Evaluation of plasma tau and neurofilament light chain biomarkers in a 12-year clinical cohort of human prion diseases

Abstract

Prion diseases are fatal neurodegenerative conditions with highly accurate CSF and imaging diagnostic tests, but major unmet needs for blood biomarkers. Using ultrasensitive immuno-assays, we measured tau and neurofilament light chain (NfL) protein concentrations in 709 plasma samples taken from 377 individuals with prion disease during a 12 year prospective clinical study, alongside healthy and neurological control groups. This provides an unprecedented opportunity to evaluate their potential as biomarkers. Plasma tau and NfL were increased across all prion disease types. For distinguishing sCJD from control groups including clinically-relevant "CJD mimics", both show considerable diagnostic value. In sCJD, NfL was substantially elevated in every sample tested, including during early disease with minimal functional impairment and in all follow-up samples. Plasma tau was independently associated with rate of clinical progression in sCJD, while plasma NfL showed independent association with severity of functional impairment. In asymptomatic PRNP mutation carriers, plasma NfL was higher on average in samples taken within 2 years of symptom onset than in samples taken earlier. We present biomarker trajectories for nine mutation carriers healthy at enrolment who developed symptoms during follow-up. NfL started to rise as early as 2 years before onset in those with mutations typically associated with more slowly progressive clinical disease. This shows potential for plasma NfL as a "proximity marker", but further work is needed to establish predictive value on an individual basis, and how this varies across different PRNP mutations. We conclude that plasma tau and NfL have potential to fill key unmet needs for biomarkers in prion disease: as a secondary outcome for clinical trials (NfL and tau); for predicting onset in at-risk individuals (NfL); and as an accessible test for earlier identification of patients that may have CJD and require more definitive tests (NfL). Further studies should evaluate their performance directly in these specific roles.

Fig. 1. Plasma tau and NfL concentrations (shown on log scales) across symptomatic prion disease and control groups.

Plots show individual data points, with median, upper and lower quartiles overlaid. Alzheimer’s disease (AD), frontotemporal dementia (FTD), inherited prion disease (IPD), sporadic, variant and iatrogenic Creutzfeldt-Jakob disease (sCJD, vCJD and iCJD).

Fig. 2. Receiver Operating Characteristic (ROC) curves.

Plasma tau (purple) and NfL (green), for distinguishing sCJD from (A) Healthy controls, (B) all neurological disease controls (AD, FTD and prion mimics), and (C) prion mimics only. Areas under curves are given in Table 2 in the main paper.

Fig. 3. Relationship of plasma tau and NfL to measures of disease progression in sporadic CJD.

A Relationship of plasma tau and NfL to number of days from death in sporadic CJD, in cross-sectional dataset (earliest available sample from each patient; n = 231—left panels) and longitudinal dataset (all samples from patients with more than one available; n = 114 samples from 47 individuals—right panels). Red continuous, dashed and dotted lines show 50th, 75th and 90th centiles of healthy controls respectively. Blue lines on the longitudinal charts show the linear mixed effects model fits for each individual (see text). Patients with very rapidly progressive disease, typically enrolled to the Cohort at a late stage of disease, are less likely to have more than one blood sample available, so are under-represented in the longitudinal sample set. B Relationship of plasma tau and NfL to measures of disease progression in sporadic CJD. Left panels show biomarkers (log scale) plotted against MRC Scale, which measures severity of functional impairment. Right panels show biomarkers (log scale) plotted against MRC Slope, which provides a measure of the rate of clinical progression. The earliest sample from each patient with necessary data available is included. As MRC Slope cannot be modelled for patients with a baseline MRC Scale < 5, fewer patients are represented in MRC Slope charts, and the missing cases are skewed towards those enrolled at an advanced stage of disease, and those with very rapid clinical progression.

Fig. 4. Plasma tau and NfL in PRNP mutation carriers.

A Plasma tau and NfL (log scales) at different stages in PRNP mutation carriers. NB—Multiple samples from some individuals are included within and between different groups in these charts, and proportions of different PRNP mutations vary between groups. Plots show individual data points, with median, upper and lower quartiles overlaid. B Individual plasma tau and NfL trajectories for PRNP mutation carriers that were asymptomatic at the time of their first sample, and subsequently developed symptoms of IPD during follow-up in the Cohort study. “Slow IPD”: P102L × 5, 5OPRI × 1, 6OPRI × 1. “Fast IPD”: E200K × 1, D178N × 1. Note that serial datapoints from individuals are joined using simple linear interpolation, which does not necessarily represent the true biomarker trajectory between the observed datapoints. Healthy Control values and their major centiles are shown for comparison.