Telomere length as a marker of cellular ageing isassociated with prevalence and progression ofmetabolic syndrome

Dóra Révész, MSc, Yuri Milaneschi, PhD, Josine E. Verhoeven, MSc,Brenda W.J.H. Penninx, PhD

Department of Psychiatry, EMGO Institute for Health and Care Institute, VU University Medical Center,Amsterdam, the Netherlands

Context. Metabolic Syndrome (MetS) clusters risk factors for age-related conditions includingcardiovascular disease and diabetes. Shorter telomere length (TL), a cellular marker for biologicalage, may predict an individual’s deteriorating metabolic condition.

Objective. We examined whether shorter baseline TL is associated with a worse metabolic profileand with less favorable trajectories of MetS components over a six-year follow-up.

Design and setting. Participants were part of the Netherlands Study of Depression and Anxiety, anongoing prospective cohort study with six years follow-up.

Participants. This study included 2848 participants aged 18–65 years.

Main outcome measures. Baseline TL from leukocytes was determined using quantitative poly-merase chain reaction, and MetS components (waist circumference, triglycerides, high-densitylipoprotein (HDL) cholesterol, systolic blood pressure (SBP) and fasting glucose) were determinedat baseline, and after two and six years. Cross-sectional and longitudinal analyses were adjustedfor relevant sociodemographic, lifestyle and health factors.

Results. Shorter baseline TL was cross-sectionally associated with HDL (��-0.016, SE�0.008, p�.05),waist circumference (��0.647, SE�0.238, p�.007), triglycerides (��0.038, SE�0.009, p�.001), andfasting glucose (��0.011, SE�0.003, p�.001), as well as with the total number of MetS components(��0.075, SE�0.023, p�.001) and the presence of MetS (OR�1.19, 95% CI�1.07–1.33; p�.002).Although baseline differences progressively reduced over time, shorter baseline TL was still sig-nificantly associated with unfavourable scores of most MetS components at the two- or six-yearfollow-ups.

Conclusions. Cellular aging, as assessed by TL, is associated with a higher metabolic risk profilewhich maintains to be unfavorable even after a period of six years. These findings suggest thatcellular aging might play a role in the onset of various aging-related somatic diseases via its effecton metabolic alterations.

Metabolic Syndrome (MetS) is a constellation of in-terrelated factors (namely abdominal obesity, dys-

lipidemia including low HDL cholesterol and high trig-lycerides, hypertension and hyperglycemia), known to bemajor risk factors for the development of aging-relateddiseases, such as cardiovascular diseases (CVD) and dia-

betes (1). Telomere length (TL), as generally assessed inleukocytes, is a novel marker of cellular aging [2:3], andhas been associated with increased risks of morbidity andmortality (4–6). Telomeres are DNA-protein complexesthat cap chromosomal ends and promote chromosomalstability (2). Normal telomere maintenance requires the

ISSN Print 0021-972X ISSN Online 1945-7197Printed in U.S.A.Copyright © 2014 by the Endocrine SocietyReceived March 25, 2014. Accepted August 27, 2014.

Abbreviations:

O R I G I N A L A R T I C L E

E n d o c r i n e R e s e a r c h

doi: 10.1210/jc.2014-1851 J Clin Endocrinol Metab jcem.endojournals.org 1

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 08 September 2014. at 05:28 For personal use only. No other uses without permission. . All rights reserved.

cellular enzyme telomerase that adds telomeric DNA, thuspreserving TL and healthy cell function (2). During eachcell division, DNA loses telomeric repeats with an esti-mated shortening rate of 25 base pairs per year, eventuallycausing replicative cell senescence (7). Further, acceleratedshortening might be the consequence of increased expo-sure to oxidation and inflammatory mediators (8). MeanTL is therefore often used as cellular marker for biologicalage, with shorter telomeres indicating increased biologicalage. However, it is still largely unknown how telomeremaintenance might influence disease processes.

Several studies have shown significant associations be-tween shorter TL and dysregulated MetS components (9–20), although some studies did not confirm this [5:21:22].Previous research has often been performed with smallsamples (N � 150) [5:9:10:16:19:23:24], or were re-stricted to males or females [5:11:15:16:20], a somaticdiseased population [17:21] or a small age range [13:17:21]. Above all, several studies did not adjust for importantconfounding factors, such as lifestyle and clinical factors,and have investigated only cross-sectional associations.To date only two studies showed a prospective associationbetween TL and obesity-related measures and insulin, butnot with the other MetS components [23:25]. Further-more, three studies also looked at TL and investigated thelink with anthropometric properties measured earlier inlife, and reported an inverse relation between TL at thetime of measurement and weight gain throughout the re-ported period [20:26:27].

Whether TL predicts a wider range of metabolic alter-ations such as in the lipid, glucose or hypertension spectraover an extensive period of time is unknown. The presentstudy examined the cross-sectional and longitudinal rela-tionship between TL and MetS components in a large-scale cohort study, while adjusting for important time-varying confounders. We hypothesized that shorter TL atbaseline is cross-sectionally associated with a worse met-abolic profile and with less favorable trajectories of MetScomponents over a six-year follow-up.

Materials and Methods

Study sampleParticipants were part of the Netherlands Study of Depression

and Anxiety (NESDA), a large on-going longitudinal cohortstudy among 2981 adults (18–65 years); a detailed descriptionof the study rationale, design and method is given elsewhere (28).Briefly, respondents were recruited between September 2004and February 2007 from community, primary care, and special-ized mental health care. Baseline data collection consisted of amedical examination, a blood draw, self-report questionnairesand a detailed interview. Participants were evaluated again attwo-year (2006–2009) and six-year (2010–2013) follow-up vis-

its. Of the entire cohort, 2842 subjects (95%) had complete dataon TL and baseline MetS indicators, and were included in ourcross-sectional and longitudinal analyses. Participants withavailable baseline data did not differ in sociodemographic char-acteristics with those not included in the current analyses. Of theselected participants, 2098 (74%) had complete data at the two-year follow-up, and 1802 (63%) at the six-year follow-up. Theresearch protocol was approved by the ethical committee of par-ticipating universities and all respondents provided written in-formed consent.

