Repeated exposure to transient obstructive sleep apnea – related conditions causes an atrial ﬁ brillation substrate in a chronic rat model

BACKGROUND High night-to-night variability in obstructive sleep apnea (OSA) is associated with atrial ﬁ brillation (AF). Obstructive apneas conditions resulted in AF substrates and was associated with increased AF susceptibility. Mild-to-moderate OSA with high night-to-night variability may deserve intensive management to prevent atrial substrate development.


Introduction
Obstructive sleep apnea (OSA) is prevalent in up to 70% of all patients with atrial fibrillation (AF) and affects the efficacy of pharmacological and catheter-based strategies for rhythm control. [1][2][3] In nonrandomized observational clinical trials, treatment of severe OSA improved AF catheter ablation outcomes. 1,2 Most clinical and preclinical research mainly focused on the role of severe OSA. However, whether nonsevere OSA leads to an AF substrate remains unclear.
OSA episodes are characterized by high-frequency desaturation-reoxygenation, and intrathoracic pressure swings occur during ineffective inspiration against the occluded upper airways, which increases transmural pressure gradients and may expose the thin-walled atria to increased stretch. 4,5 Additionally, despite the described high prevalence of OSA in patients with AF, longitudinal assessment of sleep apnea severity in patients with AF and cardiovascular implantable electronic devices in the VARIOSA-AF study (Variability of Sleep Apnea Severity and Risk of Atrial Fibrillation) indicated that most patients with AF do not demonstrate severe sleep apnea every day, but rather mild or moderate sleep apnea and considerable intraindividual night-to-night variability. [6][7][8] This creates a scenario of transient exposure to OSA-related conditions, characterized by intermittent hypoxia and intrathoracic pressure swings. Such effect of transient exposure to OSA-related stressors have not been considered in recent chronic animal models, which mainly exposed the animals to severe OSA every day. 9 We hypothesized that transient exposure to obstructive respiratory events characterized by intrathoracic pressure changes leads to atrial damage and that repeated exposure to these transient OSA-associated conditions results in a structural substrate for AF.
In this study, we developed a novel rat model for OSA in order to study the transient and acutely reversible effects of high-frequency desaturation-reoxygenation in combination with negative intrathoracic pressure changes induced by intermittent negative upper airway pressure (INAP) simulating obstructive respiratory events. Additionally, the development of an atrial arrhythmogenic substrate as a cumulative consequence of repeated exposure to transient and acutely reversible conditions induced by INAP (simulating longterm mild-to-moderate OSA with high night-to-night variability) was characterized.

Methods
For detailed methods, please see Online Supplemental Methods.

Animal model
All animal studies were performed in accordance with the German law for the protection of animals. The investigation conforms to the Guide for the Care and Use of laboratory Animals published by the US National Institutes of Health (eighth edition; revised 2011). The study was approved by the regional Animal Welfare Inspectorate (Saarl€ andisches Landesamt f€ ur Verbraucherschutz: #18/2014).

Anesthesia and killing
During experiments, rats were anesthetized with 2% isoflurane and 98% oxygen administered via a nebulizer. For killing, rats were sedated with nebulized isoflurane (2% plus 98% oxygen) and a single intraperitoneal injection of ketamine hydrochloride (80 mg/kg BW) and xylazine hydrochloride (6 mg/kg BW). The thorax was opened bilaterally along the midaxillary line, and the diaphragm was removed. For immediate killing and tissue sampling, the heart was harvested and dissected in left and right atria and ventricles.

Application of hypoxia combined with INAP
INAP was applied via customized animal masks (#E5-5100, Breezer, Medico-Lab, Winsen, Germany) by a negative pressure device, which consists of a 50-L negative pressure container and a vacuum pump controlled by a manometer (Manometer Type 831, WIKA, Alexander Wiegand SE & Co. KG, Germany). The animal mask was connected via a flexible tube to the vacuum container, and a solenoid valve was opened and the tube and the vacuum container created a closed system ( Figure 1A). The Mueller maneuver, which is defined as forced inspiration against airway obstruction, is used in the clinical setting to simulate conditions occurring during obstructive respiratory events, in particular negative thoracic pressure. 10 As a modification of this maneuver, we noninvasively applied a defined negative pressure of 240 hPa. During INAP, respiration efforts (inhalation and expiration) and pressure leakage of any kind were detectable in realtime recordings via a pressure sensor (Isotec/Sigma-Aldrich Chemie GmbH, Munich, Germany) with a data acquisition and analysis software (Notocord-hem Evolution, Notocord, Croissy, France) ( Figure 1B). Vital parameters were visualized by a multichannel monitor (Vismo PVM-2701k, Nihon Kohden, Tokyo, Japan), and an SpO 2 probe was attached to the lower limb of the rat to record oxygen saturation ( Figure 1C).

