Author + information
- Received November 25, 2015
- Revision received March 10, 2016
- Accepted March 12, 2016
- Published online May 31, 2016.
- Juan M. Fernández-Alvira, PhDa,
- Valentin Fuster, MD, PhDa,b,∗ (, )
- Beatriz Dorado, PhDa,
- Nora Soberón, PhDc,
- Ignacio Flores, PhDa,
- Mercedes Gallardo, PhDc,
- Stuart Pocock, PhDa,d,
- María A. Blasco, PhDc and
- Vicente Andrés, PhDa,∗∗ ()
- aCentro Nacional de Investigaciones Cardiovasculares Carlos III (CNIC), Madrid, Spain
- bMount Sinai School of Medicine, New York, New York
- cTelomeres and Telomerase Group, Molecular Oncology Program, Spanish National Cancer Research Centre (CNIO), Madrid, Spain
- dLondon School of Hygiene & Tropical Medicine, London, United Kingdom
- ↵∗Reprint requests and correspondence:
Dr. Valentin Fuster, Mount Sinai School of Medicine, Cardiovascular Institute, One Gustave Levy Place, Box 1030, New York, New York 10029-6500.
- ↵∗∗Dr. Vicente Andrés, Centro Nacional de Investigaciones Cardiovasculares Carlos III (CNIC), Melchor Fernández Almagro 3, Madrid 28029, Spain.
Background Leucocyte telomere length (LTL) shortening is associated with cardiovascular ischemic events and mortality in humans, but data on its association with subclinical atherosclerosis are scarce. Whether the incidence and severity of subclinical atherosclerosis are associated with the abundance of critically short telomeres, a major trigger of cellular senescence, remains unknown.
Objectives The authors conducted a cross-sectional exploration of the association between subclinical atherosclerosis burden and both average LTL and the abundance of short telomeres (%LTL<3 kb).
Methods Telomere length was assessed by high-throughput quantitative fluorescence in situ hybridization in circulating leukocytes from 1,459 volunteers without established cardiovascular disease (58% men, 40 to 54 years of age) from the PESA (Progression of Early Subclinical Atherosclerosis) study. Subclinical atherosclerosis was evaluated by coronary artery calcium scan and 2-dimensional/3-dimensional ultrasound in different aortic territories. Statistical significance of differences among multiple covariates was assessed with linear regression models. Independent associations of telomere parameters with plaque presence were evaluated using general linear models.
Results In men and women, age was inversely associated with LTL (Pearson’s r = −0.127, p < 0.001) and directly with %LTL<3 kb (Pearson’s r = 0.085; p = 0.001). Short LTL reached statistical significance as a determinant of total and femoral plaque in men, but not in women. However, this association was not sustained after adjustment for age or additional adjustment for cardiovascular risk factors. No significant independent association was found between %LTL<3 kb and plaque burden. Serum-oxidized low-density lipoprotein levels were directly associated with %LTL<3 kb in men (p = 0.008) and women (p < 0.001).
Conclusions In a cross-sectional study of a middle-aged population, average LTL and short telomere load are not significant independent determinants of subclinical atherosclerosis. Longitudinal follow-up of PESA participants will assess long-term associations between telomere length and progression of subclinical atherosclerosis.
Atherosclerosis and associated myocardial infarction or stroke are the major cause of morbidity and mortality in industrialized countries, and the economic, social, and human costs associated with treating cardiovascular disease (CVD) are painfully high (1). Our previous results revealed highly prevalent subclinical atherosclerosis in almost one-half of 4,066 participants in the PESA (Progression of Early Subclinical Atherosclerosis) study, all of whom were free from established CVD; moreover, noninvasive imaging demonstrated frequent atherosclerosis in subjects at low risk according to traditional risk scores (2). To improve diagnosis and prevention, it is thus urgent to refine classic CVD risk scores and identify new biomarkers of atherosclerotic disease initiation and early progression.
