Does oxidative stress shorten telomeres?

Abstract

Oxidative stress shortens telomeres in cell culture, but whether oxidative stress explains variation in telomere shortening in vivo at physiological oxidative stress levels is not well known. We therefore tested for correlations between six oxidative stress markers and telomere attrition in nestling birds (jackdaws Corvus monedula) that show a high rate of telomere attrition in early life. Telomere attrition was measured between ages 5 and 30 days, and was highly variable (average telomere loss: 323 bp, CV = 45%). Oxidative stress markers were measured in blood at age 20 days and included markers of oxidative damage (TBARS, dROMs and GSSG) and markers of antioxidant protection (GSH, redox state, uric acid). Variation in telomere attrition was not significantly related to these oxidative stress markers (|r| ≤ 0.08, n = 87). This finding raises the question whether oxidative stress accelerates telomere attrition in vivo. The accumulation of telomere attrition over time depends both on the number of cell divisions and on the number of base pairs lost per DNA replication and, based on our findings, we suggest that in a growing animal cell proliferation, dynamics may be more important for explaining variation in telomere attrition than oxidative stress.

1. Introduction

Telomeres are terminal DNA–protein complexes that act as ‘protective caps’ of linear chromosomes [1]. Telomeres shorten with age and telomere length has been found to predict remaining lifespan (e.g. [2,3]). Identifying the mechanisms causing telomere attrition is therefore of interest, because it may contribute to the identification of physiological processes underlying variation in health and lifespan. Mechanisms of telomere attrition have been well studied in cell culture, revealing that oxidative stress is a key factor that accelerates telomere attrition in a dose-dependent manner [4,5]. However, cell cultures are not organisms, and oxidative stress levels in vitro are difficult to scale to oxidative stress in whole organisms. This raises the question whether oxidative stress also shortens telomeres in vivo. We are aware of six recent studies of the link between oxidative stress and telomere attrition in vivo, with mixed results, but sample sizes were modest and they included few oxidative stress parameters that were measured after the telomere attrition had already occurred, i.e. not in the period between the baseline and follow-up telomere sample [6–11]. The role of oxidative stress in telomere attrition in vivo is therefore currently unclear.

We measured telomere attrition between ages 5 and 30 days in free-living jackdaw nestlings in which we previously demonstrated a high rate of telomere attrition over this period, making jackdaw nestlings a suitable study system [12]. We collected an additional blood sample at an intermediate age (day 20) to measure six commonly used oxidative stress markers that have been implicated to be important in growth and fitness, to investigate the extent to which oxidative stress and telomere dynamics were related. This panel included markers of oxidative damage (TBARS and dROMs, both markers of lipid peroxidation [13,14]; GSSG, oxidized glutathione [13,15]) and markers of antioxidant capacity (GSH, reduced glutathione [13,15,16]; uric acid, a compound with antioxidant capacity [17]; redox state, the ratio of oxidized over total glutathione [13,15]). We measured oxidative stress in blood, while telomere shortening occurs in the haematopoietic stem cells, but blood and tissue oxidative stress variables are highly correlated across the body [18].

2. Material and methods

We studied a natural population of jackdaws in 2014 in five nest-box colonies in the vicinity of Groningen, The Netherlands (53.1708° N, 6.6064° E). General field procedures were as previously described [19]. We collected ±70 µl blood samples during early (day 5) and late development (day 30, just before fledging; hatching of the oldest chick = day 1) to measure telomere dynamics in the nestling period (n = 87 nestlings from 40 broods). In addition, we collected a larger blood sample at day 20 (±1 ml) for oxidative stress measurements. These larger blood samples were stored at −80°C after centrifugation.

(a). Telomere measurements

Telomere length was measured in erythrocytes with pulsed-field gel electrophoresis [20]. DNA was extracted from erythrocyte nuclei using the CHEF Genomic DNA Plug kit (Bio-Rad, Hercules, CA, USA). DNA was digested overnight using proteinase K (50°C) and then subsequently digested with a combination of HindIII (60 U), HinfI (30 U) and MspI (60 U) for 18 h at 37°C in NEB2 buffer. Digested DNA was separated using pulsed-field gel electrophoresis at 14°C for 24 h (3 V cm^–1, switch times: 0.5–7.0 s). Dried gels were hybridized overnight at 37°C with ³²P-end-labelled oligo (5′-CCCTAA-3′)₄ that binds to the 3′ end-cap telomere overhang. Radioactive signal was quantified with the PerkinElmer cyclone storage phosphor system. Individual telomere length size distributions were quantified with densitometry using grey intensity values in ImageJ v. 1.38x, obtaining the mean telomere length of the sample [20]. The within-individual repeatability of telomere length, estimated using the day 5 and 30 samples with birdID as random effect, was 84.5%. This value underestimates the true repeatability of the measurements because telomeres had shortened in the 25-day period and shortening rate varied among individuals. When correcting day 30 telomere length for the average shortening between days 5 and 30, we estimate the repeatability to be 97.3%, corresponding to the measurement repeatability of previous telomere projects in our laboratory.

(b). Oxidative stress measurements

We measured TBARS, reactive oxygen metabolites (ROMs), protein carbonyl, uric acid and triglyceride concentrations in blood plasma, and reduced and oxidized glutathione concentrations in whole blood. For details of the methods, see the electronic supplementary material S1.

Oxidative stress measurements were done using 96-well plates and between-plate-variation was normalized by subtracting the respective plate mean values from the individual oxidative stress values prior to statistical analyses. All variables were subsequently log-transformed to normalize the distribution, and subsequently transformed to a standard normal distribution to enable a direct comparison of model coefficients between oxidative stress variables. To allow a log transformation, distributions were made numerically positive by adding a fixed number per variable to make the lowest value in that variable greater than zero. We verified that the raw data produced highly similar results.

