[Skip to Content]
Sign In
Individual Sign In
Create an Account
Institutional Sign In
OpenAthens Shibboleth
[Skip to Content Landing]
Figure 1.
Cosinor Analysis of Percentage of the Mean for β-Amyloid (Aβ) 40 and 42 Measured by Enzyme-Linked Immunosorbent Assay (ELISA) and Mass Spectrometry (MS)
Cosinor Analysis of Percentage of the Mean for β-Amyloid (Aβ) 40 and 42 Measured by Enzyme-Linked Immunosorbent Assay (ELISA) and Mass Spectrometry (MS)

Group-averaged Aβ40 and Aβ42 measured in serial cerebrospinal fluid (CSF) by ELISA and MS in 39 amyloid-negative (A-D) and 38 amyloid-positive (E-H) individuals over 48 hours. Concentration values were converted to percentages of the mean for all participants and then averaged for each group. The solid line indicates cosinor fit; error bars indicate SDs. Quantitation of MS resulted in narrower SDs (all P ≤ .001) and therefore provided a more precise fit to the cosine wave compared with ELISA. A, Aβ40 measured by ELISA in amyloid-negative individuals. B, Aβ40 measured by MS in amyloid-negative individuals. C, Aβ42 measured by ELISA in amyloid-negative individuals. D, Aβ42 measured by MS in amyloid-negative individuals. E, Aβ40 measured by ELISA in amyloid-positive individuals. F, Aβ40 measured by MS in amyloid-positive individuals. G, Aβ42 measured by ELISA in amyloid-positive individuals. H, Aβ42 measured by MS in amyloid-positive individuals. For amyloid-negative individuals, the SD for Aβ40 by ELISA was 8.158% greater than by MS and was 4.881% greater for Aβ42. For amyloid-positive individuals, the SD for Aβ40 by ELISA was 5.215% greater than by MS and was 2.021% greater for Aβ42.

Figure 2.
Association Between β-Amyloid (Aβ) Amplitude and Linear Rise With Age
Association Between β-Amyloid (Aβ) Amplitude and Linear Rise With Age

A, Aβ42 amplitude (picomolars [pM]) vs age (years) (amyloid negative: r = −0.43; P = .006; amyloid positive: r = −0.22; P = .18). B, Aβ42 linear rise (pM/h) vs age (years) (amyloid negative: r = −0.38; P = .02; amyloid positive: r = 0.15; P = .37). C, Aβ40 amplitude (pM) vs age (years) (amyloid negative: r = −0.21; P = .21; amyloid positive: r = −0.4; P = .01). D, Aβ40 linear rise (pM/h) vs age (years) (amyloid negative: r = −0.21; P = .19; amyloid positive: r = 0.03; P = .85).

Figure 3.
Association Between β-Amyloid (Aβ) Amplitude and Linear Rise With Aβ42:Aβ40 Ratio
Association Between β-Amyloid (Aβ) Amplitude and Linear Rise With Aβ42:Aβ40 Ratio

A, Aβ42 amplitude (picomolars [pM]) vs Aβ42:Aβ40 ratio (amyloid negative: r = 0.38; P = .02; amyloid positive: r = 0.11; P = .49). The horizontal dashed line is at 15 pM. B, Aβ42 linear rise (pM/h) vs Aβ42:Aβ40 ratio (amyloid negative: r = 0.26; P = .12; amyloid positive: r = 0.06; P = .70). C, Aβ40 amplitude (pM) vs Aβ42:Aβ40 ratio (amyloid negative: r = 0.16; P = .34; amyloid positive: r = 0.27; P = .10). The horizontal dashed line is at 200 pM. D, Aβ40 linear rise (pM/h) vs Aβ42:Aβ40 ratio (amyloid negative: r = 0.14; P = .41; amyloid positive: r = 0.1; P = .53). For all panels, the vertical dashes lines are at Aβ42:Aβ40 = 0.1, Aβ42:Aβ40 = 0.12, and Aβ42:Aβ40 = 0.16. Participants with the highest CSF Aβ42:Aβ40 ratios (>0.16) were previously reported to have normal Aβ stable isotope labeling kinetics (SILK). Aβ SILK alterations become progressively more pronounced as the CSF Aβ42:Aβ40 ratio decreases from 0.16 to 0.1 and then <0.1. For this reason, these cutoffs were used in the figures with Aβ42:Aβ40 ratio.

Figure 4.
Association Between β-Amyloid (Aβ) Amplitude and Linear Rise With Fractional Turnover Rate (FTR) Aβ42/40
Association Between β-Amyloid (Aβ) Amplitude and Linear Rise With Fractional Turnover Rate (FTR) Aβ42/40

A, Aβ42 amplitude (picomolars [pM]) vs FTR Aβ42/40 ratio (pools/h) (amyloid negative: r = 0.33; P = .04; amyloid positive: r = 0.049; P = .96). The horizontal dashed line is at 15 pM. B, Aβ42 linear rise (pM/h) vs FTR Aβ42/40 ratio (pools/h) (amyloid negative: r = 0.32; P = .04; amyloid positive: r = 0.25; P = .13). C, Aβ40 amplitude (pM) vs FTR Aβ42/40 (pools/h) (amyloid negative: r = 0.099; P = .55; amyloid positive: r = 0.14; P = .39). The horizontal dashed line is at 200 pM. D, Aβ40 linear rise (pM/h) vs FTR Aβ42/40 ratio (pools/h) (amyloid negative: r = 0.18; P = .28; amyloid positive: r = 0.27; P = .10). The vertical dashes line marks FTR Aβ42/40 = 1.1.

