Accelerated Telomere Shortening in Leukocyte Subpopulations of Patients With Coronary Heart Disease
Role of Cytomegalovirus Seropositivity
Background— Shortening of mean telomere length (TL) in white blood cells is correlated with the development of coronary heart disease (CHD) and with increased mortality due to infectious disease. The goal of the present study was to investigate whether telomere shortening in CHD is restricted to specific peripheral blood lymphocyte and/or myeloid cell subpopulations. Results were correlated to TL in CD34+ hematopoietic peripheral blood stem cells and progenitor cells obtained from the same individual patients.
Methods and Results— TL was measured by multicolor flow cytometry–fluorescent in situ hybridization in 12 leukocyte subpopulations after immunomagnetic bead sorting. We investigated TL in 14 young (mean age 25 years) and 13 older (mean age 65 years) healthy male volunteers and in 25 age-matched patients with CHD (mean age 65 years). We show that TL in granulocytes and monocytes mirrors TL of CD34+ peripheral blood stem cells and progenitor cells extremely well (r=0.95, P<0.0001) in patients and in healthy adults. TL was ≈0.5 kilobases (kb) shorter in leukocytes from patients with CHD than in their age-matched control subjects. This difference was identical for CD34+ peripheral blood stem cells and progenitor cells, monocytes, granulocytes, B lymphocytes, and CD4+ T cells, including their memory and naïve subpopulations. Surprisingly, only in cytotoxic CD8+ T lymphocytes, we found a substantially increased TL deficit of 1.0 kb in CHD patients as opposed to control subjects. Further analysis revealed that TL shortening was particularly pronounced in CD8+CD28− T cells obtained from cytomegalovirus-seropositive CHD patients, whereas such a difference was not observed in healthy cytomegalovirus-positive as opposed to cytomegalovirus-negative control subjects. Finally, TL shortening of CD8+CD45RA+ T cells was correlated with the decrease in left ventricular function in CHD patients (r=0.629, P=0.001).
Conclusions— Telomere shortening in patients with CHD could potentially be attributed to either inherited TL shortening or acquired accelerated telomere shortening restricted to the hematopoietic system, which affects the baseline TL of all peripheral blood cell populations, including peripheral blood stem cells and progenitor cells. In addition, cytomegalovirus-seropositive patients but not healthy control subjects exhibited further shortening of their cytotoxic T lymphocytes. Surprisingly, TL shortening of CD8+ T lymphocytes in CHD patients demonstrated a very strong correlation with cardiac dysfunction, which suggests a mechanistic link between CHD and immunosenescence.
Received January 28, 2009; accepted July 28, 2009.
Individuals with shorter telomeres of their peripheral blood leukocytes carry a higher risk for dying of infectious and cardiovascular disease.1 Naturally, both disease entities appear to represent different mechanisms in which telomere biology is involved. On closer examination, inflammation plays a crucial role in the development of atherosclerosis.2–4 Next to oxidized cholesterol, one of the proposed stimuli of inflammation is infectious pathogens such as cytomegalovirus (CMV). Despite a high prevalence, the association between the presence of serum antibodies against CMV and atherosclerosis in humans is still debated.5–7 It is also highly debated whether telomere shortening in patients with coronary heart disease (CHD) is an acquired or an inherited condition. In vitro studies suggest that increased oxidative stress, such as that which is present in patients with CHD, leads to accelerated telomere shortening.8 Brouilette and coworkers,9 however, found that leukocyte telomere shortening already existed in healthy offspring of patients with CHD compared with age-matched control subjects. This single study provides strong arguments in favor of a genetic link for CHD-associated telomere shortening, called the “telomere hypothesis.” Van der Harst and coworkers10 have recently shown that telomere length (TL) is shorter in patients with congestive heart failure than in age-matched control patients. Telomere shortening was also related to the severity of disease. We have shown previously that ischemic cardiomyopathy is associated with selective impairment of progenitor cell function in the bone marrow and in the peripheral blood, which may contribute to an unfavorable left ventricular remodeling process.11
Clinical Perspective on p 1372
The goal of the present study was to gain a deeper insight into the mechanisms involved and the cell populations responsible for telomere shortening in CHD. We used a different approach than Brouilette et al by determining TL separately for all relevant leukocyte subpopulations, including CD34+ hematopoietic peripheral blood stem and progenitor cells (PBPCs), and compared them between patients with CHD and age-matched healthy control subjects.
Please see the online-only Data Supplement for details of the methods used in the present study.