Predictor

Telomere lengthTelomere length (TL) was determined at baseline. Fasting

blood was drawn from participants in the morning between 8:30and 9:30 am. Cells were lysed and DNA samples were stored ina –20°C freezer afterwards. Subsequently, early 2012, TL wasdetermined at the laboratory of Telomere Diagnostics, Inc.(TDx, Menlo Park, CA, USA), using quantitative polymerasechain reaction (PCR) (qPCR) as described elsewhere (29). AllqPCRs were carried out on a Roche Lightcycler 480 realtimePCR machine with 384-tube capacity (Roche Diagnostics Cor-poration, Indianapolis, IN). Telomere sequence copy number ineach patient’s sample (T) was compared to a single-copy genecopy number (S), relative to a reference sample. The resulting T/Sratio is proportional to mean TL [29:30]. To control for inter-assay variability, eight control DNA samples were included ineach run. In each batch, the T/S ratio of each control DNA wasdivided by the average T/S for the same DNA from 10 runs toobtain a normalizing factor. This was done for all eight controlsamples and the average normalizing factor for these sampleswas used to correct the participant DNA samples to obtain thefinal T/S ratio. The T/S ratio for each sample was measured twice.If the duplicate T/S value and the initial value varied by more than7%, the sample was run a third time and the average of the twoclosest values was reported. The reliability of the assay was ad-equate: the included quality control (QC) DNA samples on eachPCR run illustrated a small intra-assay coefficient of variation(CV � 5.1%), and the interassay CV was also sufficiently low(CV � 4.6%), as well as for the telomere (CV � 2.04%) and thesingle-gene assays (CV � 1.58%) separately.

To compare T/S ratios to telomere restriction fragments(TRF) reported by other studies using Southern blot analysis, weused the conversion formula to convert T/S ratios to kilo basepairs: kilo base pairs � 3.274 � 2.413 x ((T/S-0.0545)/1.16), aspreviously described (31). A continuous measure of telomericbase pairs (per SD decrease; SD � 0.61 kilo base pairs) was usedin analyses; TL was also divided into quartiles in specific analysesexplore linearity in trends (TL lowest quartile: 3.85–5.04kbp,second quartile: 5.04–5.35kbp; third quartile: 5.35–5.78kbp;highest quartile: 5.78–7.39kbp).

Outcomes

Metabolic syndrome componentsMetabolic indicators were measured at baseline (Y0) and at

two- (Y2) and six-year (Y6) follow-up. Waist circumference wasmeasured with a measuring tape at the central point between thelowest front rib and the highest front point of the pelvis, overlight clothing. HDL cholesterol, triglycerides, and glucose levelswere determined from the fasting blood samples using routine

2 Telomere length and metabolic syndrome J Clin Endocrinol Metab

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 08 September 2014. at 05:28 For personal use only. No other uses without permission. . All rights reserved.

standardized laboratorial methods. The continuous measureswere adjusted for medication use based on the estimated effectsof the medication. According to the standards of medical care indiabetes, the goal of antidiabetic medication should be to lowerthe fasting glucose level to � 7.0 mmol/L (32). In agreement withthese standards, for persons using antidiabetic medication (Y0N � 93; Y2 N � 88; Y6 N � 85) when glucose level was � 7.0mmol/L, a value of 7.0 mmol/L was assigned. According to theaverage decline in triglycerides and increases in HDL cholesterolin fibrate trials (33), 0.10 mmol/L was subtracted from the HDLcholesterol level and 0.67 mmol/L was added to the triglyceridelevel of persons using fibrates (Y0 N � 6; Y2 N � 5; Y6 N � 5).Since none of the participants was using nicotinic acid, no ad-justments were needed for this. Systolic and diastolic blood pres-sure (BP) (SBP, DBP) were measured twice during supine rest onthe right arm with the Omron M4-I, HEM 752A (Omron,Healthcare Europe BV, Hoofddorp, The Netherlands) and wereaveraged over the two measurements. For persons using antihy-pertensive medication (Y0 N � 419; Y2 N � 383; Y6 N � 384),10 mm Hg was added to the SBP and 5 mm Hg to the DBPaccording to the average decline in BP in antihypertensive trials(34).

At baseline the presence of metabolic syndrome (MetS) wasdetermined based on The US National Cholesterol EducationProgram–Adult Treatment Panel III guidelines, requiring thepresence of three or more of the following criteria: 1) waist cir-cumference � 102 cm in men and � 88 cm in women; 2) trig-lycerides � 1.7 mmol/L or medication for hypertriglyceridemia;

3) high-density lipoprotein (HDL) cholesterol � 1.03 mmol/L inmen and � 1.30 mmol/L in women or medication for reducedHDL cholesterol; 4) BP: systolic � 130 and/or diastolic � 85 mmHg or antihypertensive medication; 5) fasting plasma glucose �

5.6 mmol/L or antidiabetic medication (1). We also used thenumber of MetS abnormalities as a summarizing variable thatresembles the severity of MetS in subjects (35).

CovariatesCovariates were assessed at baseline and two- and six-year

follow-up. Sociodemographic factors included sex, age, andyears of attained education. Lifestyle variables included alcoholconsumption (no drinker, mild-moderate drinker 1–14 (women)/ 1–21 drinks per week (men), heavy drinker � 14 (women) / �21(men) drinks per week (36)), smoking (never, former, current),and physical activity (International Physical Activity Question-naire (37), expressed in 1000 metabolic equivalent (MET) min-utes in the past week). The presence of CVD or diabetes, forwhich medical treatment was received, was ascertained byself-report.

We considered high levels of distress and the presence of psy-chopathology as one of the underlying mechanisms that lead toshortened TL, as earlier research has pointed out that depressionand anxiety disorders are associated both with shorter TL (38)and with MetS (35). As this is a cohort which over-recruitedsubjects diagnosed with depression and anxiety disorders, wehad the opportunity to check whether the association between

Table 1.Sample characteristics in subjects with complete data at baseline, two- or six-years follow-up

Baseline (n � 2842) Year 2 (n � 2098) Year 6 (n � 1802)

Telomere lengthT/S ratio, M (SD) 1.11 (0.30) - -Base pairs, M (SD) 5464 (614) - -

DemographicsAge (yrs), M (SD) 41.9 (13.0) 44.5 (13.1) 48.5 (13.0)Sex (female), % 66.4 64.1 64.5Years of education, M (SD) 12.2 (3.3) 12.5 (3.3) 12.8 (3.3)

Lifestyle and health factorsSmoking, n (%)

Never 798 (28.1) 698 (33.3) 542 (30.1)Former 940 (33.1) 728 (34.7) 752 (41.7)Current 1104 (38.8) 667 (31.8) 506 (28.1)

