Abstract
BackgroundPhysical activity has been traditionally used to maintain health, promote total well-being, and prescribed as part of holistic approach to disease treatments such as cardiovascular disease. In ageing adults, physical activity has purported benefits ranging from improvements in aerobic capacity and skeletal muscle function. However, ageing is frequently accompanied by other changes in the body systems such as loss of skeletal muscle mass, frailty, cardiorespiratory functions and dysmetabolism, which produces heterogenous effects of physical activity on aged adults. These heterogeneities often influence targets used to measure the effectiveness of exercise on health outcomes. For example, body mass index (BMI) is frequently used as a target of exercise regimens. However, BMI in aged adults may be reduced due to loss of skeletal muscle mass as part of sarcopenic processes in ageing, rendering BMI less accurate as a conventional outcome for exercise. Lipids profiles alone are also insufficient to address the impact of ageing on health outcomes, given that ageing of the cardiovascular system for instance, may march on, independent of lipid levels in the blood. Therefore, there is a lack of a converging marker that can be used to measure/assess the effect of lifestyle habits such as physical activity, commonly recommended as a longevity tool, which is suitable for aged adults. In the field of cardiovascular ageing, there is a critical need to identify suitable biomarkers that can be used as exercise targets are necessary to measure the effects of exercise on aged adults. Since ageing is a life course phenomenon, dynamic lifestyle factors such as physical activity, alcohol use and food intake can alter the course of physical ageing. Given that these dynamic factors all converge upon the human metabolome, metabolomics might provide a comprehensive and integrated picture of these lifelong environmental exposures, alongside exercise as a frequently practised intervention to alter the course of ageing.
Objective
The principal aim of the proposed research is to examine how metabolomics may be used as measurable biomarkers that represents the convergence of all physiological processes that occur with ageing, which accounts for the effect of exercise on the sum of these processes.
Research Questions
1.Would metabolomics biomarkers identified from blood samples of older adults be associated with cardiovascular ageing, as defined by changes in cardiac structure and function?
2.Would metabolomics biomarkers identified in (1) differentiate between physical activity levels, i.e., high versus low physical activity practices, among older adults with cardiovascular ageing?
3.Is there a better measure of cardiovascular health outcome, compared to traditional markers such as body mass index?
4.Recognising the need to incorporate multiple biological inputs, would an expansive machine learning (ML) approach help rank key factors that determine healthy cardiovascular health in ageing?
Methodology
To answer these questions, we will use data from a cohort study of older adults recruited from community population. The Cardiac Ageing Study (CAS) is a community-based study of middle aged to older adults (mean age 72±4 years) examined in 2014-2017 who did not have clinical cardiovascular disease (CVD) at baseline. In CAS, we characterised CV structure and function using novel cardiovascular imaging techniques. We found that these imaging markers defined individuals with worse structural and functional alterations that likely represent cardiovascular ageing. In conjunction with physical activity, skeletal muscle mass, dietary capture and circulating metabolites in this population, this cohort will provide the data to answer these research questions. Furthermore, apart from cross-sectional analytical approaches, we will include biomarker samples obtained at time points over longitudinal follow-up to chart changes in CV longevity over time. In such an endeavour that involves multiple biological inputs, an expansive machine learning (ML) approach will additionally help identify key factors that determine healthy cardiovascular longevity. We will use machine learning techniques to analyse these multiple inputs. The automatic feature detection of machine learning will efficiently detect the association between the combination of metabolomics features, exercises and cardiac health.
Prior publications that support this thesis:
•Koh AS, Gao F, Leng S, Kovalik JP, Zhao X, Tan RS, Fridianto KT, Ching J, Chua SJ, Yuan JM, Koh WP and Zhong L. Dissecting Clinical and Metabolomics Associations of Left Atrial Phasic Function by Cardiac Magnetic Resonance Feature Tracking. Sci Rep. 2018;8:8138-26456.
This paper integrates clinical and metabolomics signals for left atrial phasic function in older adults. We found that left atrial function alterations were a marker of cardiovascular ageing in older adults and medium and long chain acylcarnitines including amino acids such as serine, citrulline and valine were associated with phases of left atrial function. By integrating these clinical and metabolomics signals of left atrial function, metabolite signals may be useful for advancing mechanistic understanding of LA disease in future studies.
•Gao F, Kovalik JP, Zhao X, Chow VJ, Chew H, Teo LL, Tan RS, Leng S, Ewe SH, Tan HC, Tan TY, Lee LS, Ching J, Keng BM, Zhong L, Koh WP and Koh AS. Exacerbation of cardiovascular ageing by diabetes mellitus and its associations with acyl-carnitines. Aging (Albany NY). 2021;13:14785-14805.