Attrition of telomeric DNA is assumed to be a biomarker of aging and health status (3). Because age is a crucial determinant of CVD (4), the possibility that telomere attrition might predict cardiovascular risk has awakened tremendous interest. Telomeres are capping deoxyribonucleic acid (DNA)–nucleoprotein complexes that protect the ends of all eukaryotic chromosomes from degradation, unwanted recombination, or fusion (5). In most somatic cells, including leukocytes, telomeric DNA is shortened with every cell division throughout life from right after birth, in part due to inactivation of the telomere-elongating polymerase telomerase. When telomeric DNA becomes critically short, cells trigger an irreversible DNA damage response that leads to cellular senescence and, eventually, to apoptosis, thus compromising tissue repair capacity and function (6). Although human leukocyte telomere length (LTL) is predominantly determined by inheritance, it can be influenced by stresses during pregnancy and exhibits high interindividual variability at the same age (7,8). Telomeric dysfunction and attrition can be caused by genetic mutations (9) or cardiovascular risk factors (e.g., inflammation, oxidative stress, lack of estrogens, insulin resistance, smoking, obesity, alcohol abuse, sedentary life-style, psychological stress) (10–12).
Numerous epidemiological studies have investigated the relation between LTL and cardiovascular ischemic events, but the outcomes are often conflicting or too heterogeneous, as recently revealed in 2 independent meta-analyses (13,14). In their meta-analysis of 27 studies, D’Mello et al. (13) found a significant association between shortened LTL and myocardial infarction, type 2 diabetes mellitus, and stroke. In their meta-analysis of 24 studies, many overlapping those of D’Mello et al. (13), Haycock et al. (14) also showed an inverse association between LTL and nonfatal myocardial infarction, coronary heart disease death, and coronary revascularization, but the association with cerebrovascular disease was less certain. Paradoxically, D’Mello et al. (13) did not find a significant association between LTL and coronary artery disease (a secondary outcome defined as angina and nonfatal ischemic heart disease), suggesting that LTL attrition is a marker of plaque rupture and associated ischemic events.
An interesting question is whether shortened LTL is a prognostic biomarker for the early identification of subjects at high risk of developing CVD before symptoms appear. Compared with the wealth of information available on short LTL and established CVD (e.g., myocardial infarction or stroke), few epidemiological studies have focused on possible associations between telomere length (TL) and subclinical (asymptomatic) atherosclerosis. The Asklepios cohort study is the largest reported to date (2,509 volunteers free from established CVD, approximately 35 to 55 years of age) (15). The main conclusion of the Asklepios study was that shortened LTL is not a substantial determinant of preclinical atherosclerosis in either sex, and that the association between LTL and CVD cannot be explained by increased predisposition to atherosclerosis as a consequence of shorter inherited telomeres (15). Atherosclerosis in the Asklepios study was assessed in carotid and femoral arteries by intima-media thickness (IMT) and by vascular echography (absence/presence of plaque), but the potential association between LTL and plaque volume was not assessed. Moreover, to the best of our knowledge, the possible association between CVD and short telomere load has not yet been investigated, despite firm evidence showing that the accumulation of a few critically short telomeres in a cell can trigger cellular senescence (16–20). Here, we performed a cross-sectional study using high-throughput quantitative fluorescence in situ hybridization (HT-QFISH) (18) to quantify the percentage of short telomeres (%LTL <3 kb) and average LTL in a subset of 1,459 healthy volunteers in the PESA cohort, and investigated the possible association of these parameters with plaque burden measured by 3-dimensional (3D) sonography in different vascular territories.
Study design and population
The PESA study rationale and design were described elsewhere (21). In brief, PESA-CNIC-Santander is a prospective cohort study including 4,066 asymptomatic employees at the Banco de Santander Headquarters in Madrid (Spain). Participants were 40 to 54 years of age at recruitment (between June 2010 and February 2014) and were free of established CVD. Peripheral blood samples suitable for HT-QFISH (see the “Telomere Measurements in Leukocytes” section) were available only from the last 1,500 participants recruited to the PESA study. Of these, complete data were available for 1,459, who were therefore included in the present study. This subgroup shares similar main characteristics with the overall PESA study cohort (e.g., sex, age) and exhibited slightly lower subclinical atherosclerosis prevalence (3% to 7% lower). Employees were invited to participate during their regular annual medical checkup by the company’s medical services. Subjects with established CVD or any other condition reducing life expectancy were excluded from the study. The study visits include a clinical interview, physical examination (anthropometric measures, blood pressure, and heart rate), a fasting blood draw and urine sample, life-style questionnaires, accelerometry assessment of physical activity, electrocardiogram, and assessment of subclinical atherosclerosis by noninvasive vascular imaging tests, including 2-dimensional (2D)/3D vascular ultrasound and computed tomography (CT) for coronary artery calcium score (CACS) measurement. The ethics committee of Instituto de Salud Carlos III in Madrid, Spain, approved the study protocol, and written informed consent was obtained from each participant before enrollment.
