Aging is a very complex biological phenomenon characterized by a gradual decline in physiological function, accumulation of damage, the development of age-related diseases, increased frailty and susceptibility to infections and eventually death. Genetics, environmental factors and lifestyle factors all influence the aging process, the lifespan and healthspan.
Table of Contents
What is chronological age?
Chronological age is defined as the number of years a person has lived since birth. It has traditionally been the primary method used to assess an individual’s age. Chronological age serves as a standardized measure in society, establishing boundaries and determining eligibility for various activities such as driving, voting, or consuming alcohol. However, when it comes to understanding an individual’s health, lifespan, and healthspan, relying solely on chronological age becomes less informative and useful.
What is biological age?
Biological age refers to the physiological state of aging and encompasses various biomarkers and health indicators that offer a more comprehensive understanding of an individual’s overall health and potential lifespan.
Assessing biological age is very important as it provides valuable insights into an individual’s health status beyond chronological age. It allows for the identification of individuals who may be at higher risk of age-related diseases or have a shorter lifespan despite being younger in years. By evaluating biological age, healthcare professionals can tailor preventive and therapeutic interventions to specifically address age-related conditions and improve overall health outcomes. The concept of biological age acknowledges that individuals age at different rates due to a complex interplay of genetic, lifestyle, and environmental factors.
Are there biomarkers of biological aging?
There are numerous biomarkers that have demonstrated a correlation with the aging process, showing consistent patterns of change as individuals grow older. While these biomarkers can provide insights into an individual’s chronological age, accurately determining biological age remains a challenge. It is important to note that no single or combination of biomarkers definitively defines biological age or guarantees the efficacy of interventions aimed at extending lifespan and healthspan.
The identification of current biomarkers has been largely derived from long-term studies that have monitored specific parameters in individuals over several decades. By leveraging large datasets and employing machine learning algorithms, researchers have made significant strides in extrapolating biomarkers associated with aging and comprehending the trajectory of these parameters over time.
Some of the biomarkers currently utilized to assess biological age and predict lifespan and healthspan include:
- Telomere Length: Telomeres are protective caps at the ends of chromosomes that shorten with each cell division. Telomere length is considered a marker of cellular aging and can reflect the overall aging process.
- Epigenetic Modifications: Epigenetic changes, such as DNA methylation, histone modifications, and microRNA expression, can influence gene activity without altering the DNA sequence. Specific epigenetic patterns have been associated with aging and age-related diseases.
- Inflammatory Markers: Chronic low-grade inflammation, often measured by markers like C-reactive protein (CRP), interleukin-6 (IL-6), and tumor necrosis factor-alpha (TNF-alpha), is associated with aging and age-related diseases.
- Hormone Levels: Hormone dysregulation can occur with aging. For example, declining levels of growth hormone, estrogen, testosterone, and DHEA-S (dehydroepiandrosterone sulfate) are associated with aging processes and age-related conditions.
- Metabolic Markers: Various metabolic parameters, including insulin resistance, fasting glucose levels, lipid profiles, and markers of oxidative stress, can provide insights into an individual’s metabolic health and aging status.
- Immune System Markers: Age-related changes in the immune system, such as alterations in immune cell populations, reduced immune response, and increased inflammation, can serve as biomarkers of biological aging.
- Physical Performance Measures: Assessments of physical performance, such as grip strength, gait speed, and muscle mass, can indicate age-related declines in musculoskeletal function and overall physical fitness.
What are biological age clocks and deep aging clocks?
While many terms are often used interchangeably as if they had the same meaning, they do not.
A biological age clock typically refers to a model or algorithm that uses various biomarkers and health indicators to estimate an individual’s biological age. The biological age clock combines these markers to calculate a numerical value that represents the individual’s estimated biological age relative to their chronological age.
On the other hand, a deep aging clock refers specifically to a type of biological age clock that utilizes advanced machine learning techniques, such as deep learning algorithms with multiple layers of neural networks, to analyze complex patterns and relationships within large datasets. These deep aging clocks are trained on extensive data and can extract high-dimensional features that would not be readily apparent or quantifiable through traditional methods by a human being. By leveraging the power of deep learning, these clocks can potentially capture more nuanced and subtle aspects of biological aging, and analyze multiple biomarkers, leading to more accurate predictions of an individual’s biological age.
It is important to note that terms like “methylation clock” or “epigenetic clock” are often used improperly as generic references to aging clocks. In reality, they are specific types of biological age clocks that focus on analyzing epigenetic modifications, such as DNA methylation patterns.
The following are some of the most popular biological age clocks.
Horvath’s epigenetic clock (aka DNAm age – Methylation clock)
Dr. Steve Horvath’s epigenetic clock, also known as DNAm Age or the methylation clock, revolutionized longevity medicine when introduced in 2013. This biological clock utilizes methylation, a DNA modification process, as an aging metric. Embryos and induced Pluripotent Stem Cells (iPSCs) begin with virtually no DNA methylation, which accumulates as cells divide. The clock studies methylation patterns across specific genomic sites (CpG), effectively estimating biological age and revealing deviations indicative of accelerated or decelerated aging. Validated across various tissues and species, Horvath’s clock is an interesting tool for understanding aging, age-related health risks, and the effects of lifestyle on biological age.
Other notable methylation clocks include DNAm Grim Age and DNAm PhenoAge.
Peters et al transcriptomic clock
The Peters et al transcriptomic clock utilizes transcriptomics, an approach focused on analyzing the complete set of RNA transcripts produced by the genome, to provide an age predictor based on gene expression profiles in peripheral blood and potentially uncover deviations indicative of age-related diseases.. The team identified 1,497 genes whose expression levels were significantly correlated with age, offering valuable insights into the biological pathways and processes associated with aging.