Drinking, n (%)Non-drinker 484 (17.0) 319 (15.2) 315 (17.5)Mild-moderate drinker 1996 (70.2) 1493 (71.2) 1263 (70.1)Heavy drinker 362 (12.7) 239 (11.4) 167 (9.3)

Physical activity (*1000 MET-min/wk), median (IQR) 2.8 (3.4) 3.1 (4.1) 2.9 (4.0)Diabetes, n (%) 91 (3.2) 88 (4.2) 83 (4.6)Cardiovascular disease, n (%) 119 (4.2) 92 (4.4) 96 (5.3)

Metabolic Syndrome indicatorsWaist (cm), M (SD) 89.1 (14.0) 89.8 (14.0) 92.5 (13.8)Triglycerides, M (SD) * 1.12 (1.68) 1.15 (1.68) 1.15 (1.69)HDL cholesterol, M (SD) 1.62 (0.44) 1.54 (0.42) 1.55 (0.44)Systolic blood pressure, M (SD) 135.9 (20.2) 134.4 (19.0) 135.3 (20.1)Glucose, M (SD) * 5.12 (1.17) 5.28 (1.17) 5.49 (1.17)Number of MetS abnormalities, median (IQR) a 1.0 (2.0) 1.0 (3.0) 2.0 (2.0)Metabolic syndrome (ATP-III), n (%) b 604 (21.3) 532 (25.4) 523 (29.0)

Footnote: M � Mean; SD � Standard deviation; MET � Metabolic Equivalent; IQR � Interquartile range; * natural logarithm-transformed factorspresented back-transformed;a Abnormalities defined as: 1) waist circumference�102 cm (men) and �88 cm (women); 2) triglycerides �1.7 mmol/liter (150 mg/dL) ormedication for hypertriglyceridemia; 3) high-density lipoprotein (HDL) cholesterol �1.03 mmol/liter (40 mg/dL, men) and �1.30 mmol/liter (50 mg/dL, women) or medication for reduced HDL cholesterol; 4) BP: systolic �130 and/or diastolic �85 mm Hg or antihypertensive; 5) fasting plasmaglucose �5.6 mmol/liter (100 mg/dL) or antidiabetic medication; b Metabolic syndrome defined as having � 3 abnormalities; b Metabolicsyndrome defined as having � 3 abnormalities

doi: 10.1210/jc.2014-1851 jcem.endojournals.org 3

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 08 September 2014. at 05:28 For personal use only. No other uses without permission. . All rights reserved.

TL and MetS components was modified by psychopathology.The presence of psychiatric diagnosis was evaluated with theclinical interview including the lifetime version of the CompositeInterview Diagnostic Instrument (CIDI version 2.1) according toDiagnostic and Statistical Manual of Mental Disorders, FourthEdition (DSM-IV) criteria, administered by specially trained re-search staff.

Statistical analysesMeans (� standard deviation) and proportions of character-

istics of the study sample were described at baseline, two- andsix-year assessments. Due to skewed distributions, natural log-arithm-transformations were used in analyses for triglyceridesand glucose and estimated values were presented back-trans-formed in tables and figures. In cross-sectional analyses, fullyadjusted linear and logistic regression models were used to ex-amine associations between TL (per SD decrease) as a predictorwith the five separate MetS components, the overall number ofMetS abnormalities (treated as a measure of MetS severity), andthe presence of MetS as outcomes.

Additional analyses were performed by testing TL-by-sex in-teraction terms, as females are often suggested to have longer TLthan males (39). Also, since this sample has an overrepresenta-tion of psychiatric patients, the presence of diagnosed anxiety ordepressive disorder was evaluated as a potential mediator oreffect modifier by, respectively, adding psychopathology in themodel and by testing TL-by-psychopathology interaction terms.Furthermore, in order to explore whether associations weredriven by present somatic disease status, patients with baselineCVD or diabetes were excluded in sensitivity analyses.

In longitudinal analyses, associations of baseline TL withchanges in MetS components over the six year course were an-alyzed by generalized estimating equations (GEE) with an ex-changeable correlation structure that takes into account thewithin-person correlations when examining multiple observa-tions per subject and can handle missing subjects (40). In addi-tion to TL and time variables, appropriate TL-by-year2 and TL-by-year6 interaction terms were included in the models toestimate rates of change in MetS components according to base-line TL at each follow-up point. From the same models, we de-rived the parameters of the association between baseline TL andMetS component at specific follow-up. All models were adjustedfor time-varying covariates. All analyses were conducted usingSPSS version 20.0 (IBM Corp., Armonk, NY, USA) and SAS (v.

9.1, SAS Institute, Inc., Cary, NC). Significance level was set atP � .05, two-tailed.

Results

At baseline, the 2842 subjects were on average 41.9 years(SD � 13.0), 66.4% was female and they had attained12.2 years (SD � 3.3) of education (Table 1). The averageT/S ratio was 1.11 (SD � 0.30), corresponding to an av-erage TL of 5.46 kilo base pairs (SD � 0.61). Of the sub-jects, 604 (21.3%) had MetS at baseline. Overall, aftertwo years (n � 2098) and six years (n � 1802), the levelsof MetS components were only slightly deteriorated, butoverall there were relatively less current smokers andheavy drinkers throughout the follow-up years.

Shorter TL (per SD decrease) was cross-sectionally as-sociated with a higher number of MetS abnormalities(��0.075, SE � 0.023, P � .001), higher waist circum-ference (��0.647, SE � 0.238, P � .007), triglycerides(��0.038, SE � 0.009, P � .001) and fasting glucose(��0.011, SE � 0.003, P � .001), and with lower HDLcholesterol (��-0.016, SE � 0.008, P � .05) (Table 2).However, systolic BP was not associated with TL (P �.21). Furthermore, at baseline each SD decrease in TL wasassociated with a 1.19-fold higher odds (95%CI � 1.07–1.33; P � .002) of having MetS. We have derived Figure1 from this model in order to show that the probability ofhaving MetS increased with shorter TL.