Life-style and CVD risk factors
Traditional risk factors, such as smoking habits and a diagnosis of hypertension, diabetes, or dyslipidemia, were collected as part of each participant’s medical history. Blood pressure was measured at rest with an automatic oscillometric sphygmomanometer (OMRON HEM-907, OMRON Healthcare, Kyoto, Japan). Anthropometric measures were obtained following a standardized procedure. Body mass index was calculated as body mass divided by the squared height (kg/m2). Fasting blood and urine samples were collected after ≈8 h of fasting and tested for serum glucose, triglycerides, total cholesterol, high-density lipoprotein cholesterol, low-density lipoprotein cholesterol, C-reactive protein, whole-blood glycosylated hemoglobin (HbA1c), and markers of inflammation, such as oxidized low-density lipoprotein (ox-LDL) and urinary isoprostanes. Dietary intake was assessed with a computerized questionnaire (Dietary History-Enrica), conducted by trained dieticians, designed to record habitual food intake over the previous year.
All ultrasound recordings were analyzed at the PESA Core Imaging Laboratory at the CNIC. Imaging studies included vascular 2D ultrasound of carotid arteries, infrarenal aorta, and iliofemoral arteries; 3D ultrasound of carotid and femoral arteries; and CACS by CT scan. Vascular ultrasound was performed with a Philips iU22 ultrasound station (Philips Healthcare, Bothell, Washington), with adapted scanning protocols (22). A 2D high-resolution linear array transducer was used to assess the presence of plaques, and a 3D high-resolution transducer was used to assess plaque volume and, subsequently, to estimate the total atherosclerotic burden. Plaque burden was quantified with a semiautomatic 3D plaque analysis system engineered by Philips (Xcelera system and QLAB-VPQ [vascular plaque quantification]). Plaques were defined as any focal protrusion of more than 0.5 mm or more than 50% thicker than the surrounding intima-media (23). The QLAB-VPQ system estimated plaque burden by summing plaque areas and plaque volumes from successive cross-sectional scans of vascular segments containing atherosclerotic lesions. CACS was estimated by noncontrast electrocardiographic-gated prospective acquisition with a 16-slice Brilliance CT scanner (Philips Healthcare, Andover, Massachusetts).
Telomere measurements in leukocytes
Peripheral blood leukocytes were isolated with Histopaque (Sigma-Aldrich, St. Louis, Missouri), and cells were frozen in freshly prepared 90% FBS-10% DMSO and stored in liquid nitrogen until use. HT-QFISH in 96-well-plate format was performed on interphase nuclei of leukocytes to determine mean LTL and %LTL<3 kb (24). This technique uses high-throughput confocal microscopy to detect telomeres labeled with a fluorescent peptide nucleic acid probe against telomeric repeats. All samples were measured in duplicate, and mean LTL and %LTL<3 kb were calculated for each sample. Previous studies have used <3 kb as the cutoff for “short telomeres” in humans (18,25).
The intra-assay coefficient of variation was 6.3% for LTL and 10.5% for %LTL<3 kb. For 152 randomly chosen samples, duplicate measurements were repeated in a different plate to calculate the interassay coefficient of variation, which was 5.6% for LTL and 12.8% for %LTL<3 kb. The sample size (1,459) was sufficient to detect a standardized difference of 0.171 in LTL between groups with and without plaque with a statistical power of 0.9.
