MitoQ counteracts telomere shortening and elongates lifespan of fibroblasts under mild oxidative stress

Oxidative damage is thought to be a major causal factor for replicative senescence and human aging (Harman, 1994). Leakage of superoxide from the mitochondrial respiratory chain is an important source of oxidative stress (Raha & Robinson, 2000). Targeting antioxidants to mitochondria is an efficient way to attenuate oxidative damage in mitochondria due to the production of reactive oxygen species (ROS) in isolated mitochondria (Kelso et al., 2001; Echtay et al., 2002) and in mitochondria within cells (Kelso et al., 2001; Hwang et al., 2001). Therefore, by determining the effect of these antioxidants it should be possible to establish whether oxidative damage has a role in telomere shortening. It has been shown that selective targeting of a potent antioxidant to mitochondria is achieved by attaching the redox active moiety of ubiquinol to the decyltriphenylphosphonium (DPPT) cation, resulting in the mitochondria-specific antioxidant mitoQ [10-(6′-ubiquinonyl) decyltriphenylphosphonium bromide] (Kelso et al., 2001). Ubiquinol acts as an antioxidant by donating a hydrogen atom from one of its hydroxyl groups to a lipid peroxyl radical, which decreases lipid peroxidation within the mitochondrial inner membrane (Ingold et al., 1993). MitoQ reduces oxidative damage and decreases ROS-induced apoptosis in short-term experiments (Kelso et al., 2001; Echtay et al., 2002). As mitoQ is predominantly located within mitochondria in cells due to its accumulation by the mitochondrial membrane potential, its effects in cells are thought to be largely due to the prevention of mitochondrial oxidative damage (Kelso et al., 2001) and there is also evidence that mitoQ decreases the release of ROS from mitochondria (Hwang et al., 2001). The possibility of such effects being due to non-specific interactions with mitochondria within cells can be discounted by the use of control compounds such as DPPT, which are also accumulated within mitochondria driven by the membrane potential but which do not act as antioxidants. Therefore, the blocking of a process by mitoQ but not by DPPT indicates a role for ROS production in the process and is consistent with the increased ROS production being primarily mitochondrial. Telomeres act as 'mitotic clocks' in human fibroblasts because they shorten with each round of replication due to both the inability of DNA polymerases to replicate the very ends of chromosomes (Olovnikow, 1973) and the specific accumulation of stress-induced DNA damage (von Zglinicki, 2002). Eventually, telomere dysfunction triggers replicative senescence (Bodnar et al., 1998). Although intense stress can cause a senescence-like arrest without involvement of telomeres (Chen et al., 2001; Gorbunova et al., 2002), one possibility is that ROS production accelerates replicative senescence via its contribution to telomere shortening under conditions of mild stress. Therefore, we wanted to find out whether mitoQ could prolong the replicative lifespan of human fibroblasts under mild stress conditions, and whether this would correlate with a reduction in the rate of telomere shortening. MitoQ in micromolar concentrations selectively blocks mitochondrial oxidative damage and prevents apoptosis induced by acute treatments with hydrogen peroxide (Kelso et al., 2001). When mitoQ was incubated with MRC-5 fibroblasts, concentrations above 50–100 nm were cytostatic in long-term culture, and even for concentrations of 10–20 nm an adaptation period of at least one week under normoxic conditions was necessary before beneficial effects on growth could be seen (data not shown). Such an adaptation period seems to be a characteristic effect of powerful antioxidants and may reflect the involvement of ROS in a multitude of cellular signal transduction chains (Forman et al., 2002). Neither DPPT nor mitoQ had a significant effect on the intracellular peroxide content under normoxic culture conditions as measured by 2′,7′-dichlorofluorescein fluorescence. However, mitoQ, but not DPPT, abolished nearly half of the rise in peroxides induced by hyperoxic culture in untreated cultures (Fig. 1a). Chronically increased oxidative stress exerted by culture under mild hyperoxia (40% oxygen partial pressure) shortens the replicative lifespan of MRC-5 fibroblasts down to few population doublings (von Zglinicki et al., 1995; von Zglinicki, 2002). This premature aging phenotype is indistinguishable from replicative senescence under standard culture conditions (von Zglinicki et al., 1995; Saretzki et al., 1998; Toussaint et al., 2000). Treatment of MRC-5 cells under these conditions with mitoQ significantly elongated the