In sensitivity analyses, no significant TL-by-sex inter-actions were found, indicating that associations betweenTL and MetS components were consistent for men andwomen. Moreover, additional adjustment for the presenceof psychopathology did not substantially reduce the as-sociation between TL and MetS components and TL-by-psychopathology interaction terms subsequently enteredin the models were nonsignificant, implying that the as-sociations between TL and MetS components were not

Table 2. Cross-sectional associations (betas (standard errors)) between telomere length (TL, per standard deviationdecrease) and continuous metabolic syndrome (MetS) components (n � 2842)

Waist (ln) TriglyceridesHDL

cholesterol Systolic BP (ln) GlucoseMetS

abnormalitiesb

TL (Per SD decrease)a 0.647 (0.238) ** 0.038 (0.009) *** �0.016 (0.008) * �0.437 (0.345) 0.011 (0.003) *** 0.075 (0.023) **TL: Highest quartile (ref)c Ref Ref Ref Ref Ref Ref3rd quartile �0.073 (0.642) 0.023 (0.025) �0.023 (0.021) �0.829 (0.931) 0.018 (0.008) * 0.081 (0.061)2nd quartile 0.785 (0.651) 0.073 (0.026) ** �0.022 (0.021) �0.213 (0.944) 0.029 (0.008) *** 0.150 (0.062) *Lowest quartile 0.937 (0.670) 0.082 (0.027) ** �0.031 (0.022) �1.796 (0.972) � 0.028 (0.008) *** 0.144 (0.064) *

Footnote: Adjusted for age, sex, education, smoking, alcohol, physical activity; � P � 0.10; * P � 0.05; ** P � 0.01; *** P � 0.001;a Standard deviation of TL � 614bp; b Abnormalities defined as: 1) waist circumference�102 cm (men) and �88 cm (women); 2) triglycerides�1.7 mmol/liter (150 mg/dL) or medication for hypertriglyceridemia; 3) high-density lipoprotein (HDL) cholesterol �1.03 mmol/liter (40 mg/dL,men) and �1.30 mmol/liter (50 mg/dL, women) or medication for reduced HDL cholesterol; 4) BP: systolic �130 and/or diastolic �85 mm Hg orantihypertensive; 5) fasting plasma glucose �5.6 mmol/liter (100 mg/dL) or antidiabetic medication; c LTL lowest quartile: 3.85–5.04kbp, secondquartile: 5.04–5.35kbp; third quartile: 5.35–5.78kbp; highest quartile: 5.78–7.39kbp

4 Telomere length and metabolic syndrome J Clin Endocrinol Metab

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 08 September 2014. at 05:28 For personal use only. No other uses without permission. . All rights reserved.

different when a subject was healthy or had a current orremitted anxiety or depressive disorder. Furthermore, theassociations did not change when baseline CVD (N � 119)and diabetes (N � 91) patients were excluded from theanalyses, illustrating that results were not due to the pres-ence of overt clinical diseases. At last, the associationsbetween TL and glucose, lipid and BP components werenot altered by incorporating medication information (an-tidiabetics, fibrates, antihypertensives) in the final analy-ses, as findings were very similar when we did not incor-porate this information.

Longitudinal associations between baseline TL andchanges in MetS components over the six-year follow-upperiod were examined using GEE models using all 2842subjects and adjusting for time-varying covariates (Table3). Overall, as illustrated by the significant time effects, anincrease in waist circumference, triglycerides, glucose andnumber of MetS abnormalities, and a decrease in HDLcholesterol and SBP were observed during the follow-upperiod. When looking at TL-by-year2 interaction terms,we found significantly different rates of change at Year 2within triglycerides, SBP and the number of MetS abnor-malities. For the TL-by-year6 interaction terms, signifi-cantly different rates of change at Year 6 were found forwaist circumference, triglycerides and fasting glucose. Thenegative coefficients of these interaction terms show thatshorter LTL was associated with a less steep increase ofMetS components, suggesting that the differences found atbaseline progressively reduced over the follow-up. How-ever, shorter baseline TL was still associated at Year 2 withhigher waist circumference (��0.530, 95% CI: 0.041 –1.018, P � .034), glucose (��0.008, 95% CI: 0.002 –0.014, P � .009) and lower SBP (��-0.998, 95% CI:–1.670 – –0.326, P � .004) and at year 6 with lower HDLcholesterol (��-0.026, 95% CI: –0.043 – –0.009, P �.002) and a higher number of MetS components

(��0.054, 95% CI: 0.000 – 0.107, P � .049). These re-sults are listed in Supplemental Table 1 for all componentsat the different time points. In order to better illustrate therelationship between baseline TL and trajectories of MetScomponents over time, we plotted the estimated means ofMetS components at each follow–up across different TLvalues from the GEE models (Figure 2). This figure clearlyillustrated that – although differences diminished overtime – persons with short baseline TL still had a worse levelfor many of the MetS components even after an extendedperiod of time.

Discussion

This study has shown that shorter leukocyte telomerelength is cross-sectionally associated with abdominal obe-sity, dyslipidemia, hyperglycemia, and the presence andseverity of metabolic syndrome. Although baseline differ-ences progressively reduced over time, shorter baseline TLwas still significantly associated with unhealthy MetSstate for most components throughout the six-year follow-up. This clearly illustrates that cellular aging is not just ashort-term state marker of deteriorated metabolic health,but that short TL is likely reflective of an underlying mech-anism illustrating chronic subsequent deterioration of aperson’s metabolic health.

Within the current study, we have found that cellularaging is associated with abdominal obesity and dyslipide-mia (low HDL cholesterol and high triglycerides). Severalprevious cross-sectional studies have found shorter TL insubjects with a disadvantageous weight [9–13:18:24:41]or lipid profile [14–18:24], whereas few studies have notfound these associations [5:19:21:22].

Shorter TL was also associated with higher fasting glu-cose levels, as seen in earlier cross-sectional studies [15:

Figure 1. : Cross-sectional association between Metabolic Syndrome at baseline and TL: –2 standard deviations (SD � 4.23 kbp), -1SD (�4.85kbp), the mean (�5.46 kbp), �1SD (�6.08 kbp) and �2SD (�6.70 kbp). All analyses were adjusted for age, sex, education, smoking, alcohol,physical activity.

doi: 10.1210/jc.2014-1851 jcem.endojournals.org 5

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 08 September 2014. at 05:28 For personal use only. No other uses without permission. . All rights reserved.

24:41], whereas others have not found these associations[5:12:19:21:22]. The discrepancies between studies areprobably attributable to differences in the study popula-tions (age, clinical vs population-based) and in the assess-ment of metabolic parameters and TL. Our study adds tothis growing literature by showing that TL is associatedwith a wide range of metabolic parameters and predictsless favorable trajectories of these factors over time in alarge-scale cohort study, while adjusting for importanttime-varying confounders.