We analyzed 25 men (mean age 64.8±3.4 years) with angiographic documentation of CHD and healed myocardial infarction (at least 3 months after infarction). All patients included in the study had impaired left ventricular systolic function (ejection fraction <55%) as confirmed by echocardiography. None of the patients were affected by neoplastic, autoimmune, or chronic infectious disease. All subjects with recent infections were also excluded.
Fourteen young (mean age 26.7±1.8 years) and 13 older (65.1±2.1 years) healthy volunteers were included in the present study after informed consent was obtained. The health status of young and elderly control subjects was determined on the basis of both medical history questionnaire and blood tests, which comprised complete blood cell counts, glucose, hemoglobin A1c, urea, creatinine, electrolytes, high-sensitivity C-reactive protein, liver function tests, high-density lipoprotein and low-density lipoprotein cholesterol, triglycerides, and N-terminal pro-brain natriuretic peptide (NT-proBNP). Additionally, all elderly control subjects underwent extensive clinical testing that included 2-dimensional, M-mode, and Doppler echocardiography, as well as stress electrocardiography for the exclusion of subclinical heart disease.
Determination of CMV Status
The CMV serostatus of 13 elderly healthy control subjects and 20 CHD patients was determined from serum samples with the Enzygnost Anti-CMV/IgG enzyme immunoassay (Dade Behring, Marburg, Germany) according to the manufacturer’s instructions.12 For the determination of human CMV DNA load in peripheral blood, we used a TaqMan real-time quantitative polymerase chain reaction assay.13
Isolation of Leukocyte Subpopulations
Peripheral blood EDTA samples (100 mL) were collected from CHD patients and healthy volunteers (EDTA S-Monovette 9 mL, catalog No. 02-1066-001, Sarstedt, Nümbrecht, Germany). Peripheral blood mononuclear cells were obtained after density gradient centrifugation with Ficoll-Hypaque (catalog No. L6115, Biochrom, Berlin, Germany). After 2 washes with PBS, peripheral blood mononuclear cells were resuspended in ice-cold MACS buffer (Miltenyi Biotec, Bergisch Gladbach, Germany) that contained PBS with 0.5% FCS and 2 mmol/L EDTA (Titriplex III 1.1%). The cells were counted in trypan blue with a Neubauer hemocytometer. For isolation of different mononuclear cell subpopulations, we designed a cell-fractionation strategy based on immunomagnetic sorting (MACS; see online-only Data Supplement Figure I). All separation steps were performed with LS columns (catalog No. 130-042-401) placed in the QuadroMACS Separator (MACS MultiStand, catalog No. 130-042-303).
Flow–Fluorescent In Situ Hybridization and Multicolor Flow–Fluorescent In Situ Hybridization
For analysis of mean TL in subpopulations of peripheral leukocytes, we used either the standard flow cytometry–fluorescent in situ hybridization (flow-FISH) protocol (for purified CD14+ monocytes, CD8+CD28+CD56− T cells, and CD8+CD28−CD56− T cells) or a multicolor flow–FISH approach (CD15+ granulocytes, CD45RO+ T cells, CD45RA+ T cells, and CD19+ B cells; for CD34+ progenitor cells, see the online-only Data Supplement, Figure II) as previously described with minor modifications.14,15 Mean TL in subpopulations of peripheral blood leukocytes was measured with flow-FISH in accordance with previously published protocols.16 All monoclonal antibodies used for multicolor flow–FISH in the present study were conjugated with heat-stable Alexa Fluor 647 (AF647) fluorochrome. The interindividual variability of TL was 5.5%, and intraindividual variability was 1.0% (online-only Data Supplement Figure III).
Telomerase Enzyme Activity Measurement
Telomerase enzyme activity was measured with a commercially available polymerase chain reaction–based assay (TRAPeze ELISA telomerase detection kit, Chemicon International, Temecula, Calif; No. S7750) according to the manufacturer’s protocol. Telomeric repeat-amplification protocol assays were performed with biotin-labeled TS primers as described previously.17
Statistical Power Analysis
We calculated the effect size (Δ) between CHD patients and age-matched control subjects that could be detected with 80% probability at a 5% significance level for each of the main cell populations: CD14 0.66 kilobases (kb), CD15 0.64 kb, CD34 0.69 kb, CD19 0.93 kb, CD3 0.88 kb, CD4 0.81 kb, and CD8 1.02 kb. From these calculations, we concluded that because the actual differences in TL were smaller (0.5 kb), the lack of a significant difference between TL of patients and age-matched control subjects for most cell populations could have been due to the limited sample size rather than the lack of a biological difference between the groups.
