Introduction

Telomeres, the DNA-protein complexes that cap the ends of chromosomes and protect against genomic instability, contain variable-length tracts of telomeric DNA. Leukocyte telomere length (LTL) is increasingly recognized as an index of biological age that predicts incidence of age-related diseases [1], [2], as well as all-cause and disease-specific mortality in diverse cohorts of older adults [1], [3], [4], [5], [6]. In fact, older adults with below average LTL have more than threefold increased risk for early mortality [4], [6]. Long LTL is also a predictor of years of healthy living, an outcome that integrates self-perceived functional status and duration of survival [7]. Importantly, telomere shortening is not just an index, but also a potential mechanism underlying some aspects of biological aging because short telomere length can lead to cessation of mitosis with consequent loss of ability for cell replenishment and, in some cellular settings, genomic instability, end-to-end chromosome fusion, and apoptosis as well as harmful DNA damage signaling and mitochondrial dysfunction [8], [9], [10], [11]. Notwithstanding the utility of LTL in predicting mortality and longevity, there is little clinical evidence regarding modifiable factors associated with and potentially driving leukocyte telomere shortening.

Inflammatory activity is frequently proposed as a contributor to biological aging in general, and leukocyte telomere shortening in particular [5], [12], [13], [14]. Accumulating evidence suggests a causal role for inflammation in the pathogenesis of multiple age-related diseases including cancer, atherosclerosis, autoimmune disorders, diabetes, and neurodegenerative diseases [15], [16], [17], [18], [19], [20], [21]. Elevated inflammatory activity could accelerate leukocyte telomere shortening by promoting cell turnover and replicative senescence [22], and by inducing the release of reactive oxygen species that damage telomeric DNA via oxidative stress [23]. One complication is that inflammatory cytokines including tumor necrosis factor-α (TNF-α) can both inhibit and promote activity of telomerase [24], [25], [26], the cellular enzyme primarily responsible for lengthening telomeres [27]. On the other hand, in vitro research on various cell types, but not including normal leukocytes, indicates that an accumulation of senescent cells with critically short LTL may produce pro-inflammatory factors [28], [29], hich could in turn contribute to an increased inflammatory load.

Observations of short LTL in small samples of patients with inflammatory diseases including hepatitis C, liver cirrhosis, chronic kidney disease and chronic obstructive pulmonary disease provide preliminary support for the proposal that inflammation accelerates leukocyte telomere shortening [5], [14], [30], [31]. Specifically, in these studies, patients with inflammatory disorders exhibited shorter LTL cross-sectionally compared with healthy individuals [31], [32], and shorter telomeres in cells proximal to disease-related inflammatory activity [14], [30]. Additionally, elevations in disease-specific and systemic inflammatory markers have been associated with short LTL in patients [5]. Although there are plausible pathways by which inflammation could accelerate telomere shortening even in adults without chronic disease, little is known about in vivo relationships between markers of inflammation and LTL in healthy older adults.

In the present study, we examined if elevated inflammatory activity, indexed by high levels of the systemic inflammatory cytokines interleukin-6 (IL-6) and TNF-α and the acute phase protein C-reactive protein (CRP), is associated with increased risk for short LTL in the ‘Health, Aging and Body Composition’ (Health ABC) cohort, a large sample of well-functioning men and women aged 70–79 years. Our sample included 1,962 Health ABC participants who had complete data available for our primary analyses. We hypothesized that participants with high levels of IL-6 and/or TNF-α and/or CRP would be more likely than participants with lower levels of these inflammatory markers to have short LTL, and that the combination of high IL-6, high TNF-α and high CRP, as an index of a high cumulative inflammatory load, would confer the greatest odds for short LTL.

Results

In spite of the restricted age range of the sample, age was significantly associated with shorter LTL (r = −.07, p = .002) and higher levels of TNF-α (r = .07, p = .003). However, age was negatively associated with levels of CRP (r = −.07, p = .001) and was not associated with levels of IL-6 (r = .03, p = .19). Inflammatory markers were significantly and positively associated with one another. However, while the effect size for the relationship between IL-6 and CRP was medium to large (r = .47, p<.001), relationships between IL-6 and TNF-α (r = .27, p<.001) as well as between TNF-α and CRP (r = .11, p<.001) had smaller effect sizes. Table 1 summarizes the baseline characteristics of the study population stratified by LTL tertile. As predicted, participants in the bottom tertile for LTL had the highest level of IL-6 and TNF-α, but there were no differences between groups in levels of CRP.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Sample characteristics by leukocyte telomere length (LTL) tertile.

