High diagnostic performance of plasma and cerebrospinal fluid beta‐synuclein for sporadic Creutzfeldt–Jakob disease

ABSTRACT Beta‐synuclein is a promising cerebrospinal fluid and blood biomarker of synaptic damage. Here we analysed its accuracy in the discrimination between sporadic Creutzfeldt–Jakob disease (n = 150) and non‐prion rapidly progressive dementias (n = 106). In cerebrospinal fluid, beta‐synuclein performed better than protein 14‐3‐3 (AUC 0.95 vs. 0.89) and, to a lesser extent, than total tau (AUC 0.92). Further, the diagnostic value of plasma beta‐synuclein (AUC 0.91) outperformed that of plasma tau (AUC 0.79) and neurofilament light chain protein (AUC 0.65) and was comparable to that of cerebrospinal fluid biomarkers. Beta‐synuclein might represent the first highly accurate blood biomarker for the diagnosis of sporadic Creutzfeldt–Jakob disease.

Patients with Alzheimer's disease (rapidly progressive form) were diagnosed at autopsy (n=4) or according to the international criteria (n=43), 11 including the presence of a characteristic Alzheimer's disease CSF biomarker profile according to our in-house cutoff values. 12Further, the diagnosis of rpAD required at least one of the following: (1) rapid cognitive decline with or without motor signs leading to the clinical suspicion of prion disease, (2) CSF t-tau > 1100 pg/ml. 13,14oxic/metabolic encephalopathies 10 1 9 Central nervous system malignancies 5 4 1

RPD = rapidly progressive dementia
Biomarker analyses CSF t-tau was analysed using a commercially available enzyme-linked immunosorbent assay (ELISA) kit (INNOTEST htau-Ag, Innogenetics/Fujirebio Europe, Ghent, Belgium) according to the manufacturer's instructions. 15The 14-3-3 gamma isoform was measured using a commercially available ELISA assay kit (Circulex 14-3-3 gamma ELISA kit, MBL, Woburn, MA) according to the manufacturer's instructions. 15asma tau and NfL were measured on the Simoa SR-X analyzer platform (Quanterix, Billerica, Massachusetts, USA) using Simoa Human t-tau and the Simoa NF-light advantage kits, respectively. 16ta-synuclein CSF levels were measured with an in-house established sandwich ELISA, as previously described. 17In brief, Nunc Maxisorp 96-well plates (Thermo Fisher Scientific, Massachusetts, USA) were coated with 100 μL of a capture antibody (EP1646Y, Abcam, Cambridge, UK), diluted in 100 mM bicarbonate-carbonate buffer pH 9.6 at a concentration of 3.3ug/mL.After overnight incubation at +4°C, the coating solution was removed, and the plate was blocked with 320 μL 2% bovine serum albumin (BSA) in phosphate-buffered saline (PBS) solution with 0.05% Tween 20 for 2 hours at room temperature (RT).For the calibrator preparation, we purchased recombinant beta-synuclein from rPeptide (Watkinsville, Georgia, USA) and determined the exact protein concentration of the stock solution by amino acid analysis (Alphalyse A/S, Odense, Denmark).After removing the blocking buffer, we loaded in each well 100 μL calibrators ranging from 10 to 1000 pg/mL and 100 μL CSF samples diluted in blocking buffer.CSF samples were dilution stable between dilutions of 1:4 up to 1:20.Plates were shaken for 2 min on a small MR1 rocker (Biosan, Riga, Latvia) and then incubated at RT for 1.5 hours without shaking.After sample incubation, each well was washed with 300 μL washing buffer (PBS with 0.05% Tween 20) three times.We used as the detection antibody an antibeta-synuclein monoclonal antibody purchased from Abcam (EP1537Y, Abcam, Cambridge, UK), biotinylated in a 40:1 ratio according to the biotinylation protocol provided by Thermo Fisher Scientific (Massachusetts, USA).After removing samples and calibrators and washing the plate with 300 μL washing buffer, we added per well 100 μL of previously biotinylated detection antibody at a concentration of 0.66 μg/mL and incubated the plate for 30 min at RT.After a third washing step, 100 μL of a streptavidinhorseradish peroxidase (Vector laboratories, California, USA) solution was added to each well and incubated at RT for 1 hour.Solution was removed, the plate washed and 100 μL 3,3′,5,5′-Tetramethylbenzidin (ThermoFisher Scientific, Massachusetts, USA) were added and incubated for 5.5 min at RT.The reaction was stopped with 100 μL 1M hydrochloric acid per well.Plates were measured at 450 nm and 570 nm reference wavelength.Concentrations were obtained using a 4-parameter standard curve.
Beta-synuclein blood levels were measured with a novel digital ELISA assay. 2 We used a monoclonal antibody specifically recognizing β-synuclein (EP1537Y) as a capture antibody and a monoclonal antibody against αand β-synuclein (EP1646Y, both Abcam, Cambridge, UK) as a detection antibody.The biotinylation of the detection antibody is described above.The capture antibody was coupled to carboxylated paramagnetic beads (Quanterix) according to the manufacturer's protocol (coating concentration 0.2 mg/mL).A total of 500,000 beads were applied per replicate, of which 75% were Helper Beads (Quanterix).We biotinylated the monoclonal detection antibody with 40× molar excess of biotin and used in a concentration of 0.5 μg/mL.
Buffers for beads and detector were phosphate-buffered saline-Tween 0.5% and phosphate-buffered saline-Tween 0.05%, respectively.Streptavidin-β-galactosidase (SBG) (Quanterix) was diluted in SBG buffer (Quanterix) to a final concentration of 150 pM.We used, as a substrate, resorufin β-D-galactopyranoside (Quanterix).Calibrators ranging from 0.625 to 100 pg/mL were prepared with recombinant β-synuclein, and 400 μL of each calibrator was added to a 96-well plate (Quanterix).Plasma samples were diluted 1:2 and shaken for 10 minutes at room temperature at 1,200 rpm.The digital ELISA was run on the Simoa HD-X platform (Quanterix, Billerica, Massachusetts, USA).Diluted sample (220 μL) was added to each well.Beads, detector, SBG, plate, and substrate were placed into the Simoa platform, and a 3-step custom assay was started.
The mean intra-and inter-assay coefficients of variation were ≤5% and <15% for all CSF and blood biomarkers.