Data depicted in box-and-whisker plots (Figures 1 and 2⇓) are reported as median values, lower and upper quartiles, and 10%/90% percentiles to visualize dispersion. In the text, data are reported as mean±SE. Comparison of 3 means was performed by ANOVA, followed by a Tukey post hoc test for comparison of any 2 groups (patients versus age-matched control subjects and young versus elderly healthy volunteers). Comparison of 2 groups was calculated with an unpaired t test if normal probability plots (P-P plots) demonstrated approximate normality. For multiple testing (ANOVA, 2-factor design), the general linear model tool in SPSS (SPSS Inc, Chicago, Ill) was used. The linear relationship between 2 variables was calculated by the bivariate correlation procedure with the Pearson correlation coefficient. C-reactive protein and NT-proBNP levels were analyzed as continuous variables and were log transformed to take into account their skewed distribution. Comparison of slopes and intercepts in Figure 3 was performed by the ANCOVA procedure in GraphPad Prism version 5 for Macintosh (http://www.graphpad.com). All statistical tests except ANCOVA were performed with SPSS version 15.0 for Windows. Power analysis was performed with StatMate 2 for Macintosh by GraphPad Software Inc (San Diego, Calif).
We measured TL in 3 different study populations. To monitor the influence of chronological aging on TL in leukocyte subpopulations, we first analyzed 14 young (27±2 years old) and 13 elderly (65±2 years old) completely healthy volunteers (Table). Healthy volunteers were all subjected to blood tests, echocardiography (elderly volunteers), and a full ECG stress test to exclude individuals with structural or subclinical heart disease. Twenty-five patients with CHD were age matched (65±3 years old) with these volunteers and underwent extensive testing for laboratory chemistry, echocardiography, and coronary angiography. All patients had had a myocardial infarction in the past.
TL in Myeloid Blood Cell Populations Is Shorter in Patients With CHD
We isolated CD14+ monocytes by immunomagnetic bead sorting and subjected these cells to the flow-FISH protocol. In addition, we used the multicolor flow–FISH protocol to determine TL in CD15+ granulocytes. Owing to the very limited number of CD34+ cells in immobilized peripheral blood, 100 mL of peripheral blood had to be taken to first enrich cells with CD34+ PBPCs and then process them with the multicolor flow–FISH, with a second antibody raised to a different epitope. As expected, TL of all 3 populations was 1.2 kb shorter in older than in younger healthy volunteers (P<0.001; Figure 1A through 1C). In addition, TL in CHD patients was ≈0.5 kb shorter in monocytes (P=0.085; Figure 1A) and granulocytes (P=0.059; Figure 1B) than in age-matched healthy control subjects. We also found an excellent correlation between TL in CD34+ PBPCs and their peripheral blood derivates, monocytes (Pearson’s correlation coefficient 0.96, P<0.001; Figure 1D) and granulocytes (Pearson correlation coefficient 0.95, P<0.001; Figure 1E), independent of age and CHD. This proves that the TL of peripheral blood granulocytes and monocytes reflects the TL of CD34+ PBPCs.
Lymphocyte Telomere Shortening in CD8+ T Cells Is Accelerated in CHD
There was no difference in the absolute number of lymphocytes, CD4 helper cells, or CD8 cytotoxic lymphocytes among the 3 groups (online-only Data Supplement Table I). Although not statistically significant, we noted a trend toward an increased CD4/CD8 ratio with age and CHD (P=0.25). TL in all lymphoid subpopulations was 1.8 kb shorter in the older control group than in the younger volunteers (P<0.001; Figures 2A through 2D and 3A and 3⇑B). Similar to our findings in myeloid cells, TL in CD4+ T cells (Figure 2C) and CD19+ B lymphocytes (Figure 2A) was ≈500 base pairs (bp) shorter in CHD patients than in control subjects, although this difference did not reach statistical significance. Surprisingly, the TL difference in CD8+ cytotoxic T lymphocytes between CHD patients and healthy control subjects exceeded that in other populations (1.0 kb, P=0.017; Figure 2D). This was mainly attributed to TL shortening in “senescent” CD8 T cells, which lack the costimulatory receptor CD28 (CD8+CD28−; P=0.02; Figure 2F). Finally, we compared naïve and memory CD4 T cells in patients with CHD and young and elderly control subjects (Figure 3A through 3D). Memory CD4 T cells (CD4+CD45RO+) were all ≈2 kb shorter than their naïve counterparts (CD4+CD45RA+), independent of the presence of CHD or age.
