Examination of plasma biomarkers of amyloid, tau, neurodegeneration, and neuroinflammation in former elite American football players

Abstract

Introduction: Blood-based biomarkers offer a promising approach for the detection of neuropathologies from repetitive head impacts (RHI). We evaluated plasma biomarkers of amyloid, tau, neurodegeneration, and inflammation in former football players.

Methods: The sample included 180 former football players and 60 asymptomatic, unexposed male participants (aged 45-74). Plasma assays were conducted for beta-amyloid (Aβ) 40, Aβ42, hyper-phosphorylated tau (p-tau) 181+231, total tau (t-tau), neurofilament light (NfL), glial fibrillary acidic protein (GFAP), interleukin-6 (IL-6), Aβ42/p-tau181 and Aβ42/Aβ40 ratios. We evaluated their ability to differentiate the groups and associations with RHI proxies and traumatic encephalopathy syndrome (TES).

Results: P-tau181 and p-tau231(p_adj = 0.016) were higher and Aβ42/p-tau181 was lower(p_adj = 0.004) in football players compared to controls. Discrimination accuracy for p-tau was modest (area under the curve [AUC] = 0.742). Effects were not attributable to AD-related pathology. Younger age of first exposure (AFE) correlated with higher NfL (p_adj = 0.03) and GFAP (p_adj = 0.033). Plasma GFAP was higher in TES-chronic traumatic encephalopathy (TES-CTE) Possible/Probable (p_adj = 0.008).

Discussion: Plasma p-tau181 and p-tau231, GFAP, and NfL may offer some usefulness for the characterization of RHI-related neuropathologies.

Highlights: Former football players had higher plasma p-tau181 and p-tau231 and lower Aβ42/ptau-181 compared to asymptomatic, unexposed men. Younger age of first exposure was associated with increased plasma NfL and GFAP in older but not younger participants. Plasma GFAP was higher in participants with TES-CTE possible/probable compared to TES-CTE no/suggestive.

Keywords: National Football League; chronic traumatic encephalopathy; college football; concussion; football; head trauma; neurodegenerative disease; plasma biomarkers; repetitive head impacts; subconcussion; traumatic brain injury; traumatic encephalopathy syndrome.

FIGURE 1

Estimated marginal means plot—pT181, pT231. ANCOVA compared concentrations of plasma biomarkers in football players and asymptomatic UE participants. Above figure shows the differences in estimated marginal means between football players and asymptomatic UE exposure groups for plasma p‐Tau181, p‐Tau231, and the Aβ42/p‐Tau181 ratio. All models adjusted for age, race, BMI, rFSRP, and total AUDIT score. Error bars at 95% CI. (a) p‐Tau181. (b) p‐Tau231. (c). Aβ42/p‐Tau181. ANCOVA, analysis of covariance; AUDIT, alcohol use disorders identification test; BMI, body mass index; CI, confidence interval; p‐tau, hyper‐phosphorylated tau; rFSRP, revised Framingham Stroke Risk Profile; UE, unexposed

FIGURE 2

Plasma biomarker concentrations. Concentrations of plasma biomarkers after log‐transformation by ln() (if applicable) between football players (RHI) and asymptomatic UE participants (No RHI) exposure groups shown as boxplots. Individual data points are shown. Stars denote statistically significant results. Plasma biomarker concentrations are shown on the y‐axis. pg/dL, picograms per deciliter; pg/mL, picograms per milliliter, RHI, repetitive head impacts; UE, unexposed.

FIGURE 3

Plasma biomarker ROC curve analysis. ROC curve analysis of plasma biomarkers differentiating capability between football exposure groups (former players vs. asymptomatic UE participants). Binomial regression models were performed. Adjusted analysis included age, race, BMI, rFSRP and total AUDIT score. AUDIT, alcohol use disorders identification test; BMI, body mass index; rFSRP, revised Framingham Stroke Risk Profile; ROC, receiver operating characteristic; UE, unexposed.

FIGURE 4

GFAP concentrations and TES‐CTE certainty. Concentrations of plasma GFAP after log‐transformation by ln() between TES‐CTE possible/probable (left) and TES‐No/TES‐CTE Suggestive (right). Individual data points are shown. Plasma GFAP concentrations are shown on the y‐axis. GFAP, glial fibrillary acidic protein; pg/dL, picograms per deciliter; TES‐CTE, traumatic encephalopathy syndrome‐chronic traumatic encephalopathy.

FIGURE 5

Association between plasma NfL, GFAP, and AFE. Results of regression models for starting age of football on both plasma NfL (left, blue; NfL Δ R² = 0.033, B = −0.011, p_adj = 0.03) and GFAP (right, maroon; GFAP Δ R² = 0.022, B = −0.008, p_adj = 0.033). These partial regression plots show the results with covariates taken into account and include a regression line with 95% CI. Covariates include age, race, BMI, rFSRP, total AUDIT score and total years of football play. AFE, age of first exposure; AUDIT, alcohol use disorders identification test; BMI, body mass index; CI, confidence interval; GFAP, glial fibrillary acidic protein; NfL, neurofilament light; rFSRP, revised Framingham Stroke Risk Profile.