Table 1.  
Associations Between Aβ42 Amplitude and Linear Rise With Age, Amyloid Status, and Aβ Kinetics
Associations Between Aβ42 Amplitude and Linear Rise With Age, Amyloid Status, and Aβ Kinetics
Table 2.  
Associations Between Aβ40 Amplitude and Linear Rise With Age, Amyloid Status, and Aβ Kinetics
Associations Between Aβ40 Amplitude and Linear Rise With Age, Amyloid Status, and Aβ Kinetics
1.
Prince  M, Wimo  A, Guerchet  M, Ali  G-C, Wu  Y-T, Prina  M.  World Alzheimer Report 2015: The Global Impact of Dementia: An Analysis of Prevalence. London, England: Alzheimer’s Disease International; 2015.
2.
Patterson  BW, Elbert  DL, Mawuenyega  KG,  et al.  Age and amyloid effects on human central nervous system amyloid-beta kinetics.  Ann Neurol. 2015;78(3):439-453.PubMedGoogle ScholarCrossref
3.
Meyer-Luehmann  M, Stalder  M, Herzig  MC,  et al.  Extracellular amyloid formation and associated pathology in neural grafts.  Nat Neurosci. 2003;6(4):370-377.PubMedGoogle ScholarCrossref
4.
Kang  J-E, Lim  MM, Bateman  RJ,  et al.  Amyloid-β dynamics are regulated by orexin and the sleep-wake cycle.  Science. 2009;326(5955):1005-1007.PubMedGoogle ScholarCrossref
5.
Huang  Y, Potter  R, Sigurdson  W,  et al.  Effects of age and amyloid deposition on Aβ dynamics in the human central nervous system.  Arch Neurol. 2012;69(1):51-58.PubMedGoogle ScholarCrossref
6.
Roh  JH, Huang  Y, Bero  AW,  et al.  Disruption of the sleep-wake cycle and diurnal fluctuation of β-amyloid in mice with Alzheimer’s disease pathology.  Sci Transl Med. 2012;4(150):150ra122.PubMedGoogle ScholarCrossref
7.
Lucey  BP, Gonzales  C, Das  U,  et al.  An integrated multi-study analysis of intra-subject variability in cerebrospinal fluid amyloid-β concentrations collected by lumbar puncture and indwelling lumbar catheter.  Alzheimers Res Ther. 2015;7(1):53.PubMedGoogle ScholarCrossref
8.
Lucey  BP, Bateman  RJ.  Amyloid-β diurnal pattern: possible role of sleep in Alzheimer’s disease pathogenesis.  Neurobiol Aging. 2014;35(suppl 2):S29-S34.PubMedGoogle ScholarCrossref
9.
Xie  L, Kang  H, Xu  Q,  et al.  Sleep drives metabolite clearance from the adult brain.  Science. 2013;342(6156):373-377.PubMedGoogle ScholarCrossref
10.
Kress  BT, Iliff  JJ, Xia  M,  et al.  Impairment of paravascular clearance pathways in the aging brain.  Ann Neurol. 2014;76(6):845-861.PubMedGoogle ScholarCrossref
11.
Morris  JC.  The Clinical Dementia Rating (CDR): current version and scoring rules.  Neurology. 1993;43(11):2412-2414.PubMedGoogle ScholarCrossref
12.
Berg  L, Miller  JP, Storandt  M,  et al.  Mild senile dementia of the Alzheimer type 2: longitudinal assessment.  Ann Neurol. 1988;23(5):477-484.PubMedGoogle ScholarCrossref
13.
Mintun  MA, Larossa  GN, Sheline  YI,  et al.  [11C]PIB in a nondemented population: potential antecedent marker of Alzheimer disease.  Neurology. 2006;67(3):446-452.PubMedGoogle ScholarCrossref
14.
Mawuenyega  KG, Kasten  T, Ovod  V, Lucey  B, Sigurdson  W, Bateman  RJ. Immuno-based-LC/SRM as a diagnostic tool for protein dynamics of amyloid β isoforms instead of ELISA in the clinical laboratory. Paper presented at: 62nd Conference on Mass Spectrometry and Allied Topics; June 15-19, 2014; Baltimore, MD.
15.
Bateman  RJ, Munsell  LY, Morris  JC, Swarm  R, Yarasheski  KE, Holtzman  DM.  Human amyloid-β synthesis and clearance rates as measured in cerebrospinal fluid in vivo.  Nat Med. 2006;12(7):856-861.PubMedGoogle ScholarCrossref
16.
Fagan  AM, Mintun  MA, Mach  RH,  et al.  Inverse relation between in vivo amyloid imaging load and cerebrospinal fluid Abeta42 in humans.  Ann Neurol. 2006;59(3):512-519.PubMedGoogle ScholarCrossref
17.
Potter  R, Patterson  BW, Elbert  DL,  et al.  Increased in vivo amyloid-β42 production, exchange, and loss in presenilin mutation carriers.  Sci Transl Med. 2013;5(189):189ra77.PubMedGoogle ScholarCrossref
18.
Elbert  DL, Patterson  BW, Bateman  RJ.  Analysis of a compartmental model of amyloid beta production, irreversible loss and exchange in humans.  Math Biosci. 2015;261:48-61.PubMedGoogle ScholarCrossref
19.
Vlassenko  AG, McCue  L, Jasielec  MS,  et al.  Imaging and cerebrospinal fluid biomarkers in early preclinical alzheimer disease.  Ann Neurol. 2016;80(3):379-387.PubMedGoogle ScholarCrossref
20.
Bateman  RJ, Wen  G, Morris  JC, Holtzman  DM.  Fluctuations of CSF amyloid-β levels: implications for a diagnostic and therapeutic biomarker.  Neurology. 2007;68(9):666-669.PubMedGoogle ScholarCrossref
21.
Li  J, Llano  DA, Ellis  T,  et al.  Effect of human cerebrospinal fluid sampling frequency on amyloid-β levels.  Alzheimers Dement. 2012;8(4):295-303.PubMedGoogle ScholarCrossref
22.
Moghekar  A, Goh  J, Li  M, Albert  M, O’Brien  RJ.  Cerebrospinal fluid Aβ and tau level fluctuation in an older clinical cohort.  Arch Neurol. 2012;69(2):246-250.PubMedGoogle ScholarCrossref
23.
Slats  D, Claassen  JA, Spies  PE,  et al.  Hourly variability of cerebrospinal fluid biomarkers in Alzheimer’s disease subjects and healthy older volunteers.  Neurobiol Aging. 2012;33(4):831.e1-831.e9.PubMedGoogle ScholarCrossref
24.
Dobrowolska  JA, Kasten  T, Huang  Y,  et al.  Diurnal patterns of soluble amyloid precursor protein metabolites in the human central nervous system.  PLoS One. 2014;9(3):e89998.PubMedGoogle ScholarCrossref
25.
Ooms  S, Overeem  S, Besse  K, Rikkert  MO, Verbeek  M, Claassen  JA.  Effect of 1 night of total sleep deprivation on cerebrospinal fluid β-amyloid 42 in healthy middle-aged men: a randomized clinical trial.  JAMA Neurol. 2014;71(8):971-977.PubMedGoogle ScholarCrossref
26.
Van Broeck  B, Timmers  M, Ramael  S,  et al.  Impact of frequent cerebrospinal fluid sampling on Aβ levels: systematic approach to elucidate influencing factors.  Alzheimers Res Ther. 2016;8(1):21.PubMedGoogle ScholarCrossref
Original Investigation
February 2017