Telomere Shortening in CHD Affects All Cells Equally
Figure IVA in the online-only Data Supplement and Figure 4A and 4B herein summarize TL shortening in healthy volunteers across all main leukocyte subpopulations, which suggests that TL decreases at a similar rate in B and T lymphocytes (≈50 bp/year). Telomere attrition in both granulocytes and monocytes occurs more slowly with age (≈30 bp/year). Thus far, the results of the present study strongly suggest that in healthy males, age-dependent telomere shortening is primarily dependent on the TL of myeloid and lymphoid progenitors. Figure IVB in the online-only Data Supplement demonstrates that leukocyte telomere shortening in patients with CHD is essentially due to identical attrition rates in granulocytes, monocytes, B lymphocytes, and CD4 T cells. Interestingly, cytotoxic T lymphocytes represented the only exception to this pattern (Figure 4B). When naïve CD4 T cells were used as an internal standard, the difference in TL between these cells and cytotoxic CD8+ T lymphocytes (CTLs) increased from 0.79±0.81 kb in age-matched healthy volunteers to 1.49±0.65 kb in CHD patients (P=0.009), which suggests a different mechanism for telomere shortening in CD8 CTLs from patients with CHD.
CMV-Positive Patients and Control Subjects Show Increased Numbers of CD28-Negative T Cells
Recent publications have shown that chronic infection with CMV can lead to T-cell memory inflation and hence an accumulation of CMV-specific CD8+CD28− T cells with shorter telomeres.18 Because we had found accelerated telomere shortening specifically in CD8+CD28− T cells from patients with CHD (Figure 2F), we retrospectively analyzed frozen plasma from patients and control subjects in the present study for CMV IgG antibody. Thirteen of 20 patient samples were positive for CMV, whereas 5 of 13 control samples appeared positive (χ2 test P=0.05; online-only Data Supplement Table II). Acute or reactivated CMV infection was ruled out by analysis of CMV-specific DNA copy number, which was negative in all participants. Age distribution was similar among CMV-negative and -positive subjects. The CD4/CD8 ratio was lower in CMV-positive control subjects (1.3±0.9 versus 2.1±1.0) and patients (1.8±1.2 versus 2.8±1.2, P=0.044; online-only Data Supplement Table II). CMV-positive patients also had a higher C-reactive protein level (1.45±1.5 versus 0.37±0.35 mg/dL, P=0.03; online-only Data Supplement Table II) and a lower ejection fraction (25.5±9.5% versus 37.1±11.5%, P=0.04; online-only Data Supplement Table II). In both groups, seropositive patients and control subjects revealed an increase in the absolute number of CD28-negative T-helper cells and in cytotoxic T lymphocytes (Figure 5A). CMV-seropositive patients with CHD showed a significant increase in CD4+CD28− cells (75±85 versus 4±2 cells/μL, P=0.01), CD8+ T cells (566±298 versus 331±137 cells/μL, P=0.03), and CD8+CD28− CTLs (320±243 versus 67±39 cells/μL, P=0.003; Figure 5A). Using 2-way ANOVA testing, we found a CMV-dependent increase in CD8+ T cells (P=0.009), CD8+CD28− CTLs (P=0.002), and CD4+CD28− cells (P=0.006) but not in total CD4+ T cells (P=0.82; online-only Data Supplement Table III), which reflects the general increase in CD28-negative T cells among CMV-seropositive individuals, both patients and control subjects.
CMV-Seropositive Patients but Not Control Subjects Show TL Attrition in CD8+ CTLs
We reanalyzed our existing data on TL in different leukocyte populations in reference to CMV status (Figure 5B). Surprisingly, healthy control subjects did not show any difference in TL from lymphocyte subsets between seropositive and seronegative subjects, as Figure 5B shows. In contrast, CMV-positive CHD patients had significantly shorter telomeres in their CD8+ T cells (5.18±0.99 versus 6.00±0.55 kb, P=0.025; Figure 5B) and CD8+CD28− CTLs (4.37±0.98 versus 5.52±0.57 kb, P=0.006; Figure 5B) but not in other lymphocyte populations, including CD8+CD28+ CTLs. TL in CD34+ progenitor cells and in CD15+ granulocytes was unaffected in CMV-positive patients (Figure 5B).