All statistical analyses were performed using the Statistical Package for the Social Sciences (version 20.0, SPSS, IBM Corp., Armonk, New York). The 3D ultrasound data on plaque burden were categorized into tertiles, and CACS was dichotomized into <1 and ≥1 Agatston units. Pearson correlation coefficients were calculated to assess the associations between telomere parameters and age. Linear regression lines were fitted to visually assess the associations between LTL and %LTL<3 kb with age and cubic regression lines for the associations between LTL and %LTL<3 kb. The chi-square test and Student t test were applied to assess differences in plaque presence and plaque burden between men and women. The association of other participant characteristics with mean LTL, %LTL<3 kb, or total plaque burden were assessed with analysis of variance and linear regression models. Independent associations of telomere parameters with plaque burden were assessed with linear regression models adjusted for other covariates.
Distinct circulating leukocyte subsets have different mean TLs (26–28). However, because analysis of our samples revealed no association between the percentage of different leukocyte subtypes and plaque burden, this variable was not taken into account as a confounder in the statistical analysis.
The baseline prevalence of subclinical atherosclerotic lesions in the 1,459 study participants is summarized in Table 1. Subclinical atherosclerosis was determined by vascular ultrasound as the presence of plaques in the infrarenal aorta, carotid arteries, or iliofemoral arteries (plaque presence and total plaque burden), or by CT determination of CACS. Plaques in all territories and CACS ≥1 were markedly more prevalent in men (p < 0.001). Moreover, total plaque burden in participants with observed plaques was much higher in men (p < 0.001). As expected, age in both sexes was inversely associated with LTL (men: Pearson’s r = −0.108; p = 0.002; women: Pearson’s r = −0.138; p = 0.001) (Figure 1A), and directly associated with %LTL<3 kb (men: Pearson’s r = 0.071; p = 0.039; women: Pearson’s r = 0.084; p = 0.038) (Figure 1B), and LTL was significantly longer in women than in men (women: 10.19 kb ± 1.69 vs. men: 9.96 kb ± 1.84; p = 0.019). We also found an inverse association between LTL and %LTL<3 kb in men (Pearson’s r = −0.762; p < 0.001) and women (Pearson’s r = −0.751; p < 0.001) (Figure 1C).
Table 2 shows the stratification of CVD risk factors in the study population by sex and plaque burden. A higher plaque burden was associated with a worse CVD risk-factor profile, especially among men. Unadjusted LTL differed significantly across the plaque burden categories in men, with the highest mean LTL value found in men with no plaques (p-trend = 0.033). LTL showed no significant association with plaque burden among women. The %LTL<3 kb did not differ across plaque burden categories in men or women.
Sex- and age-adjusted associations between telomere parameters and CVD risk factors are shown in Table 3. The only direct significant associations after these adjustments were between %LTL<3 kb and urinary isoprostanes and serum ox-LDL in men (p = 0.007 and 0.008, respectively), and between %LTL<3 kb and serum ox-LDL in women (p = 0.001).
Associations between telomeres and atherosclerosis burden in the femoral and carotid arteries were investigated in 3 regression models: unadjusted, adjusted for age (and sex in nonstratified models), and also adjusted for 8 additional CVD risk factors (Table 4). The analysis considered plaque burden in these territories individually and in combination (total plaque burden). In the unadjusted model, LTL reached statistical significance as a determinant of femoral plaque and total plaque burden in men, and in the total sample. However, the association did not hold after adjustment for age and sex (Model 1) or additional CVD risk factors (Model 2). For %LTL<3 kb, none of the regression models showed significant associations for men, women, or the total sample.
An association between short LTL and the occurrence of cardiovascular ischemic events has been described in several studies (13,14,29); however, information is still scarce on the relationship between LTL and subclinical atherosclerosis in subjects free from established CVD. The short telomere load correlates with the number of senescent cells (17), suggesting that an association between CVD and the number of short telomeres might be better than average LTL at predicting aging and associated atherosclerosis. However, previous studies have not investigated this possible association. Here, we analyzed LTL, as well as short telomere load (%LTL<3 kb), in a subset of 1,459 middle-aged CVD-free participants from the PESA cohort (2,21) (Central Illustration). Our analysis shows that neither high %LTL<3 kb nor short mean TL in circulating leukocytes is an independent determinant of subclinical atherosclerosis after adjusting for age or other well-known cardiovascular risk factors.
