tools for the mechanistic study of disease pathophysiology and inform clinical practice and treatment. Recent data have shown that fragmentation of sleep in mice dysregulates sleep/wake neuropeptidergic signaling in the hypothalamus that enhances HSC proliferation, hematopoiesis, and the generation of monocytes that seed the vasculature to enlarge atheromata⁶. Moreover, animal studies have uncovered a role for sleep in HSC epigenetic programming and diversity^7,8, the maintenance of the vascular endothelium, and the regulation of metabolic balance. These murine studies have been replicated in humans^8,9 and have undoubtedly advanced our knowledge of sleep-mediated atherosclerosis; however, the use of animals to study sleep remains limited. Sleep is the only universal animal behavior, but sleep architecture, timing and abundance vary greatly within the mammalian class and across the animal kingdom. For example, large herbivores require very little sleep, 2–6 h, whereas rodents typically sleep for more than 10 h per day. Some animals perform single-hemisphere sleep whereas others lack certain sleep stages all together. Mice and rats are the most common laboratory animals in modern biomedical research and even their sleep is different to that of humans. Rodents are nocturnal, although about 20% of their sleep time is spent during their ‘active’ (dark) phase. Rodent sleep is more fragmented than humans and they cycle through stages more quickly. Even among inbred strains of mice, sleep is variable and dependent on genetic background. Laboratory conditions including social housing, diet, food timing, bedding and nesting, temperature, humidity, lighting and noise influence rodent sleep. Although these nuances may be viewed as impediments to using rodents as models of sleep, these environmental and genetic factors reflect the wide variation in human sleep. There is substantial variability in human sleep timing and duration driven by cultural and societal norms, socioeconomics, genetics, environment, chronotype, disease and lifestyle choices. Human sleep is also highly dependent on developmental stage and aging. The use of animal models to study the mechanistic pathophysiology of sleep-mediated ASCVD should therefore be embraced, refined, improved and contextualized. Harnessing the variability of animal and human sleep will lead to the discovery of new underlying biology and disease pathology (Fig. 1).

Fig. 1 |. Sleep and cardiovascular health.

A combination of animal models and human trials will lead to a better understanding of the influence of sleep on cardiovascular diseases such as ASCVD.

Rodent models of experimental sleep manipulation (disruption, restriction or extension) largely fall into broad categories: physical, physiological, pharmacological, genetic and opto-/chemogenetic. Physical disruption through a rotating wheel or platform, an automated moving bar, or gentle handling, offer the ability to curtail sleep in virtually any mouse model or strain and in combination with other interventions, with effects on stress examined. Physiological interventions such as intermittent hypoxia are widely used to model sleep apnea. Pharmacological or genetic manipulation of sleep-mediating genes and pathways are broad interventions that can restrict or augment sleep, but organ or tissue specificity, circadian timing and developmental influences need to be cautiously considered. Opto- and chemogenetic systems are the most precise tools to control sleep, wakefulness and underlying neural circuitry. The delivery of light to the brain to stimulate or inhibit specific neuronal populations enables millisecond control of neural firing and meticulous temporal command of sleep circuits over the circadian day. Naturally, each model has varied advantages, disadvantages and specificity, and their suitability over months-long atherogenesis must be considered. Crucially, each model can be applied to Apoe^−/−, Ldlr^−/− or PCSK9-AAV-injected mice fed an atherogenic diet to manipulate sleep in animals with dyslipidemia, vascular and systemic inflammation, and varying stages of atherosclerosis. A combination of sleep models and careful consideration of alternative mechanistic hypotheses are needed. Undoubtedly, sleep maintains a bidirectional relationship with metabolism, circadian rhythms, energy balance and in particular, stress. Several mouse models, accurate tools and cautious data interpretation will dissect the contributions of specific pathways and systems to sleep-mediated experimental outcomes. Disruption of the core molecular clock machinery (such as Clock, Arntl, Nr1d1 and Nr1d2) robustly interrupts circadian rhythms but these are parallel and distinct models that encompass a plethora of complex physiological changes beyond sleep abundance and quality. Further work is also necessary to uncover new neural circuits that maintain sleep and wakefulness to develop novel opto- and chemogenetic systems, harness sleep-regulating genes and pathways, and expand the experimental toolbox. In summary, current models are effective, successful and rigorous, but their refinement, the discovery of new fundamental mechanisms, and the combinatorial use of several models will advance our understanding of sleep and its influence on cardiovascular health and ASCVD.

