Early-onset sleep alterations found in patients with amyotrophic lateral sclerosis are ameliorated by orexin antagonist in mouse models

Lateral hypothalamic neurons producing melanin-concentrating hormone (MCH) and orexin/hypocretin are involved in sleep regulation. Both MCH and orexin neurons are altered in amyotrophic lateral sclerosis (ALS), the most common adult-onset motor neuron disease. However, sleep alterations are currently poorly characterized in ALS, and could represent either early symptoms or late consequences of disease progression. We characterized sleep architecture using polysomnography in cohorts of both early ALS patients without respiratory impairment and presymptomatic carriers of mutations leading to familial ALS. We observed sleep alterations, including increased wake and decreased deep sleep (non-rapid eye movement—NREM3) and increased wake correlated with worse cognitive performance, in particular, verbal fluency in both cohorts. These changes in sleep architecture were replicated in three mouse models of familial ALS, Sod1G86R, FusΔNLS/+ and TDP-43Q331K mice. Altered sleep structure in mice was fully rescued by per os administration of a dual-orexin receptor antagonist, and partially rescued by intracerebroventricular MCH.

Methods
ALS patients were recruited from the inpatient and outpatient clinics of the neurologic department of the University Hospital of Ulm, Germany. The inclusion criteria for ALS patients included a diagnosis of definite ALS based on the revised El Escorial criteria. Presymptomatic carriers of fALS genes were recruited through the study centre of the Neurological University Hospital, through which first-degree relatives of confirmed familial ALS patients receive longitudinal follow-up and counselling. Controls were recruited from the general population at the neurology clinic, and matched to ALS patients based on age, sex, and geographical location; the requirement for this group was the exclusion of neurodegenerative diseases. All individuals in the control group were unrelated to ALS or familial ALS.

The study in the ALS patient cohort was approved by the Ethics Committee of the University of Ulm (reference 391/18), as well as the study in the presymptomatic carriers which was also approved by the Ethics Committee of the University of Ulm (reference 68/19), in compliance with the ethical standards of the current version of the revised Helsinki Declaration. All participants gave informed consent prior to enrolment.

Medical history was documented. For ALS patients, the ALSFRS-r and characteristics of disease progression were documented (site of first paresis/atrophy, date of onset). All participants also completed validated daytime sleepiness and sleep quality questionnaires, namely the Epworth Sleepiness Scale (ESS) and the Pittsburgh Sleep Quality Index (PSQI).

The exclusion criteria were intended to exclude all possible circumstances that might otherwise alter sleep architecture. For this reason, participants who had an apnoea-hypopnea index (AHI) above 20 per hour or participants who had a periodic limb movement index (PLMSI) above 50 per hour were excluded. In particular, we intended to exclude respiratory insufficiency in ALS patients. Respiratory insufficiency develops earlier or later in the progression of ALS, depending on the individual course, but is generally present in advanced stages, and is known to influence sleep architecture. For this reason, ALS patients received transcutaneous capnometry in addition to polysomnography.

Cognition was measured with the German version of the Edinburgh Cognitive and Behavioral ALS Screen (ECAS) by trained psychologists. The ECAS addresses cognitive domains of language, verbal fluency, executive functions (ALS-specific functions) and memory and visuospatial functions (ALS non-specific functions). Age and education-adjusted cut-offs were used. Behavioural changes were assessed by patient caregiver/1st-degree relative interviews on disinhibition, apathy, loss of sympathy/empathy, perseverative/stereotyped behaviour, hyperorality/altered eating behaviour and psychotic symptoms.

All participants, ALS patients, healthy controls, fALS gene carriers and fALS controls underwent a single night full polysomnography, involving monitoring of various physiological parameters including electroencephalogram (EEG), surface electromyogram (EMG), electrooculogram (EOG), respiratory effort and flow, pulse and oxygen saturation. All measurements were conducted according to the criteria of the American Academy of Sleep Medicine (AASM) guidelines. The EEG electrodes were placed according to the international 10-20 system, the following electrodes were used in each subject: Fz, C3, C4, Cz, P3, P4, Pz, O1, O2, A1, and A2. The sampling rate was 512 Hz in each case. The individually different points in time at which the participant turned off the lights and tried to sleep were marked with a "lights off" marker in each recording.

Analyses were performed using available Python packages (only compatible with Python 3.10 or newer, Python Software Foundation. Python Language Reference, version 3.12. Available at http://www.python.org) relying on MNE package. EEGs were first de-identified using the open-source Prerau Lab EDF De-identification Tool (Version 1.0; 2023) in Python (Prerau Lab EDF De-identification Tool [Computer software], 2023, Retrieved fromhttps://sleepeeg.org/edf-de-identification-tool). De-identified EEGs were then notch-filtered to remove the 50Hz powerline. Independent component analysis was performed to remove all remaining artefacts from the signal. Analyses were limited to both sensorimotor cortices (C3 and C4), which are known to be impaired at the onset of the disease, as well as nearby interhemispheric sulci (Fz, Cz and Pz). Sleep staging was performed on a 6-hour window with 30-second epochs, starting when lights were turned off, using YASA deep learning algorithm (v0.6.4), as well as the spectral analysis. Time in bed and total sleep time were calculated over the whole recording period. The automated sleep staging, hypnograms, and spectrograms were performed using Welch’s method. Sleep pressure was determined using the area under the curve (AUC) of Delta power (0.5-4Hz) of the first hour of the 6-hour window. Simpson’s rule was used to compute the AUC. REM efficiency was computed by dividing Theta power (4-8Hz) by Delta power (0.5-4Hz) specifically during REM epochs. Sleep staging and analysis were performed following the AASM’s guidelines.