Prediction of telomere length and telomere attrition using a genetic risk score: The multi-ethnic study of atherosclerosis (MESA)

Abstract

Background: Short telomere length (TL) and telomere attrition (TA) have been associated with age-related diseases. Objective: We assessed whether a genetic risk score for short TL (GRS-TL) combining seven TL-associated genetic variants identified in a European-ancestry genome-wide association study (GWAS) was associated with TL and TA over 10 years. Methods: Relative TL (T/S ratio) was measured by the quantitative polymerase chain reaction method for a sample of white, African American, and Hispanic participants, who attended Exam 1 and/or 5 of the Multi-Ethnic Study of Atherosclerosis (MESA). Our final sample included 1,227 participants for the TL analysis and 1,138 for the TA analysis. Participants were 45-84 years at Exam 1. We used a linear mixed effects model and adjusted for age, sex, and population structure. Models were stratified by race/ethnicity. Results: In the TL analysis, higher GRS-TL significantly predicted shorter TL (estimates = -0.18 [S.E. = 0.08], p = 0.02 for white; -0.18 [0.07], p < 0.01 for African American; and -0.13 [0.05], p = 0.02 for Hispanic) in fully adjusted models. In the TA analysis, no association between GRS-TL and TA over 10 years was found. Conclusion: Although GRS-TL was developed in European-ancestry populations, it was significantly associated with TL (but not TA) in all three race/ethnic groups examined.

Keywords: PCR; genetic risk score; race/ethnicity; telomere attrition; telomere length.

FIGURE 1

Plots of GRS-TL and TL (T/S ratio) at Exams 1 and 5 by race/ethnicity (white, African American, and Hispanic) and for the whole sample combined. Telomere length as a function of the GRS-TL in Exam 1 (red circles, n = 1162) and Exam 5 (blue circles, n = 1203). Telomere length is plotted as log T/S ratio. Panel (A): GRS-TL and log-transformed TL for whites. Panel (B): GRS-TL and log-transformed TL for Blacks. Panel (C): GRS-TL and log-transformed TL for Hispanics. Panel (D): GRS-TL and log-transformed TL for the whole sample.

FIGURE 2

Violin Plots of the GRS-TL distributions stratified by race/ethnicity. Violin Plots of the GRS-TL distributions including a box plot with median, and quartiles stratified by race/ethnicity for those at or below the 25th percentile of TL (shorter TL) and those above the 25th percentile of TL (longer TL). Panel (A) GRS by TL at 25 Percentile by Race-Ethnicity at Exam 1; Panel (B) GRS by TL at 25 Percentile by Race-Ethnicity at Exam 5.