Biomarking the Alzheimer landscape: could melatonin be the key to detect subclinical epileptiform activity?

Worldwide, more than 55 million people suffer from dementia, with Alzheimer’s disease (AD) being its most prevalent cause (60–70%) (1). Furthermore, dementia has become the most important cause of death in Belgium since 2019 (2). Despite a decrease in dementia incidence in high-income countries, the prevalence will continue to increase in the aging population (3). Although the emergence of disease-modifying therapies has led to a revolution in the world of AD recently, their expected effects on disease progression are at most moderate (4,5). Therefore, the search for alternative treatment strategies in AD continues. There is also an ongoing search for new biomarkers with potential pathophysiological and clinical relevance in AD, as these biomarkers might aid in improved early and correct differential diagnosis and monitoring of disease progression (6). In this dissertation we evaluate subclinical epileptiform activity (SEA) and melatonin disruptions in AD, as both might serve either as biomarker or as potential new treatment/treatable phenomenon, depending on further research.

Melatonin is an endogenous produced neurohormone that is also readily available over-the-counter as exogenous sleep-promoting drug with no to only little and mild adverse effects (7). Multiple anti-AD properties have been attributed to melatonin, amongst which not only its free radical scavenging, antioxidant and anti-inflammatory properties(8), but also its potential to inhibit amyloid beta (Aβ) fibril formation, to tilt amyloid precursor protein processing towards the non-amyloidogenic pathway and
to reduce glycogen synthase kinase 3β activity leading to less tau hyperphosphorylation (9–12). Moreover, endogenous melatonin levels seem to be decreased in AD. Firstly, we highlight the alterations of melatonin levels in different biofluids in AD patients, based on data from the systematic review we conducted (6). Melatonin levels decrease with age, and according to the majority of available evidence cerebrospinal fluid (CSF) and blood night-time melatonin levels decrease even more in AD patients as compared to controls. Literature was not conclusive on alterations in daytime blood melatonin, as well as on saliva and urinary melatonin levels changes in AD (6). Furthermore, we concluded that melatonin might become a novel biomarker for AD disease progression, as a moderate inverse correlation between CSF melatonin levels and Braak stages had been described (13).

In a second step, we evaluated daytime blood and CSF melatonin levels in AD in a retrospective study (14). We could not find any significant differences in daytime melatonin levels between patients with dementia due to AD, mild cognitive impairment due to AD and healthy controls, nor could we find an association between daytime melatonin levels on cognitive decline thereafter or a correlation between daytime CSF or blood melatonin levels and baseline Mini Mental State Examination (MMSE) scores.
We hypothesised that this might be due to the fact that daytime melatonin levels do not contribute enough to the total 24-hour melatonin production. However, we did find a good correlation between spinal CSF and blood melatonin levels in this retrospective study, suggesting valid use of blood melatonin levels to study melatonin disruptions in AD (14).