Associations Between β-Amyloid Kinetics and the β-Amyloid Diurnal Pattern in the Central Nervous System

Author Affiliations
  • 1Department of Neurology, Washington University School of Medicine, St Louis, Missouri
  • 2Hope Center for Neurological Disorders, Washington University School of Medicine, St Louis, Missouri
  • 3Department of Medicine, Washington University School of Medicine, St Louis, Missouri
  • 4Department of Biomedical Engineering, Washington University, St Louis, Missouri
  • 5Knight Alzheimer’s Disease Research Center, Washington University School of Medicine, St Louis, Missouri
 

Copyright 2016 American Medical Association. All Rights Reserved.

JAMA Neurol. 2017;74(2):207-215. doi:10.1001/jamaneurol.2016.4202
Key Points

Question  What is the association between β-amyloid (Aβ) day/night patterns and Aβ production and clearance rates?

Findings  In this study of 77 older adults, after controlling for amyloid deposition, Aβ concentration and production rates were associated with Aβ diurnal amplitude, while Aβ fractional turnover rates (eg, half-life or clearance) were associated with Aβ linear rise.

Meaning  These findings implicate Aβ concentration, production rates, and clearance rates in the cerebrospinal fluid as important correlates of Aβ day/night patterns even in the absence of plaques and may be important for the designs of future secondary prevention trials for Alzheimer disease.

Abstract

Importance  Recent studies found that the concentration of amyloid-β (Aβ) fluctuates with the sleep-wake cycle. Although the amplitude of this day/night pattern attenuates with age and amyloid deposition, to our knowledge, the association of Aβ kinetics (ie, production, turnover, and clearance) with this oscillation has not been studied.

Objective  To determine the association between Aβ kinetics, age, amyloid levels, and the Aβ day/night pattern in humans.

Design, Setting, and Participants  We measured Aβ concentrations and kinetics in 77 adults aged 60 to 87 years with and without amyloid deposition by a novel precise mass spectrometry method at the Washington University School of Medicine in St Louis, Missouri. We compared findings of 2 orthogonal methods, enzyme-linked immunosorbent assay and mass spectrometry, to validate the day/night patterns and determine more precise estimates of the cosinor parameters. In vivo labeling of central nervous system proteins with stable isotopically labeled leucine was performed, and kinetics of Aβ40 and Aβ42 were measured.

Interventions  Serial cerebrospinal fluid collection via indwelling lumbar catheter over 36 to 48 hours before, during, and after in vivo labeling, with a 9-hour primed constant infusion of 13C6-leucine.

Main Outcomes and Measures  The amplitude, linear increase, and other cosinor measures of each participant’s serial cerebrospinal fluid Aβ concentrations and Aβ turnover rates.

Results  Of the 77 participants studied, 46 (59.7%) were men, and the mean (range) age was 72.6 (60.4-87.7) years. Day/night patterns in Aβ concentrations were more sharply defined by the precise mass spectrometry method than by enzyme-linked immunosorbent assay (mean difference of SD of residuals: Aβ40, −7.42 pM; P < .001; Aβ42, −3.72 pM; P < .001). Amyloid deposition diminished day/night amplitude and linear increase of Aβ42 but not of Aβ40. Increased age diminished day/night amplitude of both Aβ40 and Aβ42. After controlling for amyloid deposition, amplitude of Aβ40 was positively associated with production rates (r = 0.42; P < .001), while the linear rise was associated with turnover rates (r = 0.28; P < .05). The amplitude and linear rise of Aβ42 were both associated with turnover (r = −0.38; P < .001) and production (r = 0.238; P < .05) rates.

Conclusions and Relevance  Amyloid deposition is associated with premature loss of normal Aβ42 day/night patterns in older adults, suggesting the previously reported effects of age and amyloid on Aβ42 amplitude at least partially affect each other. Production and turnover rates suggest that day/night Aβ patterns are modulated by both production and clearance mechanisms active in sleep-wake cycles and that amyloid deposition may impair normal circadian patterns. These findings may be important for the designs of future secondary prevention trials for Alzheimer disease.

Introduction

Alzheimer disease (AD) is a neurodegenerative disorder characterized pathologically by the extracellular deposition of amyloid-β (Aβ) in senile plaques and intracellular neurofibrillary tangles of tau, leading to neuronal loss and progressive cognitive impairment. Alzheimer disease represents a current and growing public health threat, with the worldwide prevalence of the disease projected to increase from 46.8 million people in 2015 to 131.5 million in 2050.1 Age and Aβ deposition are major risk factors for AD. Age slows the half-life and clearance of soluble Aβ, while Aβ deposition causes irreversible loss of soluble Aβ42 clearance.2 Because aggregation and deposition of Aβ into insoluble extracellular plaques is concentration-dependent,3 factors that affect Aβ concentration through changes in production and clearance are potential therapeutic targets for the prevention of AD.