Left Ventricular Function Strongly Correlates With TL in Lymphocytes
Finally, we attempted to evaluate whether shortening of TL from CMV-seropositive patients could have any effect on the underlying disease in patients with CHD. Because all patients had had a myocardial infarction in the past, we wanted to know whether TL in any cell population was associated with the extent of left ventricular dysfunction in this group. To our surprise, TL from all leukocyte populations, which presumably contained CMV-specific lymphocytes, held a positive correlation with ejection fraction (online-only Data Supplement Table IV). In contrast, TL of myeloid cell populations or age both did not show any correlation with ejection fraction. The greatest correlation between TL and ejection fraction was seen for the CD8+CD45RA+ lymphocyte population (Pearson correlation coefficient 0.629, P=0.001; Figure 5C), which largely overlaps with the CD8+CD27−CD28−CD45RA+ cell population in CMV-positive patients, the phenotype of CMV-specific CTLs.19 These results suggest a link between CD8+ T cells and the deterioration of myocardial function in CMV-seropositive patients with CHD. Interestingly, CD8+CD28− cells could not reactivate telomerase on repeated ligation of the T-cell receptor with anti-CD3 (Figure 5D).
Our investigations on TL distribution in leukocyte subpopulations of patients with CHD as opposed to young and elderly healthy control subjects yielded 3 important novel results. First, leukocyte TL in patients with CHD was 500 bp shorter than in age-matched control subjects, a difference that appeared to be highly conserved throughout the hematopoietic system. In addition, TL in cytotoxic CD8+ T cells (CTLs) was 1000 bp shorter than in control subjects. We found the most pronounced degree of telomere shortening in CD8+CD28− T cells in CMV-seropositive CHD patients, but not in CMV-positive healthy volunteers. Finally, TL shortening of CD8+ T cells demonstrated a strong correlation with the decrease in left ventricular function in CHD patients. The results of the present study strongly suggest that telomere shortening in leukocyte subpopulations from patients with CHD is either inherited or reflects accelerated telomere shortening in common precursors of the leukocyte subpopulations, including PBPCs, the hematopoietic stem cell compartment. Whether in patients with CHD, the copresence of CMV accelerates immunosenescence and imposes on the course of CHD itself will need to be demonstrated in future studies.
We present for the first time a detailed analysis of TL in 12 leukocyte subsets, including CD34+ PBPCs, from young healthy volunteers and from elderly patients with CHD and their age-matched healthy control subjects. The results of this investigation could ultimately enable us to define the primary affected cell population accountable for accelerated telomere shortening in patients with established CHD. To achieve this goal, a large amount of peripheral mononuclear cells had to be gathered to harbor a sufficient number of CD34+ PBPCs that could be subjected to TL measurement. Furthermore, our previously used flow-FISH method20,21 had to be refined to allow measurement of TL in conjunction with determination of surface antigens (multicolor flow–FISH). To limit the number of study participants required, we included only male subjects and performed extensive noninvasive testing for each volunteer before their inclusion in the study, which greatly facilitated characterization and homogeneity of the study populations. The results revealed a number of unexpected findings, all relevant to the interpretation of TL data in patients with CHD. The majority of telomere studies performed thus far on patients with cardiovascular disease have used terminal restriction fragment analysis methods on the total leukocyte population, which does not allow for a sophisticated analysis of leukocyte subpopulations.
Age-Dependent Influence on Leukocyte Subpopulations
It has been shown in multiple studies that leukocyte TL declines with age.22–24 We and others have also shown previously that TL shortening in lymphocytes occurs at a higher rate than in granulocytes.25–27 We confirm in the present study that lymphoid cell populations clearly demonstrate a faster rate of telomere shortening than myeloid cells during the same life span. Interestingly though, B lymphocytes, CD4 T cells, and CD8 T cells all showed the same rate of telomere shortening in healthy individuals. These results suggest that despite the complexity of the immune system, with clonal expansion and contraction of subpopulations, there is a common denominator for basic TL. The results of the present study strongly suggest that telomere shortening of a common progenitor of B lymphocytes and T lymphocytes in a healthy person determines basic TL. The generation of CD4 memory T cells leads to a substantial shortening of 2000 bp, which corresponds to ≈20 cell divisions. This TL difference was almost identical in young, elderly, and unhealthy study participants and was independent of the basic TL. These findings suggest that telomere shortening in CD4 memory cells reflects a general mechanism that is solely dependent on proliferation kinetics rather than on disease or age.
CD34+ PBPCs in CHD
Vaziri and coworkers22 showed that candidate human stem cells with a CD34+CD38lo phenotype that were purified from adult bone marrow have shorter telomeres than cells from fetal liver or umbilical cord blood. Together with findings by Notaro and coworkers28 that demonstrated that concomitant with their proliferation, hematopoietic progenitor cells lose TL, these findings indicate that the origin of leukocyte telomere shortening occurs at the level of their bone marrow–residing precursors. Therefore, it would appear very attractive to determine the TL of hematopoietic progenitor cells by obtaining TL of peripheral leukocyte populations. Although 1 study has found a strong correlation between PBPCs and peripheral blood leukocytes, these data were generated in mobilized patients with malignancies.29 We have shown previously that TL in mononuclear bone marrow cells correlates extremely well with peripheral blood granulocytes in patients with CHD.21 In the present study, we were able to find a very strong correlation (r>0.95) between TL in PBPCs and monocytes or granulocytes. Our findings in completely healthy volunteers were also independent of age, showing for the first time that TL of peripheral myeloid cells can be used as an indicator for TL in hematopoietic progenitor cells in healthy persons.
