Decreased myelin proteins in brain donors exposed to football-related repetitive head impacts

Abstract

American football players and other individuals exposed to repetitive head impacts can exhibit a constellation of later-life cognitive and neuropsychiatric symptoms. While tau-based diseases such as chronic traumatic encephalopathy can underpin certain symptoms, contributions from non-tau pathologies from repetitive head impacts are increasingly recognized. We examined cross-sectional associations between myelin integrity using immunoassays for myelin-associated glycoprotein and proteolipid protein 1 with risk factors and clinical outcomes in brain donors exposed to repetitive head impacts from American football. Immunoassays for myelin-associated glycoprotein and proteolipid protein 1 were conducted on dorsolateral frontal white matter tissue samples of 205 male brain donors. Proxies of exposure to repetitive head impacts included years of exposure and age of first exposure to American football play. Informants completed the Functional Activities Questionnaire, Behavior Rating Inventory of Executive Function-Adult Version (Behavioral Regulation Index), and Barratt Impulsiveness Scale-11. Associations between myelin-associated glycoprotein and proteolipid protein 1 with exposure proxies and clinical scales were tested. Of the 205 male brain donors who played amateur and professional football, the mean age was 67.17 (SD = 16.78), and 75.9% (n = 126) were reported by informants to be functionally impaired prior to death. Myelin-associated glycoprotein and proteolipid protein 1 correlated with the ischaemic injury scale score, a global indicator of cerebrovascular disease (r = -0.23 and -0.20, respectively, Ps < 0.01). Chronic traumatic encephalopathy was the most common neurodegenerative disease (n = 151, 73.7%). Myelin-associated glycoprotein and proteolipid protein 1 were not associated with chronic traumatic encephalopathy status, but lower proteolipid protein 1 was associated with more severe chronic traumatic encephalopathy (P = 0.03). Myelin-associated glycoprotein and proteolipid protein 1 were not associated with other neurodegenerative disease pathologies. More years of football play was associated with lower proteolipid protein 1 [beta = -2.45, 95% confidence interval (CI) [-4.52, -0.38]] and compared with those who played <11 years of football (n = 78), those who played 11 or more years (n = 128) had lower myelin-associated glycoprotein (mean difference = 46.00, 95% CI [5.32, 86.69]) and proteolipid protein 1 (mean difference = 24.72, 95% CI [2.40, 47.05]). Younger age of first exposure corresponded to lower proteolipid protein 1 (beta = 4.35, 95% CI [0.25, 8.45]). Among brain donors who were aged 50 or older (n = 144), lower proteolipid protein 1 (beta = -0.02, 95% CI [-0.047, -0.001]) and myelin-associated glycoprotein (beta = -0.01, 95% CI [-0.03, -0.002]) were associated with higher Functional Activities Questionnaire scores. Lower myelin-associated glycoprotein correlated with higher Barratt Impulsiveness Scale-11 scores (beta = -0.02, 95% CI [-0.04, -0.0003]). Results suggest that decreased myelin may represent a late effect of repetitive head impacts that contributes to the manifestation of cognitive symptoms and impulsivity. Clinical-pathological correlation studies with prospective objective clinical assessments are needed to confirm our findings.

Keywords: cerebrovascular disease; chronic traumatic encephalopathy; myelin; repetitive head impacts; white matter.

Graphical abstract

Figure 1

Association between years of American football play and PLP concentrations. (A) Distribution of years of American football play by PLP concentrations dichotomized into high and low based on the sample median split. The bar is the mean and the error bars are 95% CI. The figure shows unadjusted associations. Analysis of covariance controlling for age, education level, arteriolosclerosis and CTE stage showed the difference to be significant (F = 6.90, mean diff. = 1.79, 95% CI = 0.45–3.13, P = 0.01). (B) Scatter plot of the association between years of American football play and PLP as a continuous variable. The figure shows unadjusted associations. Multivariable linear regression models showed that more years of football play was associated with lower PLP concentrations (unstandardized beta = −2.45, 95% CI = −4.52, −0.38, t = −2.34, P = 0.02), adjusting for age at death, level of education, arteriolosclerosis and CTE stage.

Figure 2

Distribution of PLP and MAG concentrations by FAQ Status among older brain donors. The x-axis for both is the FAQ dichotomized into impaired (score of 9 or higher) and unimpaired (score <9). The bar is the mean and the error bars are 95% CI. The sample is restricted brain donors who were 50 years or older at the time of death. The figure shows unadjusted associations. Analysis of covariance controlling for age, education level, arteriolosclerosis and CTE stage show differences to be statistically significant: (A) F = 4.48, mean diff. = 33.71, P = 0.036 and (B) F = 4.43, mean diff. = 66.58, P = 0.037.