Acute and chronic trials
In the acute test series (ATS), INAP was applied for 1 minute followed by a 4-minute resting period. INAP (INAP-ATS, n 5 8) sequence was repeated 48 events/4 h (apnea-hypopnea index [AHI] 12 events/h). In pilot experiments, a 4-minute resting period between 2 consecutive INAP applications ensured an optimal recovery of vital parameters. Rats were killed immediately after the last INAP application (isoflurane anesthesia 2% plus 98% oxygen) and an intraperitoneal injection of ketamine hydrochloride (80 mg/kg BW) and xylazine hydrochloride (6 mg/kg BW). As a control group (CTR-ATS, n 5 9), rats were anesthetized during the time period of the experiments but no INAP was applied. Additional rats underwent the same ATS protocol but were killed after 24 hours of the resting period (CTR-ATS-REC: n 5 5; INAP-ATS-REC: n 5 5) ( Figure 1D). After killing, blood and cardiac tissue for biochemical analysis were preserved. Wet lung tissue was weighed and dried for 48 hours in a 60 C incubator and afterward weighed again for evaluation of fluid accumulation due to lung edema.
In the chronic test series (CTS), INAP was applied for 1 minute followed by a 9-minute resting period. The INAP (INAP-CTS, n 5 10) sequence was repeated 24 events/4 h (AHI 6 events/h) every second day throughout 3 weeks to simulate high night-to-night variability. As a control group (CTR-CTS: n 5 10), rats were anesthetized during the time period of the experiments but INAP was not applied. Twenty-four hours after the last INAP application, rats were killed ( Figure 1D).
In CTS rats, blood pressure (by telemetry), left ventricular (LV) hemodynamics (by invasive pressure measurements), and AF inducibility (by transesophageal atrial stimulation) were determined. Cardiac tissue was harvested for histologi-cal, immunohistochemical, biochemical, and microarray analyses, which were performed to characterize the development of an AF substrate.

AF inducibility and duration
A surface electrocardiogram (ECG) was recorded using Notocord-hem Evolution. A lead with the optimal P-wave amplitude was selected to allow an optimal visualization of atrial activation. Susceptibility to AF was tested by using transesophageal atrial burst stimulation. In sedated rats, a 4-F catheter was placed in the esophagus in the height of the left atrium. An initial stimulation (cycle length 200 ms; amplitude twice the diastolic threshold; stimulus width 8.6 ms) was performed to ensure atrial capture in the surface ECG. Repetitive 3-second bursts of stimuli (cycle length 10 ms; amplitude twice the diastolic threshold; stimulus width 8.6 ms) were applied. When the surface ECG showed indistinguishable P waves and irregularly alternating RR intervals for at least 1 second, AF was diagnosed. Each subject was burst paced 20 times independent of successful AF induction.

Statistical analysis
Data are expressed as mean 6 SEM. An unpaired Student t test (2-tailed) was used for statistical analysis comparing 2  independent groups, a paired analysis was used for intrasubject analysis. P values ,.05 were regarded as statistically significant. Statistical analysis was performed using GraphPad Prism version 8.0.1 (GraphPad Software, La Jolla, CA).

ATS: Transient short-term effects of INAP
INAP significantly increased respiratory rate compared with sole isoflurane deprivation ( Figure 2A) and led to increases in respiratory efforts ( Figure 2B). In the ATS-REC groups, mean oxygen desaturation did not differ from that in the ATS groups, demonstrating valid reproducibility during the ATS protocols ( Figure 2C).
Left ventricular expression of brain natriuretic peptide  Figures 4A and 2C). Heart rate, mean arterial blood pressure, and arterial pulse pressure measured by telemetry measured during 3 weeks of the INAP protocol did not show any differences between the CTR and INAP groups ( Figures 4C-4E). Blood pressure measured using the tail-cuff method during the CTS protocol confirmed the finding of telemetry-based measurements (Online Supplemental Table 1).

LV hemodynamics and remodeling
Chronic application of INAP had no effect on heart weight/ tibia length ratio, nor on LV cardiomyocyte diameter, GSH/GSSG ratio, or BNP mRNA expression as compared with CTR application (Online Supplemental  Figure 2). Invasive LV hemodynamic measurements revealed no significant changes in diastolic or systolic parameters (Online Supplemental Table 2).

Atrial remodeling
In INAP-CTS, atrial antioxidative capacity (GSH/GSSG ratio) measured 24 hours after the last applied INAP maneuver did not differ between groups ( Figure Table 1). However, AF inducibility and durations were significantly increased in INAP-CTS (Figure 7). Telemetry recordings did not reveal any spontaneous AF periods throughout the CTS.

Discussion
To investigate the effect of mild-to-moderate OSA with high night-to-night variability on AF substrates, we developed and described a novel rat model for OSA-related atrial remodeling, which represents a unique approach (1)

Major findings
Four hours of simulated OSA transiently increased atrial oxidative stress, which completely reversed within 24 hours. Mild-to-moderate OSA with high night-to-night variability simulated by repeated exposure to these transient biological responses related to OSA for 3 weeks resulted in an arrhythmogenic substrate for AF.