In humans, the largest randomized controlled trials assessing the benefit of sleep-specific interventions in ASCVD are on the use of continuous positive airway pressure (CPAP) therapy in non-sleepy patients with OSA. Individuals with OSA randomized to receive CPAP, with real-world adherence to therapy, did not experience a significant reduction in ASCVD events¹⁰. However, subgroup analyses among those individuals adherent to CPAP demonstrate ASCVD benefit, and retrospective studies continue to show CPAP improves risk of hospitalization and all-cause mortality. This lack of effect is probably not because CPAP fails to treat OSA but instead due to heterogeneous ASCVD risk and pathology in individuals with OSA. Substantial knowledge gaps remain on how disrupted sleep and hypoxia in combination with CPAP therapy modulate atherosclerosis and inflammation in patients with OSA. Leveraging animal models and translational studies to identify new atherogenic processes in patients will help to explain the mixed ASCVD responses to apnea therapy. Pragmatic studies are also needed to assess multimodal interventions that improve adherence and test personalized combination therapies for cardiovascular benefit. Such work would help refine the current monolith approach of treating all patients with CPAP, and identify patients with OSA and CPAP-responsive ASCVD.

Therapeutic options for OSA focus on relieving obstructive apneas and hypopneas, whereas for insomnia, sleep consolidation and extension through cognitive behavioral therapy or sedative hypnotic medications are first line. Little attention has been paid to the inflammatory consequences of OSA events and the short or fragmented sleep of insomnia. This is despite an understanding that inflammatory biomarkers, including monocytes, CRP, IL-6 and TNF, are increased in human OSA, insomnia and sleep deprivation¹¹. Anti-TNF agents have been trialed in OSA and insomnia, and have demonstrated improvements in sleep and quality of life¹². However, whether anti-inflammatory therapy reduces ASCVD risk in patients with sleep disorders is unknown and merits exploration. Translational work bridging animal models and human participants will furnish new interventions targeting inflammatory ASCVD in OSA and poor sleep.

Quantification and monitoring of sleep in rodents and humans is vital to understanding links with ASCVD. We do not fully know how ASCVD or its dietary, genetic and lifestyle risk factors alter sleep and how these changes in turn contribute to disease. The gold standard of human and rodent sleep assessment is electroencephalogram (EEG) and electromyography (EMG) recording and analysis. Although accurate, this method requires invasive surgical electrode implantation in mice or attachment of electrodes to the scalp of humans and highly trained experts to score EEG and EMG data. For rodent studies, alternatives are emerging — including automated EEG and EMG scoring, video monitoring, motion sensing and machine learning — that offer robust and high-throughput sleep analysis. In humans, automated EEG assessment has been validated¹³ and frontal-only EEG allows for increased ease-of-use and accurate at-home participant-applied automated polysomnography (PSG) scoring¹⁴. Peripheral arterial tonometry (PAT), an alternative to PSG, captures surges in sympathetic tone as a consequence of respiratory events and/or arousals and has been specifically validated in patients with OSA. Actigraphy is another accurate and unobtrusive method for quantifying sleep fragmentation and duration, and circadian rhythm. Modern consumer-wearable devices infer sleep time and quality from actigraphy, heart rate, blood pressure and blood oxygenation. As these technologies proliferate, abundant sleep data will be ripe for high-level analysis to identify population trends and those at risk of ASCVD. Modern machine and deep learning analysis of this ubiquitous wearable data will enrich human sleep–ASCVD trial design and scientific output.

To advance our knowledge and treatment of sleep-mediated ASCVD, an integrated translational approach is needed. Each vicissitude of sleep — fragmentation, deprivation, extension, variable timing or disorders, such as OSA, and their treatment — is likely to have unique biological outcomes, contributions to disease, and therapeutic opportunity. It is therefore crucial to amalgamate experimental approaches in animals and humans to uncover new biology at a fundamental level. Scientific and methodological integration will progress understanding, improve therapy and expand successful public health initiatives that alleviate the prodigious global burden of insufficient sleep and cardiovascular disease.