Network Science Reveals Which Existing Drugs May Target the Hallmarks of Aging
Researchers from Northeastern and Harvard mapped 6,442 drugs against aging hallmark genes using protein interaction networks to identify longevity candidates.
Summary
Scientists have developed a network-based method to identify existing approved drugs that may slow aging. Using a protein interaction map called the interactome, researchers from Northeastern University and Harvard linked 1,250 aging-related genes to the hallmarks of aging, then tested 6,442 known compounds for proximity to those gene clusters. Drugs whose protein targets sit close to hallmark gene modules are flagged as candidates for slowing aging processes. This approach sidesteps the decades-long wait needed for direct lifespan studies, using network proximity as a proxy signal. Published in Nature Aging, the study offers a systematic, data-driven shortcut for repurposing already-approved drugs as potential longevity therapeutics, dramatically narrowing the field of candidates worth testing in future trials.
Detailed Summary
Aging is driven by thousands of interacting genes, making it extraordinarily difficult to identify which drugs might slow the process. A new study published in Nature Aging from Northeastern University and Harvard proposes an elegant shortcut: use the mathematics of protein interaction networks to predict which existing, approved drugs are most likely to interfere with the biological machinery of aging.
The research team, led by physicist Albert-László Barabási and postdoctoral researcher Bnaya Gross, built on the established framework of network medicine. This approach maps proteins as nodes in a vast interconnected graph — the human interactome — containing over 500,000 experimentally supported interactions. Disease-related genes tend to cluster into identifiable modules within this network. The team applied this same logic to the hallmarks of aging, treating each hallmark as its own disease-like module.
Starting from the OpenGenes database, the researchers identified 2,358 aging-linked genes, of which 1,250 could be assigned to at least one hallmark. They then assessed 6,442 compounds from DrugBank, measuring each drug's network proximity — essentially how close its protein targets fall to hallmark gene clusters. Shorter distances suggest a drug is more likely to meaningfully perturb aging pathways.
Key insights include the finding that many genes span multiple hallmarks, highlighting shared molecular machinery across aging processes. The gene TP53 alone bridges seven hallmarks. This interconnectedness is precisely what network analysis captures well, revealing that aging is not a random pile of genetic events but an organized, mappable system.
For health-conscious individuals and clinicians, the practical implication is significant: this framework could accelerate the repurposing of existing drugs for longevity applications, bypassing lengthy drug development timelines. However, network proximity is a computational proxy, not a clinical guarantee. Validation in animal models and human trials will be essential before any specific drug can be recommended for aging intervention.
Key Findings
- 1,250 aging-related genes were mapped onto the human interactome and organized by hallmarks of aging.
- 6,442 existing approved compounds were scored for network proximity to hallmark gene clusters as longevity candidates.
- TP53 spans seven hallmarks, illustrating how shared molecular machinery links multiple aging processes.
- Network proximity provides a faster computational proxy for identifying anti-aging drug candidates than direct lifespan studies.
- The method builds on 15 years of network medicine research previously applied to heart disease, asthma, and COVID-19.
Methodology
This is a research summary of a peer-reviewed study published in Nature Aging, a high-credibility journal. The evidence basis is computational, using network proximity analysis across curated gene and drug databases. No clinical trial data is presented; findings are hypothesis-generating and require experimental validation.
Study Limitations
The study relies on computational network proximity as a proxy for biological efficacy, which has not yet been validated in human longevity outcomes. Only 26 of 2,358 aging genes carry the highest confidence level, meaning much of the gene set rests on weaker evidence. Primary source should be consulted for full drug candidate lists and statistical thresholds used.
Enjoyed this summary?
Get the latest longevity research delivered to your inbox every week.
Enter your email to subscribe: