Influence of lifestyle modifications on age-related free radical injury to brain

Abstract

Importance

The Healthy Brain Initiative seeks to optimize brain health as we age. Free radical injury is an important effector of molecular and cellular stress in aging brain that derives from multiple sources.

Objective

Identify potentially modifiable risk factors associated with increased markers of brain oxidative stress.

Design, Setting, and Participants

Our study consisted of 320 research volunteers (178 women) aged 21 to 100 years old who were medically healthy and cognitively normal.

Measures

Free radical injury to brain was assessed using cerebrospinal fluid (CSF) F₂-isoprostanes (IsoPs) and correlated with age, gender, race, cigarette smoking, body mass index (BMI), inheritance of the ε4 allele of apolipoprotein E gene (APOE), and cerebrospinal fluid biomarkers of Alzheimer’s disease (AD).

Results

CSF F₂-IsoP concentration increased with age by approximately 10% from age 45 to 71 years in medically healthy cognitively normal adults. CSF F₂-IsoP concentration increased by an average of >10% for every 5 kg/m² increase in BMI. Current smokers had approximately three-fold greater effect than age on CSF F₂-IsoPs. Women had greater average CSF F₂-IsoP concentration than men at all ages after adjusting for other factors. Neither CSF AD biomarkers nor inheritance of APOE ε4 allele were associated with CSF F₂-IsoP concentration in this group of medically healthy cognitively normal adults. Association between CSF F₂-IsoP concentrations and race was not significant after controlling for effect of current smoking status.

Conclusions & Relevance

Our results are consistent with an age-related increase in free radical injury in human brain, and uniquely suggest that this form of injury may be greater in women than in men. Our results also highlighted two lifestyle modifications that would have even greater impact on suppressing free radical injury to brain than would suppressing processes of aging. These results inform efforts to achieve success in the Healthy Brain Initiative.

INTRODUCTION

The Healthy Brain Initiative 2013–2018¹ formulated by the Centers for Disease Control and Prevention and the Alzheimer’s Association is a challenge to researchers, health care providers, and public health officials to devise and deliver approaches that optimize brain health as we age. Free radical injury, an important effector of molecular and cellular stress in the central nervous system (CNS) as we age, derives from multiple sources including processes of aging, genetics, environmental factors, and latent Alzheimer’s disease (AD). Clarification of the sources of age-related free radical injury to CNS is important because those that are approachable through lifestyle modification or treatment offer an opportunity to optimize brain heath across the human life span. This study focuses on several potential drivers of age-related free radical injury in CNS in 320 research volunteers aged 21 to 100 years who were normal by medical examination and extensive cognitive testing by determining associations of genetic, environmental, demographic, and cerebrospinal fluid (CSF) AD biomarker data with CSF F₂- isoprostanes (IsoPs), a widely-used biomarker of free radical injury to brain.^2–9

METHODS

The institutional review boards of the participating institutions approved all procedures and all subjects provided written informed consent prior to study enrollment. 320 cognitively normal subjects were recruited from AD Centers from 2002–2009; were healthy by medical examination; had Mini-Mental State Examination scores ≥ 26, Clinical Dementia Rating (CDR) score = 0; and had no evidence or history of cognitive or functional decline. The neuropsychologic tests included in this study examined multiple aspects of cognition.¹⁰ The Wechsler Memory Scale-Revised Logical Memory Immediate and Delayed paragraph recall tasks measure verbal episodic memory.^11,12 The Trail Making A task measures psychomotor speed, visuospatial function, and visual attention; Trail Making B adds a set-shifting element and captures executive function.¹³ Category fluency (animals) is a measure of semantic memory and language.¹⁴ CSF was collected in the morning after an overnight fast, frozen immediately, and stored at −80°C until quantified for F₂-IsoPs, Aβ₄₂, total Tau, and Tau-P181 as described previously.^15,16 All CSF protein, glucose, and cell values were within normal ranges (not shown).³ Apolipoprotien E genotype (APOE) was determined by a restriction digest method.¹⁶