Recent studies in animal models and humans have found that (1) the concentration of Aβ in the central nervous system fluctuates with the sleep-wake cycle as a day/night (ie, diurnal) pattern, (2) Aβ concentration in the cerebrospinal fluid (CSF) increases linearly with serial sampling, which is mitigated with amyloid deposition, and (3) the amplitude of this day/night pattern is attenuated with age and amyloid deposition.4-6 The findings of the diurnal pattern and linear increases of Aβ levels in human CSF have been replicated across multiple studies.7 Decreased Aβ production and increased clearance during sleep have been hypothesized to drive the day/night oscillation.8,9 Sleep has been linked to a mechanistic flushing of extracellular components, such as Aβ, which appears to be a normal glymphatic clearance mechanism.9,10 The relationship of sleep, the day/night pattern, and the clearance of Aβ (potentially by this glymphatic system) is not well understood. In this study, we assess the association between Aβ production, clearance, day/night patterns, and the linear rise as it relates to normal and abnormal Aβ production and clearance mechanisms. These findings will inform the design of future studies and may lead to approaches to maintain normal physiological control of Aβ concentrations through sleep-mediated changes in the day/night pattern.

Methods
Participants

Seventy-seven participants serving as research volunteers in the longitudinal studies of the Knight Alzheimer’s Disease Research Center and its affiliated studies at Washington University in St Louis, Missouri, were recruited to participate in this study. These 77 individuals ranged in age from 60 to 87 years; 46 (59.7%) were men (mean [range] age, 72.6 [60.4-87.7] years) and 31 (40.3%) were women (mean [range] age, 72.6 [63.8-85.2] years). All 77 participants were assessed clinically with a standard protocol that included the Clinical Dementia Rating Sum of Boxes score, which ranged from 0 (no impairment) to 18 (maximal impairment).11,12 Thirty-three participants (42.9%) had a score of 0 and 44 participants (57.1%) had a score greater than 0. Participant demographic information is shown in eTable 1 in the Supplement.

The study protocol was approved by the Washington University Institutional Review Board and the General Clinical Research Center Advisory Committee. All participants provided written informed consent and were compensated for their participation in the study.

Determining Amyloid Status

Amyloid status was established for each participant, as previously described.2 For the 44 participants with positron emission tomography (PET) and carbon II–labeled Pittsburgh compound B ([11C]PiB) scans, participants were considered amyloid positive if their mean cortical binding potential was greater than 0.18.13 For the 33 participants without PET (11C)PiB scans, a Aβ42:Aβ40 ratio less than 0.12 in the cerebrospinal fluid defined amyloid-positive.

Aβ Concentration

Previous studies of the Aβ concentrations measured in day/night patterns used Aβ concentrations determined by enzyme-linked immunosorbent assay (ELISA), which has a relatively high variance of measurement (typical coefficient of variation percent of 10% to 20%). Recent advances in mass spectrometry (MS) allow for the precise simultaneous quantitation of Aβ isoform concentrations in the CSF and Aβ stable isotope labeling kinetics (SILK).

From each time point collected, Aβ40 and Aβ42 levels were measured using ELISA, as previously described.5 All samples from each participant were measured together on the same ELISA plate to avoid interplate variation, and each sample was assessed in duplicate. Mass spectrometry Aβ SILK and absolute quantitation of Aβ levels were performed simultaneously. All samples were processed and measured, as previously described.2,14

Aβ SILK

The procedure for stable isotope amino acid tracer administration, sample collection, Aβ SILK tracer protocol, and compartmental modeling analysis of Aβ kinetics for all 77 participants were previously reported.2,15 Analysis of Aβ SILK kinetics generated 3 key measures of Aβ kinetics, representing the fractional turnover rates (directly associated with the half-life), absolute production rates, and an exchange process or delay component.

Cosinor Analysis

Cosinor analysis was used to fit a cosine wave to each individual’s serial CSF Aβ values measured by both ELISA and MS; as previously described, a 24-hour period was set as the default circadian cycle.7 The y-intercept (equivalent to the mesor or midline of the Aβ oscillation), amplitude (distance between the peak and mesor), acrophase (time corresponding to the peak of the curve), and slope of the linear rise were calculated for each participant.

To measure the uncertainty of the cosine fit between the ELISA and MS measurements, we converted all 77 participants’ serial CSF Aβ40 and Aβ42 concentrations to a percentage of the mean to control for differences between the assays. The data transformation was calculated separately for Aβ40 and Aβ42. Then, we calculated the SD of the residuals (SDR) for each cosine fit to determine how well the fitted cosine wave compared between ELISA and MS data. Cosinor analysis determines the curve that minimizes the sum of squares of the distances between individual data points and the best-fit line. The SDR is equal to the square root of the sum of squares divided by degrees of freedom and is expressed in the same units as Y (ie, percentage of the mean). Degrees of freedom, in this case, equals the number of data points minus the number of parameters fit. The greater the SDR, the greater the uncertainty of the true best-fit curve. Cosine fit SDRs were also compared with the SDRs of a straight-line fit.

Statistics

SPSS version 23 (IBM) was used for all statistical analyses. Graphpad Prism version 6.0b (Graphpad Software) for Mac computer (Apple) was used to calculate the parameters of the cosinor analysis for Aβ40 and Aβ42. Statistical significance was set at P < .05.