replicative lifespan by an average of 40% (ranging from 15 to 70%) in four independent experiments whereas the lifespan of DPPT-treated control cells remained unchanged (Fig. 1b). This is in agreement with effects of other potent antioxidants on the replicative lifespan of human cells, e.g. the spin trap α-phenyl-N-t-butyl nitrone (Chen et al., 1995; von Zglinicki et al., 2000), its derivative N-t-butyl hydroxylamine (Atamna et al., 2000) or an oxidation-resistant vitamin C derivative (Asp-2-O-phosphate (Furumoto et al., 1998). MitoQ decreases the cellular peroxide content and prolongs replicative lifespan of MRC-5 cells under hyperoxia. Cells (PD between 20 and 28) were left untreated (no) or were pretreated for one week with 10–20 nm (data pooled) of either mitoQ or DPPT and then subjected to 40% hyperoxia until cessation of proliferation (4–6 weeks). MitoQ or DPPT, respectively, were prepared as stock solutions in 100% alcohol and added to the medium every other day. (a) Cell peroxide content under normoxia (filled bars) and after one week of hyperoxia (open bars) was measured by 2′,7′-dichlorofluorescein using flow cytometry as described (Lorenz et al., 2001). Data are mean ± SEM from 3 to 5 independent experiments. (B) Replicative lifespan under hyperoxia (in population doublings, PD) was measured in nine (no treatment), three (DPPT) or four (MitoQ) independent experiments. Box plots indicate the median, upper and lower quartiles and percentiles. Asterisks indicate a significant difference (P < 0.05) between mitoQ and either DPPT or no treatment. To examine the involvement of telomeres, we measured telomere length at the start point of the treatment and at up to four different time points under hyperoxic culture by non-radioactive Southern blotting as described (Petersen et al., 1998; Serra et al., 2002). MitoQ treatment minimized telomere shortening under hyperoxia (Fig. 2a). Quantitative evaluation of data from four independent experiments (Fig. 2b) indicated that hyperoxia increased the rate of telomere shortening per population doubling (PD), similar to that seen in DPPT-treated and untreated control cells. This is in close agreement with earlier data (von Zglinicki, 2002). However, mitoQ treatment completely prevented this rise in telomere shortening rate due to hyperoxia and instead gave a negligible rate of telomere shortening. Again, this is in agreement with the reported protective effect of PBN (von Zglinicki et al., 2000), Asp-2-O-phosphate (Furumoto et al., 1998) or superoxide dismutase overexpression (Serra et al., 2002) on telomere maintenance. MitoQ counteracts telomere shortening under hyperoxia. Cells were grown under normoxia (NO) and, after one week pretreatment with either DPPT or mitoQ, under 40% hyperoxia (HO) and DNA was sampled at the times and PD as indicated. (A) Telomere Southern blot including λHindIII marker (M). Blots were scanned with a FUJI LAS-100 luminescence imager. Average telomere length was calculated using the AIDA software and is indicated by white lines. (B) Telomere shortening rates per PD (mean ± SEM) under normoxia (filled bar) and hyperoxia (open bars). Cells were untreated (no) or treated with either DPPT or mitoQ. Data were calculated by linear regression of triplicate measurements from three or four independent experiments. Telomere shortening rate in mitoQ-treated cells under hyperoxia is not significantly different from zero, but is significantly (P < 0.05, anova) smaller than in both untreated or DPPT-treated cells. Together, these data indicate that minimizing oxidative stress significantly slows down telomere shortening and prolongs replicative lifespan. Moreover, they suggest that it is accelerated telomere shortening in response to increased mitochondrial ROS production that induces premature senescence-like arrest under conditions of mild stress such as chronic hyperoxia. Intense, acute stress, which leads to an immediate arrest in the vast majority of cells, is of course telomere length-independent (Chen et al., 2001; Gorbunova et al., 2002). Such stress will damage DNA all over the genome and thus trigger arrest before replication and concomitant telomere shortening occurs. On the other hand, low levels of interstitial DNA damage are compatible with cell proliferation, but damage will contribute to telomere shortening because of the low telomere-specific efficiency of single-strand break repair (Petersen et al., 1998). Thus, telomeres appear to act as cellular sentinels for oxidative damage under near-physiological conditions by limiting cell proliferation if and when stress and thus greater mutational risks accumulate (von Zglinicki, 2002). This work was supported by MRC grant G0100140 57030 to T.v.Z.