Max-Planck-Partner Group on Stem Cell aging, Key laboratory of Human Diseases Comparative Medicine Ministry of Health Institute of Laboratory Animal Sciences, Chinese Academy of Medical Sciences, Beijing, China
Hans A. Kestler, Research group of Bioinformatics and Systems Biology, Institute of Neural Information Processing, University of Ulm, Albert-Einstein-Allee 11, 89081 Ulm, Germany. Tel.: +49 731 5024248; fax: +49 731 5024156 ; e-mail: firstname.lastname@example.org K. Lenhard Rudolph, Department of Molecular Medicine, Max-Planck Research Group on Stem Cell Aging, University of Ulm, Albert-Einstein-Allee 11, 89081 Ulm, Germany. Tel.: 0049 731 5036 100; fax: 0049 731 5036 102 ; e-mail: Lenhard.Rudolph@uni-ulm.de
Cellular aging is characterized by telomere shortening, which can lead to uncapping of chromosome ends (telomere dysfunction) and activation of DNA damage responses. There is some evidence that DNA damage accumulates during human aging and that lifestyle factors contribute to the accumulation of DNA damage. Recent studies have identified a set of serum markers that are induced by telomere dysfunction and DNA damage, and these markers showed an increased expression in blood during human aging. Here, we investigated the influence of lifestyle factors (such as exercise, smoking, body mass) on the aging-associated expression of serum markers of DNA damage (CRAMP, EF-1α, stathmin, n-acetyl-glucosaminidase and chitinase) in comparison with other described markers of cellular aging (p16INK4a upregulation and telomere shortening) in human peripheral blood. The study shows that lifestyle factors have an age-independent impact on the expression level of biomarkers of DNA damage. Smoking and increased body mass indices were associated with elevated levels of biomarkers of DNA damage independent of the age of the individuals. In contrast, exercise was associated with an age-independent reduction in the expression of biomarkers of DNA damage in human blood. The expression of biomarkers of DNA damage correlated positively with p16INK4a expression and negatively with telomere length in peripheral blood T-lymphocytes. Together, these data provide experimental evidence that both aging and lifestyle impact on the accumulation of DNA damage during human aging.
There is an ongoing effort to identify biomarkers of human aging. Biomarkers of aging could be used in clinical settings to predict disease outcomes and sensitivity of elderly patients to invasive therapies. In addition, there are growing efforts to develop molecular therapies and lifestyle interventions that slow down the aging process. However, the evaluation of the efficacy of such approaches is extremely difficult because the measurement of effects on fitness and lifespan would require long-lasting clinical studies. The development of molecular markers that not only indicate the biological age of an individual but are also sensitive to intervention and lifestyle factors would represent an important step for the future development of anti-aging therapies/interventions.
There is growing evidence that an accumulation of nuclear DNA damage can contribute to the impairments in maintenance and function of organs during aging and chronic disease (for review see Nalapareddy et al., 2008). In support of this hypothesis, it has been shown that mutations in DNA repair and checkpoint genes are the cause of some progeroid syndromes in humans (Lieber & Karanjawala, 2004). A variety of studies have revealed evidence for an increase in DNA damage in aging human tissues. Specifically, γH2AX-positive DNA damage foci increase in aging lymphocytes (Sedelnikova et al., 2008), double-strand breaks increase in sperm of aged humans (Singh et al., 2003; Schmid et al., 2007), single-strand DNA breaks increase in aging lymphocytes (Chicca et al., 1996) and muscle (Zahn et al., 1987), oxidative DNA lesions (7-methylguanine and 8-hydroxydeoxyguanosine) show an age-dependent increase in human urine (Tamae et al., 2009), and an age-dependent increase in cytogenetic abnormalities occurs in human lymphocytes (Ramsey et al., 1995).
