Characterization of the postsynaptic protein neurogranin in paired cerebrospinal fluid and plasma samples from Alzheimer’s disease patients and healthy controls

The paired CSF and plasma study included 25 AD patients (mean age: 76 ± 8 years) and 20 control subjects without dementia (mean age: 54 ± 14 years) (see Table 1 for demographics and biomarker characteristics). The AD group consisted of patients with probable AD with a high level of evidence of an AD pathophysiologic process and patients with mild cognitive impairment with a high likelihood of an underlying AD process. In the rest of this article, this group is called AD. The patients were recruited in Erlangen, Germany, and each underwent a thorough clinical investigation, including a medical history as well as a physical, neurological and psychiatric examinations. The diagnosis of AD was made according to the National Institute of Neurological and Communicative Disorders and Stroke-Alzheimer’s Disease and Related Disorders association (NINCDS-ADRDA) criteria [17]. No patient with AD had a family history of dementia suggestive of autosomal dominant AD. The control group consisted of patients with other neuropsychiatric conditions without dementia. The clinical diagnoses (including neuroimaging and neuropsychology) were combined with the results of the following CSF biomarkers: Aβ_1–42, Aβ_1–40 (to calculate the Aβ_42/40 concentration ratio), t-tau and p-tau181. The subjects were included in the study only if the clinical, neuroimaging and neuropsychological diagnoses were in accordance with the neurochemical findings as described elsewhere [18]. There were no patients included in the study with a completely negative AD biomarker pattern. Control subjects had all biomarkers within normal ranges. CSF t-tau, p-tau, Aβ_1–40 and Aβ_1–42 concentrations were determined using established ELISA methods as described elsewhere [18]. All analyses were performed at the Universitätsklinikum Erlangen. The study was conducted according to the provisions of the Declaration of Helsinki. All subjects gave their written informed consent for the use of their clinical data for research purposes, as approved by the ethics committee of the Universitätsklinikum Erlangen.

Plasma samples were collected from 13 patients with AD (mean age: 78 ± 7 years) and 17 individuals without dementia (controls; mean age: 58 ± 9 years) at the Memory Clinic, Skåne University Hospital, Sweden (see Table 1 for full biomarker characteristics and demographics). The individuals underwent brain imaging, routine laboratory testing and neurological, psychiatric and cognitive examinations. Patients diagnosed with AD fulfilled the dementia criteria of the Diagnostic and Statistical Manual of Mental Disorders, Third Edition, Revision [19], and the criteria for probable AD as defined by NINCDS-ADRDA [17]. The control individuals without dementia experienced subjective cognitive symptoms at baseline, but thorough clinical investigation as well as clinical follow-up revealed that the patients were not affected by a dementia disorder or a neurological disease. CSF t-tau, p-tau and Aβ_1–42 concentrations were determined using established ELISA methods as described elsewhere [16]. All analyses were performed at the Clinical Neurochemistry Laboratory at the Sahlgrenska University Hospital, Mölndal, Sweden. All individuals gave their informed consent to participate in research before their samples were stored in a biobank. A passive consent procedure was then used whereby consent for retrospective use of banked samples and basic data was assumed if individuals did not actively retract permission, as instructed in local press advertisements. This study procedure was approved by the local ethics committee at Lund University, Sweden.

The monoclonal mouse antibody Ng7, with an epitope including amino acids 52 to 65 [16], was used as a capturing antibody and was coated on QUICKPLEX 96-well plates (Meso Scale Discovery (MSD), Rockville, MD, USA) at a final concentration of 2.0 μg/ml (40 μl/well) in phosphate-buffered saline (PBS), followed by incubation overnight at room temperature (RT). Before being coated with the capturing antibody, the plates were washed once with 150 μl of PBS. After the plates were washed with 300 μl of PBS containing 0.05% Tween 20 (PBS-Tween) four times for 30 seconds each time, the remaining protein-binding sites were blocked with 5% MSD Blocker A solution (150 μl/well; Meso Scale Discovery) for 1 hour at RT and subsequently washed again with 300 μl of PBS-Tween four times for 30 seconds each time. Then the full-length Ng calibrators with concentrations ranging between 31.3 and 4,000 pg/ml (1:2 dilution; CSF samples) or 31.3 to 23,000 pg/ml (1:3 dilution; plasma samples), blanks and samples (50 μl/well) were incubated in duplicate together with the detecting antibody, polyclonal Ng anti-rabbit antibody (ab23570; Upstate Biotechnology, Lake Placid, NY, USA) diluted 1:20,000 in 0.1% bovine serum albumin (BSA) in PBS-Tween overnight at RT (50 μl/well). Plasma samples were diluted 1:40 in PBS. After washing four times for 30 seconds each time in PBS-Tween, a MSD SULFO-TAG goat anti-rabbit antibody (25 μl/well; Meso Scale Discovery), concentration 0.5 μg/ml diluted in 0.1% BSA in PBS-Tween, was added and incubated for 2 hours at RT while being shaken at 700 rpm. After washing with 300 μl of PBS-Tween four times for 30 seconds each time, 150 μl of 2× MSD Read Buffer T with surfactant (Meso Scale Discovery) were used for immediate reading with a QUICKPLEX SQ 120 reader (Meso Scale Discovery). A fitted four-parameter logistic model with relative weighting (1/y ²) was used as the calibration curve, and the blank was included as zero concentration of Ng. All samples were analyzed without knowledge of clinical diagnosis.