This paper highlights the work we did to define relationships between acylcarnitines and cardiovascular function in ageing. We found that distinct alterations in fuel oxidation pathways in short chain and long chain acyl-carnitines, di-carboxyl and hydroxylated acyl-carnitines. These links between fuel oxidation pathways in older adults were associated with impairments in myocardial relaxation and worse left atrial function, likely reflecting early disturbances in diastolic function.
•Kovalik JP, Zhao X, Gao F, Leng S, Chow V, Chew H, Teo LLY, Tan RS, Ewe SH, Tan HC, Wee HN, Lee LS, Ching J, Keng BMH, Koh WP, Zhong L and Koh AS. Amino acid differences between diabetic older adults and non-diabetic older adults and their associations with cardiovascular function. J Mol Cell Cardiol. 2021;158:63-71. doi: 10.1016/j.yjmcc.2021.05.009.:63-71.
This paper highlights the work we did to define relationships between amino acids and cardiovascular function in ageing. We found correlations between metabolites in the one-carbon and nitrogen handling pathways and ageing heart functions. These findings point to a potential role for changes in nitrogen handling in the pathogenesis of heart failure in older subjects.
•Koh AS, Gao F, Tan RS, Zhong L, Leng S, Zhao X, Fridianto KT, Ching J, Lee SY, Keng BMH, Yeo TJ, Tan SY, Tan HC, Lim CT, Koh WP and Kovalik JP. Metabolomic correlates of aerobic capacity among elderly adults. Clin Cardiol. 2018;41:1300-1307.
Combining echo-based and CMR-based imaging techniques to characterise cardiac ageing, this paper investigated metabolomics markers in relation to aerobic capacity. We found that low physical activity, associated with deleterious changes in cardiovascular structure and function, was distinguished by a metabolomic signature of wide-spectrum acylcarnitines and several amino acids. Combined cardiac and metabolomics phenotyping may be useful for tracking future interventions related to physical activity among community cohorts.
•YH Tan, JP Lim, WS Lim, F Gao, LLY Teo, SH Ewe, BMH Keng, RS Tan, WP Koh, Koh AS. Obesity in Older Adults and Associations with Cardiovascular Structure and Function. Obesity Facts, 2022.
This paper evaluated body mass index versus percentage fat mass in determining cardiovascular structure and function in older adults. Waist circumference, rather than body mass index, identified higher prevalence of obesity. Across body mass index categories, waist circumference identified more adverse measurements in myocardial relaxation, aerobic capacity and left atrial structure.
•Loh DR, Yeo SY, Tan RS, Gao F, Koh AS. Explainable Machine-Learning Predictions To Support Personalised Cardiology Strategies. European Heart Journal - Digital Health. 2022;3:49-55.
This paper tested a method in Artificial Intelligence, known as Explainable Machine Learning, to identify personalised factors related to cardiovascular health state among older adults. Our work showed that machine learning could converge heterogenous features, including metabolomics and physical activity and demonstrate its effects on cardiovascular health.
Innovations and Importance of this Proposal:
This approach attempts to conglomerate the complexities of ageing. Some ageing studies have been cross-sectional and thus are somewhat limited in their ability to detect causal associations between biochemical pathways and the effects of exercise on ageing. Pre-specified cohorts that study ageing and exercise, independently of traditional risk factors are necessary. Furthermore, analysis of community cohorts that include biomarker samples obtained at multiple time points is necessary to provide future reference targets for community cohorts.
Now is the right time for this idea. Ageing is a global problem. By 2030, approximately 20% of the world population will be aged 65 years or older. There is growing awareness and practice of using exercise as a lifestyle intervention to reduce ill-health associated with ageing. Yet, there is hardly any measurable biomarker that can quantify the effect of exercise on the individual older adult at a personalised level. Without robust methods of measuring the effect of exercise on ageing, exercise advice is prescribed blindly, indiscriminately while ignoring innate differences between individuals and their corresponding responses to exercise. Metabolomic profiling is an important systems biology tool that measures large numbers of metabolites with diverse chemical properties in a quantitatively rigorous and reproducible fashion. In contrast to other ‘omics’ platforms, such as genomics, transcriptomics and proteomics, metabolomics measures the net composition of genomic, transcriptomic, and proteomic variability providing an integrated profile of an individual’s biological status. Thus, the metabolome provides a comprehensive picture of the immediate effects of exercise on the body, potentially preceding end-organ effects, exerting maximal preventative effect and personalised feedback to the user.
| Date of Award | 2024 |
|---|---|
| Original language | English |
| Awarding Institution |
|
| Supervisor | Chim Lang (Supervisor) & Anna-Maria Choy (Supervisor) |