Telomere shortening represents a cell intrinsic mechanism that can contribute to the accumulation of DNA damage during aging. Telomeres cap the chromosomal ends to preserve DNA integrity and chromosomal stability. Telomeres shorten 50–100 base pairs during each cell division because of the end-replication problem of the DNA polymerase and because of processing of telomeres during S-phase (Levy et al., 1992; Verdun & Karlseder, 2007). Telomere shortening eventually leads to telomere dysfunction (chromosome uncapping) and the induction of DNA damage checkpoints resulting in a permanent cell cycle arrest and replicative senescence (d’Adda di Fagagna et al., 2003; Takai et al., 2003). p16 and p21 represent two cell cycle inhibitors that are upregulated in senescent cells (Alcorta et al., 1996; Stein et al., 1999). Functional studies on senescent fibroblasts have shown that p21 is a downstream target of p53 and telomere dysfunction, whereas p16 appears to be upregulated in a p53 and telomere-independent manner (Herbig et al., 2004).
Both telomere shortening and the upregulation of p16 have been associated with aging. Telomere shortening occurs in the vast majority of human tissues during aging and it is conceivable that this process contributes to the accumulation of DNA damage in aging human tissues (for review see Jiang et al., 2007). Correlative studies in humans have found that short-for-age telomere length is associated with decreased longevity and an increased development of certain age-associated diseases, particularly cancer, cardiovascular disease and deleterious infection (see for example Cawthon et al., 2003; Farzaneh-Far et al., 2008; Yang et al., 2009). An upregulation of p16 has been observed in various tissues of aging mice as well as in humans (Krishnamurthy et al., 2004; Ressler et al., 2006).
Studies on telomerase knockout mice (Terc−/−) have provided the first experimental evidence that telomere dysfunction can impair the maintenance of organ systems leading to premature aging (Rudolph et al., 1999; Choudhury et al., 2007; Schaetzlein et al., 2007). Mutations of genes that are required for telomerase activity also lead to accelerated telomere shortening, organ failure, and a shortened lifespan in humans, for example in patients with dyskeratosis congenita (for review see Gu et al., 2009; Walne & Dokal, 2009). These findings suggest that human telomeres have a limited length and telomere shortening can represent a causal factor influencing human aging, the evolution of age-associated diseases and lifespan.
Together, there is growing evidence that age- and lifestyle-dependent rates of telomere shortening and DNA damage accumulation could contribute to human aging and disease. However, the measurement of telomere length and DNA damage has not yet been implemented in clinical settings, except from clinical testing of bone marrow failure and leukaemia in young adults. This is partly because of the weak correlation of telomere length with age and the complexity of the assays for telomere length measurement. We have recently identified a set of biomarkers (CRAMP, stathmin, EF-1α and chitinase enzyme activity) that are induced by telomere dysfunction or γ-irradiation (Jiang et al., 2008). Our previous study indicated that the expression of these proteins is induced by DNA damage, although the exact level of DNA damage or the exact number of dysfunctional telomeres per cell that induce the expression of these marker proteins is currently unknown. The functional role of an increased expression of these marker proteins in response to DNA damage or telomere dysfunction also remains to be investigated. However, these proteins could represent useful biomarkers because the expression of these proteins can easily be measured in human blood serum using ELISA and enzyme activity assays. The expression of these marker proteins was found to increase in human blood serum in association with aging and the occurrence of age-associated diseases (Jiang et al., 2008).
Here, we analysed the influence of lifestyle factors on the expression of these biomarkers in blood serum, and we determined its correlation with well-described markers of cellular senescence (upregulation of p16INK4a, telomere shortening) in blood cells of the same individuals. The study shows that lifestyle factors (specifically smoking, body mass index and exercise) have an age-independent impact on the expression of serum markers of telomere dysfunction and DNA damage and the expression of these markers correlates with biomarkers of cellular senescence in peripheral blood T-lymphocytes. These findings suggest that biomarkers of DNA damage and telomere dysfunction could help to evaluate the efficacy of lifestyle interventions and future therapies aiming to slow down the aging process.