Linear regression was used to assess associations between CSF F₂-IsoP concentration and age, gender, race (white versus non-white), current smoking status (yes versus no), BMI (per 5 kg/m² increase), presence/absence of APOE ε4 allele, and CSF AD biomarker concentrations (Aβ₄₂, total tau and tau-P181). Total tau and tau-P181 were strongly correlated (r=0.77, p<0.0001) and so to avoid colinearity, only tau was included in modeling. Age was modeled as a 3 degree restricted cubic spline summarized as average difference in CSF F₂-IsoP concentrations in those aged 71 (75^th percentile for age) compared to those aged 45 (25^th percentile). Each study characteristic was modeled first as an independent variable with adjustment for age (Model 1). CSF AD biomarkers were added to age in Model 2. All other variables, including, gender, race, BMI, current smoking status, and APOE ε4 allele were then added to Model 2 to determine which had an independent association with CSF F₂-IsoPs in the presence of all others (Model 3). Finally, an age by each potential effect modifier (gender, BMI, smoking status, or APOE ε4) interaction term was added to Model 3. All models were summarized using adjusted R². Analyses were carried out using R version 2.15.2¹⁷ using the rms package.¹⁸

RESULTS

Female (n=172) and male (n=148) subjects were matched well (Table 1). As expected, there was a positive association between CSF F₂-IsoP concentration and age across the adult human life span (Table 2) that was independent of CSF AD biomarkers and other variables (Model 3). This association was approximately linear, although with some suggestion of accelerated increase in advanced age after adjustment for covariates in Model 3 (nonlinear trend p=0.05). Results from all three models estimated that CSF F₂-IsoP concentration increased approximately 3 pg/ml, or about 10%, from age 45 to 71 years.

Table 1.

Characteristics of 320 cognitively normal, healthy volunteers stratified by gender

	Women (n=172)	Men (n=148)	Both (n=320)
Age (yr)	57.8 [1.3], 21–100	56.4 [1.6], 22–88	57.2 [1.0], 21–100
Education (yr)	15.8 [0.2], 10–22	16.5 [0.2], 10–27	16.1 [0.1], 10–27
Caucasian (%)	152 (88%)	130 (88%)	282 (88%)
APOE ε4 + (%)	64 (37%)	46 (31%)	110 (34%)
BMI (kg/m2)^	25.8 [0.3], 18–41	26.4 [0.3], 20–37	26.1 [0.2], 18–41
Current smoker (%)	11 (6%)	10 (7%)	21 (7%)
CSF F2-IsoPs (pg/ml)	30.6 [0.7], 13–60	28.8 [0.7], 11–65	29.8 [0.5], 11–65
CSF Aβ42 (pg/ml)*	334 [10], 60–829	322 [10], 91–786	328 [7], 60–829
CSF Tau (pg/ml)*	48.5 [1.0], 5–85	50.0 [1.2], 20–125	49.2 [0.8], 5–125
CSF Tau-P181 (pg/ml)*	31.8 [0.6], 2–63	32.1 [1.0], 10–98	31.9 [0.6], 2–98

Data are mean with [SE] and range, or number with (%).

^{^}using CDC standard weight status categories²⁷ one (female) volunteer was underweight, 132 (80 female) volunteers were normal weight, 129 (59 female) volunteers were overweight, and 53 (32 female) volunteers were obese; BMI data was missing for one (male) volunteer.

^*adjusted for ELISA batch.

Table 2.

Associations between CSF F₂-IsoP concentrations (pg/ml) and age, CSF AD biomarkers, and other genetic, environmental, and demographic variables in 320 cognitively normal subjects.

Variables	Model 1	Model 2 R2 = 0.06	Model 3 R2 = 0.20
	β ± SE, p-value	β ± SE, p value	β ± SE, p value
Age (change in pg/ml CSF F2-IsoPs from age 45 to age 71)	3.0 ± 0.8, p=0.0009 R2 = 0.04	3.1 ± 0.8, p=0.0007	3.3 ± 0.8, p=0.0001
CSF Aβ42 (per increase equivalent to the interquartile range, 240–386)	1.7 ± 0.6, p=0.004, R2 = 0.06	1.3 ± 0.6, p=0.05	1.1 ± 0.6, p=0.07
CSF total tau (per increase equivalent to the interquartile range, 40–57)	1.5 ± 0.6, p=0.01 R2 = 0.05	0.9 ± 0.7, p=.19	0.8 ± 0.6, p=0.24
Gender (female)	1.8 ± 1.0, p=0.07 R2 = 0.04		2.3 ± 0.9, p=0.02
Race (non-Caucasian)	3.4 ± 1.6, p=0.03 R2 = 0.05		1.1 ± 1.5, p=0.45
Current Cigarette Smoker (yes)	9.4 ± 2.0, p<0.0001 R2 = 0.10		8.3 ± 1.9, p<0.0001
BMI (per 5 kg/m2 increase)	3.3 ± 0.6 p<0.0001 R2 = 0.11		3.9 ± 0.7, p<0.0001
APOE ε4 allele (one or two)	0.8 ± 1.1, p=0.46 R2 = 0.04		1.1 ± 1.0, p=0.29