Patients with CHD have been shown to have shorter telomeres than age-matched healthy control subjects.30 Therefore, the finding that patients in the present study had telomere shortening of ≈500 bp compared with their age-matched control group is well in line with previous reports. It has been debated intensely whether this difference in TL is the result of increased oxidative stress over time in patients with severe atherosclerosis, which would render TL a surrogate parameter for “cumulative oxidative stress.” Nevertheless, triggered by several recent publications, another pathophysiological concept has rapidly gained importance, the so-called telomere hypothesis that shorter telomeres will contribute in part to the pathogenesis of CHD.9,31 In the first randomized case-control study, Brouilette and colleagues31 have demonstrated the correlation between TL and the risk of developing CHD. In favor of the telomere hypothesis is another of their studies showing that mean TL in healthy offspring of subjects with CHD is shorter than in offspring of healthy control subjects.9 The present data (Figure 4; online-only Data Supplement Figure IV) point in the same direction, using a very different experimental approach. We found that TL was shortened by ≈500 bp in CHD patients compared with healthy age-matched individuals in the present study. Interestingly, this difference was maintained throughout most cell types (with cytotoxic T cells as the only exception), independent of their myeloid or lymphoid origin. If the difference of 500 bp were simply due to accelerated aging, we would expect a similar constellation in TL differences, such as between young and old healthy individuals, in whom lymphoid cell populations show faster telomere attrition than myeloid cells. This was not the case, however, because we found that except for CD8+ T cells, all cell populations appeared to have similar shortening, independent of their myeloid or lymphoid origin. We believe, therefore, that the present data argue strongly in favor of leukocyte TL shortening as a predisposing parameter in CHD rather than a coincidental surrogate.
Telomere Shortening of CD8+ CTLs and Role of CMV
Accelerated telomere shortening of CD8+ T cells on top of “inherited” shortening in patients with CHD in the present study occurred mainly within the CD8+CD28− subpopulation. CD8+ T cells of this phenotype circulate at high frequencies in CMV-infected people.32 Most CMV-specific CD8+ memory T cells are reexpressing CD45RA+ and lack the costimulatory molecules CD27 and CD28.19,33 Therefore, we retrospectively analyzed the CMV status of our patients and control subjects and found that seropositive patients but not seropositive control subjects showed accelerated telomere shortening in CD8+CD28− or CD8+CD45RA+ T cells, respectively. Different possibilities arise from this finding: (1) CMV seropositivity could represent a predisposition for CHD; (2) the shorter basic TL in patients with CHD could trigger a different response to chronic CMV antigen presentation; and (3) more likely, in patients who survive the stress situation of an acute myocardial infarction, a different and stronger response of the immune system is triggered. In favor of the latter hypothesis is our finding that ejection fraction was lower in CMV-positive patients and correlated strongly with TL of CD8+CD45RA+ T cells. Interestingly, all cell populations that are thought to contain predominantly CMV-specific cells demonstrated a positive correlation with ejection fraction (online-only Data Supplement Table IV). This could suggest that patients with larger infarcts (and hence lower ejection fraction) had a stronger CMV-related reaction. In line with other studies, we also found an increased number of CD4+CD28− and CD8+CD28− T cells in CMV-seropositive control subjects and CHD patients.34 Nevertheless, because of the very small proportion of CD28-negative CD4+ T cells compared with CD8+CD28− T cells in CMV-seropositive CHD patients, the presumed telomere loss in these cells cannot affect the TL of CD4 cells to the extent it does in CD8 cells. A landmark study by Fletcher and coworkers35 investigated the phenotype (CD27−CD28−) and TL of CMV-specific CD4+ T cells. Those authors found that the latter differentiate more rapidly than other populations of memory T cells, which suggests that the extent of CMV reactivation might correlate with the extent of telomere loss.