Age correlates with expression level of serum markers of DNA damage and telomere dysfunction
In previous studies, we have identified a set of serum markers that showed an increased expression in culture medium of telomere dysfunctional cells and γ-irradiated cells compared to culture medium from young, nonirradiated cells (Jiang et al., 2008). Moreover, the expression of these serum markers showed an increased expression during human aging and age-associated diseases (Jiang et al., 2008). Here, the expression of these biomarkers was measured by ELISA (for CRAMP, EF-1α and stathmin) and enzyme activity assays (for n-acetyl-glucosaminidase, chitobiosidase and chitotriosidase) in serum samples of a previously described cohort of 170 healthy persons aged from 18 to 80 (Liu et al., 2009). For the current study, a set of 136 serum samples was available from the previous described cohort. The best combination score of serum biomarkers was calculated according to a population-based stochastic search method (particle swarm optimization algorithm), which was used to maximize the rank correlation of a linear combination of all marker subsets under the constraint of positive weight (for more details see Data S1, pages 1–3). In addition, the data were correlated with telomere length and the previously analysed expression level of p16INK4a mRNA in peripheral blood T-lymphocytes (PBTLs) of the same cohort (Liu et al., 2009).
In agreement with our previous studies, the expression of the serum markers (except CRAMP) showed a significant age-dependent increase in the current cohort (data not shown). The best calculated combination score correlating with age of individuals in the current cohort was based on the expression of stathmin, n-acetyl-glucosaminidase and chitobiosidase (Fig. 1A and Data S1, page 3). In addition to the previously identified markers, the enzyme activity of chitobiosidase and chitotriosidase (both parameters measure chitinase activity) showed a significant age-dependent increase in chitinase activity in blood serum (Fig. 1B, and data not shown for chitotriosidase). Consistent with previous studies, analysis of peripheral blood T-lymphocytes (PBTLs) revealed that expression of p16INK4a mRNA was positively correlated with age and telomere length was negatively correlated with age (Fig. 1C,D). Together, these data confirmed previous results showing an age-dependent increase in the marker proteins of DNA damage and telomere dysfunction in human blood serum.
Lifestyle factors impact on the expression level of serum markers of DNA damage and telomere dysfunction
To evaluate the role of lifestyle factors on the accumulation of DNA damage markers in human blood, a questionnaire was used evaluating tobacco smoking and physical exercise. In addition, the body mass index of all individuals was determined at the time of blood withdrawal.
We calculated a combination score of biomarkers of DNA damage and telomere dysfunction, which showed the best correlation with smoking (measured in pack years, n = 103 individuals). The best combination score was based on the expression of stathmin and chitobiosidase (Data S1, page 3). The combination score indicated a direct association between the dosage of smoking and the accumulation of DNA damage and telomere dysfunction (Fig. 2A). Of note, regression analysis including age as a confounding factor revealed an age-independent direct association between the dosage of smoking and DNA damage accumulation (Table 1A, P = 0.002). The partial correlation between the dosage of smoking (pack years) and the combination of serum markers after correction for age was 0.30179.
Table 1. Summary statistics for fitted regression models
Comb Pk/Years: n = 103
Comb exercise minutes per month: n = 80
Residual standard error is 10.34 on 100 df.
Model R2 =24%.
Residual standard error is 782.9 on 77 df.
Model R2 =19%.
Residual standard error is 0.7743 on 88 df.
Model R2 =16%.
BMI, body mass indices.
Comb BMI n = 91
The dosage of tobacco smoking also showed a significant direct association with the expression of p16INK4a mRNA and an inverse association with telomere length (Fig. 2B,C). However, this associations were weaker compared to the association between smoking dosage and the combination score of the blood biomarkers of DNA damage and telomere dysfunction (Fig. 2A).