Model 1 includes each variable separately and is adjusted only for age.

Model 2 includes both CSF AD biomarkers in addition to age.

Model 3 includes model 2 and gender, BMI, smoking status, and APOE ε4.

BMI and smoking were the strongest independent correlates of CSF F₂-IsoP concentration (p< 0.0001), which increased by an average of >10% for every 5 kg/m² increase in BMI (Model 3; Figure). Current cigarette smokers had significantly higher CSF F₂-IsoP concentration than nonsmokers (Model 3; Figure) that was approximately three-fold greater than the effect of age. Women had greater average CSF F₂-IsoP concentration than men after adjusting for other factors (Model 3; Figure), although the estimated difference between men and women was smaller than that observed for advancing age. None of these three variables was a significant effect modifier of the association between age and CSF F₂-IsoP concentration (p=0.91, 0.29, or 0.11 for gender, BMI or current cigarette smoking). The difference in CSF F₂-IsoP concentration by smoking status tended to be greatest in midlife; however, the number of current smokers in our study sample may be too few for detection of a significant effect modification.

Figure. CSF F₂-IsoP concentrations versus age.

Predicted mean (lines) and 95% confidence intervals (shaded regions) for CSF F₂-IsoP concentrations versus age in cognitively normal subjects across adult ages calculated from Model 3 and stratified by gender (A), BMI (B) and smoking status (C).

CSF AD biomarkers neither were associated with CSF F₂-IsoP concentration after adjustment for age (Model 2) nor after adjustment for other co-variates (Model 3). Similarly, there was no effect on CSF F₂-IsoP concentrations when stratified into those with versus without an APOE ε4 allele (p=0.29, Model 3). The association between CSF F₂-IsoPs concentration and age did not differ by APOE ε4 status (age by ε4 allele interaction p=0.64).

Association between CSF F₂-IsoP concentrations and race was complex and confounded by greater prevalence of cigarette smoking among non-Caucasians (24% for non-Caucasian versus 4% for Caucasians). Multivariable analysis (Model 3) did not detect a significant association between race and CSF F₂-IsoP concentrations after controlling for effect of current smoking status.

DISCUSSION

Our results from a large group of adult research volunteers who were cognitively normal and in good health are consistent with an age-related increase in free radical injury in human brain, and suggest that this form of CNS injury may be slightly but consistently greater in women than in men. This difference is consistent with other reports of a sex difference in brain reserve and clinical manifestations of AD pathologic changes,^19–21 although this association remains a point of discussion. Importantly our results also showed that both smoking and increased BMI, already know to be associated with increased free radical injury in peripheral organs,^22,23 were also associated with increased CSF F₂-isoprostanes. Thus avoidance of smoking and reducing BMI, in addition to their already established benefits, may also be beneficial in reducing molecular and cellular stress to our brains. Within the limitations of our study, our data do not implicate latent AD, inheritance of APOE ε4, or race as major contributors to this form of brain injury in healthy adults. Indeed, other factors likely are involved, as seen in our Model 3 results, where most of the F₂-IsoP variability is not explained by the defined parameters. For example, obesity is associated with a variety of other abnormalities, including diabetes and dyslipidemia, which could contribute to or modify the observed effect. Finally, it is worth noting that while CSF F₂-IsoPs are widely-used biomarkers of free radical injury,^2–9 they also activate the thromboxane A2 receptor and thereby contribute to abnormal vasoconstriction in multiple organs, including cerebrum.^24–26 Together, these results contribute to the evidence base necessary for achieving the long-term goal of the Healthy Brain Initiative: “to maintain or improve the cognitive performance of all adults.”