In 2 of 3 cross-sectional studies performed in relatively small cohorts (from 129 to 325 subjects), short LTL showed an association with asymptomatic atherosclerosis estimated by coronary artery calcification (30–32). In a 6-year prospective study of common carotid IMT in 768 participants, telomere shortening showed an association with increased subclinical carotid vascular damage and worse cardiovascular prognosis (33). In a cohort of 800 subjects, short LTL also emerged as a significant and independent predictor for the risk of advanced atherosclerosis and its complications (myocardial infarction and stroke), but was not associated with early atherogenesis (34). Similarly, the large-scale Asklepios study (2,509 relatively young volunteers free from established CVD) found no or only limited independent association between average LTL and IMT or plaque presence in carotid or femoral arteries in men and women (15). By analyzing different vascular territories in 1,459 PESA participants and performing a more extensive characterization of subclinical atherosclerosis burden, which included not only the presence/absence of plaques, but also the continuous variable “plaque volume,” our studies reinforce the notion that average LTL is not associated with subclinical atherosclerosis. In addition, we show for the first time that the abundance of short telomeres in leukocytes is not an independent determinant of subclinical atherosclerosis in carotid or femoral arteries. It would be important, however, to carry out longitudinal studies with the same participants analyzed herein to ascertain whether telomere attrition over time is associated with cardiovascular ischemic events. As expected, men in our cohort had a higher prevalence of subclinical atherosclerosis than women. This might explain the significant association of total plaque burden with LTL in men, but not in women, in the unadjusted model. Indeed, the overall young age of the study population might help to explain the lack of association between LTL and plaque burden, especially for women, who generally have a lower atherosclerosis burden than men of the same age.
Telomere dysfunction induces metabolic and mitochondrial compromise (35), and increased burden of oxidative stress can accelerate LTL attrition (12,29). Consistent with the results of multivariate regression analyses showing that femoral artery plaques are independently related to serum ox-LDL (36), we found significantly higher serum ox-LDL concentrations among participants with a higher plaque burden. Although we did not observe significant association between mean LTL and serum ox-LDL, interestingly, the %LTL<3 kb was significantly higher in men and women with higher serum ox-LDL levels. These results suggest that accumulation of short telomeres correlates with abnormal lipid accumulation in blood; therefore, longitudinal follow up of PESA participants will be of high interest.
First, our sample represents a homogeneous middle-aged and sociodemographic-specific population, and thus might not be representative of the general population; different results might be obtained in studies carried out with participants who are older (and thus have a larger extent of subclinical atherosclerosis) or who have different ethnic or demographic profiles (13,14,29). Second, our results are limited by the cross-sectional nature of the study. Prospective or longitudinal studies of large samples are clearly needed to establish with confidence whether there is an association between telomere parameters and the progression of subclinical atherosclerosis, and whether such an association could help identify subjects at high risk of developing CVD before symptoms appear (e.g., increased atherosclerosis progression, and risk of myocardial infarction and stroke in subjects with higher rates of LTL attrition and/or more accumulation of short telomeres over time). Indeed, telomere shortening and accumulation of short telomeres over time has been proposed to be predictive of mortality in birds and mammals (37–39).
The results of our cross-sectional study of middle-aged men and women free from established CVD show that short mean telomere length and the amount of short telomeres (<3 kb) in leukocytes are not significant independent determinants of subclinical atherosclerosis burden. Future longitudinal studies are warranted to determine whether variation in telomere length over time is a predictor of subclinical atherosclerosis progression or the occurrence of cardiovascular ischemic events.
COMPETENCY IN MEDICAL KNOWLEDGE: In a middle-aged population without clinical CVD, average LTL and short telomere load are not significant independent determinants of subclinical atherosclerosis.
TRANSLATIONAL OUTLOOK: Prospective studies are needed to determine whether subjects with higher rates of telomere length attrition or accumulation of short telomeres exhibit accelerated progression of atherosclerosis or an increased risk of ischemic events.