To evaluate an association of physical exercise with the expression of biomarkers of DNA damage and telomere dysfunction, exercise duration (minutes per month) was analysed in correlation with the expression of serum markers (n = 80 individuals). The best biomarker combination score correlating with physical exercise was based on the expression of stathmin, n-acetyl-glucosaminidase, chitotriosidase and CRAMP (Data S1, page 3). This score indicated an inverse association between exercise duration and the accumulation of DNA damage and telomere dysfunction (Fig. 3A). Regression analysis including age as a confounding factor revealed an age-independent inverse association between exercise duration and the accumulation of DNA damage (Table 1B, P = 0.0127). The partial correlation between exercise duration (minutes per month) and the expression of biomarkers of DNA damage after correction for age was −0.277. In addition, physical exercise duration showed an inverse association with the level of p16INK4a mRNA expression (Fig. 3B), while telomere length was not significantly correlated with exercise (Fig. 3C). Together, the data suggested that exercise may exert protective effects impairing the age-dependent accumulation of cells with increased levels of DNA damage, telomere dysfunction or p16INK4a mRNA expression.
To evaluate an association of body weight with the expression of serum markers, the body mass indices (BMI) of the individuals (n = 91) were evaluated at the time point of phlebotomy. The best biomarker combination score correlating with BMI was based on the expression of CRAMP, EF-1α and chitobiosidase (Data S1, page 3). This score showed a significant direct association between the BMI and the accumulation of DNA damage and telomere dysfunction (Fig. 4A). Regression analysis including age as a confounding factor revealed an age-independent direct association of the BMI with the accumulation of DNA damage (Table 1C, P = 0.0063). The partial correlation between BMI and the combination of biomarkers of DNA damage after correction for age was 0.2856. Other parameters (p16INK4a mRNA expression and telomere length) were not significantly correlated with the BMI of the individuals (Fig. 4B,C).
Together, these data show that lifestyle factors (smoking, exercise and body weight) have a significant age-independent effect on the expression level of some biomarkers of telomere dysfunction and DNA damage in human blood.
Blood serum markers of DNA damage and telomere dysfunction correlate with senescence markers in peripheral blood T-lymphocytes
The expression of serum markers of DNA damage and telomere dysfunction was correlated with the mRNA level of p16INK4a as well as with the telomere length in peripheral blood T-lymphocytes (PBTLs) from the same individuals.
Individual serum markers of telomere dysfunction and DNA damage (except CRAMP) showed a significant association with the mRNA level of p16INK4a as well as with the telomere length in PBTLs (data not shown). The best biomarker combination showed a significant inverse association with the telomere length (Fig. 5A and Data S1) and a direct association with the mRNA expression level of p16INK4a (Fig. 5B, and Data S1) of PBTLs.
The current study indicates that lifestyle factors have an age-independent influence on the accumulation of biomarkers of DNA damage and telomere dysfunction in human blood serum. Previous studies have suggested that lifestyle factors can influence on the accumulation of oxidative DNA damage, DNA breaks and DNA adduct formation (Niess et al., 1998; Tamae et al., 2009). Moreover, there are emerging reports that lifestyle interventions can reduce the levels of DNA damage and telomere shortening in vivo (Allgayer et al., 2008; Hofer et al., 2008; O’Callaghan et al., 2009). There is growing evidence that the accumulation of DNA damage and telomere dysfunction can contribute to the impairments in organ maintenance and function during human aging and the progression of chronic diseases (for review see Nalapareddy et al., 2008). However, the development of biomarkers that can easily be measured is currently one of the biggest hurdles for a systematic evaluation of the prospective value of levels of DNA damage and telomere dysfunction in predicting the evolution of pathophysiological conditions and disease stages during human aging and chronic diseases. In this regard, the development of blood biomarkers of telomere dysfunction and DNA damage could provide a useful tool. The predictive and diagnostic value of the blood biomarkers needs to be evaluated in clinical trials and lifestyle intervention studies.