Results

Concentrations of Aβ40 and Aβ42 measured by MS and ELISA for all 77 participants were transformed to percentages of the mean. The mean correlation coefficients between MS and ELISA for both isoforms were 0.3. Serial Aβ concentrations were fitted to a straight line and a cosine wave. Cosinor analysis fit both the MS and ELISA data with lower SDRs than a straight line (2-tailed t test; mean difference of SDRs: Aβ40, −7.42 pM; Aβ42, −3.72 pM; both P < .001), indicating a day/night pattern in Aβ concentrations by both methods (eTable 2 in the Supplement).

All 77 participants were separated into 2 groups, with 39 in the amyloid-negative group and 38 in the amyloid-positive group. Aβ40 and Aβ42 percentage of the mean values for both assays were averaged for each time point. The cosine wave fit the mean data from the MS group with significantly narrower SDs than data from the ELISA group for both Aβ40 and Aβ42 (Figure 1A-H; all P ≤ .001). We then compared the individual cosinor parameters for each group. Aβ40 and Aβ42 day/night amplitude and acrophase did not differ significantly between assays in either group (all P > .05). Aβ42 amplitude and linear rise calculated by MS was significantly decreased in amyloid-positive individuals compared with amyloid-negative individuals (Figure 1D and H; P < .002), replicating previous findings.5 However, the linear increase of Aβ40 and Aβ42 determined by data from the MS group was not as high compared with the ELISA group in both amyloid-negative and amyloid-positive groups (Figure 1A-H; all P ≤ .01).

Based on these findings, we conclude that MS more precisely fit the day/night oscillation of Aβ40 and Aβ42 using cosinor analysis compared with ELISA. Only Aβ concentrations measured by MS are analyzed throughout the remainder of this article. The cosine-fitted data for Aβ40 and Aβ42 measured by MS and ELISA are shown for all participants in the supplement (eFigures 1-77 in the Supplement).

Association of Aβ42 Day/Night Amplitude and Linear Increase With Age

Aβ40 and Aβ42 day/night amplitude and linear increase were compared with age for both amyloid-negative and amyloid-positive individuals. Aβ42 amplitude declined 0.91 pM per year in amyloid-negative individuals, while Aβ42 day/night amplitudes in the amyloid-positive group did not vary significantly with age (Figure 2A). The magnitude of the Aβ42 rise with serial CSF sampling significantly declined with age in the amyloid-negative group (−0.12 pM/h each year) (Figure 2B). In contrast, there was no linear rise for Aβ42 with serial CSF sampling in amyloid-positive individuals regardless of age (0.01 pM/h per year) (Figure 2B). For those older than 73 years, there were no significant differences between the 2 amyloid groups in the mean day/night amplitude or linear increase of Aβ42 (eTable 3 in the Supplement).

For Aβ40, in both amyloid-negative and amyloid-positive groups, there was a decrease in day/night amplitude with age of 3.0 to 3.5 pM per year, although the decline was only significant in amyloid-positive individuals (Figure 2C). There was no age-related change in Aβ40 linear increase in either amyloid-negative or amyloid-positive participants (Figure 2D).

Association of Amyloid Burden With Aβ Day/Night Amplitude and Linear Rise

Amyloid deposition decreases as Aβ linear increases but has less effect on Aβ day/night amplitude.5 To assess the effect of amyloid burden, we compared the Aβ day/night amplitude and linear increase with mean cortical binding potential and CSF Aβ42:Aβ40 concentration ratios. Aβ42 amplitude and linear increase showed a sudden discontinuous step change between amyloid-negative and amyloid-positive individuals measured by PET (11C)PiB. All amyloid-positive participants had a significant loss of amplitude variability (P < .001) and linear increase (P = .02) compared with amyloid-negative individuals (eFigure 78A and B in the Supplement). Only Aβ42 amplitude was significantly associated with mean cortical binding potential (r = −0.41; P = .006). This finding is similar to that seen with CSF Aβ42 concentrations and mean cortical binding potential in the setting of amyloid deposition.16 In contrast, Aβ40 amplitude and linear rise did not show this loss of variability with amyloid deposition (P > .05; eFigure 78C and D in the Supplement).

Similar associations were observed when comparing Aβ amplitude and linear increase with Aβ42:Aβ40 ratios. For all 77 participants, Aβ42 day/night amplitude (r = −0.53; P < .001) and linear increase (r = −0.41; P < .001) were associated with the Aβ42:Aβ40 ratio, but Aβ40 was not (P > .05). After separating participants into amyloid-negative and amyloid-positive groups, only Aβ42 amplitude in amyloid-negative individuals remained significant (Figure 3A). Then, we divided participants into 3 groups based on Aβ42:Aβ40 ratios, as previously described2: Aβ42:Aβ40 less than 0.1, Aβ42:Aβ40 between 0.1 and 0.16, and Aβ42:Aβ40 greater than 0.16. The Aβ42:Aβ40 cutoff for amyloid-positive individuals was less than 0.12. When participants’ Aβ42:Aβ40 ratio was between 0.12 and 0.16, both the amplitude and linear rise of Aβ42 declined rapidly to levels similar to amyloid-positive individuals (Figure 3A and B). Aβ40 amplitude and linear rise do not show the same pattern (Figure 3C and D).

Association of Aβ Kinetics With Aβ Day/Night Amplitude and Linear Rise

The half-life of Aβ increases by 250% between ages 30 and 80 years, and the fractional turnover rate (FTR) of Aβ42 is specifically increased relative to Aβ40 in the presence of amyloid deposits, consistent with active deposition of Aβ42 relative to Aβ40.2,17 To assess how changes in Aβ kinetics (eg, FTR) are associated with the day/night oscillation of Aβ, we assessed the Spearman correlations of FTR Aβ40, FTR Aβ42, FTR Aβ42/40 (elevated in amyloid-positive individuals), CSF Aβ40, CSF Aβ42, and Aβ production rates with Aβ40 and Aβ42 day/night amplitude and linear rise with partial correlations, controlling for age and amyloid deposition effects (Table 1 and Table 2). Complementary associations between Aβ FTRs, concentrations, production rates, and cosinor parameters are expected, since the kinetic parameters are interrelated (eMethods in the Supplement).18 In this study, CSF Aβ concentration was highly associated with the production rate but weakly with FTR (eTable 4 in the Supplement).