Statistical analyses
We used IBM SPSS Statistics V.21 (IBM), GraphPad Prism V.7 (GraphPad Software, La Jolla, California, USA), and R software V.4.0.2 (R foundation, Vienna, Austria).Depending on distribution, data were expressed as percentage, mean±standard deviation (SD), or median and interquartile range (IQR).We adopted the χ 2 test for categorical variables.For continuous variables, depending on the data distribution and number of groups, we applied the Mann-Whitney U test, t-test, Kruskal-Wallis test (followed by Dunn-Bonferroni post hoc test) or the ANOVA (followed by Tukey's post hoc test).All reported p values were adjusted for multiple comparisons.We performed multivariate linear regression models to adjust for age the differences in CSF biomarkers between the groups after the transformation of the dependent variable in the natural logarithmic scale.Spearman's correlations and uni-or multivariate regression analyses were performed to test the possible associations between variables.The diagnostic accuracy of each marker was calculated by means of Receiver operating characteristic (ROC) analyses.The optimal cut-off value for each biomarker was defined using the maximized Youden's index.The DeLong was used to compare the different area under the curve (AUC) values.
For the analysis of survival, beta-syn concentration was naturally log-transformed to fulfil the normal distribution.Univariate and multivariate Cox regression analyses tested the associations between survival and CSF or plasma beta-syn or known prognostic factors in sporadic Creutzfeldt-Jakob disease such as age at sample collection, disease duration at sample collection, PRNP codon 129 genotype and clinicopathological subtype. 16We performed survival analyses in the whole cohort of sporadic Creutzfeldt-Jakob disease cases

CSF and plasma biomarkers in the diagnostic groups
Patients with non-prion RPD were older than those with sporadic Creutzfeldt-Jakob disease (p<0.001),but there were no differences in sex distribution between the two groups.Age influenced plasma NfL levels in the Creutzfeldt-Jakob disease (r=0.298,p=0.003) and non-prion RPD (r=0.275,p=0.006) groups, but no other biomarker concentrations.At variance, sex showed no effect on blood and CSF biomarkers.Accordingly, all analyses on plasma NfL were adjusted for age.
Comparisons of these biomarkers between molecular subgroups have already been reported. 15,16ter the exclusion of probable sporadic Creutzfeldt-Jakob disease cases, higher CSF beta-syn levels were found in MM(V)1 (p=0.004) and VV2 (p=0.006)groups compared to the MV2K group.Further, MM(V)1 subjects showed higher plasma beta-syn values than VV2 (p=0.001) and MV2K (p=0.004)patients, whereas the level of the biomarker did not differ between VV2 and MV2K groups.
After considering all diagnostic subgroups, there was no significant difference in CSF and blood beta-syn levels among non-prion RPD etiologies.However, when only the two most numerous non-prion RPD etiologies were considered, patients with inflammatory RPD showed higher CSF (p<0.008) but not plasma beta-syn concentrations compared to those with neurodegenerative RPD.CSF = cerebrospinal fluid; IQR = interquartile range; NfL = neurofilament light chain; RPD = rapidly progressive dementia; sCJD = sporadic Creutzfeldt-Jakob disease; t-tau = total tau protein;

Diagnostic value of CSF and blood beta-syn and associations of beta-synuclein with survival in sCJD
In the ROC analyses, by limiting the analysis to the most frequent and rapidly progressing sporadic Creutzfeldt-Jakob disease subtypes (i.e., MM[V]1 and VV2), the performances of both CSF (AUC 0.974±0.009)and blood beta-syn (AUC 0.952±0.014)increased.Still, the comparisons with other markers remained similar.beta-syn = beta-synuclein; CI = confidence interval; CSF = cerebrospinal fluid; HR = hazard ratio and in three separate subgroups, according to the most prevalent clinicopathological subtypes: (1) sporadic Creutzfeldt-Jakob disease MM(V)1, (2) VV2, and (3) MV2K.The results are presented as hazard ratios (HRs) and 95% confidence intervals (CIs).The assumption of proportional hazard was assessed by Schoenfeld residuals.Statistical tests were two-tailed, and p values were considered statistically significant at <0.05.

Table S6 . Diagnostic accuracies of CSF and blood beta-syn in the differential diagnosis between sporadic Creutzfeldt-Jakob disease and major forms of RPD.
RPD = rapidly progressive dementia; RT-QuIC = real-time quaking-induced conversion assay; sCJD = sporadic Creutzfeldt-Jakob disease; sens = sensitivity, spec = specificity; t-tau = total tau protein;

Table S7 . Associations of CSF and blood beta-syn with survival time in the whole sporadic Creutzfeldt- Jakob disease cohort and after stratification according to the disease subtype Diagnostic group Beta-syn Univariate Cox regression
* Both multivariate Cox regression analyses included age and time from symptoms onset to LP as covariates.