Large-scale parent-child comparison confirms a strong paternal influence on telomere length

Abstract

Telomere length is documented to have a hereditary component, and both paternal and X-linked inheritance have been proposed. We investigated blood cell telomere length in 962 individuals with an age range between 0 and 102 years. Telomere length correlations were analyzed between parent-child pairs in different age groups and between grandparent-grandchild pairs. A highly significant correlation between the father's and the child's telomere length was observed (r=0.454, P<0.001), independent of the sex of the offspring (father-son: r=0.465, P<0.001; father-daughter: r=0.484, P<0.001). For mothers, the correlations were weaker (mother-child: r=0.148, P=0.098; mother-son: r=0.080, P=0.561; mother-daughter: r=0.297, P=0.013). A positive telomere length correlation was also observed for grandparent-grandchild pairs (r=0.272, P=0.013). Our findings indicate that fathers contribute significantly stronger to the telomere length of the offspring compared with mothers (P=0.012), but we cannot exclude a maternal influence on the daughter's telomeres. Interestingly, the father-child correlations diminished with increasing age (P=0.022), suggesting that nonheritable factors have an impact on telomere length dynamics during life.

Figure 1

Hypothetical pedigree illustrating the selection of duos and trios. (I) Trio selected using father, mother (in-law parent) and son (oldest child). (II) Duo not allowed as, in order to avoid confounders, only one parent born within the family tree could be used to select a duo/trio. (III) Duo selected using father (in-law parent) and son (oldest child). (IV) Duo selected using mother (in-law parent) and daughter (oldest child).

Figure 2

Age-associated telomere shortening stratified by gender. (a) Linear regression analysis including the whole cohort of 962 individuals. (b–d) Linear regression analysis restricted to individuals included in the inheritance analysis (n=444). TL shortening with age in (b) all individuals, (c) offspring, (d) parents.

Figure 3

(a–f) Telomere length correlations between parents and offspring. The statistics given are adjusted for age using Pearson's partial correlation.

Figure 4

(a–f) Telomere length correlations between parents in different age groups (at the time of blood sampling) and children. The statistics given are adjusted for age using Pearson's partial correlation.

Figure 5

R² telomere length correlations in different age groups and correlations over generations. (a) R² parent–child correlations in three different age groups, illustrating the percentages by which the variation in offspring TL may be explained by the parental TL. r² was calculated using the correlation coefficient (r) from the partial correlation in Figure 4 and Fisher z-transformation was used to investigate whether the correlation coefficients differed significantly. The P-values for the 2 × 3 possible comparisons (left: father–child; right: mother–child) are indicated on top. (b) Telomere length correlation between grandparents and grandchildren. The statistics given are adjusted for age and sex using Pearson's partial correlation.