We compared CSF Aβ day/night and linear concentrations with SILK kinetic parameters of production and clearance rates to determine the associations between hypothesized mechanisms of production and active clearance during sleep and wakefulness. The FTR is the turnover rate or soluble clearance rate of Aβ. As expected, markers of amyloid deposition, such as amyloid status, increased the Aβ42/40 turnover rate and lowered CSF Aβ42, and increased Aβ42 production was strongly associated with low Aβ42 day/night amplitudes and linear increases, regardless of age. The turnover rate of Aβ42 was not significantly associated with Aβ42 amplitude or linear rise (P > .05). The clearance of Aβ42 as measured by FTR Aβ42/40 was associated with both Aβ42 day/night amplitude and linear rise in the amyloid-negative group but not in the amyloid-positive group (Figure 4A and B). In contrast, Aβ40 day/night amplitude and linear rise were not associated with FTR Aβ42/40 for either amyloid group (Figure 4C and D).

Different associations were found for Aβ40. Aβ40 day/night amplitude and linear increase were not changed by markers of amyloid deposition (Table 1 and Table 2; all P > .05). Aβ40 amplitude was associated with age, Aβ40 production rate, and CSF Aβ40, while Aβ40 linear rise was associated with CSF Aβ40 and Aβ40 turnover rate. The associations remained significant even after controlling for age and amyloid status. These findings suggest that production and clearance mechanisms affect 2 relatively similar peptides, Aβ42 and Aβ40, differently.

Discussion

To our knowledge, we report the first comparison of the associations between Aβ production and clearance rates with CSF Aβ day/night amplitude and linear increases, accounting for aging and amyloid deposition effects. Although longitudinal follow-up studies are needed, our results suggest that amyloid deposition leads to premature loss of Aβ42 day/night patterns associated with aging, in contrast to Aβ40, which is largely driven by production rates. These results may have implications for the design of AD prevention trials targeting Aβ production (eg, β-secretase 1 inhibitors) and using CSF Aβ as a marker of target engagement. First, these results may affect the timing of therapeutic intervention. In amyloid-negative individuals, timing of antiamyloid intervention and age of participants may be critical factors. Adults younger than 73 years may benefit from antiamyloid therapy during the day or waking hours when production and concentrations are highest. For amyloid-positive individuals, the timing of antiamyloid therapy may be irrelevant because there are no significant time-of-day differences for Aβ to aggregate into insoluble plaque. Second, sleep disturbances have been implicated in AD pathogenesis,8 potentially exacerbating amyloidosis with impaired clearance mechanisms. Improving sleep quality or treating sleep disorders to reduce Aβ production and increase clearance could decrease the growth of amyloids and may prevent AD. However, this approach may not be effective in adults older than 73 years or in amyloid-positive individuals of any age. Third, CSF sampling frequency and volume need to be carefully controlled in studies to avoid sampling effects that may be caused by concentration gradients that give rise to linear rise, especially in those without amyloidosis.

The FTR includes a process of irreversible loss of Aβ. The strong associations between amyloid status, FTR Aβ42/40, Aβ42 day/night amplitude, and linear increase suggest that the loss of Aβ42 is to amyloid plaques. The changes in Aβ42 amplitude and linear rise associated with amyloid deposition were seen at borderline Aβ42:Aβ40 ratios between 0.12 and 0.16 when participants were classified as amyloid negative. This finding complements recently published work showing that CSF Aβ42 levels are highly associated with cortical amyloid load and decrease markedly prior to passing the threshold for amyloid positive on PET (11C)PiB.19 For Aβ40, the linear rise was associated with FTR Aβ40 but not with FTR Aβ42/40 or amyloid status. The etiology of Aβ40 loss is unclear and may be caused by clearance across the blood-brain barrier, degradation, formation of higher-order Aβ structures, or other causes, including clearance by more frequent sampling of higher volumes of CSF. It is possible that collecting 6 mL per hour would substantially increase the clearance of Aβ species, which can diffuse into the CSF and may represent the mechanism for the linear rise in Aβ seen in many CSF catheter studies.

In the last 10 to 15 years, multiple lumbar catheter studies have measured the effect of CSF sampling and time of day on CSF Aβ concentration5-7,20-26 (eTable 5 in the Supplement). Several of these prior studies reported that the linear increase and day/night variability in CSF Aβ levels during serial collection depends on the sampling frequency and/or volume,7,21,26 possibly caused by shifting CSF flow toward the lumbar space with repeated draws. However, none of these studies controlled for amyloid deposition. Our group previously reported that the amplitude of CSF Aβ oscillation decreased with age, while amyloid deposition markedly decreased linear rise.5 We have extended these findings to show that the linear rise of CSF Aβ also demonstrates an age-dependent effect, and amplitude is decreased in individuals with amyloid deposition regardless of age when the draw frequency and volume is uniform. Further, these amyloid effects were observed in Aβ42 rather than Aβ40. This finding has important implications in study designs using lumbar catheters in order to control for these effects.

Limitations

This study had limitations. A major weakness of our study is the lack of sleep-wake monitoring. Decreased Aβ production from neuronal activity and increased clearance via bulk fluid flow during sleep are 2 mechanisms hypothesized to drive the Aβ day/night pattern. We hypothesize that both Aβ amplitude and linear increase in amyloid-negative individuals are likely dependent on the sleep-wake cycle and other factors. However, deposition into amyloid plaques acts as a sink and is the dominant factor affecting the Aβ42 amplitude and linear rise in amyloid-positive individuals; since Aβ42 is more likely to aggregate into insoluble plaques than Aβ40, Aβ40 amplitude and linear rise are not significantly changed. Without concurrent sleep studies, we cannot determine if variability of Aβ amplitude and linear rise in individuals younger than 73 years old is affected by alterations in total sleep time or other sleep parameters.