https://doi.org/10.1371/journal.pone.0019687.t001

Inflammatory Activity and LTL

In our primary analytic model, we found that the odds of having short LTL (i.e., LTL in the bottom tertile or ≤4260 bp) were significantly higher for those participants who had high (i.e., top tertile) levels of either IL-6 (≥2.39 pg/mL), TNF-α (≥3.72 pg/mL) or CRP (≥2.51 mg/L). In these analyses, the results of which are summarized in Table 2, groups were not mutually exclusive, such that the same individual could be represented in multiple groups if he or she had high levels of IL-6, TNF-α and CRP. Adjusting for a wide range of potential confounds and covariates, the odds for short LTL were significantly higher in those who had either IL-6 levels in the top tertile (OR = 1.3, 95% CI = 1.1–1.7) or TNF-α levels in the top tertile (OR = 1.5, 95% CI = 1.2–1.9). The highest odds for short LTL were observed in those participants who had high levels of both IL-6 and TNF-α (OR = 1.8, CI = 1.3–2.4). In contrast, the addition of high levels of CRP did not confer increased odds for short LTL, as indexed by the lack of association between CRP and LTL (OR = 1.1, 95% CI = 08–1.4), and by the roughly equal odds of having short LTL for those who had high levels of CRP in addition to high levels of IL-6 and TNF-α (OR = 1.7, CI = 1.1–2.6).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Systemic inflammatory markers as predictors of short leukocyte telomere length.

https://doi.org/10.1371/journal.pone.0019687.t002

We additionally ran the same models using the established high-risk clinical cutoff (>3 mg/L) for CRP [33]. While individuals who had CRP above this clinical cutoff had marginally higher odds for short LTL, this was not significant in any of the models, including when we only adjusted for telomere batch (OR = 1.3, CI = 1.0–1.7). Thus, individuals with CRP in either the top tertile in the sample or above the recognized clinical cutoff did not appear to have significantly increased odds for short LTL.

To further examine the observed associations of IL-6 and TNF-α with short LTL and to compare mutually exclusive groups instead of potentially overlapping groups, we excluded CRP from our model and compared the odds for short LTL among four mutually exclusive groups of participants: those with both high IL-6 and high TNF-α; only high IL-6; only high TNF-α; and neither high IL-6 nor high TNF-α. These analyses were conducted while statistically controlling for established risk factors and potential confounds and with both high IL-6 and high TNF-α as our reference group. Results indicated that the combination of both high IL-6 and high TNF-α levels was associated with roughly double the odds for short LTL compared with having neither high IL-6 nor high TNF-α (OR = 0.52, CI = 0.37–0.72). Furthermore, the combination of high IL-6 and high TNF-α was associated with significantly higher odds for short LTL compared with having only IL-6 high (OR = 0.57, CI = 0.39–0.83) or only TNF-α high (OR = 0.67, CI = 0.46–0.99). Finally, mean LTL adjusted for site, age, gender, and race was 201 base pairs shorter in participants with high levels of both IL-6 and TNF-α (M = 4701.93 bp, SE = 68.07) compared with that of participants who did not have high levels of both of these (M = 4902.21 bp, SE = 28.20), and this difference in absolute LTL was significant, F(1,1937) = 7.35, p = .007.

Discussion

Short LTL, a potential index of biological age, is associated with increased risk for age-related diseases as well as early all-cause and disease-specific mortality [3], [4], [5], [6]. The present research study is to date the largest to demonstrate an association between elevated inflammatory activity and short LTL. Moreover, the present study is the first to indicate that the combination of high IL-6 and TNF-α is associated with increased odds for short LTL, over and above the odds associated with having high levels of only one of these markers. Of note, having high levels of both IL-6 and TNF-α conferred almost twice the odds of being in the bottom tertile for LTL compared with having lower levels of both markers. In contrast, having high levels of CRP, as indexed by CRP levels either in the top tertile for the sample or above the established clinical cutoff, was not associated with increased odds for short LTL. Importantly, all of our results held when controlling for the contribution of numerous established risk factors for short LTL and potential confounds. Together with the observed causal relationships between inflammatory activity and LTL in animal and in vitro studies, these data provide preliminary evidence that adjunct anti-inflammatory therapies could potentially prevent accelerated leukocyte telomere shortening or protect against the negative effects of short LTL in older adults.