(c). Statistical analyses

We used mixed-effects models with restricted maximum log-likelihood estimation in R-lme4 [21] and included birth nest as random factor to account for the within-brood dependence of siblings due to shared early environmental and genetic effects. When testing oxidative stress effects on telomere attrition, each model included only the oxidative stress variable as fixed effect in addition to the random effect.

3. Results

Nestling telomere lengths at ages 5 and 30 days were strongly correlated (r = 0.97). Nestlings lost on average 323 bp (s.d. = 146) over this period (figure 1), but telomere attrition varied strongly between nestlings (CV = 45%). We tested the associations between telomere attrition and our panel of oxidative stress variables, but none of the oxidative stress variables were associated with telomere attrition (figure 2). The strongest correlation was close to zero and in the opposite direction from what we expected (dROMs, r = −0.08, p = 0.49). We further tested whether the inclusion of sex, brood size and nestling mass on day 20 and their interactions affected the relationship between oxidative stress and telomere attrition, but this did not change the results. Given our sample size, we would have been able to detect an association of R² > 0.078 with power 80%, suggesting that an existing relationship that we did not detect is likely to be weaker than this limit.

Figure 1.

Telomere length at age 30 days plotted against telomere length at age 5 days in jackdaw nestlings (r = 0.97, n = 87). The dashed line represents equal values of day 5 and 30 telomere lengths and hence the distance below this line reflects the telomere shortening.

Figure 2.

Association between oxidative stress variables and telomere shortening (n = 87). The oxidative stress variables were transformed to a standard normal distribution and units on the x-axis therefore represent standard deviations. d-ROMS and uric acid were corrected for plasma triglycerides and handling time, respectively (see the electronic supplementary material SI for details). The Pearson correlation coefficients (r) are based on the data as shown, but p-values are from the mixed models. (Online version in colour.)

We used principal components analysis (singular value decomposition) to assess whether covariation among the six oxidative stress variables could predict telomere attrition. The first two principal components predicted a substantial amount of variation (51 and 20%), but neither predicted telomere attrition (p > 0.52).

4. Discussion

We tested the hypothesis that oxidative stress accelerates telomere attrition in vivo, but we detected no significant associations between telomere attrition and a panel of six oxidative stress variables despite (i) large variation in telomere attrition, (ii) a sample size that allowed us to identify modest effect sizes with reasonable power, and (iii) using a panel of six oxidative stress markers. The most parsimonious view emerging from our data is therefore that oxidative stress has little effect on telomere attrition in vivo, at least in growing nestlings. At the same time, it is a truism that the absence of a process cannot be proven, and there are of course limitations to our study that potentially obscured oxidative stress effects on telomere attrition. For example, oxidative stress is challenging to assess reliably [22] and we cannot rule out that other oxidative stress markers or tissues would have yielded a different result. Furthermore, oxidative stress levels could have fluctuated over time and it remains unclear whether a snapshot at day 20 is representative of the total oxidative stress in the period over which telomere attrition was determined. Finally, we did not measure telomerase activity and hence it would be informative to verify the extent to which oxidative stress and telomerase activity are correlated.

The difficulty in quantifying or manipulating oxidative stress hampers convincing refutation of any oxidative stress–related hypothesis, which in turn makes it more difficult to publish negative results [23]. For example, among the four published studies relating telomere attrition to oxidative stress, the non-confirmative results were not presented as a main research finding [8–11], which contrasted with papers with confirmative results [6,7], which also contained non-confirmative results that could be given equal weight. This difference creates a bias in the perception of the role of oxidative stress in telomere attrition. Clearly, more studies on the link between oxidative stress and telomere attrition in vivo are needed and we should in particular welcome the unbiased publication of confirmative and non-confirmative findings.

What explains variation in telomere attrition when it is not oxidative stress? Cumulative telomere attrition is the product of the number of cell divisions and the number of base pairs that is lost per cell division. While cell culture studies express telomere attrition per cell population doubling time, and hence control for the rate of cell proliferation, in vivo studies investigate the cumulative telomere attrition. We hypothesize that the variation in telomere attrition we observed was largely caused by variation in cell proliferation rate, with only a modulating effect of, for example, oxidative stress. The among-individual variation in the rate of cell division could be of particular importance when studying telomere attrition during development, when cells proliferate most. To determine the relative importance of variation in cell proliferation rate versus oxidative stress as sources of telomere attrition in vivo, studies are required that measure both quantities simultaneously. Few studies have reported a positive relationship between nestling growth and telomere attrition (e.g. [12,24], but see [8]) and it would be informative to investigate the extent to which this pattern reflects a relationship between cell proliferation rate and telomere attrition.

Ethics

This work was approved by the University of Groningen animal ethics committee under licence number ‘6832 A’ before conducting the research.

Data accessibility

The dataset is included as the electronic supplementary material (Boonekamp_data_ESM).

Authors' contributions

J.J.B. and S.V. designed the study, conducted the fieldwork, analysed the data and wrote the first draft of the manuscript. Measurements and analyses were done by C.B. (telomere length) and E.M. together with J.J.B. (oxidative stress). All authors contributed to editing the manuscript and gave final approval for publication.

Competing interests

We declare we have no competing interests.

Funding

Cor Dijkstra contributed to the fieldwork. J.J.B. was supported by an NWO grant no. 823.01.009 awarded to S.V. C.B. was supported by a DFG fellowship BA 5422/1-1.