The authors thank all the participants in the PESA study, and to the PESA technical staff for their excellent work. The authors are thankful to their team of imaging technicians (Alberto Ávila, Aurora del Barrio, Ángel Macías, Rosario Pérez, Braulio Pérez), nurse (Maite Rodríguez), and radiologists (Ana Álvarez, Estefanía Fernández, Susana Linares) for their outstanding work on the high-quality data acquisition, special care of participants, and supervision of the imaging studies. The authors also thank the epidemiology technicians (Carolina Rojas, Estrella Rubio, María J. Diego, Natalia Serrano) for their excellent work on inclusion and retrieval of information. Laboratory data were available thanks to the notable work of Ricardo Ponce, Juan J. Álvarez, and Marta Rodríguez. Skillful work on the development of the PESA database (Sergio Cárdenas), and further statistical analyses and data discussion (Belén Oliva, José L. Peñalvo, José M. García) is also very much appreciated. Special thanks are extended to the medical team at Banco Santander Health Services (Agustín Morcoroa, Alicia C. Muñoz, Baltasar Furones, Catalina Ortega, Juan Muñoz, Laura Gómez, María A. Serrano) for their crucial work in collecting participants’ health status. The authors also thank the administrative team at Banco Santander Health Services (Emilio Marín, Esther Graciano, José I. Leirana, Magdalena Inmaculada), and the CNIC (Ana Gutiérrez) for ensuring optimal coordination between the research teams and the PESA participants. The authors are grateful to Alberto Sanz (CNIC General Manager) for his enthusiasm and collaboration, the CNIC administrative team for their continuous support, Simon Bartlett for English editing, and Anna de Andrés for help preparing the Central Illustration.
This study was supported by Instituto de Salud Carlos III (Grant RD12/0042/0028 to Dr. Andrés), with cofunding from Fondo Europeo de Desarrollo Regional (FEDER), and by a noncompetitive, unrestricted grant shared equally between the CNIC and Banco de Santander. The PESA study is a noncommercial study that is independent of the health care and pharmaceutical industry. The Spanish Ministry of Economy and Competitiveness (MINECO) and the Pro CNIC Foundation support the CNIC. The MINECO RETOS Program and the Fundación Botín fund work in Dr. Blasco′s group. The CNIC and CNIO are recipients of Centro de Excelencia Severo Ochoa awards from MINECO (SEV-2015-0505 and SEV-2015-0510, respectively). The authors have reported that they have no relationships relevant to the contents of this paper to disclose. Peter M. Nilsson, MD, served as Guest Editor for this paper.
- Abbreviations and Acronyms
- %LTL<3 kb
- percent short telomeres (<3 kb)
- coronary artery calcium score
- computed tomography
- cardiovascular disease
- deoxyribonucleic acid
- high-throughput quantitative fluorescence in situ hybridization
- intima-media thickness
- leucocyte telomere length
- oxidized low-density lipoprotein
- telomere length
- Received November 25, 2015.
- Revision received March 10, 2016.
- Accepted March 12, 2016.
- American College of Cardiology Foundation
- Mozaffarian D.,
- Benjamin E.J.,
- Go A.S.,
- et al.,
- American Heart Association Statistics Committee and Stroke Statistics Subcommittee
- Fernández-Friera L.,
- Peñalvo J.L.,
- Fernández-Ortiz A.,
- et al.
- Sanders J.L.,
- Newman A.B.
- Kovacic J.C.,
- Moreno P.,
- Hachinski V.,
- et al.
- Fuster J.J.,
- Andrés V.
- D'Mello M.J.,
- Ross S.A.,
- Briel M.,
- et al.
- Haycock P.C.,
- Heydon E.E.,
- Kaptoge S.,
- et al.
- De Meyer T.,
- Rietzschel E.R.,
- De Buyzere M.L.,
- et al.,
- Asklepios Study Investigators
- Sillesen H.,
- Muntendam P.,
- Adourian A.,
- et al.
- Stein J.H.,
- Korcarz C.E.,
- Hurst R.T.,
- et al.
- Canela A.,
- Vera E.,
- Klatt P.,
- et al.
- Rufer N.,
- Brümmendorf T.H.,
- Kolvraa S.,
- et al.
- Spyridopoulos I.,
- Erben Y.,
- Brummendorf T.,
- et al.
- Spyridopoulos I.,
- Hoffmann J.,
- Aicher A.,
- et al.
- Willeit P.,
- Willeit J.,
- Brandstätter A.,
- et al.
- Langlois M.R.,
- Rietzschel E.R.,
- De Buyzere M.L.,
- et al.,
- Asklepios Investigators
- Heidinger B.J.,
- Blount J.D.,
- Boner W.,
- et al.