Relationship between telomere shortening and age in Korean individuals with mild cognitive impairment and Alzheimer’s disease compared to that in healthy controls

Participant characteristics

In total, 437 individuals were enrolled. The mean patient age was 69.9 years, and 58.6% of the participants were female (Table 1). The average TL was longer in female participants than in male participants (8.40 vs. 7.85 kbp; P=0.015). Seventy-five (17.2%) individuals were diagnosed with AD, with female participants showing a higher rate of AD than male participants (21.1% vs. 11.6%, respectively; P=0.003). No significant differences were observed in the prevalence of hypertension, diabetes, hyperlipidemia, heart disease, and stroke between male and female participants. Further detailed information is provided in Table 1.

Overall association between age and telomere length

There was an overall significant negative correlation between age and TL (Figure 1). Multivariable linear regression analysis showed an approximate decrease of 37 bp of TL per 1-year increase in age in participants (B=-0.037; P=0.002; Table 2).

Figure 1. Scatterplot with a linear regression line and 95% confidence interval showing the association between age and telomere length in all participants. *P<0.05.

Associations between age and telomere length according to sex

When the study participants were divided according to sex, we observed negative correlations between age and TL in both male and female participants (B=-0.048 [r=-0.202] and P=0.006; B=-0.056 [r=-0.201]; P=0.001, respectively; Figure 2). We observed that female participants had an overall longer TL than male participants and that the rate of telomere shortening was slightly faster in female participants than in participants. However, only female participants showed a significant negative correlation between age and TL in multivariable linear regression analysis (B=-0.052; P=0.002; Table 2). The boxplot shows a statistically significant longer TL in women than in men in the second age quartile (64–70 years; P=0.031; Figure 3A). In the older age group (age ≥65 years), female participants also showed a significantly longer TL than male participants among those (P=0.042; Figure 3B).

Figure 2. Scatterplot with linear regression lines with 95% confidence intervals showing the associations between age and telomere length according to sex (A), age and telomere length in men (B); and age and telomere length in women (C). *P<0.05.

Figure 3. Boxplots with dot plots of telomere length. (A) telomere length classified by quartile age group according to sex; (B) telomere length classified by the 65-year age group according to sex. *P<0.05.

Association between age and telomere length in healthy individuals and in individuals with MCI and AD

We observed a negative relationship between age and TL among individuals with MCI and AD and healthy individuals (Figure 4). Overall, individuals with AD demonstrated a significantly steeper negative correlation between age and TL than those with MCI and healthy individuals (B=-0.050 [r=-0.267]; P=0.020; Figure 4A). However, there was no overall significant difference in the mean TL between individuals with AD and healthy individuals (Supplemental Figure 1). Among male participants, only those with AD showed a significant negative correlation between age and TL (B=-0.064 [r=-0.446]; P=0.043; Figure 4B). However, we did not observe significant associations between age and TL in healthy individuals and individuals with MCI and AD (Figure 4C). Multivariable linear regression analysis showed that the mean rate of telomere shortening was 60 bp per year in individuals with AD (B=-0.060; P=0.006) (Table 3). We also observed a linear association between age and TL according to amyloid-beta positron emission tomography (PET) (Supplemental Figure 2). PET-negative and PET-positive individuals showed similar negative associations between age and TL.

Figure 4. Scatterplot with linear regression line showing the associations between age and telomere length among healthy individuals and individuals with MCI and AD (A), age and telomere length among the three groups in men (B), and age and telomere length among the three groups in women (C). MCI=mild cognitive impairment; AD=Alzheimer’s disease. *P<0.05.