Zhavoronkov’s blood test clock
By feeding a deep neural network with millions of routine blood tests that include common values such as cholesterol, cell count, and inflammatory markers, Dr. Zhavoronkov managed to create an algorithm capable of predicting chronological age and suggest premature or delayed aging in individuals.
Galkin’s microbiome’s clock
Dr. Galkin’s microbiome clock is a development in the field of aging biology. Recognizing that our gut flora undergoes changes as we age, Dr. Galkin and his team have developed a sophisticated deep aging clock. This tool predicts an individual’s chronological age by sequencing the genome of their gut bacteria and processing this data through a specialized algorithm. This innovative approach provides new insights into the aging process, potentially offering novel ways to study age-related changes and their impact on human health, proposing new possible therapeutic interventions targeting the gut microbiota, and suggesting links between specific microbe species and their impact on lifespan and healthspan.
Bone X Ray clock (BoneXpert)
The bone X-ray clock represents the AI-driven evolution of a traditional method used to estimate a child’s age, which involves comparing X-ray images of a child’s hand with a reference atlas containing X-ray images from children at different ages. In this innovative approach, a deep learning algorithm was trained with over 30,000 X-ray images, each manually rated by multiple experienced radiologists. This extensive training data enabled the creation of a reliable and highly accurate automated method to estimate age from X-rays, achieving an impressive accuracy within a six-month range.
Brain MRI clock (BrainAGE)
MRI (Magnetic Resonance Imaging) scans provide detailed images of the brain’s structure. Specific alterations in this structure over time have been linked to the aging process, as well as to various neurodegenerative pathologies.
Deep machine learning has been used to analyze these images and identify patterns that may not be easily discernible by humans. By training these deep learning models on large datasets of brain MRIs paired with subjects’ chronological ages, symptoms, and conditions it’s possible to develop a model that can predict a person’s biological age based on their brain MRI. Furthermore, this technique could possibly predict the risk of developing certain diseases or facilitate early diagnosis.
PhotoAgeClock
The PhotoAgeClock by Bobrov et al. analyses pictures of the corner of the eye to predict the age of an individual with an accuracy of 1.9 years.
Limitations and controversies on deep aging clocks
While deep aging clocks and biological clocks represent significant advancements in our understanding of the aging process, they are not without controversies, limitations and challenges.
As indicated by my exploration of the hallmarks of aging, no single mechanism governs aging. Instead, it’s a complex interplay of intertwined factors. Therefore, an intervention targeting a single mechanism—analyzed by one aging clock—might not register any change when viewed from the perspective of other aging clocks. This brings us to the issue of self-validation of results: biomarkers and proposed interventions are often designed to satisfy each other, a concern raised by experts in the field.
Take telomere length, for example. While it’s correlated with chronological age—allowing us to predict chronological age within a certain margin of error by checking telomere length—it doesn’t necessarily reflect biological age. If we therapeutically lengthen telomeres and an aging clock suggests we are 40 years younger than our chronological age, it unfortunately doesn’t imply that we are genuinely 40 years younger in terms of health status, remaining healthspan, or lifespan.
Additionally, most of these models have been trained on specific demographic groups, and their ability to accurately predict age or disease risk or health outcomes in other populations remains a question.
Furthermore, the field is burgeoning with numerous aging clocks and new ones continually emerging. However, they currently don’t predict health outcomes with reliable accuracy. This is largely due to a lack of comprehensive studies; most of the available evidence comes from longitudinal studies, which, while valuable, cannot provide the complete picture.
Finally, the ethical implications of these clocks are a topic of discussion. If an individual’s predicted biological age or disease risk significantly differs from their chronological age, how should this information be used? Could it lead to discrimination or unequal access to resources?
As we continue to develop and refine these tools, it’s important to acknowledge these limitations and work towards addressing them. Despite these challenges, deep aging clocks and biological clocks offer promising avenues for research and have the potential to revolutionize how we approach aging and age-related diseases. As we continue to refine these tools and address their limitations, they could significantly contribute to the development of personalized medicine and health interventions to increase healthspan and lifespan.
Sources
- Biohorology and biomarkers of aging: Current state-of-the-art, challenges and opportunities.
F Galkin – Ageing Research Reviews, Jul 2020
DOI: https://doi.org/10.1016/j.arr.2020.101050 - DNA methylation age of human tissues and cell types.
S Horvath – Genome Biology, Oct 2013
DOI: https://doi.org/10.1186%2Fgb-2013-14-10-r115 - The transcriptional landscape of age in human peripheral blood.
MJ Peters – Nature Communications, Oct 2015
DOI: https://doi.org/10.1038/ncomms9570 - Human Gut Microbiome Aging Clock Based on Taxonomic Profiling and Deep Learning.
F Galkin – iScience, Jun 2020
DOI: https://doi.org/10.1016/j.isci.2020.101199 - Deep Aging Clocks: The Emergence of AI-Based Biomarkers of Aging and Longevity.
A Zhavoronkov – Trends in Pharmacological Sciences, Jul 2019
DOI: https://doi.org/10.1016/j.tips.2019.05.004 - Deep biomarkers of aging and longevity: from research to applications.
A Zhavoronkov – Aging, Nov 2019
DOI: https://doi.org/10.18632%2Faging.102475 - Accuracy and self-validation of automated bone age determination.
DD Martin – Nature Scientific Reports, Apr 2022
DOI: https://doi.org/10.1038/s41598-022-10292-y
