2.1 Participants
2.1.1 PRECISION study cohort
Blood was collected from ten healthy volunteers twice a day at two-day intervals over a five-day period. Participant characteristics are presented in Supplemental Table 1 (Additional File 1). On each day, blood was collected in the morning between 7 - 10 am and in the evening between 4 - 7 pm. The average time between blood draws on the same day was 8.4 ± 1.0 hours. Participants had fasted for at least 10 hours prior to morning blood draws (average fasting period: 11.2 ± 1.0 hours), but fasting was not requested for blood collection done in the evening. Blood was collected into 8 mL ethylenediaminetetraacetic acid (EDTA) treated tubes by peripheral venipuncture with a 21G butterfly needle. Plasma was processed as quickly as possible, and the average time between venipuncture to plasma isolation was 19 minutes. Specimens were centrifuged at 2,000 x g for 10 minutes at 4 °C, and the supernatant was carefully removed and aliquoted into 500 μL lots in 1.5 mL Eppendorf protein lobind tubes and stored at -80 °C. The study was conducted in accordance with local clinical research regulations and approved by the Partners Institutional Review Board (Walsh, BWH2017P000259), and all participants gave written informed consent.
2.1.2 Down syndrome and control participants
Plasma samples were taken from 100 people with Down syndrome. Samples were obtained from three centers. The bulk of the samples (n=82) were collected prospectively at the Massachusetts General Hospital Down Syndrome Program. Additional samples came from the Center for Neuroimaging of Aging and Neurodegenerative Disease of the Massachusetts General Hospital (n=13), and the LonDownS Consortium at King’s College London (n=5).
Massachusetts General Hospital Down Syndrome Program
Participants with Down Syndrome and/or their legal guardians were consented during outpatient visits at the Massachusetts General Hospital Down Syndrome Program. The protocol was approved by the Partners Human Research Committee. Blood was drawn from the antecubital vein with a 21G needle and collected into EDTA treated tubes. Plasma was isolated by centrifugation at 2000 x g at 4°C for 10 minutes, and aliquots processed and stored as described for the PRECISION cohort.
Center for Neuroimaging of Aging and Neurodegenerative Disease of the Massachusetts General Hospital
Participants with DS were recruited from the MGH Down syndrome research database. Experimental procedures were explained and signed informed consent/assent was obtained prior to participation. The protocol was approved by the Institutional Review Board (2018P000898). Blood collection and processing was performed as described for the PRECISION study cohort.
LonDownS Consortium
Participants with DS were also recruited as a part of the LonDownS Consortium’s cohort study of Alzheimer’s disease [25]. Ethical approval was obtained from the North West Wales Research Ethics Committee (13/WA/0194). Written informed consent was obtained if participants could consent for themselves, otherwise a consultee was asked to approve the individual’s inclusion. Blood samples were collected in EDTA tubes and processed within 2 hours. Plasma was prepared by centrifuging samples for 10 minutes at 2,200 x g at 4°C; the supernatant was aliquoted and stored at -80 °C.
Boston Children’s Hospital and Partners Biobank
Specimens from age- and sex-matched control subjects were obtained from the Precision Link Biobank at Boston Children’s Hospital (BCH) and the Biobank at Partners Healthcare in Boston (MA, USA). Controls had no history of diseases of the central nervous system. Specifically, there was no evidence of brain tumor, normal pressure hydrocephalus, stroke, severe brain trauma, brain surgery, epilepsy, encephalitis or dementia. Included participants were free of acute infectious disease.
Precision Link Biobank participants (n=67) were enrolled throughout the hospital, across diverse clinical settings. In-person informed consent was obtained from all participants enrolling in the Biobank and provides permission to (1) access electronic health record data for research, (2) collect and use of residual specimens produced as by-products of routine care, and (3) share de-identified data and specimens outside of the institution. Participants provided a 4 mL blood sample for research use. Whole blood was collected in EDTA tubes, centrifuged at 2000 x g for 10 minutes at room temperature, with plasma removed and aliquoted into 0.5 mL microcentrifuge tubes. Aliquots were stored at -80 °C in the Biobank Core Lab facility until requested. The Precision Link Biobank initiative is approved by the BCH Institutional Review Board (P00000159).
Additional samples (n=33) and health information were obtained from the Partners HealthCare Biobank, a biorepository of consented patient samples at Partners HealthCare. The Partners HealthCare Biobank is approved by the Institutional Review Board (2009P002312). Blood was collected, and plasma generated as described above.
