S3.4 Genetic and environmental influences on the markers of biological age
Chair: Juulia Jylhävä
Markers of biological age have recently proven to be strong mortality predictors independent of other risk factors. As individuals with the same chronological age can vary substantially for their biological age it has become a pressing issue to find the determinants for biological age. This symposium presents findings on the genetic and environmental influences to the following markers of biological age: self-rated health, the frailty index, the epigenetic clock and telomere length. Using unique longitudinal twin cohorts with wide age ranges, we are also able to assess the changes in the contributing sources variation – i.e., genetics and shared and unique environments. The first paper demonstrates how the heritability of self-rated health is influenced by financial strain. The second paper examines if the polygenic risk scores for well-being are predictive of objective (i.e. frailty and multi-morbidity) and subjective indicators of age-related health (e.g. self-perceived health). Similarly to the first paper, the third and fourth paper make use of the classic twin modeling and shed light into the genetic and environmental influences on epigenetic aging at different ages and how paternal age at conception moderates the heritability of telomere length, respectively. Taken together, these studies identify some of the factors contributing to biological aging and ultimately help to select the ideal targets for future preventive measures.
S3.4.1 Sex differences in financial strain moderation of genetic influences on subjective
Indiana University Southeast, USA
Background: Women have a higher burden of chronic disease, rate their health as poorer and have higher levels of morbidity, yet outlive men; the reasons for this sex paradox, however, are not clear. Sex differences in exposures to environmental factors associated with health may play a role. Evidence suggests that subjective experience of financial strain has a stronger association with health outcomes than actual income levels. Methods: Nine of the studies in the Interplay of Genes and Environment across Multiple Studies (IGEMS) consortium included items assessing financial strain, representing 10,756 individuals. The sample was 55% women, included 3185 MZ twins and 5228 DZ twins, and age ranged from 24 to 98. Four items assessed financial strain: extent to which money covers needs, difficulty in paying monthly bills, economic situation compared to others, and whether there is money for extras. A factor model was used to create a harmonized measure of financial strain across studies and items. Results: Twin analysis of genetic and environmental variance indicated both age and sex differences in the etiology of the harmonized financial strain score. Increased financial strain resulted in increased genetic variance in subjective health for women. In contrast, increased financial strain was associated with increased shared environmental variance for men. Conclusions: The pattern of results for women supports the diathesis-stress model, whereas the pattern of results for men likely reflects increasing salience of shared cultural conceptions of "good health."
S3.4.2 Influences of the genetic component of subjective wellbeing on frailty, multi-
morbidity and self-perceived health in an aging Swedish population: A genetic risk score approach
Karolinska Institutet, Stockholm, Sweden
Background: Simple measures of subjective well-being (SWB) are excellent predictors of an individual’s morbidity and mortality and often even superior to objective clinical assessment, suggesting that SWB is not only an indicator of current objective health status, but also a powerful predictor of future health. But what drives the predictive value of SWB over and above objective health? We explored whether higher genetic predisposition for subjective well-being would be protective of frailty and multi-morbidity and indicates better subjective health in an aging population of Swedish twins. Methods: We derived polygenic risk-scores (PRS) based on the genome-wide association results of SWB to predict age-related health as measured by self-perceived health, the Rockwood frailty index (FI), and the Cumulative Illness Rating Scale, a cumulative score based on medical problems in 13 different organ systems which reflects common problems of the elderly. Individuals with existing diseases in more than one organ system were regarded as multi-morbid (MM). MM scores and self-perceived health as well as genetic information were available for 2,615 participants from 1,692 twin families from the IGEMS consortium and the FI was available in a sub-sample of about 400 individuals. The PRS was regressed on the observed health scores to assess the degree to which variability in age-related health is explained by variability in genes underlying SWB. Results: Associations between the phenotypic variables were all highly significant (p < 0.01). Although the wellbeing PRS were significantly related to self-perceived health (p < 0.05), it was unrelated to frailty and MM. Conclusions: Genetic variants underlying SWB were not predictive of age-related health decline as measured by FI and MM, the objective indicators of age-related health status, although it was significantly associated with self-perceived health. Analyses will be replicated with additional samples to increase power and implications of the findings will be discussed.
S3.4.3 Genetic and environmental influences on the epigenetic clock across the old age
Karolinska Institutet, Stockholm, Sweden
Background: A measure of DNA methylation age, the Horvath’s epigenetic clock, has recently emerged as a promising marker to predict mortality. However, the contribution of genetic and environmental factors to the epigenetic clock at different old ages is incompletely understood. Methods: We estimated the magnitude of genetic and environmental influences to the epigenetic clock in 126 (52 MZs, 74 DZs) Swedish and Danish twin pairs at two time points: at mean age of 69 and 79 years. Using the Cholesky decomposition, we also estimated the magnitude of the sources for shared variation between the two ages. The DNA methylation data were assessed in whole blood using the Illumina 450k BeadChip and the epigenetic clock estimates were assessed using the Horvath algorithm. The analyses were adjusted for age, sex, and cell type heterogeneity. Results: Intraclass correlations for MZ and DZs were 0.259 and 0.125 at 69 years and 0.282 and 0.020 at 79 years. The cross-time correlations were 0.341 and 0.294 for MZs and DZs. A model incorporating additive genetics (A) and unique environmental influences (E) best explained the variation at both ages (E=76.6%, A=23.4% at 69 years; E=78.1%, A=21.9% at 79 years). The cross-time covariation was primarily due to additive genetics than unique environment (A=69.7%, E=30.3%). Conclusions: Our results demonstrate that most of the variation in the epigenetic clock in old age is due to environmental factors and to a lesser extent due to genetics. The genetic sources of variation are largely the same across the old ages whereas those for the environmental ones seem to differ more. This means that there is a possibility to affect the rate of epigenetic aging even at old age.
S3.4.4 Telomere length in twins: unravelling the biomarker of aging –question
University of Southern Denmark, Denmark
Background: Telomere length, a highly heritable trait, shortens by age and is longer in females and longer telomeres also associate with better survival - all features that make telomere length a viable candidate as biomarker of aging. We review some of the implications from a twin study perspective and focus on the well-established relationship that telomeres are longer in offspring of older fathers. This perplexing feature has been attributed to the longer telomeres in sperm of older men and it might be an ‘epigenetic’ mechanism through which paternal age plays a role in telomere length regulation in humans. Methods: We consider two independent (discovery and replication) twin studies from the Danish and the UK cohorts, comprising 889 twin pairs. Results: We show an increase in the resemblance of leukocyte telomere length between dizygotic twins of older fathers, which is not seen in monozygotic twins: We observe a high correlation (0.85) in monozygotic pairs independent of paternal age at conception. Conclusions: This phenomenon might result from a paternal age-dependent germ stem cell selection process, whereby the selected stem cells have longer telomeres, are more homogenous with respect to telomere length, and share resistance to aging.