Telomere length (TL) is a prognostic factor in Chronic Lymphocytic Leukemia (CLL) with short TL being a predictor of time to first treatment, progression-free survival and overall survival. However, little is known about telomere dynamics through the course of the disease. Most studies conducted on CLL patients have measured TL in unselected peripheral blood mononuclear cell populations, often at a single time point. In this context, longitudinal analysis of TL is problematic as patients who undergo disease progression and/or treatment may have a significantly altered proportion of CLL cells in their peripheral blood compared to T-cells. In order to ensure that we specifically analyzed the TL of the tumor cells, we used fluorescence-activated cell sorting (FACS) to sort populations of CD19+CD5+ CLL B-cells and CD3+T-cells from samples taken from individual patients at different time points throughout their disease (n=30). We then performed high-resolution single telomere length analysis (STELA) on these sorted subsets of cells and analyzed their telomere dynamics over time (median follow up 69 months).

We found a signifcant difference in the CLL B-cell TL (p=0.05) with a mean erosion rate of -52base pairs/year. TL change in the 18/30 patients who remained untreated at all time points was -51bp/year. In the 6/30 patients who received treatment following their initial TL measurements the mean TL change was -40bp/year. Finally, for 11/30 patients samples were only available in the post treatment setting, for these patients the TL change was lower at -29bp/year. The difference in TL erosion between these different groups was not statistically significant. These data shows that the TL erosion in CLL B-cells is modest and similar to that of an age-matched population (Steenstrup et al 2013). Furthermore, CLL B-cells showed no reduction in TL standard deviation signifying the maintenance of intraclonal diversity (p=0.78).

In contrast to the modest changes in TL observed in the CLL B-cells, TL change in the T-cell populations were much more pronounced with mean change of -119bp/year (p=0.02). Furthermore, there was a trend towards increased erosion in the treated patient group when compared with the untreated group (-230bp/year vs -85bp/year, p=0.14) suggesting that therapy may have an impact on the composition of T-cell populations in treated CLL patients. In keeping with this notion, the T-cells derived from CLL patients showed a significant reduction in TL standard deviation (P=0.02), implying that the T-cell repertoire is significantly altered during the course of the disease.

In conclusion, this study of TL in ex vivo CLL B-cell samples shows TL erosion during long-term follow-up that is comparable to that seen in non-leukemic leukocyte and lymphocyte samples (Lansdorp et al 1999). In keeping with a recent study, we showed that the erosion rate correlated with starting TL, with the longest telomeres showing the largest erosion and the shortest telomeres showing elongation (Rosenquist et al 2013). This implies that the radical telomere shortening observed in some CLL patients is an early disease event which is in keeping with our previous data demonstrating that a proportion of stage A patients possess very short dysfunctional telomeres (Lin et al 2010). Given that short TL is associated with an inferior clinical outcome, our data indicates that part of the explanation for the clinical heterogeneity seen in CLL may be telomere dependent whereby if the mutagenic event occurs in a B lymphocyte which already has shorter TL then a more aggressive disease occurs whereas if it occurs in a B lymphocyte with longer TL then the outcome is less aggressive disease.

Finally, T-cells in CLL patients show markedly more TL erosion corroborating previous studies suggesting there is extensive and abnormal T-cell proliferation in CLL. Whether the CLL cells themselves are driving T-cell TL erosion is at present unknown.


No relevant conflicts of interest to declare.

Author notes


Asterisk with author names denotes non-ASH members.