We included IL-6, TNF-α and CRP as three separate measures of systemic inflammatory activity in our study. Although all of these inflammatory markers were significantly correlated with one another, relationships between them were not strong (r's≤.47). IL-6 and CRP were the most strongly associated of the inflammatory markers, sharing 22% of variance. Given that IL-6 is a major inducer of CRP production by hepatocytes [34], [35], the larger effect size for the relationship between these markers is not unexpected. However, high levels of IL-6, but not high levels of CRP, were associated with increased odds of short LTL in the sample. There is evidence that IL-6 is necessary but not sufficient to induce CRP synthesis [36], and also evidence that human coronary artery smooth muscle cells may produce CRP and that IL-6 is not a necessary inducer of CRP synthesis by these cells [37]. Thus, although IL-6 is an inducer of CRP and their concentrations are correlated, the physiological actions of IL-6 and CRP are not expected to be identical. The present findings, which need to be further investigated, suggest that in the case of LTL, these two factors do not have the same effects. Notably with regard to our overall findings, IL-6 and TNF-α shared only 7% of variance, supporting the formulation that circulating levels of these cytokines reflect different aspects of the inflammatory response. Thus, while each of these cytokines may contribute to leukocyte telomere shortening by promoting cell turnover and replicative senescence, inducing oxidative stress, and regulating telomerase activity [22], [23], [24], [25], [26], chronic exposure to the cumulative load of both high IL-6 and high TNF-α may have the greatest impact on the rate of leukocyte telomere shortening.

Equally plausible, however, is the emerging hypothesis that participants with short LTL in our study had the greatest proportion of accumulated senescent bodily cells overall, including fibroblasts and epithelial cells, which could plausibly contribute to the observed higher levels of IL-6 and TNF-α in this group [28], [38]. Given that TNF-α and IL-6 may inhibit programmed cell death in specific cell types [39], [40], high levels of these cytokines could even contribute to the maintenance of senescent cells in the system and hence the continued production of pro-inflammatory factors. However, the cross-sectional design of our study precludes drawing conclusions about the causal direction in the relationship between inflammatory activity and LTL.

In previous research, the cumulative load of high IL-6, high TNF-α and high CRP was found to confer the greatest risk for cardiovascular events in the Health ABC cohort [7]. However, CRP did not likewise contribute to the predictive value of the inflammatory markers with regard to LTL. Our finding of no association between CRP and LTL is in line with a previous finding of no associations between CRP and LTL in post-menopausal women, and in ‘Cardiovascular Health Study’ participants who were older than 73 years [7]. However, this finding does not support our hypothesis that higher levels of all measures of inflammatory activity would be associated with shorter LTL. Previous research has demonstrated that although CRP is reliably associated with cardiovascular disease (CVD) outcomes, CRP does not appear to be causally involved in the development of CVD [41]. Furthermore, CRP is less predictive of CVD in older compared with younger adults [42], [43] and CRP levels were not predictive of cardiovascular events in a sample of high-risk Japanese adults in whom IL-6 was an independent predictor of such events [44]. However, it is possible that CRP would be associated with LTL in older populations with other diseases, such as inflammatory diseases characterized by very high levels of CRP. Our findings in relation to CRP may have particular relevance in clinical settings where CRP is the most commonly used index of inflammatory activity because they suggest that pro-inflammatory cytokine concentrations may serve as complementary indices of inflammation.

The present findings must be interpreted in the context of several limitations. First, the cross-sectional design of the present study precludes causal interpretations. Second, the narrow age range of our sample of older adults (70–79 years) is both a strength and weakness of our study. While this narrow age range allows us to maximize power to detect associations between inflammatory activity and LTL without many of the confounding factors associated with a sample of wider age range (e.g., cohort effects, hormonal effects), the findings of the present study may not be generalizable to other stages of the lifespan. Moreover, anti-inflammatory medications were common in our sample of older adults. Given the wide range of medications that have anti-inflammatory effects and the variability in dosage across individuals, our statistical control for anti-inflammatory medications is likely to be insufficient but is the best option currently available. There has been debate about the relative merits of different methods used to assay LTL and the Q-PCR method has been subject to criticism [45], [46]. However, strong associations between LTL as assessed by the Southern blot method and Q-PCR method have been documented [47], and the coefficient of variation of the LTL assay for this study was low at 4%. In addition, some findings related to LTL in Health ABC have been replicated in a cohort who had LTL measured with the traditional Southern blot technique [48]. Finally, it should be noted that effect sizes were small in this study and caution is therefore advised in the interpretation of the results pending future studies in independent cohorts.

Data from this large cohort of older adults indicates links between inflammatory activity and biological aging. In particular, the present data indicate that the cumulative load of high IL-6 and high TNF-α is accompanied by increased risk for short LTL. The present data also replicate a previous finding of no association between CRP and LTL in older adults, and suggest that systemic inflammatory cytokines may have more relevance to leukocyte telomere shortening than the more commonly used inflammatory marker CRP. In sum, older adults with high levels of inflammatory activity may be at increased risk for accelerated leukocyte telomere shortening, and those with short LTL may have increased risk for diseases with an inflammatory etiology.

Methods

Acknowledgments

The authors thank the Health ABC Study staff at the Pittsburgh and Memphis study sites and at the UCSF Coordinating Center for their assistance and patience, as well as the Health ABC Study participants for their time and efforts.