Comments
Alzheimer’s disease (AD) has a long preclinical phase, lasting years to decades, in which proteins implicated in the pathogenesis of AD can accumulate. The National Institute on Aging and Alzheimer’s Association have proposed a research framework allowing for characterization of the continuum of such brain changes even in cognitively normal older individuals (1). This AT(N) model groups fluid and imaging biomarkers into those impacting brain amyloid-β (Aβ) deposition (A), pathologic tau (T), and neurodegeneration (N). Aβ peptides are secreted from neurons and can form insoluble plaques, a process thought to be associated with decreased concentrations of soluble Aβ peptides in cerebrospinal fluid (CSF) and plasma. Tau is a microtubule protein that becomes hyper-phosphorylated, forming neurofibrillary tangles. Neurofilament light (NfL) is an axonal protein that is released in response to neural injury. Until recently, measuring the load of these proteins required PET imaging or CSF analysis, which are expensive and invasive. Due to advances in novel ultrasensitive immunoassays, assessment of AD biomarkers in plasma has now become feasible. Like patterns observed in CSF, progression to AD is characterized by decreases in levels of plasma Aβ42 and Aβ40 (A), and increases in plasma total and phosphorylated tau (T), and plasma NfL (N). Importantly, these plasma measures have been shown to track with both CSF/PET amounts and cognitive status (2), but their diurnal modulation and acute impact of obstructive sleep apnea (OSA) have not been examined in humans.
Previously, we and others have shown that self-reported OSA is associated with an earlier onset of mild cognitive impairment and that both subjective and objectively measured OSA predicts longitudinal gains in amyloid burden or tau accumulation by both brain PET imaging and CSF biomarker measures (3, 4). Further, in individuals with OSA treated with positive airway pressure (PAP), change in the AHI4% correlated with changes in CSF levels of both amyloid and tau (5). In rodents and humans, there is circadian fluctuation in levels of both amyloid and tau, with nadir levels observed after a period of sleep (6, 7). This raises the possibility that OSA could impact the overnight dynamics of fluid AD biomarkers. The current study uses a crossover PAP withdrawal paradigm in individuals preselected with severe OSA who are normally adherent to PAP at home to evaluate overnight changes in plasma AD biomarkers (secondary analyses of NCT02824263). Additionally, we evaluated whether sleep architecture or measures of hypoxemic burden between the on- and off-PAP conditions correlated with changes in plasma AD biomarkers using linear mixed-effects modeling.
We analyzed blood collected in the evening (10:40 p.m.) and next morning (6:40 a.m.) across a night of polysomnographically recorded sleep in 30 individuals with severe OSA during conditions of therapeutic PAP or the third consecutive night off PAP (acute withdrawal) in a counterbalanced fashion (8). Participants were 51.5 years old (95% confidence interval [CI], 47.3–55.6), 27% women, and 67% non-Hispanic White. Plasma samples were analyzed using SIMOA Neurology 3-Plex A and NF-Light kits (HD-1 Analyzer, Quanterix). All samples were tested in duplicate, and samples with a coefficient of variation ⩾20% were excluded from the analysis: NfL n = 25, T-Tau n = 13, Aβ42 n = 25, Aβ40 n = 27.
Sleep architecture measures including %N3 (OSA: 6.1% [3.7–8.5], PAP: 15.1% [10.6–19.6], P < 0.001), %REM (OSA: 11.8% [8.8–14.7], PAP: 20.6% [18.3–22.9], P < 0.001), arousal index (OSA: 53/h [44–62], PAP: 12/h [9–14], P < 0.001) and measures of breathing including AHI4% (OSA: 63/h [54–72], PAP: 3/h (2–4), P < 0.001), SpO2 below 90% (OSA: 20 min [14–26], PAP: 1 min [0–3], P < 0.001), and SpO2 min (OSA: 77% [74–80], PAP: 88% [86–90], P < 0.001) were all significantly different in the untreated (OSA) versus PAP treated nights. Total sleep time (TST, OSA: 381 min [365–398], PAP: 388 min [372–404], P = 0.344) and sleep efficiency (SE, OSA: 79% [76–82], PAP: 81% [65–95], P = 0.610) were not different between conditions.