Conclusions

Mass spectrometry is a novel assay that simultaneously measures absolute Aβ concentration and Aβ SILK. The profound age-dependent effect of amyloid on Aβ amplitude and linear rise, particularly for Aβ42, was not previously appreciated until the more precise MS method was used. We also found novel associations between Aβ concentration and production rates to Aβ amplitude and linear rise, as well as Aβ turnover and linear rise, independent of age and amyloid deposition. Further Aβ SILK studies in participants under different sleep conditions are needed to determine the sleep parameters that can manipulate Aβ production, clearance, and concentrations. Understanding the factors that influence Aβ physiology throughout the sleep/wake cycle could establish potential approaches and targets for the prevention or treatment of AD.

Back to top
Article Information

Corresponding Author: Brendan P. Lucey, MD, Department of Neurology, Washington University School of Medicine, 660 S Euclid Ave, Campus Box 8111, St Louis, MO 63110 (luceyb@neuro.wustl.edu).

Accepted for Publication: August 29, 2016.

Published Online: December 19, 2016. doi:10.1001/jamaneurol.2016.4202

Author Contributions: Drs Lucey and Bateman had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

Concept and design: Lucey, Bateman.

Acquisition, analysis, or interpretation of data: All Authors.

Drafting of the manuscript: Lucey, Mawuenyega, Patterson, Bateman.

Critical revision of the manuscript for important intellectual content: All Authors.

Statistical analysis: Lucey, Mawuenyega, Patterson.

Administrative, technical, or material support: Mawuenyega, Patterson, Kasten, Ovod, Bateman.

Study supervision: Lucey, Bateman.

Conflict of Interest Disclosures: Dr Lucey has consulted for AbbVie and Neurim Pharmaceuticals. He owns less than $5000 of stock in Cardinal Health. Dr Lucey also receives research support from the National Institutes of Health, the BrightFocus Foundation, and the McDonnell Center for Systems Neuroscience. Dr Mawuenyega may receive royalty income based on a patent of methods for simultaneously measuring the in vivo metabolism of 2 or more isoforms of a biomolecule, licensed by Washington University to C2N Diagnostics. Dr Patterson has provided consultations on β-amyloid peptide turnover kinetics for C2N Diagnostics. Mr Ovod may receive royalty income based on technology licensed by Washington University and tied to agreement 010395-0001. Dr Kasten receives a royalty from C2N for the cerebrospinal fluid β-amyloid patent/protocol. Neither Dr Morris nor his family owns stock or has equity interest (outside of mutual funds or other externally directed accounts) in any pharmaceutical or biotechnology company. Dr Morris has participated or is currently participating in clinical trials of antidementia drugs sponsored by the Anti-Amyloid Treatment in Asymptomatic Alzheimer’s Disease trial. Dr Morris has served as a consultant for Lilly USA and Takeda Pharmaceuticals. He receives research support from Eli Lilly/Avid Radiopharmaceuticals and is funded by National Institutes of Health grants P50AG005681, P01AG003991, P01AG026276, and UF01AG032438. Dr Bateman receives research funding from the National Institutes of Health, Alzheimer’s Association, and an anonymous foundation. He also receives grants from the DIAN Pharma Consortium (Amgen, AstraZeneca, Biogen, Eisai, Eli Lilly and Co, FORUM, Hoffman La-Roche, Pfizer, and Sanofi) and a tau consortium (AbbVie and Biogen) and has received honoraria from Roche, OECD, and Merck as a speaker and from IMI, Sanofi, and Boehringer Ingelheim as a consultant. Along with Dr Bateman, Washington University and David Holtzman, MD (chair of the Department of Neurology at Washington University), have equity stakes in C2N Diagnostics and may receive royalty income based on technology licensed by Washington University to C2N Diagnostics. In addition, Dr Bateman receives income from C2N Diagnostics for serving on the scientific advisory board. Through Washington University, Dr Bateman has also submitted the US nonprovisional patent application “Methods for Measuring the Metabolism of CNS Derived Biomolecules In Vivo.” No other disclosures were reported.

Funding/Support: This study was supported by grants R01NS065667, P50AG05681, P01AG03991, UL1 RR024992, P30DK056341, P41RR000954, P41GM103422, P60DK020579, and P30 DK020579 from the National Institutes of Health, grants UL1 TR000448 and KL2 TR000450 from the National Center for Advancing Translational Sciences via the Washington University Institute of Clinical and Translational Sciences, the Adler Foundation, and an anonymous foundation.

Role of the Funder/Sponsor: The funders had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.