To date, only two studies have investigated the longi-

tudinal association between TL [23:25] and obesity-re-lated measures. Njajou et al have reported that shorter TLassessed at baseline was associated with smaller increasesin the percentage change for both BMI and body fat be-tween baseline and a seven-year follow-up in elderly sub-jects (25). This is in line with the pattern that we found ofa less steep increase of some MetS components over timefor those with shorter baseline TL. This may be partly dueto the fact that subjects with longer telomeres at baselinemay have more room to worsen, determining therefore anapparently steeper rate of worsening (regression to the

Figure 2. : Estimated means of the five MetS components and the number of MetS abnormalities at Y0, Y2 and Y6 (N � 2842), calculated fortelomere length (TL) –2 standard deviations (SD � 4.23 kbp), -1SD (�4.85 kbp), the mean (�5.46 kbp), �1SD (�6.08 kbp) and �2SD (�6.70kbp). Betas, 95% CI and p-values of all the associations between TL (per SD decrease) and each of the components at the different time points arelisted in Supplemental Table 1.

6 Telomere length and metabolic syndrome J Clin Endocrinol Metab

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 08 September 2014. at 05:28 For personal use only. No other uses without permission. . All rights reserved.

mean) during the follow-up period. The availability ofrepeated measurements of TL may yield a more precisepicture of the longitudinal relationship with MetS com-ponents as it may also account for change in TL over time.Indeed, the study by Gardner et al based on 50 participantswith two assessments of TL, showed that increase in TLover time was associated with lower insulin resistance andlower BMI over 10 years (23). Nevertheless, despite thatbaseline differences progressively reduced over time in thecurrent study, participants with short TL still had moredisadvantageous metabolic outcomes during the six-yearfollow-up compared to subjects with long TL. This is in-deed what has been found also by the study of Njajou etal, as they reported that TL was still negatively associatedwith BMI and percentage total body fat after seven years(25).

We did not find consistent patterns of association be-tween TL and hypertension, in line with earlier research[5:21:22]. The finding of lower SBP at 2-year follow-upfor shorter baseline TL is unclear. Since this associationwas found neither at baseline nor at 6-year follow-up, wemay hypothesize that this was a spurious finding, or thatwe did not had reliable and sensitive enough assessmentsof BP over time. Alternatively, BP may reflect anotherpathway of metabolic abnormalities, as opposed to ab-normalities of obesity-induced insulin resistance and dys-lipidemia, as suggested by Melka et al (42).

Altogether, we confirmed that accelerated cellular ag-ing is linked to MetS [5:19], and with a higher number ofMetS abnormalities, representing a more severe metabolicprofile. Shortened telomeres could lead to a worsened met-abolic state through various molecular mechanisms, suchas increased inflammation, sympathetic nervous systemactivity, and oxidative damage [9:31:41:43:44], condi-tions known to be associated with the telomere attritionrate [15:22:23]. When cells with short telomeres becomesenescent, for example, they release inflammatory cyto-

kines, inducing insulin resistance and defective HDL cho-lesterol (45). Another mechanism could be that telomericDNA damage leads to compromised mitochondrial func-tioning through dampened p53 expression, eventuallycausing less fatty acid oxidation and glucose utilization,and less protection against oxidative stress (45). Addition-ally, shortened telomeres are also associated with adi-pocyte hypertrophy, which in turn are found to be linkedto poor glycemic and lipid control (9).

Some limitations of this study were the fact that we onlydetermined TL at baseline, whereas repeated measure-ments would have yielded more information regarding thecoherence of TL and the trajectory of MetS. Due to theobservational nature of the current study, no conclusionson causation can be directly drawn. In addition, we did notmeasure the dynamics of the telomere maintenance sys-tem, since we did not measure activity of the telomeraseenzyme (2). Previously, Epel et al found that glucose wasnot associated directly with TL but with telomerase, whichmay have acted as a compensatory mechanism within cellswith short telomeres (5). Even though the absence of te-lomerase has been shown to lead to progressive telomereshortening, it is still under debate how telomerase negativecells, or cells with very low levels of telomerase, can sur-vive and replicate. For instance, a study in mice has shownthat mutations in the genes encoding telomerase led toprogressive shortening of telomeres (46). However, res-toration of normal telomerase genotypes in offspring ofaffected individuals did not elongate inherited short telo-meres immediately (46), suggesting that this maintenancesystem is more complex, and that even though humans andmice both share the same telomere sequence in their DNA(TTAGGG), the telomere – telomerase system does notseem identical in mice and humans (47). Furthermore, TLwas assessed in leukocytes, as they are easily accessible andthe results are comparable with most other studies. Al-though there is an ongoing debate about the comparability

Table 3. Longitudinal associations (betas (standard error), time-associations and time-interactions) betweentelomere length (TL, per standard deviation decrease) and continuous metabolic syndrome (MetS) components (n �2842)

Waist (ln)Triglycerides HDL cholesterol Systolic BP (ln) GlucoseMetS

abnormalitiesb

TLa 0.747 (0.234)** 0.040 (0.009)*** �0.015 (0.008)� �0.408 (0.334) 0.011 (0.003)*** 0.070 (0.022)**Year 2 0.814 (0.141)*** 0.042 (0.009)*** �0.098 (0.006)*** �2.408 (0.295)*** 0.030 (0.003)*** 0.098 (0.020)***Year 6 3.408 (0.183)*** 0.046 (0.010)*** �0.083 (0.007)*** �0.967 (0.341)** 0.069 (0.003)*** 0.289 (0.025)***TL*Yr2 �0.217 (0.134) �0.031 (0.009)*** 0.004 (0.006) �0.590 (0.298)* �0.003 (0.003) �0.038 (0.020)*TL*Yr6 �0.338 (0.168)* �0.030 (0.010)** �0.011 (0.007)� 0.122 (0.330) �0.006 (0.003)* �0.016 (0.024)

Footnote: Adjusted for time, baseline age, sex, education, smoking, alcohol, physical activity; � P � 0.10; * P � 0.05; ** P � 0.01; *** P �0.001;a Standard deviation of TL � 614bp; b Abnormalities defined as: 1) waist circumference�102 cm (men) and �88 cm (women); 2) triglycerides�1.7 mmol/liter (150 mg/dL) or medication for hypertriglyceridemia; 3) high-density lipoprotein (HDL) cholesterol �1.03 mmol/liter (40 mg/dL,men) and �1.30 mmol/liter (50 mg/dL, women) or medication for reduced HDL cholesterol; 4) BP: systolic �130 and/or diastolic �85 mm Hg orantihypertensive; 5) fasting plasma glucose �5.6 mmol/liter (100 mg/dL) or antidiabetic medication

doi: 10.1210/jc.2014-1851 jcem.endojournals.org 7

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 08 September 2014. at 05:28 For personal use only. No other uses without permission. . All rights reserved.

of TL in different subtypes of leukocytes (48), leukocyteTL seems to correlate strongly with other body tissues(49). Strengths of this study were its large sample size andthe detailed measurement of MetS, taking into accountimportant medication that could influence the separatecomponents. Furthermore, MetS components and all co-variates were determined longitudinally, giving us theunique opportunity to examine the course over time andtake into account changing lifestyle and clinical factors aswell.