Overnight change, defined as the post-sleep (6:40 a.m.) measure minus the pre-sleep (10:40 p.m.) measure, for plasma NfL, total tau (T-tau), Aβ42 and Aβ40 proteins as a function of PAP condition is shown in Figure 1. We observed significant differences in the overnight change in NfL and Aβ40 between conditions, such that PAP withdrawal resulted in a relative increase in NfL and a relative decrease in Aβ40. Change in T-tau and Aβ42 were not different between conditions. Importantly, all four AD biomarkers were not different between sleep conditions at the start of the night within-subjects.
Figure 1. Overnight change in plasma Alzheimer’s disease (AD) biomarkers comparing obstructive sleep apnea (OSA) versus positive airway pressure (PAP) conditions in subjects with severe OSA. Plasma-derived AD biomarkers neurofilament light chain (NfL), total tau (T-tau), and Aβ42 and Aβ40 (amyloid β-42 and 40) were measured across a night of PAP use (PAP) or PAP withdrawal (OSA). Overnight change in plasma-NfL was significantly lower with PAP use while overnight change in plasma-Aβ40 was significantly higher with PAP use (by paired t tests).
[More] [Minimize]We next performed linear mixed-effects modeling to investigate the relationship between overnight change in NfL or Aβ40 with sleep and breathing measures. We modeled the overnight change in NfL (Table 1: Model 1) and the overnight change in Aβ40 (Table 1: Model 2) with fixed effects for sleep condition (OSA versus PAP). Subject identifiers were used as random effects. We found that the number of sleep stage transitions, a marker of sleep fragmentation, and the amount of time below 90% O2 saturation (T90) were significantly associated with ΔNfL, phenomena driven by PAP withdrawal. By contrast, no identified sleep or respiratory measures predicted the overnight change in Aβ40.
| Model 1: outcome overnight ΔNfL (pg/mL) | |||
|---|---|---|---|
| Fixed effects | β coefficient | 95% CI | P |
| SE, % | 0.003 | (−0.097, 0.047) | 0.931 |
| Arousal Index, events/h | −0.012 | (−0.014, 0.051) | 0.638 |
| Sleep stage transitions, # | 0.008 | (0.000, 0.016) | 0.047 |
| NREM 3, % | −0.004 | (−0.067, 0.043) | 0.894 |
| REM, % | −0.025 | (−0.158, 0.006) | 0.065 |
| AHI4%, events/h | 0.002 | (−0.050, 0.055) | 0.925 |
| O2min, % | −0.005 | (−0.101, 0.090) | 0.903 |
| T90, min | 0.061 | (0.013, 0.135) | 0.032 |
| Model 2: outcome overnight ΔAβ40 (pg/mL) | |||
| Fixed effects | β coefficient | 95% CI | P |
| SE, % | 0.510 | (−0.923, 1.946) | 0.476 |
| Arousal Index, events/h | −0.139 | (−1.134, 0.855) | 0.778 |
| Sleep stage transitions, # | 0.0753 | (−0.081, 0.231) | 0.337 |
| NREM 3, % | −0.386 | (−1.748, 0.976) | 0.570 |
| REM, % | 0.501 | (−1.377, 2.380) | 0.593 |
| AHI4%, events/h | 0.416 | (−0.564, 1.397) | 0.397 |
| O2min, % | −1.082 | (−2.820, 0.654) | 0.258 |
| T90, min | −0.393 | (−1.385, 0.599) | 0.429 |
Our findings suggest that OSA acutely increases plasma NfL and decreases plasma Aβ40 during overnight sleep, while PAP prevents these changes. Although not exclusive to AD, NfL levels track closely with AD progression and inversely correlate with cognitive function, even in cognitively normal individuals (9). This finding provides novel evidence that OSA may be contributing to overnight neural injury potentially via hypoxemic burden and/or sleep fragmentation. We did not identify sleep or breathing correlates of overnight change in Aβ40, though the degree to which plasma amyloid reflects brain amyloid may be assay-specific (10). Although we only tested one model for each of the two biomarkers, eight different sleep/breathing measures were used in each model. There is thus potential for false discovery in these mixed-effects models. Other limitations of this study include the potentially underpowered sample size for plasma T-tau due to high duplicate variation. Sleep deprivation increased plasma derived tau compared with natural sleep (11), and the signal that OSA may increase plasma tau warrants further investigation. Overall, this study identifies novel modulation of plasma AD biomarkers as a function of PAP withdrawal and suggests the importance of PAP-associated restoration of sleep continuity and oxygenation in plasma diurnal AD biomarker metabolism. It remains to be seen whether these changes represent a reversible risk factor for AD progression.