References
1.
Prince  M, Wimo  A, Guerchet  M, Ali  G-C, Wu  Y-T, Prina  M.  World Alzheimer Report 2015: The Global Impact of Dementia: An Analysis of Prevalence. London, England: Alzheimer’s Disease International; 2015.
2.
Patterson  BW, Elbert  DL, Mawuenyega  KG,  et al.  Age and amyloid effects on human central nervous system amyloid-beta kinetics.  Ann Neurol. 2015;78(3):439-453.PubMedGoogle ScholarCrossref
3.
Meyer-Luehmann  M, Stalder  M, Herzig  MC,  et al.  Extracellular amyloid formation and associated pathology in neural grafts.  Nat Neurosci. 2003;6(4):370-377.PubMedGoogle ScholarCrossref
4.
Kang  J-E, Lim  MM, Bateman  RJ,  et al.  Amyloid-β dynamics are regulated by orexin and the sleep-wake cycle.  Science. 2009;326(5955):1005-1007.PubMedGoogle ScholarCrossref
5.
Huang  Y, Potter  R, Sigurdson  W,  et al.  Effects of age and amyloid deposition on Aβ dynamics in the human central nervous system.  Arch Neurol. 2012;69(1):51-58.PubMedGoogle ScholarCrossref
6.
Roh  JH, Huang  Y, Bero  AW,  et al.  Disruption of the sleep-wake cycle and diurnal fluctuation of β-amyloid in mice with Alzheimer’s disease pathology.  Sci Transl Med. 2012;4(150):150ra122.PubMedGoogle ScholarCrossref
7.
Lucey  BP, Gonzales  C, Das  U,  et al.  An integrated multi-study analysis of intra-subject variability in cerebrospinal fluid amyloid-β concentrations collected by lumbar puncture and indwelling lumbar catheter.  Alzheimers Res Ther. 2015;7(1):53.PubMedGoogle ScholarCrossref
8.
Lucey  BP, Bateman  RJ.  Amyloid-β diurnal pattern: possible role of sleep in Alzheimer’s disease pathogenesis.  Neurobiol Aging. 2014;35(suppl 2):S29-S34.PubMedGoogle ScholarCrossref
9.
Xie  L, Kang  H, Xu  Q,  et al.  Sleep drives metabolite clearance from the adult brain.  Science. 2013;342(6156):373-377.PubMedGoogle ScholarCrossref
10.
Kress  BT, Iliff  JJ, Xia  M,  et al.  Impairment of paravascular clearance pathways in the aging brain.  Ann Neurol. 2014;76(6):845-861.PubMedGoogle ScholarCrossref
11.
Morris  JC.  The Clinical Dementia Rating (CDR): current version and scoring rules.  Neurology. 1993;43(11):2412-2414.PubMedGoogle ScholarCrossref
12.
Berg  L, Miller  JP, Storandt  M,  et al.  Mild senile dementia of the Alzheimer type 2: longitudinal assessment.  Ann Neurol. 1988;23(5):477-484.PubMedGoogle ScholarCrossref
13.
Mintun  MA, Larossa  GN, Sheline  YI,  et al.  [11C]PIB in a nondemented population: potential antecedent marker of Alzheimer disease.  Neurology. 2006;67(3):446-452.PubMedGoogle ScholarCrossref
14.
Mawuenyega  KG, Kasten  T, Ovod  V, Lucey  B, Sigurdson  W, Bateman  RJ. Immuno-based-LC/SRM as a diagnostic tool for protein dynamics of amyloid β isoforms instead of ELISA in the clinical laboratory. Paper presented at: 62nd Conference on Mass Spectrometry and Allied Topics; June 15-19, 2014; Baltimore, MD.
15.
Bateman  RJ, Munsell  LY, Morris  JC, Swarm  R, Yarasheski  KE, Holtzman  DM.  Human amyloid-β synthesis and clearance rates as measured in cerebrospinal fluid in vivo.  Nat Med. 2006;12(7):856-861.PubMedGoogle ScholarCrossref
16.
Fagan  AM, Mintun  MA, Mach  RH,  et al.  Inverse relation between in vivo amyloid imaging load and cerebrospinal fluid Abeta42 in humans.  Ann Neurol. 2006;59(3):512-519.PubMedGoogle ScholarCrossref
17.
Potter  R, Patterson  BW, Elbert  DL,  et al.  Increased in vivo amyloid-β42 production, exchange, and loss in presenilin mutation carriers.  Sci Transl Med. 2013;5(189):189ra77.PubMedGoogle ScholarCrossref
18.
Elbert  DL, Patterson  BW, Bateman  RJ.  Analysis of a compartmental model of amyloid beta production, irreversible loss and exchange in humans.  Math Biosci. 2015;261:48-61.PubMedGoogle ScholarCrossref
19.
Vlassenko  AG, McCue  L, Jasielec  MS,  et al.  Imaging and cerebrospinal fluid biomarkers in early preclinical alzheimer disease.  Ann Neurol. 2016;80(3):379-387.PubMedGoogle ScholarCrossref
20.
Bateman  RJ, Wen  G, Morris  JC, Holtzman  DM.  Fluctuations of CSF amyloid-β levels: implications for a diagnostic and therapeutic biomarker.  Neurology. 2007;68(9):666-669.PubMedGoogle ScholarCrossref
21.
Li  J, Llano  DA, Ellis  T,  et al.  Effect of human cerebrospinal fluid sampling frequency on amyloid-β levels.  Alzheimers Dement. 2012;8(4):295-303.PubMedGoogle ScholarCrossref
22.
Moghekar  A, Goh  J, Li  M, Albert  M, O’Brien  RJ.  Cerebrospinal fluid Aβ and tau level fluctuation in an older clinical cohort.  Arch Neurol. 2012;69(2):246-250.PubMedGoogle ScholarCrossref
23.
Slats  D, Claassen  JA, Spies  PE,  et al.  Hourly variability of cerebrospinal fluid biomarkers in Alzheimer’s disease subjects and healthy older volunteers.  Neurobiol Aging. 2012;33(4):831.e1-831.e9.PubMedGoogle ScholarCrossref
24.
Dobrowolska  JA, Kasten  T, Huang  Y,  et al.  Diurnal patterns of soluble amyloid precursor protein metabolites in the human central nervous system.  PLoS One. 2014;9(3):e89998.PubMedGoogle ScholarCrossref
25.
Ooms  S, Overeem  S, Besse  K, Rikkert  MO, Verbeek  M, Claassen  JA.  Effect of 1 night of total sleep deprivation on cerebrospinal fluid β-amyloid 42 in healthy middle-aged men: a randomized clinical trial.  JAMA Neurol. 2014;71(8):971-977.PubMedGoogle ScholarCrossref
26.
Van Broeck  B, Timmers  M, Ramael  S,  et al.  Impact of frequent cerebrospinal fluid sampling on Aβ levels: systematic approach to elucidate influencing factors.  Alzheimers Res Ther. 2016;8(1):21.PubMedGoogle ScholarCrossref
×