Participants

A total of 437 individuals were enrolled from 11 hospitals. These hospitals included nine medical centers that comprised the validation cohort of the Korean Brain Aging Study for the Early Diagnosis and Prediction of AD and two independent centers (Asan Medical Center and Hanyang University Guri Hospital). Participants were diagnosed with MCI if they met the clinical criteria established by the National Institute on Aging-Alzheimer’s Association (NIA-AA) and the modified criteria proposed by Petersen—1) Clinical Dementia Rating (CDR) scale score of 0.5; 2) memory complaints compared to the participant’s previous cognitive function by patients, caregivers, or clinicians; 3) independent activities of daily living; 4) an objective cognitive decline of >1.5 standard deviations below the age-, education-, and sex-adjusted normative means for one or more of the four neuropsychological tests (memory, visuospatial, language, or frontal-executive function) included in the Consortium to Establish a Registry for Alzheimer’s Disease; and 5) absence of dementia. Patients were diagnosed with probable AD according to the criteria proposed by the National Institute of Neurological and Communicative Disorders and Stroke-Alzheimer’s Disease and Related Disorders Association or the NIA-AA; Diagnostic and Statistical Manual of Mental Disorders 4^th edition; and CDR scale score of 0.5 or 1. The study individuals included 173 healthy controls, 123 individuals with MCI, and 75 individuals with AD. MCI and AD were diagnosed based on clinical examinations, including Mini Mental State Examination (MMSE), CDR scale, and amyloid-beta PET. Individuals who were diagnosed with other types of dementia, such as frontotemporal dementia, were excluded and classified in the not available group to avoid the risk of bias. The study was performed according to the International Conference on Harmonization Good Clinical Practice guidelines and was approved by the institutional review board of each participating center. Before participation, all participants provided written informed consent. Due to significant cognitive declines in some individuals with AD, informed consent was obtained from their legal guardians.

Clinical assessments

Individuals were determined to have hypertension if they were prescribed antihypertensive medication(s) or had systolic blood pressure >140 mmHg and diastolic blood pressure >90 mmHg. Diabetes mellitus was diagnosed based on the prescription of insulin or oral hypoglycemic medications, high plasma glucose levels (≥126 mg/dL), or high glycated hemoglobin level (≥6.5%) after 8 hours of fasting. Individuals were diagnosed with hyperlipidemia if they were prescribed lipid-lowering medications or had high levels of total cholesterol (≥200 mg/dL), low-density lipoprotein cholesterol (≥130 mg/dL), and triglycerides (≥150 mg/dL) and low levels of high-density lipoprotein cholesterol (<40 mg/dL). Heart disease was defined as a history of coronary artery bypass grafting, percutaneous coronary intervention, or atrial fibrillation or use of heart-related drugs. Stroke was defined as a medical history of cerebral infarction, intracerebral hemorrhage, or subarachnoid hemorrhage.

Amyloid-beta positron emission tomography

A total of 325 participants underwent amyloid-beta PET at baseline. The standardized uptake value ratio (SUVR) was calculated using ¹⁸F-flutemetamol PET and ¹¹C-PiB PET, which were adjusted using the pons and the cerebellar gray matter as reference regions, respectively. The composite SUVR was calculated by obtaining the region of interest activity of the frontal, temporal, parietal, occipital, anterior cingulate, and posterior cingulate/precuneus cortices and dividing them by the mean intensity of the reference regions. Amyloid-beta PET positivity was determined based on the composite SUVR (≥0.634 for ¹⁸F-flutemetamol PET and >1.21 for ¹¹C-PiB PET).

Telomere length assay

At baseline, whole blood from individuals was collected and separated into plasma and buffy coat. The leukocyte DNA was extracted from the buffy coat using D-DEX^™ II b RBC lysis buffer and D-DEX^™ II b Cell lysis buffer (Intron, MA, USA). The DNA was then hydrated with 300 μL of DNA hydration solution (QIAGEN, Hilden, Germany). TLs were measured using a nonradioactive TeloTAGGG TL Assay (Roche Boehringer-Mannheim, Grenzach-Wyhlen, Germany) according to the manufacturer’s instructions. Briefly, 2–4 μg of DNA was fragmented using a Hinf I/RsaI enzyme mix and separated using agarose gel electrophoresis. The DNA fragments were then transferred to a nylon membrane (Millipore, Bedford, MA, USA) and incubated with digoxigenin (a digoxigenin-labeled probe), which specifically attaches to telomeric repeats. Next, the membranes were incubated with secondary antibodies conjugated with alkaline phosphatase. TLs were visually measured using chemiluminescence and an image analyzer (ImageQuant LAS 4000, GE Healthcare, Little Chalfont, UK). TLs were determined by comparing them to molecular weight standards.

Statistical analyses

The chi-square test and Student’s t-test were used to identify differences between the two groups. Missing values were replaced with “99” and included in the statistical analysis [28]. Box plots with dot plots were used to assess the differences in TLs according to age groups and sex.

Scatterplots with a regression line were constructed to visualize the associations between age and TL. Multivariable linear regression was performed to evaluate the independent association between age and TL according to sex among individuals with MCI and AD and healthy individuals.

P-values <0.05 were considered statistically significant. All statistical analyses were performed using R software version 3.6.3 and IBM SPSS Statistics for Windows, version 24.0 (Armonk, NY).