Future studies should attempt to shed more light on thecomplex relation between the telomere maintenance sys-tem and the MetS components by measuring TL at moretime points, and applying an experimental design to testwhether TL is a cause, a consequence or an epiphenome-non of age-related metabolic dysregulations. If lifestyleand externally modifiable factors are predicting telomereshortening, there might also be an opportunity to improvemaintenance and even lengthen telomeres with behavioralinterventions, such as diet, exercise and stress manage-ment strategies [50:51]. We believe that the current find-ings provide an initial input for future research (especiallyincludingexperimental andbench studies) aimedat testingwhether TL may be one of the causal factors, among andin interactions with other biological processes, for age-related metabolic dysregulations. We also want to pointout that aging, with or without metabolic dysregulations,is a complex phenomenon, and is accompanied by manybiological changes. On the other hand, researchers shouldalso weigh in mind that telomere maintenance is an im-portant homeostasis system for the uncontrolled prolifer-ation of tumor cells, demonstrating the fragile balancewithin this system (52–54).

To conclude, we showed that cellular aging, as assessedby TL, is associated with a higher metabolic risk profileand with less favorable trajectories of metabolic biomark-ers over time. These findings suggest that TL might play arole in the onset of various aging-related somatic diseasesvia its effect on metabolic alterations. Understanding ofthe possible causal pathways between telomere shorteningand metabolic alterations requires further research. If suchmechanisms will be shown to be accessible to effectiveinterventions, reduction of telomere attrition may becomea target to prevent metabolic dysregulations and their del-eterious effect on health.

Acknowledgments

The infrastructure for the NESDA study (www.nesda.nl) isfunded through the Geestkracht program of the NetherlandsOrganisation for Health Research and Development (ZonMW,grant number 10–000–1002) and is supported by participating

universities and mental health care organisations (VU UniversityMedical Center, GGZ inGeest, Arkin, Leiden University MedicalCenter, GGZ Rivierduinen, University Medical Center Gro-ningen, Lentis, GGZ Friesland, GGZ Drenthe, IQ Healthcare,Netherlands Institute for Health Services Research (NIVEL) andNetherlands Institute of Mental Health and Addiction (Trim-bos). All authors declare that they have no conflicts of interest.BP, DR, JV and telomere length assaying were supportedthrough a NWO-VICI grant (number 91811602). The authorsthank Jan Willem van Rooij for his help with the figures.

Address all correspondence and requests for reprints to: Cor-responding author: Dóra Révész, MSc, Visiting address: A.J.Ernststraat 1187, room M1.06, 1081 HL Amsterdam, The Neth-erlands, Correspondence: PO Box 74077, 1070 BB Amsterdam,The Netherlands, Phone: �31 20 788 4596, Fax: �31 20 7885664, E-mail: D.Revesz@ggzingeest.nl.

Disclosure summary: All authors declare that they have noconflicts of interest.

This work was supported by .

References

1. Third Report of the National Cholesterol Education Program(NCEP) Expert Panel on Detection, Evaluation, and Treatment ofHigh Blood Cholesterol in Adults (Adult Treatment Panel III) finalreport. Circulation 2002 December 17;106: 3143–421.

2. Blackburn EH. Switching and signaling at the telomere. Cell. 2001September;106:661–7321.

3. Olovnikov AM. Telomeres, telomerase, and aging: origin of thetheory. Exp Gerontol. 1996 July;31:443–8.

4. Epel E. How. “reversible” is telomeric aging? Cancer Prev Res(Phila). 2012 October;5:1163–8.

5. Epel ES, Lin J, Wilhelm FH, Wolkowitz OM, Cawthon R, Adler NE,Dolbier C, Mendes WB, Blackburn EH. Cell aging in relation tostress arousal and cardiovascular disease risk factors. Psychoneu-roendocrinology. 2006 April;31:277–87.

6. Cawthon RM, Smith KR, O’Brien E, Sivatchenko A, Kerber RA.Association between telomere length in blood and mortality in peo-ple aged 60 years or older. Lancet. 2003 February;361:393–51.

7. Muezzinler A, Zaineddin AK, Brenner H. A systematic review ofleukocyte telomere length and age in adults. Ageing Res Rev. 2013March;12:509–19.

8. Sahin E, Colla S, Liesa M, Moslehi J, Muller FL, Guo M, Cooper M,Kotton D, Fabian AJ, Walkey C, Maser RS, Tonon G, Foerster F,Xiong R, Wang YA, Shukla SA, Jaskelioff M, Martin ES, HeffernanTP, Protopopov A, Ivanova E, Mahoney JE, Kost-Alimova M, PerrySR, Bronson R, Liao R, Mulligan R, Shirihai OS, Chin L, DePinhoRA. Telomere dysfunction induces metabolic and mitochondrialcompromise. Nature. 2011 February;470:359–6517.

9. Monickaraj F, Gokulakrishnan K, Prabu P, Sathishkumar C, AnjanaRM, Rajkumar JS, Mohan V, Balasubramanyam M. Convergenceof adipocyte hypertrophy, telomere shortening and hypoadi-ponectinemia in obese subjects and in patients with type 2 diabetes.Clin Biochem. 2012 November;45:1432–8.

10. Zannolli R, Mohn A, Buoni S, Pietrobelli A, Messina M, ChiarelliF, Miracco C. Telomere length and obesity. Acta Paediatr. 2008July;97:952–4.

11. Valdes AM, Andrew T, Gardner JP, Kimura M, Oelsner E, CherkasLF, Aviv A, Spector TD. Obesity, cigarette smoking, and telomerelength in women. Lancet. 2005 August;366:662–420.

