TL;DR
Researchers have developed a suite of 7 organ-specific MRI-based "aging clocks" (MRIBAGs) that measure the biological age gap of the brain, heart, liver, adipose tissue, spleen, kidney, and pancreas. By integrating imaging data from over 313,000 individuals with proteomics, metabolomics, and genetics, the study identifies 9 druggable genes and demonstrates that these structural aging markers can predict disease risk, mortality, and patient response in Alzheimer’s clinical trials.
Abstract
Leveraging clinical phenotypes, neuroimaging, proteomics, metabolomics, and epigenetics, biological aging clocks across organ systems and tissues have advanced our understanding of human aging and disease. In this study, we expand this biological aging clock framework to multi-organ magnetic resonance imaging (MRI) by developing 7 organ-specific MRI-based biological age gaps (MRIBAGs), including the brain, heart, liver, adipose tissue, spleen, kidney, and pancreas. Leveraging imaging, genetic, proteomic, and metabolomic data from 313,645 individuals curated by the MULTI consortium, we link the 7 MRIBAGs to 2,923 plasma proteins, 327 metabolites, and 6,477,810 common genetic variants. These associations reveal organ-specific and cross-organ interconnection landscapes, identifying distinct molecular signatures related to organ aging. Genome-wide associations identify 53 MRIBAG-locus pairs (P < 5×10⁻⁸). Genetic correlation and Mendelian randomization analyses further support organ-specific and cross-organ interconnections with phenotype-based, proteome-based, and metabolome-based aging clocks, as well as 525 disease endpoints. Through functional gene mapping and Bayesian colocalization analysis linking evidence from genetics, proteomics, and metabolomics, we prioritize 9 druggable genes as targets for future anti-aging treatments. Finally, we demonstrate the clinical relevance of the 7 MRIBAGs in predicting disease endpoints (e.g., diabetes mellitus), all-cause mortality, and capturing differential and heterogeneous cognitive decline trajectories over 240 weeks of treatment with the Alzheimer's disease drug (Solanezumab). Sex differences are evident across multiple organ systems, manifesting at structural, molecular, and genetic levels. In summary, we developed 7 MRI-based aging clocks that enhance the existing multi-organ biological aging framework, offer multi-scale insights into aging biology, and demonstrate clinical potential to advance future aging research.
Introduction
Aging is not a monolithic process; different organs in the same body can age at vastly different rates. While we have seen the rise of "epigenetic clocks" or "brain age" models, most existing biological aging frameworks are either limited to a single organ or rely solely on molecular data like blood markers. There is a significant gap in our ability to systematically "digitize" aging across multiple organ systems using non-invasive structural imaging. Furthermore, we haven't fully understood how the structural changes seen on an MRI scan relate to underlying genetic and metabolic landscapes. This paper aims to bridge that gap by creating a multi-organ MRI framework that connects macroscopic structural changes to microscopic biological drivers.
Methodology and Data Integration
The authors developed MRIBAGs (MRI-based biological age gaps) for seven organs: the brain, heart, liver, adipose tissue, spleen, kidney, and pancreas. The core workflow involved training age prediction models using Lasso regression and Linear Support Vector Regressors (SVR) to predict chronological age from Imaging-Derived Phenotypes (IDPs). The "age gap" is defined as the difference between the predicted biological age and the actual chronological age.
The study leveraged a massive amount of data curated by the MULTI consortium, primarily from the UK Biobank (UKBB) (~313,645 individuals). Independent replication of proteomic signals was performed using the Baltimore Longitudinal Study of Aging (BLSA), and clinical utility was tested using the Anti-Amyloid Treatment in Asymptomatic Alzheimer’s (A4) Study. Disease endpoint statistics were sourced from FinnGen and the Psychiatric Genomics Consortium (PGC).
Molecular Signatures of Organ Aging
Through Proteome-Wide (ProWAS) and Metabolome-Wide (MetWAS) association studies, the researchers linked structural gaps to 2,923 plasma proteins and 327 metabolites. This revealed distinct molecular signatures for organ aging. While some proteins like GDF15 appear to be systemic aging markers, others are highly specific, such as VCAM1 for the spleen and PLA2G1B for the pancreas.
Key Results and Clinical Utility
- Prediction Accuracy: The models achieved a Mean Absolute Error (MAE) of approximately 5 years across all 7 organs in independent test sets.
- Genetic Discovery: GWAS identified 53 significant MRIBAG-locus pairs.
- Druggable Targets: By triangulating evidence from MRI, genetics, and proteomics, the study prioritized 9 druggable genes (including ALDH2 and CDKN2A), suggesting structural MRI clocks can identify new therapeutic targets.
- Disease Prediction: MRIBAGs significantly predicted the incidence of diabetes mellitus (Hazard Ratios between 1.44 and 1.70) and all-cause mortality.
- Clinical Trial Stratification: The brain MRIBAG successfully stratified participants in the Solanezumab Alzheimer’s trial. "Decelerated agers" had higher cognitive scores than "accelerated agers," demonstrating the potential for these clocks in patient selection for future trials.
Discussion and Future Directions
The study highlights that sex differences in aging manifest at structural, molecular, and genetic levels across multiple organ systems. However, several limitations remain. The analysis is primarily cross-sectional, necessitating longitudinal studies to understand aging progression. Furthermore, imaging features for abdominal organs (spleen, kidney, pancreas) are currently less developed than those for the brain and heart, calling for more advanced feature engineering. Finally, the genetic risk scores currently explain a small portion of variance, limiting their standalone clinical use. Despite these challenges, the MRIBAG framework provides a robust foundation for multi-scale aging research and personalized clinical interventions.
Summary updated