Our previous work has shown that CRAMP, EF-1α, stathmin and chitinase enzyme activity are induced in response to telomere dysfunction and DNA damage and these markers show an age-dependent increase in human blood (Jiang et al., 2008). The current study shows that smoking and elevated body mass correlate with increased expression of the blood biomarkers of DNA damage, whereas exercise is correlated with reduced levels of expression. However, in most of the tested variables, only a subset of the serum biomarkers was induced. A possible explanation indicates that different types of DNA lesion may be induced by different lifestyle factors and may thus lead to a differential induction of different subtypes of markers (Ramsey et al., 1995; Niess et al., 1998; Toyooka & Ibuki, 2009). It is also possible that the blood biomarkers that were used in this study are induced by other cellular stresses that are not directly linked to DNA damage or telomere dysfunction, e.g. protein oxidation or mitochondria damage. However, the current study shows that the expression of serum markers of DNA damage correlated with two prominent markers of cellular senescence in PBTLs (telomere shortening and upregulation of p16INK4a mRNA expression). These data support the concept that the blood markers represent biomarkers for the determination of telomere dysfunction and DNA damage – two-well known inductors of cellular senescence.
To our knowledge, the here described biomarkers represent the first protein markers of DNA damage and telomere dysfunction that can easily be detected in serum and show a strong correlation with markers of cellular senescence. The technical difficulties in the in vivo determination of senescence, telomere dysfunction and DNA damage have hampered its use in clinical trials and lifestyle intervention studies. The current study indicates that such studies could be conducted by the use of novel serum markers of telomere dysfunction and DNA damage. Such studies could boost the performance of clinical and lifestyle intervention trials aiming to reduce the accumulation of telomere dysfunction and DNA damage to prevent age-related declines in cellular and organismal functions.
Blood serum was obtained from healthy volunteers at a central site on the campus of the University of North Carolina Hospitals in Chapel Hill, NC. In addition, each volunteer completed a questionnaire about their health, health behaviours, such as smoking and exercise, and demographics. Expression of p16INK4a was determined as previously described (Liu et al., 2009). Genomic DNA was prepared from buffy coats using Puregene Blood Core Kit (Qiagen, Hilden, Germany) according to manufacturer’s instructions. De-identified serum and DNA samples were sent to Ulm for analysis by investigators blinded to sample identification. All studies on human samples were carried out in accordance with the declaration of Helsinki and approved by the UNC Institutional review board. Extraction of RNA was described previously (Liu et al., 2009). Briefly, RNA was prepared from PBTLs using RNeasy mini kit (Qiagen) according to manufacturer’s directions. One microgram of purified RNA was used for reverse transcription with ImProm-II™ RT system (Promega, Madison, WI, USA) according to manufacturer’s instructions. Expression of p16INK4a mRNA was determined as previously described (Liu et al., 2009). Physical activity was determined by three questions: (1) How often do you exercise per week? (2) How long do you exercise each time? (3) How many hours do you exercise per month?
Enzyme-linked immunosorbent assay
The assay was previously described (Jiang et al., 2008). Stathmin and EF-1α serum levels were determined by ELISA. Briefly, microwell plates (NUNC) were coated with 100 μL serum (four times dilution) and standard protein at 4°C over night. For stathmin, a commercially available standard protein was used (EBP07115, Everest biotech). For EF-1α, serum from young healthy controls was used as a standard. After incubation, the coated plates were washed three times with PBS-Tween and then incubated with the first antibodies for 2 h at room temperature anti-stathmin (Everest Biotech, EB07115, 1:1000, Upper Heyford, UK) and anti-EF1-α (Millipore, 05–235, 1:1000, Billerica, MA, USA). After washing (4× with PBS-Tween), the plates were coated with secondary antibodies (for EF-1α: anti-mouse HRP; for stathmin: anti-goat HRP) and incubated at room temperature for 1 h. After washing the plates four times with PBS-Tween, 100 μL of TMB peroxide-based substrate solution (R&D systems, DY999, Minneapolis, MN, USA) was added to each well, 30 min later 50 μL stop solution (R&D systems, DY994) was added to stop the reaction. The absorbencies were read in the microplate reader in dual-wavelength mode (450–540 nm). Human CRAMP levels were determined with the ELISA kit from Cell Sciences (Cat. #HK321) according to manufacturer’s instructions.
Chitinase activity was measured with the Chitinase assay kit (Sigma, CS1030, St. Louis, MO, USA) using 2 μL of serum per reaction. The chitinase assay kit was used to detect N-acetyl-glucosaminidase, chitobiosidase and chitotriosidase activity. The assay was performed according to manufacturer’s instructions.
Quantitative real-time PCR for telomere length
DNA was prepared from peripheral blood T-lymphocytes (PBTLs) using Puregene Blood Core Kit (Qiagen) according to the manufacturer’s instructions. Quantitative RT-PCR was performed according to previously published method (Cawthon, 2002). The relative telomere length was calculated with slight modifications according to the previously published method (Cawthon, 2002). Briefly, nine human cell lines were used as reference DNA samples. The mean telomere length of these cell lines was determined by Southern blot (see Data S1, page 5). The same reference DNA was analysed by quantitative PCR to detect telomeric sequences (T) and a single-copy gene (S), which wasβ globin gene (HBG). The correlation of the mean telomere length determined by Southern blotting with the T/S ratio was determined. These results were used as y standard for telomere length measurement of the test samples (see Data S1 page 5). Quantitative RT-PCR assay was performed on 96-well PCR plates in a total volume of 25 μL. Genomic DNA (20 ng) was mixed with iTaq SYBR Green supermix (Bio-Rad, Hercules, CA, USA) and 100 nm forward and 900 nm reverse primers for telomeres in a separate well. For HBG as single-copy gene, the reaction final concentration was 300 nm for forward primer and 700 nm for reverse primer. The reaction was run for 40 cycles at 60°C in an RT-PCR cycler system (7300 Real Time PCR System, Applied Biosystems, Foster City, CA, USA). Primer sequences were for telomeric DNA:
for HBG: fw 5′-GCTTCTGACACAACTGTGTTCACTAGC-3′,
rv 5′-CACCAACTTCATCCA CGTTCACC-3′.
Statistical analyses were performed with Statistica (v. 7) and R (version 2.10, http://www.r-project.org). Simple and multiple linear regression analyses were performed using the standard least squares estimator. Best combination scores were calculated from the set of serum markers: n-acetyl-glucosaminide, chitobiosidase, chitotriosidase, stathmin, CRAMP and EF-1α. A population-based stochastic search method (particle swarm optimization) algorithm was used to maximize the rank correlation of a linear combination of all marker subsets under the non-negative weight constraint. See Data S1 for details (page 1–3).
The authors thank Norman E. Sharpless for providing samples and expertise. This work was supported from the European Community’s 7th Framework Programme (FP7/2007-2013) under grant agreement no 202230, acronym ‘GENINCA’, the EU-TELOMARKER consortium and the Deutsche Forschungsgemeinschaft (DFG) (RU745-13-1, RU745-10-1). Y.L, C.T and P.D were funded from NIA grant AG024379. HK was funded by and supported by the German Science Foundation (SFB 518, Project C5), the Stifterverband für die Deutsche Wissenschaft (HAK), the German Federal Ministry of Education and Research (HAK) and the Graduate School of Mathematical Analysis of Evolution, Information and Complexity (HAK). This project was partly funded by the federal ministry of education and research (BMBF) within the framework of the programme of medical genome research (PaCa-Net; project ID PKB-01GS08). The responsibility for the content lies exclusively with the authors.