To minimize nonspecific binding of other proteins during the immunoprecipitation, all plasma samples were pretreated twice with 20 μl of Novex Dynabeads Protein G (Life Technologies, Carlsbad, CA, USA) per milliliter of plasma and incubated for 1 hour at RT before the beads were magnetically removed as described previously, with some modifications [20]. HI-MS analysis of Ng peptides was performed as described elsewhere [16]. In brief, 4 μg of the monoclonal anti-Ng antibodies Ng2 and Ng3 were separately added to 25 μl of Dynabeads M-280 sheep anti-mouse immunoglobulin G (Life Technologies) and cross-linked as previously described [21]. Beads coated with Ng2 and Ng3 were used for immunoprecipitation of plasma to which n-octyl-β-d-glucopyranoside (10%, end concentration 0.1%; Fluka Chemie, Buchs, Switzerland) had been added. The beads and sample were transferred to a KingFisher magnetic particle processor (polypropylene tubes; Thermo Scientific, Waltham, MA, USA) for automatic washing and elution of the Ng peptides and protein. Eluted Ng was collected and dried in a vacuum centrifuge and redissolved in 5 μl of 0.1% formic acid (FA) in 20% acetonitrile and subsequently analyzed using an ultrafleXtreme matrix-assisted laser desorption/ionization (MALDI) time-of-flight/time-of-flight (TOF/TOF) MS (Bruker Daltonics, Bremen, Germany). All solvents used were of high-performance liquid chromatography (HPLC) quality.

In both the paired cohort and validation study where MALDI-TOF/TOF MS was used, the custom Ng peptide RKKIKSGERGRKGPGPGGPGGAGVARGGAGGGP (corresponding to Ng_43–75), with glycine labeled with double ¹³C (theoretical molecular mass: 3,011 Da; CASLO, Lyngby, Denmark), was added to the plasma prior to sample preparation and was used as an internal standard. Samples were analyzed blinded (that is, without knowledge of clinical diagnosis).

The identities of the Ng peptides were confirmed using a top-down approach as previously described [22], but with updated instrumentation. Briefly, analysis was performed using nanoflow liquid chromatography (Dionex UltiMate 3000; Thermo Scientific) coupled to an electrospray ionization Q Exactive Hybrid Quadrupole Orbitrap tandem mass spectrometer (Thermo Scientific). An Acclaim PepMap 100 C8 trap column (length: 20 mm; inner diameter: 75 μm; particle size: 3 μm; pore size: 100 Å (Thermo Scientific)) was used for online desalting, and a reversed-phase Acclaim PepMap rapid separation liquid chromatography column (length: 500 mm, inner diameter: 75 μm; particle size: 2 μm; pore size: 100 Å (Thermo Scientific)) was used for high-resolution separation. Mobile phases were 0.1% aqueous FA (v/v) (A) and 0.1% FA in 84% acetonitrile in water (v/v) (B). The separation was performed at a flow rate of 150 nl/min by applying a linear gradient of 3% to 40% B for 50 minutes. A column oven was used and set to 40°C. All solvents used were of HPLC quality. The Q Exactive system was used in positive ion mode to acquire full-scan spectra in the 300 to 1,200 mass-to-charge (m/z) ratio range at a resolution setting of 70,000. Subsequent tandem mass spectra (MS/MS) were produced using the higher-energy collisional dissociation fragmentation cell and acquired at the same resolution setting. An inclusion list was employed to ensure that we obtained MS/MS data of the compounds of interest. Mass spectral raw files were then charge-deconvoluted by using Mascot Distiller software (Matrix Science, Boston, MA, USA) and subsequently submitted to a Mascot database search (Matrix Science).

Plasma and CSF samples were depleted of native Ng by three consecutive immunoprecipitations before full-length recombinant Ng (1,000 pg/ml) was added to the samples, followed by 24-hour incubation at RT. After the incubation, samples were analyzed by HI-MS to screen for degradation products from the full-length Ng protein in the form of newly formed Ng peptides.

Pooled CSF samples with or without a stabilizing agent (sodium azide at a final concentration of 0.1% in samples), were kept at RT, 4°C or −20°C for 1 to 7 days before they were placed in long-term storage at −80°C pending analysis. As a reference sample, a freshly prepared aliquot, with or without stabilizer, was stored at −80°C on the day of the preparation (that is, day 0). All other samples were normalized against the Ng content of the respective reference sample for either group. Ng content was analyzed by MSD.

For statistical analysis, Prism 6 for Windows software (GraphPad Software, La Jolla, CA, USA) was used. Because biomarker values were skewed, nonparametric tests were used. Differences between groups were assessed using the Mann–Whitney U test, and significance was determined at P < 0.05. Correlations were determined using Spearman’s correlation.