| 1. | Jack CR Jr, Bennett DA, Blennow K, Carrillo MC, Dunn B, Haeberlein SB, et al.; Contributors. NIA-AA Research Framework: toward a biological definition of Alzheimer’s disease. Alzheimers Dement 2018;14:535–562. |
| 2. | Rauchmann BS, Schneider-Axmann T, Perneczky R; Alzheimer’s Disease Neuroimaging Initiative (ADNI). Associations of longitudinal plasma p-tau181 and NfL with tau-PET, Aβ-PET and cognition. J Neurol Neurosurg Psychiatry 2021;92:1289–1295. |
| 3. | Sharma RA, Varga AW, Bubu OM, Pirraglia E, Kam K, Parekh A, et al. Obstructive sleep apnea severity affects amyloid burden in cognitively normal elderly. A longitudinal study. Am J Respir Crit Care Med 2018;197:933–943. |
| 4. | Bubu OM, Pirraglia E, Andrade AG, Sharma RA, Gimenez-Badia S, Umasabor-Bubu OQ, et al.; Alzheimer’s Disease Neuroimaging Initiative. Obstructive sleep apnea and longitudinal Alzheimer’s disease biomarker changes. Sleep (Basel) 2019;42:zsz048. |
| 5. | Ju YS, Zangrilli MA, Finn MB, Fagan AM, Holtzman DM. Obstructive sleep apnea treatment, slow wave activity, and amyloid-β. Ann Neurol 2019;85:291–295. |
| 6. | Lucey BP, Hicks TJ, McLeland JS, Toedebusch CD, Boyd J, Elbert DL, et al. Effect of sleep on overnight cerebrospinal fluid amyloid β kinetics. Ann Neurol 2018;83:197–204. |
| 7. | Holth JK, Fritschi SK, Wang C, Pedersen NP, Cirrito JR, Mahan TE, et al. The sleep-wake cycle regulates brain interstitial fluid tau in mice and CSF tau in humans. Science 2019;363:880–884. |
| 8. | Chopra S, Rathore A, Younas H, Pham LV, Gu C, Beselman A, et al. Obstructive sleep apnea dynamically increases nocturnal plasma free fatty acids, glucose, and cortisol during sleep. J Clin Endocrinol Metab 2017;102:3172–3181. |
| 9. | He L, Morley JE, Aggarwal G, Nguyen AD, Vellas B, de Souto Barreto P; MAPT/DSA Group. Plasma neurofilament light chain is associated with cognitive decline in non-dementia older adults. Sci Rep 2021;11:13394. |
| 10. | Janelidze S, Teunissen CE, Zetterberg H, Allué JA, Sarasa L, Eichenlaub U, et al. Head-to-head comparison of 8 plasma amyloid-β 42/40 assays in Alzheimer disease. JAMA Neurol 2021;78:1375–1382. |
| 11. | Benedict C, Blennow K, Zetterberg H, Cedernaes J. Effects of acute sleep loss on diurnal plasma dynamics of CNS health biomarkers in young men. Neurology 2020;94:e1181–e1189. |
‡ Equal contribution as senior author.
Supported by AASM Foundation Focused Projects Grant JI-19–5 (K.K.), National Heart, Lung, and Blood Institute R01HL135483 (J.J.), National Institute on Aging R01AG056031 (R.S.O.), R01AG056531 (R.S.O.), R01AG056682 (A.W.V.), and R01AG066870 (A.W.V.).
Author Contributions: Conception and study design: K.K., J.J., R.S.O., and A.W.V. Data acquisition, analysis, and interpretation: K.K., J.J., A.P., O.M.B., A.E.M., C.G., L.P., T.M.W., D.M.R., I.A., R.S.O, and A.W.V. Manuscript preparation: K.K., J.J., A.P., O.M.B., R.S.O, and A.W.V.
Originally Published in Press as DOI: 10.1164/rccm.202202-0262LE on June 13, 2022
Author disclosures are available with the text of this letter at www.atsjournals.org.