2.2 Single molecule array (Simoa) immunoassays
Levels of Aβ42, NT1 tau, and NfL were quantified using Simoa assays. All assays were performed by the same operator and conducted on the same automated HD-1 analyzer (Quanterix, Billerica, MA). Consumables and reagents other than certain antibodies were obtained from Quanterix. The NT1 tau assay was developed in-house [19], and commercial kits were used to measure Aβ42 and NfL.
NT1
This is a 3-step assay capable of detecting all forms of tau which contain residues 6 – 198 [19]. BT2 (194-198, Thermo, Waltham, MA, USA) was conjugated onto paramagnetic beads at 2 mg/mL and used for capture. Biotinylated Tau12 (6-13, Merck Millipore, Darmstadt, Germany) was used for detection.
The optimal plasma dilution (1:4) to minimize matrix effects was determined previously [19]. Plasma samples were thawed on ice, centrifuged at 14,000 x g for 4 minutes, and the upper 90 % of the supernatant transferred to Eppendorf protein lobind tubes and then diluted 1:4 with Tau 2.0 sample diluent reagent (Quanterix, Billerica, MA). Samples, standards and blanks were analyzed at least in duplicate.
The lower limit of quantification (LLoQ) defined as the lowest standard: (i) with a signal higher than the average signal for the blank plus 9 SDs, and (ii) allowed a percent recovery ≥100 ± 20 %. In 4 runs over 4 days the LLoQ was 0.25 pg/mL. The average % coefficient of variation (%CV) of all samples measured in the study was 13.8%. Assay characteristics including selectivity, dilution linearity, and spike and recovery were reported previously [19].
Aβ42 and NfL
The Simoa® Aβ42 Advantage (Quanterix, Billerica, MA, USA) and Simoa™ NF-light® Advantage (Quanterix, Billerica, MA, USA) kits were used according to the manufacturer’s instructions. Reagents from a single lot were used for analysis of all specimens from either the PRECISION study cohort or the Down syndrome study cohort. To determine the optimal dilution factor, plasma samples from 73 healthy donors were diluted 1:4 and 1:8 and analyzed for Aβ1-42 and NfL, and the highest dilution factor that allowed reliable quantification of samples was used. Thereafter, specimens were diluted 1:8 for Aβ42 and 1:4 for NfL. As with specimens for the NT1 assay, plasma was centrifuged at 14,000 x g for 4 minutes, the upper 90 % transferred to a new Eppendorf protein lobind tube and diluted with sample diluent provided in the kits. LLoQs for the Aβ1-42 and NfL assays were calculated as described for the NT1 assay, and were 0.41 pg/mL and 0.47 pg/mL, respectively. The average %CV for all samples measured in the study was 5.5% for the Aβ1-42 and 7.5% for the NfL assay.
2.3 Statistical analyses
Statistical analyses were carried out using GraphPad Prism, version 8 (LaJolla, CA, USA) and Stata, version 15.1 (Stata Corp., College Station, TX, USA). The effect of daytime (morning vs. evening blood draws) and different weekdays (Day 1 vs. Day 3 vs. Day 5) on plasma biomarker levels was assessed using linear mixed models (“mixed”). Both variables were included as fixed affects, whereas subjects were used as random effects. To assess goodness-of-fit we included both variables in a stepwise process and compared Akaike Information criterion (AIC) of the respective models (“estat ic”). Tests on estimated coefficients were performed using the Wald test postestimation command (“test”). Differences of biomarker levels between DS and controls for specified age-ranges were assessed using a paired t-test (for normally distributed data) or a Wilcoxon signed-rank test (for non-normally distributed data). p-values were adjusted for multiple testing by Bonferroni correction. Additionally, piecewise linear regression using hockey-stick regression was applied to model the effect of age on plasma biomarker levels in DS and controls (and differences between both groups) [26]. The slopes of both regression lines and the change-point where both regression lines meet were estimated using the “nl hockey” function. From specimens of DS and age-matched control participants, 10 out of 600 measurements (1.7%, 7 for the NT1 tau and 3 for the NfL assay, see Additional File 1: Supplemental Table 2 for raw data) were below the LLoQ of the respective assay (determined as described above), and for statistical analysis these samples were assigned values equal to the LLoQ of the assay. Statistical analysis was repeated with the actual readings of these 10 measurements, but produced similar results. The significance threshold was set to a two-sided p ≤ 0.05.