AI-Driven Tool Links Lung Fibrosis to Accelerated Aging in IPF and Aging Research

Artificial intelligence (AI) is opening new frontiers in the fight against idiopathic pulmonary fibrosis (IPF).

In a study published in Aging on Aug. 8, 2025, researchers at Insilico Medicine unveiled AI-powered models that uncover shared biological pathways between IPF and the aging process, pointing to novel opportunities for drug discovery, biomarker development and personalized treatment strategies for patients with the deadly lung disease.

Carol Ann Satler, M.D.

“It is totally appropriate to treat IPF as its own disease and target the fibrotic processes that can happen only in the lung,” Carol Ann Satler, M.D., senior vice president of clinical development for the company and leader of the program, told Managed Healthcare Executive. “But we think that such a strategy needlessly narrows down the range of applications. Instead, by treating IPF as one manifestation of aging and targeting the processes that take place in all people, in every tissue, you get a chance to discover a drug that will treat many other diseases and could even improve the lifespan in general.”

In the study, the fibrosis-aware aging clock and IPF-P3GPT models showed strong predictive and generative capabilities.

“Aging clocks and omics are getting progressively more popular in clinical trials,” Alex Zhavoronkov, Ph.D., M.S., founder and CEO of Insilico Medicine, said. “In real-life clinical settings, however, these technologies are still more of a novelty, apart from some cases when omics are essential to select an effective therapy, such as tumor genomics.” In research- and discovery-oriented fields, AI tools are already quite popular, he said. These technologies let scientists analyze vastly larger amounts of information.

“Even something as vanilla as LLM [large language model] chats are great tools for literature reviews and staying up to date with the latest developments,” Zhavoronkov said. “For something more low-level, like models that work with omics directly, the greatest challenge is narrow utility. To generate biological knowledge, one needs to combine lots of dissimilar data types, which cannot be represented as text-only without losing important information.”

So, the AI tools, aging clocks, and other models are already helping a great deal, but they can do so much more once these problems are solved.

Extracellular matrix remodeling

The researchers identified four central pathways (TGF-β signaling, oxidative stress, inflammation and extracellular matrix [ECM] remodeling) as key in both aging and IPF but with different gene-level involvement. Satler believes ECM remodeling holds the most immediate promise for therapeutic intervention.

“But that is only because in a fibrotic disease these proteins are the end point of all other pathological processes,” she said. “You probably can't prevent faulty ECM remodeling without affecting the other three. All these processes are interconnected, and splitting them in four is just a way to formalize the molecular biology of the disease.” With Insilico’s success in accelerating preclinical drug discovery timelines, these models can be used to prioritize the targets based on which of the preselected genes it considers more important. Additionally, researchers could assess the effect a treatment has on the pace of aging in trial participants with a model that is not biased by being trained primarily on disease-free individuals.

“For prognosis, we need more data linked to IPF outcomes,” Satler said. “Much more data than is currently available in the public domain. For diagnosis, AI models could be used to notice pre-onset IPF (and other diseases), but such applications would be most efficient when combining multiple data sources, such as genotyping, medical imaging and personal health-related data.”

Looking ahead, the researchers plan on extending their research to other fibrotic and age-related diseases.

“Some diseases, such as T2D and IPF, are so well aligned with the hallmarks of aging that they can serve as the models of organismal aging,” Satler said. “IPF, while being a devastating disease, is particularly important for aging studies since the molecular processes that drive it are so intense that they effectively compress decades of aging into several years. We believe that any restorative IPF drug, or even one that can arrest its progression, would most likely enact its benefits via the hallmarks of aging and thus be easy to translate to any other aging-related disease.”