12. Zee RYL, Castonguay AJ, Barton NS, Germer S, Martin M. Mean

8 Telomere length and metabolic syndrome J Clin Endocrinol Metab

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 08 September 2014. at 05:28 For personal use only. No other uses without permission. . All rights reserved.

leukocyte telomere length shortening and type 2 diabetes mellitus: acase-control study. Transl Res. 2010 April;155:166–9.

13. Buxton JL, Walters RG, Visvikis-Siest S, Meyre D, Froguel P,Blakemore AIF. Childhood obesity is associated with shorter leu-kocyte telomere length. J Clin Endocrinol Metab. 2011 May;96:1500–5.

14. Nordfjall K, Eliasson M, Stegmayr B, Melander O, Nilsson P, RoosG. Telomere length is associated with obesity parameters but witha gender difference. Obesity (Silver Spring). 2008 December;16:2682–9.

15. Aviv A, Valdes A, Gardner JP, Swaminathan R, Kimura M, SpectorTD. Menopause modifies the association of leukocyte telomerelength with insulin resistance and inflammation. J Clin EndocrinolMetab. 2006 February;91:635–40.

16. Demissie S, Levy D, Benjamin EJ, Cupples LA, Gardner JP, HerbertA, Kimura M, Larson MG, Meigs JB, Keaney JF, Aviv A. Insulinresistance, oxidative stress, hypertension, and leukocyte telomerelength in men from the Framingham Heart Study. Aging Cell. 2006August;5:325–30.

17. Harte AL, da Silva NF, Miller MA, Cappuccio FP, Kelly A, O’HareJP, Barnett AH, Al-Daghri NM, Al-Attas O, Alokail M, Sabico S,Tripathi G, Bellary S, Kumar S, McTernan PG. Telomere lengthattrition, a marker of biological senescence, is inversely correlatedwith triglycerides and cholesterol in South Asian males with type 2diabetes mellitus. Exp Diabetes Res. 2012;2012:895185.

18. Lee M, Martin H, Firpo MA, Demerath EW. Inverse associationbetween adiposity and telomere length: The Fels LongitudinalStudy. Am J Hum Biol. 2011 January;23:100–6.

19. Satoh M, Ishikawa Y, Takahashi Y, Itoh T, Minami Y, NakamuraM. Association between oxidative DNA damage and telomereshortening in circulating endothelial progenitor cells obtained frommetabolic syndrome patients with coronary artery disease. Athero-sclerosis. 2008 June;198:347–53.

20. Strandberg TE, Saijonmaa O, Tilvis RS, Pitkala KH, Strandberg AY,Miettinen TA, Fyhrquist F. Association of telomere length in oldermen with mortality and midlife body mass index and smoking. JGerontol A Biol Sci Med Sci. 2011 July;66:815–20.

21. Fitzpatrick AL, Kronmal RA, Gardner JP, Psaty BM, Jenny NS,Tracy RP, Walston J, Kimura M, Aviv A. Leukocyte telomere lengthand cardiovascular disease in the cardiovascular health study. Am JEpidemiol. 2007 January;165:14–211.

22. Bekaert S, De Meyer T, Rietzschel ER, De Buyzere ML, De BacquerD, Langlois M, Segers P, Cooman L, Van Damme P, Cassiman P,Van Criekinge W, Verdonck P, De Backer GG, Gillebert TC, VanOostveldt P. Telomere length and cardiovascular risk factors in amiddle-aged population free of overt cardiovascular disease. AgingCell. 2007 October;6:639–47.

23. Gardner JP, Li S, Srinivasan SR, Chen W, Kimura M, Lu X, BerensonGS, Aviv A. Rise in insulin resistance is associated with escalatedtelomere attrition. Circulation. 2005 May;111:2171–73.

24. Monickaraj F, Aravind S, Gokulakrishnan K, Sathishkumar C,Prabu P, Prabu D, Mohan V, Balasubramanyam M. Acceleratedaging as evidenced by increased telomere shortening and mitochon-drial DNA depletion in patients with type 2 diabetes. Mol CellBiochem. 2012 June;365:343–50.

25. Njajou OT, Cawthon RM, Blackburn EH, Harris TB, Li R, SandersJL, Newman AB, Nalls M, Cummings SR, Hsueh WC. Shorter telo-meres are associated with obesity and weight gain in the elderly. IntJ Obes (Lond). 2012 September;36:1176–9.

26. Cui Y, Gao YT, Cai Q, Qu S, Cai H, Li HL, Wu J, Ji BT, Yang G,Chow WH, Shu XO, Zheng W. Associations of leukocyte telomerelength with body anthropometric indices and weight change in Chi-nese women. Obesity (Silver Spring). 2013 December;21:2582–8.

27. Kim S, Parks CG, DeRoo LA, Chen H, Taylor JA, Cawthon RM,Sandler DP. Obesity and weight gain in adulthood and telomerelength. Cancer Epidemiol Biomarkers Prev. 2009 March;18:816–20.

28. Penninx BWJH, Beekman ATF, Smit JH, Zitman FG, Nolen WA,

Spinhoven P, Cuijpers P, de Jong PJ, Van Marwijk HWJ, AssendelftWJJ, van der Meer K, Verhaak P, Wensing M, de Graaf R,Hoogendijk WJ, Ormel J, van Dyck R. The Netherlands Study ofDepression and Anxiety (NESDA): rationale, objectives and meth-ods. Int J Methods Psychiatr Res. 2008;17:121–40.

29. Cawthon RM. Telomere measurement by quantitative PCR. Nu-cleic Acids Res. 2002 May;30:e4715.

30. Aviv A, Hunt SC, Lin J, Cao X, Kimura M, Blackburn E. Impartialcomparative analysis of measurement of leukocyte telomere length/DNA content by Southern blots and qPCR. Nucleic Acids Res. 2011November;39:e1341.

31. Révész D, Verhoeven JE, Milaneschi Y, de Geus EJ, Wolkowitz OM,Penninx BWJH. Dysregulated physiological stress systems and ac-celerated cellular aging. Neurobiol Aging. 2014;35:1422–30.

32. Standards of medical care in diabetes–2013. Diabetes Care 2013January; 36 Suppl 1: S11–S66.

33. Packard CJ. Overview of fenofibrate. Eur Heart J 1998 February; 19Suppl A: A62–A65.

34. Mancia G, Parati G. Office compared with ambulatory blood pres-sure in assessing response to antihypertensive treatment: a meta-analysis. J Hypertens. 2004 March;22:435–45.

