The polygenic nature of telomere length and the anti-ageing properties of lithium

Abstract

Telomere length is a promising biomarker for age-related disease and a potential anti-ageing drug target. Here, we study the genetic architecture of telomere length and the repositioning potential of lithium as an anti-ageing medication. LD score regression applied to the largest telomere length genome-wide association study to-date, revealed SNP-chip heritability estimates of 7.29%, with polygenic risk scoring capturing 4.4% of the variance in telomere length in an independent cohort (p = 6.17 × 10^-5). Gene-enrichment analysis identified 13 genes associated with telomere length, with the most significant being the leucine rich repeat gene, LRRC34 (p = 3.69 × 10^-18). In the context of lithium, we confirm that chronic use in a sample of 384 bipolar disorder patients is associated with longer telomeres (p = 0.03). As complementary evidence, we studied three orthologs of telomere length regulators in a Caenorhabditis elegans model of lithium-induced extended longevity and found all transcripts to be affected post-treatment (p < 0.05). Lithium may therefore confer its anti-ageing effects by moderating the expression of genes responsible for normal telomere length regulation. This is supported by our bipolar disorder sample, which shows that polygenic risk scores explain a higher proportion of the variance in telomere length amongst chronic lifetime lithium users (variance explained = 8.9%, p = 0.01), compared to non-users (p > 0.05). Consequently, this suggests that lithium may be catalysing the activity of endogenous mechanisms that promote telomere lengthening, whereby its efficacy eventually becomes limited by each individual's inherent telomere maintenance capabilities. Our work indicates a potential use of polygenic risk scoring for the prediction of adult telomere length and consequently lithium's anti-ageing efficacy.

Fig. 1

Genetic correlations with telomere length and individualised risk prediction. a Genetic correlations between single nucleotide polymorphisms predictive of increased telomere length and age-related phenotypes. b Left: Output from PRSice displaying a range of p-value thresholds (P_T) tested, including the optimal P_T as shown in the tallest bar at threshold P_T = 0.013, which explained ~4.4% of the variance (p = 6.174 × 10⁻⁵). Right: A scatterplot showing the positive correlation between polygenic risk scores for telomere length (PRS-TL; adjusted for 3 PCs) and relative telomere length (RTL; adjusted for age, sex and BMI), Pearson (r) = 0.205, p ≤ 0.0001

Fig. 2

Lithium affects telomere length. a Scatterplot showing a positive association between lithium treatment duration and relative telomere length (RTL; adjusted for age, sex, BMI, PCs 1-3 and current lithium use) in chronic lifetime lithium users. b Data summarisation results using Stouffer’s sum of z method. Table includes previous studies assaying the effect of chronic lithium duration on RTL, the direction of effect observed (effect), sample size (n) and p-value (p), as well as a weighted effect combining results from all studies

Fig. 3

Genetic regulators of telomere length and effects of lithium. a Manhattan plot showing results from telomere length gene-enrichment analyses, indicating which genes are most important in affecting telomere length. The dashed line represents the threshold for genome-wide significance. b Three orthologs were assessed at the mRNA level in a C. elegans model of lithium-induced extended longevity, *p ≤ 0.05, **p ≤ 0.01

Fig. 4

Lithium use and the effect of PRS-TL. Scatterplots showing the effects of polygenic risk scores for telomere length (PRS-TL; adjusted for PC’s 1-3), on relative telomere length (RTL; adjusted for age, sex, BMI and current lithium use), in a lithium-naive BD patients, b short-term lifetime lithium users and c chronic lifetime users. Significant correlations are indicated with a red line of best fit