Study Maps Existing Drugs to the Hallmarks of Aging

A new study suggests a way to predict whether existing drugs can extend human lifespan. This method uses a network approach that detects longevity signals in protein interactions [1].

Slowing aging with existing drugs?

Finding drugs that can slow aging, a multifactorial process involving thousands of genes, is hard. Directly measuring the effect of a prospective drug on human aging would take decades, so scientists need proxy markers, such as epigenetic clocks. Another hurdle is regulation, as the system is geared towards single-disease indications. On the other hand, thousands of drugs for these diseases have already been approved. What if some of them also affect the rate of aging, and how can we unlock this potential?

A new study from Northeastern University and Harvard, published in Nature Aging, offers an interesting way forward. “As someone whose hair turned gray years ago, I share the universal wish that there might one day be a pill that slows aspects of aging,” said Albert-László Barabási, Distinguished University Professor of Physics at Northeastern who oversaw the study. “The challenge is figuring out which drugs are worth testing.”

Hitting the hallmarks

The study is based on network medicine, a framework that the Barabási lab has developed over fifteen years. Its core idea is that proteins do not act in isolation but form a giant graph (the interactome) that shows which proteins interact physically or functionally. Genes underlying a given disease tend to clump together into a “disease module.” Once such a module is outlined, scientists can ask which drugs have targets in the vicinity of that “neighborhood” and are thus candidates for perturbing it. This approach has been used before for asthma, heart disease, and COVID-19 [2].

The Hallmarks of Aging has been a defining paradigm in geroscience for many years. It includes crucial biological features and processes that get disrupted with age, such as DNA stability and intercellular communication. This study is built on the premise that each hallmark of aging behaves not unlike a disease module and that the same machinery that network medicine applies to diseases can therefore be used with these hallmarks.

“You have genes related to aging by some definition or by some reasoning, but it feels like you just have a very big pile of genes related to aging,” said Bnaya Gross, a postdoctoral researcher in Barabási’s lab at Northeastern and lead author of the study. “Networks allow us to organize them, saying, OK, it’s not just a pile of genes. They are connected to each other. They form some sort of organization. It’s not a random process.”

The researchers started from the OpenGenes database, a manually curated resource linking 2,358 genes to aging/longevity. Each gene is tagged with a confidence level from 1 (changing the gene’s activity actually extends mammalian lifespan) to 5 (lowest, weak association). Notably, only 26 genes sit at confidence level 1. 1,250 of the genes could be assigned to at least one hallmark of aging; the remaining 1,108 are still aging-linked but couldn’t be pinned to a specific hallmark. 860 genes belong to a single hallmark and 390 to multiple hallmarks, with TP53 spanning the most (seven).

The fact that many genes are related to several hallmarks shows their interconnectedness via shared molecular machinery. The network approach is quite good at catching this complex web of interactions, while also showing that the hallmarks occupy distinct nodes in it. The team then validated that their 1,250-gene set is largely related to aging and mapped the genes onto the human interactome: a network of more than 500 thousand experimentally supported interactions among proteins.

The team then took 6,442 compounds from DrugBank and, for each hallmark module, measured each drug’s network proximity – the average shortest-path distance from the drug’s protein targets to the nearest hallmark genes. Drugs whose targets sit significantly closer than chance are predicted to perturb that hallmark. However, they found that some of the hits acted in the opposite direction – for instance, induced rather than lowered cellular senescence. So, proximity alone correctly identifies drugs that act on a hallmark but not whether they help or harm.

Turning the pAGE

The researchers devised a metric that accounts for directionality and called it pAGE. They outlined a Systematic Hallmark-based Aging Repurposing Pipeline (SHARP) based on proximity and pAGE, and they validated it against drugs tested in mammalian longevity studies. For instance, out of the eight compounds that increased mouse lifespan in the Intervention Testing Program (ITP) trials and also had interactomic data, all had a positive pAGE for at least one hallmark. Of the drugs that failed in the ITP, less than half did. Interestingly, some pro-longevity drugs were beneficial for some hallmarks but harmful for others, highlighting possible trade-offs.

Another test came from drugs currently in human longevity trials, such as metformin and rapamycin. Of the 17 compounds, 11 had significant proximity. Interestingly, aspirin was mapped to six hallmarks and dasatinib to five, whereas rapamycin hit only one: intercellular communication. Finally, the team successfully tested their method on 10 compounds from a parallel study whose results appeared after their predictions [3] – the closest thing to a prospective test.

The researchers then ran SHARP across all hallmarks to generate candidates. They identified 370 drugs that are proximal to at least one hallmark, and 83 of them are “network drugs” that do not directly target any aging gene but affect a hallmark module through the network topology. These would be invisible to any method that looks solely at direct drug-target relationships, and they are the strongest argument for the network approach. The team showed their predictions are mechanistically interpretable by analyzing how exactly oxymetazoline – a nasal/topical decongestant and rosacea drug – affects longevity-related genes.

Literature

[1] Gross, B., Ehlert, J., Gladyshev, V. N., Loscalzo, J., & Barabási, A. L. (2026). Network-driven discovery of repurposable drugs targeting hallmarks of aging. Nature Aging, 1-16.

[2] Morselli Gysi, D., Do Valle, Í., Zitnik, M., Ameli, A., Gan, X., Varol, O., … & Barabási, A. L. (2021). Network medicine framework for identifying drug-repurposing opportunities for COVID-19. Proceedings of the National Academy of Sciences, 118(19), e2025581118.

[3] Shindyapina, A. V., Tyshkovskiy, A., Bozaykut, P., Castro, J. P., Gerashchenko, M. V., Trapp, A., … & Gladyshev, V. N. (2025). Molecular signatures of longevity identify compounds that extend mouse lifespan and healthspan. bioRxiv.