35. van Reedt Dortland AK, Giltay EJ, van VT, Zitman FG, PenninxBW. Metabolic syndrome abnormalities are associated with severityof anxiety and depression and with tricyclic antidepressant use. ActaPsychiatr Scand. 2010 July;122:30–91.

36. Stuurgroep Multidisciplinaire Richtlijnontwikkeling. Stoornissen inhet gebruik van alcohol: Richtlijn voor de diagnostiek en behandel-ing van patiënten met een stoornis in het gebruik van alcohol. GGZ:Utrecht, The Netherlands; 2009.

37. Craig CL, Marshall AL, Sjostrom M, Bauman AE, Booth ML, Ain-sworth BE, Pratt M, Ekelund U, Yngve A, Sallis JF, Oja P. Interna-tional physical activity questionnaire: 12-country reliability and va-lidity. Med Sci Sports Exerc. 2003 August;35:1381–95.

38. Verhoeven J, Révész D, Epel E, Lin J, Wolkowitz O, Penninx B.Major depressive disorder and accelerated cellular aging: resultsfrom a large psychiatric cohort study. Mol Psychiatry. 2013 No-vember 12; -doi: 10.1038/mp. 2013.151.

39. Gardner M, Bann D, Wiley L, Cooper R, Hardy R, Nitsch D, Mar-tin-Ruiz C, Shiels P, Sayer AA, Barbieri M, Bekaert S, Bischoff C,Brooks-Wilson A, Chen W, Cooper C, Christensen K, De Meyer T,Deary I, Der G, ez Roux A, Fitzpatrick A, Hajat A, Halaschek-Wiener J, Harris S, Hunt SC, Jagger C, Jeon HS, Kaplan R, KimuraM, Lansdorp P, Li C, Maeda T, Mangino M, Nawrot TS, NilssonP, Nordfjall K, Paolisso G, Ren F, Riabowol K, Robertson T, RoosG, Staessen JA, Spector T, Tang N, Unryn B, van der Harst P, WooJ, Xing C, Yadegarfar ME, Park JY, Young N, Kuh D, Von ZglinickiT, Ben-Shlomo Y. Gender and telomere length: Systematic reviewand meta-analysis. Exp Gerontol. 2014 March;51:15–27.

40. Twisk JWR. Longitudinal data analysis. A comparison betweengeneralized estimating equations and random coefficient analysis.Eur J Epidemiol. 2004;19:769–76.

41. Ma D, Zhu W, Hu S, Yu X, Yang Y. Association between oxidativestress and telomere length in Type 1 and Type 2 diabetic patients. JEndocrinol Invest. 2013 December;36:1032–7.

42. Melka MG, Abrahamowicz M, Leonard GT, Perron M, Richer L,Veillette S, Gaudet D, Paus T, Pausova Z. Clustering of the meta-bolic syndrome components in adolescence: role of visceral fat.PLoS One. 2013;8:e82368.

43. Licht C, de Geus E, Penninx B. Dysregulation of the autonomicnervous system predicts the development of the Metabolic Syn-drome. J Clin Endocrinol Metab. 2013 April 3.

44. Bakker SJ, IJzerman RG, Teerlink T, Westerhoff HV, Gans RO,Heine RJ. Cytosolic triglycerides and oxidative stress in central obe-sity: the missing link between excessive atherosclerosis, endothelialdysfunction, and beta-cell failure? Atherosclerosis. 2000 January;148:17–21.

45. Fyhrquist F, Saijonmaa O, Strandberg T. The roles of senescence and

doi: 10.1210/jc.2014-1851 jcem.endojournals.org 9

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 08 September 2014. at 05:28 For personal use only. No other uses without permission. . All rights reserved.

telomere shortening in cardiovascular disease. Nat Rev Cardiol.2013 March 12.

46. Chiang YJ, Calado RT, Hathcock KS, Lansdorp PM, Young NS,Hodes RJ. Telomere length is inherited with resetting of the telomereset-point. Proc Natl Acad Sci U S A. 2010 June;107:10148–531.

47. Calado RT, Dumitriu B. Telomere dynamics in mice and humans.Semin Hematol. 2013 April;50:165–74.

48. Lin J, Epel E, Cheon J, Kroenke C, Sinclair E, Bigos M, WolkowitzO, Mellon S, Blackburn E. Analyses and comparisons of telomeraseactivity and telomere length in human T and B cells: insights forepidemiology of telomere maintenance. J Immunol Methods. 2010January;352:71–8031.

49. Daniali L, Benetos A, Susser E, Kark JD, Labat C, Kimura M, DesaiK, Granick M, Aviv A. Telomeres shorten at equivalent rates insomatic tissues of adults. Nat Commun. 2013 March;4:159719.

50. Ornish D, Lin J, Chan JM, Epel E, Kemp C, Weidner G, Marlin R,Frenda SJ, Magbanua MJ, Daubenmier J, Estay I, Hills NK, Chai-

nani-Wu N, Carroll PR, Blackburn EH. Effect of comprehensivelifestyle changes on telomerase activity and telomere length in menwith biopsy-proven low-risk prostate cancer: 5-year follow-up of adescriptive pilot study. Lancet Oncol. 2013 October;14:1112–20.

51. Garcia-Calzon S, Moleres A, Marcos A, Campoy C, Moreno LA,zcona-Sanjulian MC, Martinez-Gonzalez MA, Martinez JA, ZalbaG, Marti A. Telomere Length as a Biomarker for Adiposity Changesafter a Multidisciplinary Intervention in Overweight/Obese Adoles-cents: The EVASYON Study. PLoS One. 2014;9:e89828.

52. Li Y, Tergaonkar V. Noncanonical functions of telomerase: impli-cations in telomerase-targeted cancer therapies. Cancer Res. 2014March;74:1639–4415.

53. Artandi SE, DePinho RA. Telomeres and telomerase in cancer. Car-cinogenesis. 2010 January;31:9–18.

54. Wu XQ, Huang C, He X, Tian YY, Zhou DX, He Y, Liu XH, Li J.Feedback regulation of telomerase reverse transcriptase: new insightinto the evolving field of telomerase in cancer. Cell Signal. 2013December;25:2462–8.

10 Telomere length and metabolic syndrome J Clin Endocrinol Metab

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 08 September 2014. at 05:28 For personal use only. No other uses without permission. . All rights reserved.