Telomeres are elongated in older individuals in a hibernating rodent, the edible dormouse (Glis glis)

Abstract

Telomere shortening is thought to be an important biomarker for life history traits such as lifespan and aging, and can be indicative of genome integrity, survival probability and the risk of cancer development. In humans and other animals, telomeres almost always shorten with age, with more rapid telomere attrition in short-lived species. Here, we show that in the edible dormouse (Glis glis) telomere length significantly increases from an age of 6 to an age of 9 years. While this finding could be due to higher survival of individuals with longer telomeres, we also found, using longitudinal measurements, a positive effect of age on the rate of telomere elongation within older individuals. To our knowledge, no previous study has reported such an effect of age on telomere lengthening. We attribute this exceptional pattern to the peculiar life-history of this species, which skips reproduction in years with low food availability. Further, we show that this "sit tight" strategy in the timing of reproduction is associated with an increasing likelihood for an individual to reproduce as it ages. As reproduction could facilitate telomere attrition, this life-history strategy may have led to the evolution of increased somatic maintenance and telomere elongation with increasing age.

Figure 1. Effect of age on relative telomere length (RTL) in edible dormice (n = 49, N = 158).

The effect of age was best described by a quadratic polynomial (age coefficient −0.14, SE 0.07; P = 0.033; age² coefficient 0.019; SE 0.007; P = 0.005). Component and residuals plot from a linear mixed effects model. The r² of the full model was 0.53; the r² of random effects was 0.40.

Figure 2. Effect of age on the change of relative telomere length (RTL) in edible dormice (n = 49, N = 109).

The effect of age was best described by a quadratic polynomial (age coefficient −0.18, SE 0.08; P = 0.020; age² coefficient 0.022; SE 0.007; P = 0.003). Component and residuals plot from a linear mixed effects model. The r² of the full model was 0.62; the r² of random effect was 0.43.

Figure 3

Differences in RTL changes among younger (panels a,b) and older dormice (panels c,d). There were no significant between-subject effects in either age group (panels a,c). There was, however, significant RTL shortening within younger dormice (b) and significant RTL elongation within older dormice (d) Age groups were split at a mean age of 5.3 years, which minimized the combined sum of squares.

Figure 4. Effect of age on the probability of female dormice to reproduce (n = 1529, N = 816).

Black circles show means for each age. The sigmoidal prediction line was obtained from a mixed effects logistic regression and was averaged over all observation years and individuals. The vertical lines close to the upper and lower x-axes each indicate an observation of reproduction (1) or no reproduction (0), respectively.

Figure 5. Relative telomere length (RTL) of two qPCR runs with identical samples showing the low inter-run variability of the qPCR method used in this study.