A significant number of clinical studies have been published in the past that investigated the controversial association between CHD and CMV infection.36–38 Interestingly, the coexistence of inflammation in addition to seropositivity for CMV, reflected by elevated serum levels of either C-reactive protein or interleukin-6, represented the most important predictor of survival. Patients who were CMV seropositive without coexisting inflammation basically failed to show an adverse prognosis. In the present study, patients were proven to already have rather advanced disease, reflected by previous myocardial infarction, severely reduced left ventricular function, and elevated NT-proBNP serum levels. In addition, high-sensitivity C-reactive protein levels were higher in the CMV-seropositive patients. Our understanding is, therefore, that CMV seropositivity in otherwise healthy individuals does not confer a higher risk for the development of CHD or associated complications unless there is ongoing and persistent reactivation of the virus. In this case, one would expect either increased levels of C-reactive protein or interleukin-6 or other evidence of systematic inflammation. The present study investigated the TL of CD8+ T cells and their subpopulations. Accelerated telomere shortening in lymphocytes could be the result of persistent (eventually lifelong) reactivation/proliferation of these cells and therefore represents an ideal parameter to understand cumulative CMV burden. Future prospective studies with sufficient statistical power will be required to prove our hypothesis.
The lack of a significant difference between the TL of patients and age-matched control subjects for several cell populations could have been due to the limited sample size rather than the lack of biological difference between the groups. The study also was not powered for the discrimination of TL between CMV-seropositive and -seronegative subpopulations. Future studies must include patients on a larger scale to gain more evidence, potentially including CMV-specific CTLs. Only approximately 70% of our measurements of CD34+ PBPCs gated for TL analysis were performed on more than 1000 events, a number assumed to be necessary to obtain a reliable TL measure (online-only Data Supplement Figure V). Nevertheless, the lack of correlation between the small amount of gated cells and a potential increase/decrease in the TL of PBPCs does not suggest largely incorrect assumptions of TL in the residual 30% of measurements on fewer than 1000 events.
The assistance of Carmen Schön and Natalja Reinfeld is greatly appreciated.
Sources of Funding
This work was supported by grants of the Deutsche Forschungsgesellschaft (DFG Sp-502/4-2) and by the European Union (EVGN). Dr Brümmendorf is supported by a grant from the Eppendorfer Krebs- und Leukämiehilfe.
Bentz GL, Yurochko AD. Human CMV infection of endothelial cells induces an angiogenic response through viral binding to EGF receptor and beta1 and beta3 integrins. Proc Natl Acad Sci U S A. 2008; 105: 5531–5536.
Serra V, von Zglinicki T, Lorenz M, Saretzki G. Extracellular superoxide dismutase is a major antioxidant in human fibroblasts and slows telomere shortening. J Biol Chem. 2003; 278: 6824–6830.
Brouilette SW, Whittaker A, Stevens SE, van der Harst P, Goodall AH, Samani NJ. Telomere length is shorter in healthy offspring of subjects with coronary artery disease: support for the telomere hypothesis. Heart. 2008; 94: 422–425.
Kissel CK, Lehmann R, Assmus B, Aicher A, Honold J, Fischer-Rasokat U, Heeschen C, Spyridopoulos I, Dimmeler S, Zeiher AM. Selective functional exhaustion of hematopoietic progenitor cells in the bone marrow of patients with postinfarction heart failure. J Am Coll Cardiol. 2007; 49: 2341–2349.
Beier F, Balabanov S, Buckley T, Dietz K, Hartmann U, Rojewski M, Kanz L, Schrezenmeier H, Brummendorf TH. Accelerated telomere shortening in glycosylphosphatidylinositol (GPI)-negative compared with GPI-positive granulocytes from patients with paroxysmal nocturnal hemoglobinuria (PNH) detected by proaerolysin flow-FISH. Blood. 2005; 106: 531–533.
Haendeler J, Hoffmann J, Diehl JF, Vasa M, Spyridopoulos I, Zeiher AM, Dimmeler S. Antioxidants inhibit nuclear export of telomerase reverse transcriptase and delay replicative senescence of endothelial cells. Circ Res. 2004; 94: 768–775.
Plunkett FJ, Franzese O, Finney HM, Fletcher JM, Belaramani LL, Salmon M, Dokal I, Webster D, Lawson AD, Akbar AN. The loss of telomerase activity in highly differentiated CD8+CD28-CD27- T cells is associated with decreased Akt (Ser473) phosphorylation. J Immunol. 2007; 178: 7710–7719.
Spyridopoulos I, Haendeler J, Urbich C, Brummendorf TH, Oh H, Schneider MD, Zeiher AM, Dimmeler S. Statins enhance migratory capacity by upregulation of the telomere repeat-binding factor TRF2 in endothelial progenitor cells. Circulation. 2004; 110: 3136–3142.
