Longitudinal decline of leukocyte telomere length in old age and the association with sex and genetic risk

Abstract

Telomeres are DNA-protein structures at the ends of chromosomes. Leukocyte telomere length (LTL) shortening has been associated with advanced age. However, most studies use cross-sectional data, hence, the aim of our study was to model longitudinal trajectories of LTL attrition across 20 years at old age. Assessments of LTL were done by qPCR in SATSA (Swedish Adoption/Twin Study of Aging; N=636 individuals). Cross-sectional and longitudinal associations with age were estimated, the latter using latent growth curve analysis. A genetic risk score (GRS) for LTL was further assessed and included in the models. We confirmed an inverse cross-sectional association of LTL with age (B=-0.0022 T/S-ratio; 95% CI: -0.0035, -0.0009, p-value=0.0008). Longitudinal LTL analyses adjusted for sex (1598 samples; ≤5 measurements) suggested modest average decline until 69 years of age but accelerating decline after 69 years, with significant inter-individual variation. Women had on average ~6% T/S-ratio units longer LTL at baseline, and inclusion of the GRS improved the model where four risk alleles was equivalent to the effect size difference between the sexes. In this cohort of old individuals, baseline LTL varied with age, sex and genetic background. The rate of change of LTL accelerated with age and varied considerably between individuals.

Keywords: aging; genetic risk score; latent growth curve; longitudinal decline; telomere length.

Figure 1

Plots of all leukocyte telomere length (LTL) samples across measurement points for the longitudinal analyses for men (left panel) and for women (right panel). The x-axis represents age at sample testing, and the y-axis represents the plate-adjusted residuals of LTL re-scaled back to T/S-ratio. A loess smoothing line calculated from a weighted regression over age for all samples is shown.

Figure 2

Latent growth curve analysis of telomere length attrition with a group intercept (I) and two slopes (S1) and (S2) as well as sex as covariate (C). Observed data are denoted by Y0 through Y4. MI: mean intercept; MS1: mean slope 1; MS2: mean slope 2; MC: mean covariate. V stands for variance of each component respectively. r denotes the correlation between the intercept and the slopes. B1t through B2t represent the age basis coefficients with change over time t. U0 through U4 represent random components from the telomere length measurements, constrained to be equal for each assessment.

Figure 3

Predicted trajectories for men and women based on parameter estimates from the two-slope model of leukocyte telomere length (LTL) including sex and genetic risk score (GRS) effects. Male sex and addition of risk alleles in the GRS each result in shorter LTL. A decline after the centering age of 69.3 years is apparent for both men and women. Age in years is on the x-axis, and the plate-adjusted residuals of LTL re-scaled back to T/S-ratio is on the y-axis. The dashed line indicates centering age and the left corner panel is a zoomed version.

Figure 4

Individual relative leukocyte telomere length (LTL) change in the longitudinal cohort. The difference in LTL measurement between any two time-points in the same individual is on the x-axis. The frequency is on the y-axis. Telomere elongation is exhibited in samples with delta LTL>0, and telomere attrition is exhibited in samples with delta LTL<0. The sample distribution shows an overall elongation in 46%.