Comparison with previous animal studies
Several mechanisms have been identified to contribute to AF in the setting of severe OSA. Acute apnea-associated atrial electrophysiological changes and increased occurrence of triggers due to intermittent hypoxemia followed by reoxygenation, intrathoracic pressure changes during ineffective breathing attempts, and arousal-related sympathovagal activation create a complex and dynamic substrate for AF during sleep. 4,14 Additionally, long-term severe OSA has been shown to be associated with atrial enlargement, voltage reduction, and site-specific conduction abnormalities in humans 15 and with connexin dysregulation and increased AF inducibility in rodent models. 9 In recently described animal models for sleep apnea (SA) with daily orotracheal intubation during deep sedation, 9 implantable tracheal balloons, 16  transiently increased atrial oxidative stress, which was reversible within 24 hours. Pilot data in our rat model suggest that repetitive mechanical cardiac pressure alterations as in this model induced by INAP is crucial for a prominent decrease in antioxidative capacity in the atria, which could not be induced by intermittent hypoxia alone. when atrial oxidative stress was not increased anymore. Our results provide insights into the underlying mechanisms of increased AF inducibility including the development of a structural atrial remodeling process predominantly characterized by reduced connexin 43 expression. Importantly, the development of an atrial arrhythmogenic substrate was independent of the development of hypertension or overt LV diastolic or systolic dysfunction. Intrathoracic pressure swings, as observed during INAP, result in cyclical atrial stretch, which mechanically stress atrial tissue. 4,18 Rodent models confirm that repeated stretch transiently activates stretch-sensitive and oxidative stress sensitive profibrotic signaling pathways and that stretch in combination with low doses of H 2 O 2 sensitize cells to early afterdepolarizations, which may further contribute to increased AF susceptibility. 19 Affymetrix analysis in our study identified a differential expression of multiple components of the adenosine monophosphate-regulatory protein kinase signaling pathway in rats with INAP, which has been shown to be involved in the initiation, maintenance, and progression of atrial arrhythmias. 11 INAP resulted in a differential mRNA regulation of members of the insulin signaling pathway, such as insulin-like growth factor 1 receptor and ETS transcription factor 1, both shown to play an important role in the pathogenesis of AF. 12

Study limitations
In this novel animal model for OSA, we applied stable and reproducible INAP maneuvers at 240 hPa, which was also measured in patients with obstructive OSA. 10 The actual pressure in the thoracic cavity, however, was not measured, which might differ from the negative pressure applied during INAP maneuvers. In order to address the effect of experimental conditions, such as volatile sedation, 21 on the transient increase in oxidative stress, we introduced a CTR group in our study. Affymetrix analysis (Affymetrix, Santa Clara, USA), was performed in atria of rats killed 24 hours after the last INAP maneuver. Similar to the transient increase in atrial oxidative stress, gene expression levels may be partially reversible within this period, which may have contributed to an inhomogeneous gene expression pattern in the INAP group. Inflammatory cytokines and infiltration of macrophages were studied in this model, indicating increased systemic inflammation, which is associated with OSA. 21 However, precise inflammasome signaling was not studied in depth. Moreover, neither ion channel expression nor function was studied. In this model, we investigated the isolated effects of OSA on atrial remodeling. Because of the small size of rat atria, an analysis of the spatial distribution of atrial remodeling processes was not possible. Moreover, additional classical contributory factors related to common comorbidities in patients with AF and sleep apnea, including hypertension or obesity, 22 were not incorporated in the present rat model. Whether treatment of OSA, without targeting concomitant risk factors within an aggressive risk factor modification program, can prevent progression of AF or reverse structural remodeling processes in the atrium is unknown. 22 Clinical investigation is needed to validate the applicability of our findings in humans.

Clinical perspectives
OSA is highly prevalent in patients with AF and affects the efficacy of pharmacological and catheter-based strategies for rhythm control. 1,2,15,23 Although clinical assessment of OSA-severity and OSA treatment initiation is primarily guided by the AHI, representing the number of apneas and hypopneas per hour of sleep, a more detailed characterization of the dynamicity of apneas in patients with AF may result in a better disease-based assessment of OSA in patients with AF. Moreover, our results have relevance to therapeutic approaches. Although treatment of OSA is mainly recommended in patients with AF and severe sleep disordered breathing, 24 our findings suggest that even mildto-moderate OSA with high night-to-night variability is sufficient to result in an arrhythmogenic substrate and may therefore warrant consequent management. Identification of mild-to-moderate OSA with high night-to-night variability may easily be missed by a single night sleep study spot assessment of sleep disordered breathing. High night-to-night variability in OSA severity may require a long-term OSA screening approach to capture the daily OSA pattern and severity. 6

Conclusion
Simulated acute OSA by INAP transiently increased atrial oxidative stress. Although acute INAP-induced oxidative stress was completely reversible, cumulative exposure to transient INAP-related conditions every second day resulted in an arrhythmogenic AF substrate. Not just severe, but even mild-to-moderate OSA with high night-to-night variability may represent a treatment target to prevent the progression of AF substrates.