The telomere lengthening conundrum--artifact or biology?

Abstract

Recent longitudinal studies of age-dependent leukocyte telomere length (LTL) attrition have reported that variable proportions of individuals experience LTL lengthening. Often, LTL lengthening has been taken at face value, and authors have speculated about the biological causation of this finding. Based on empirical data and theoretical considerations, we show that regardless of the method used to measure telomere length (Southern blot or quantitative polymerase chain reaction-based methods), measurement error of telomere length and duration of follow-up explain almost entirely the absence of age-dependent LTL attrition in longitudinal studies. We find that LTL lengthening is far less frequent in studies with long follow-up periods and those that used a high-precision Southern blot method (as compared with quantitative polymerase chain reaction determination, which is associated with larger laboratory error). We conclude that the LTL lengthening observed in longitudinal studies is predominantly, if not entirely, an artifact of measurement error, which is exacerbated by short follow-up periods. We offer specific suggestions for design of longitudinal studies of LTL attrition to diminish this artifact.

Figure 1.

Theoretical curves showing the distribution of measured change in LTL because of measurement error. Red shaded areas denote the percentage of individuals who will be misclassified as LTL gainers. Left panels assume a loss of 300 bp and right panels assume a loss of 200 bp during follow-up periods, which could correspond to 10 years of follow-up with an attrition of either 30 or 20 bp/year, respectively. Top panels assume an SD of the measured difference of 150 bp, and bottom panels assume an SD of the measured difference of 212 bp, which could correspond to an SD of the measurement error of 150 bp for two independent measurements or just a single measurement, respectively. As we consider the measured change in LTL to be positive for LTL gainers, μ (200 or 300 bp) is positioned to the left of 0 on the x-axis.

Figure 2.

Percentage of LTL gainers versus follow-up time based on studies presented in Table 1. The number in each symbol denotes the reference of the study. *A subset of the same study; **this percentage of LTL gainers is an underestimate, as it does not include half the subgroup with ‘unchanged LTL’ (which was not provided by the authors); ***this percentage of LTL gainers is an overestimate, as it includes the entire ‘unchanged LTL’ subgroup.

Figure 3.

Theoretical curves showing the corresponding percentage of LTL gainers for a given follow-up period for different inter-assay CVs. Upper panels are based on two independent LTL measurements at baseline and two independent measurements at follow-up, and lower panels on one measurement at baseline and follow-up. Left panels display the curves for an average rate of LTL attrition of 30 bp/year. Right panels display the curves for an average rate of LTL attrition of 20 bp/year. Keys for trajectories of different CVs are shown in the insets.