Spyridopoulos I, Erben Y, Brummendorf TH, Haendeler J, Dietz K, Seeger F, Kissel CK, Martin H, Hoffmann J, Assmus B, Zeiher AM, Dimmeler S. Telomere gap between granulocytes and lymphocytes is a determinant for hematopoietic progenitor cell impairment in patients with previous myocardial infarction. Arterioscler Thromb Vasc Biol. 2008; 28: 968–974.
Vaziri H, Dragowska W, Allsopp RC, Thomas TE, Harley CB, Lansdorp PM. Evidence for a mitotic clock in human hematopoietic stem cells: loss of telomeric DNA with age. Proc Natl Acad Sci U S A. 1994; 91: 9857–9860.
Kimura M, Hjelmborg JV, Gardner JP, Bathum L, Brimacombe M, Lu X, Christiansen L, Vaupel JW, Aviv A, Christensen K. Telomere length and mortality: a study of leukocytes in elderly Danish twins. Am J Epidemiol. 2008; 167: 799–806.
Rufer N, Brummendorf TH, Kolvraa S, Bischoff C, Christensen K, Wadsworth L, Schulzer M, Lansdorp PM. Telomere fluorescence measurements in granulocytes and T lymphocyte subsets point to a high turnover of hematopoietic stem cells and memory T cells in early childhood. J Exp Med. 1999; 190: 157–167.
Notaro R, Cimmino A, Tabarini D, Rotoli B, Luzzatto L. In vivo telomere dynamics of human hematopoietic stem cells. Proc Natl Acad Sci U S A. 1997; 94: 13782–13785.
Kuijpers TW, Vossen MT, Gent MR, Davin JC, Roos MT, Wertheim-van Dillen PM, Weel JF, Baars PA, van Lier RA. Frequencies of circulating cytolytic, CD45RA+CD27-, CD8+ T lymphocytes depend on infection with CMV. J Immunol. 2003; 170: 4342–4348.
Prelog M, Schwarzenbrunner N, Sailer-Hoeck M, Kern H, Koppelstaetter C, Wurzner R, Zimmerhackl LB, Brunner J. Indications for a disturbed peripheral T-cell homeostasis in juvenile idiopathic arthritis (JIA): absent expansion of CD28 T-cells and no decrease of naive T-cells in cytomegalovirus-positive patients with JIA. J Rheumatol. 2008; 35: 520–527.
Fletcher JM, Vukmanovic-Stejic M, Dunne PJ, Birch KE, Cook JE, Jackson SE, Salmon M, Rustin MH, Akbar AN. Cytomegalovirus-specific CD4+ T cells in healthy carriers are continuously driven to replicative exhaustion. J Immunol. 2005; 175: 8218–8225.
Blankenberg S, Rupprecht HJ, Bickel C, Espinola-Klein C, Rippin G, Hafner G, Ossendorf M, Steinhagen K, Meyer J. Cytomegalovirus infection with interleukin-6 response predicts cardiac mortality in patients with coronary artery disease. Circulation. 2001; 103: 2915–2921.
Muhlestein JB, Horne BD, Carlquist JF, Madsen TE, Bair TL, Pearson RR, Anderson JL. Cytomegalovirus seropositivity and C-reactive protein have independent and combined predictive value for mortality in patients with angiographically demonstrated coronary artery disease. Circulation. 2000; 102: 1917–1923.
Ridker PM, Hennekens CH, Stampfer MJ, Wang F. Prospective study of herpes simplex virus, cytomegalovirus, and the risk of future myocardial infarction and stroke. Circulation. 1998; 98: 2796–2799.
Previous studies have provided evidence of an inverse association between telomere length in peripheral blood leukocytes and cardiovascular morbidity or mortality. Using a complex cell-fractionation protocol, we were able to measure telomere length in leukocyte subpopulations of patients with coronary heart disease and age-matched control subjects. Our results show that in patients with coronary heart disease, almost all leukocyte populations, including peripheral blood stem cells and progenitor cells, are affected equally by telomere shortening, which argues in favor of an inherited rather than an acquired cause. The mechanism by which shorter telomere length and the risk of developing coronary heart disease are linked remains to be investigated. It can be speculated whether impaired telomere biology in bone marrow–residing precursors could impose on vascular repair mechanisms. Finally, telomere length specifically in CD8+ cytotoxic T cells was shorter in cytomegalovirus-seropositive than cytomegalovirus-seronegative patients with coronary heart disease and was associated with decreased left ventricular function. Future research should investigate whether in patients with coronary heart disease, the coexistence of cytomegalovirus accelerates immunosenescence and imposes on the course of coronary heart disease.
The online-only Data Supplement is available with this article at http://circ.ahajournals.org/cgi/content/full/CIRCULATIONAHA.109.854299/DC1.