Furthermore, we evaluated the role of subclinical epileptiform activity (SEA) in AD. Seizures and SEA are known to be more prevalent in probable AD patients as compared to controls (15–18). As hallmark AD proteins might play a role in producing neuronal hyperactivity (19–21), we wanted to evaluate the presence of SEA in the earliest stages of AD. To this end, we recruited participants based on 2011 National Institute of Aging-Alzheimer’s Association research criteria, also including preclinical AD subjects in our study (22). The prevalence of SEA was significantly increased in the AD continuum as compared to healthy controls, with the prevalence increasing through the different disease stages: 25% in preclinical AD subjects, 27% in mild cognitive impairment (MCI) patients and 50% in patients with dementia. We found SEA most commonly in left (fronto)temporal regions and, based on longterm-electroencephalography (LTM-EEG), most often during non-rapid eye movement (NREM) sleep stage II. In accordance with the (fronto)temporal origin of SEA found in our study, seizures in AD are thought to arise from the mesial temporal lobe which is one of the first structures affected by AD
pathology and is the region in the brain most prone to develop focal epilepsy (23). As the mesial temporal lobe is located mesial and far from the surface EEG electrodes, electrical activity might not propagate to EEG scalp electrodes (24). Therefore, we evaluated the use of different non-invasive neurophysiological techniques in detecting SEA in AD, comparing LTM-EEG with 128-electrode highdensity-EEG (hd-EEG) and magnetoencephalography (MEG). LTM-EEG has the advantage of covering a whole day and night (with sleep), hd-EEG has the advantage of including an inferior temporal chain which could lead to increased coverage of electrical activity arising from the mesial temporal lobe (25), and MEG has different sensitivities as compared to EEG as it is more sensitive to tangential sources (26–28). There were no differences regarding the prevalence of SEA between our different techniques (AD group: 19% with LTM-EEG, 19% with hd-EEG and 25% with MEG). MEG could detect significantly more interictal epileptiform discharges than LTM-EEG and hd-EEG per 50-minutes. However, when looking at the 10 participants belonging to the AD group in whom SEA was found and in whom both EEG (either LTM-EEG or hd-EEG) and MEG were available, EEG was the only modality to detect SEA in 6 participants, MEG the only modality to detect SEA in only 1 participant and both detected SEA in 3 participants. As there were no significant differences in detecting SEA between LTM-EEG and hd-EEG, amplification of a short-term (e.g. 50-minute) EEG with an inferior temporal chain seemed ideal, because recording and readout time are substantially lower with hd-EEG than LTM-EEG. However, in those AD participants in whom LTM-EEG and hd-EEG were available, LTM-EEG did detect SEA in participants in whom hd-EEG did not and vice versa. Therefore, the different techniques remain complementary in detecting SEA in AD. Adding the inferior temporal chain to the standard LTM-EEG caps/nets could be a potential way forward, concluding two examinations into one. We furthermore tried to characterise those AD patients at risk for SEA, and found those AD patients with SEA to score
worse on the RBANS attention and visuospatial subset and to have higher left frontal, (left) temporal and (left and right) entorhinal cortex volumes, which needs further investigations but might be attributed to gliosis. The identification of AD patients with SEA might become increasingly important, as neuronal hyperactivity and SEA might lead to disease progression in AD (17–20,29).

Melatonin might, next to its suggested anti-AD effects, have anti-epileptic effects as well.(30) In the final chapter, we firstly found a good correlation between saliva and plasma melatonin in AD. We furthermore found decreased plasma night-time melatonin levels at 23:00 in the whole AD continuum as compared to controls, with trends towards lower plasma melatonin levels at 22:00 and 00:00 in the AD continuum as compared to controls. We could not find significant differences regarding saliva melatonin levels between AD continuum participants and healthy controls. Melatonin did, furthermore, not seem to be good biomarker for disease stage in AD. Nevertheless, it might have potential to become an interesting biomarker for SEA in AD. Given melatonin’s potential antiepileptic effects (30), and the fact that baseline melatonin levels were lower in patients with mesial temporal lobe epilepsy and children with seizures (31,32), we hypothesised that melatonin levels would be even
lower in AD participants with SEA as compared to those without SEA. However, we found the exact opposite with increased plasma melatonin levels at 04:00 and saliva melatonin levels at 01:00 in AD participants with SEA as compared to those without. As these values were higher in AD participants with SEA, one blood sample or saliva sample taken during the night might potentially guide us to decide whether or not to perform more time-consuming and potentially expensive neurophysiological
examinations in an AD patient. The reason of this increase in melatonin levels is not known. A compensatory feedback mechanism against SEA might be a plausible line of thought (33), as melatonin levels do seem to increase after seizures in patients with mesial temporal lobe epilepsy and children with seizures (31,32). On the other hand, some literature states that melatonin might potentially have pro-epileptic effects, in humans as well as animal models (34–37). Whether the melatonin surge in AD participants with SEA is rather a cause or consequence of this SEA, merits further investigation.

The question we have not answered yet is whether SEA in AD should be treated, and if melatonin could be of added value in treating SEA in AD, or whether it could rather serve as a potential novel biomarker for SEA in AD. Further research on these interesting topics is warranted before we will be able to draw definite conclusions.