Stanford researchers report that a major driver of aging across multiple organs is not damage accumulation per se but a failure of garbage collection. Tissue-resident macrophages, the long-lived immune cells embedded in liver, brain, heart, lung and spleen, normally eat and dispose of the roughly 100 billion neutrophils the body produces daily. With age, a receptor called EP2 (the receptor for prostaglandin E2, a product of the COX-2 inflammatory pathway) becomes overexpressed on these macrophages and shuts down their metabolism and their ability to engulf cargo. Senescent neutrophils then pile up in tissues, where they degranulate, release DNA nets, and stress neighboring cells. Deleting EP2 from tissue-resident macrophages in aged mice, or blocking it with a drug for two months starting at 22 months of age, restored clearance and was accompanied by better memory, more muscle, less fat, lower frailty scores, better cardiac function, and a plasma protein profile resembling that of young mice.
For decades, aging research has focused on damage: broken DNA, worn telomeres, misfolded proteins, exhausted stem cells. A study published in Science in July 2026 argues for a different framing. The problem may be less that the body accumulates junk, and more that it stops picking it up.
The junk in question is neutrophils. These are the most abundant white blood cell and the shortest lived, surviving only 8 to 12 hours in circulation. Humans make about 100 billion of them a day. Old neutrophils that are not removed promptly become dangerous, spilling digestive enzymes and sticky webs of DNA into surrounding tissue. The cells responsible for removing them are tissue-resident macrophages, a distinct and often overlooked population that seeds organs before birth and then quietly maintains itself there for the rest of the animal’s life. They make up 60 to 90 percent of all macrophages in the brain, liver, heart and kidney.
The Stanford team, led by Katrin Andreasson, found that aged macrophages become progressively deaf to their own maintenance job because of a single receptor. EP2 receives signals from prostaglandin E2, an inflammatory lipid made by the COX-2 enzyme, the same pathway targeted by ibuprofen and celecoxib. In aged mice, EP2 activity rises in tissue-resident macrophages specifically, suppressing their mitochondrial function and disabling the molecular grip they use to grab and swallow dying cells.
When the researchers genetically removed EP2 from these macrophages, aged mice looked considerably younger by many measures at once. They solved spatial memory mazes as well as young animals. They had more limb muscle, less visceral fat, stronger grip, lower frailty scores, better heart ejection fraction and dramatically less cardiac fibrosis. Liver enzymes and blood inflammatory markers fell back toward youthful values. Of 71 blood proteins that changed with age, 59 were pushed back toward a young profile.
Critically, the same benefit was partly reproduced with a drug. Two months of PF-04418948, an experimental EP2 blocker, given to 22-month-old mice restored liver macrophage clearance capacity to nearly youthful levels, though effects in spleen and bone marrow were weaker.
Human tissue data are supportive but correlational: aged and diseased human livers and hearts also show fewer resident macrophages, elevated EP2 expression in those macrophages, and expanded senescent-like neutrophils. Nobody has tested this in a living human. And notably, the mice were never followed to death, so whether any of this extends life remains unknown.
Actionable Insights
Bluntly: there is no validated human action here. This is an all-male mouse study with no survival data. What it offers is a framework.
The underlying defect is large. Macrophages from old mice engulfed senescent neutrophils about 10 times less efficiently than young ones (roughly 60 percent uptake in young versus 6 percent in old). Removing EP2 recovered about 45 percent of that lost capacity for senescent neutrophils and 77 to 90 percent for ordinary dying cells. Senescent neutrophil burden in the liver fell from roughly 68 percent of neutrophils to 18 percent, about 87 percent of the way back to young. These are not marginal effects; treated and untreated groups barely overlap.
The tempting inference is that COX-2 inhibitors like celecoxib or daily ibuprofen would replicate this. Resist it. This paper blocked one of four PGE2 receptors in one cell type, whereas blanket COX inhibition also suppresses beneficial prostaglandins and carries real cardiovascular, renal and gastrointestinal risk. PF-04418948 is a laboratory compound with no human safety data.
What is reasonable now: track neutrophil-driven inflammation. Neutrophil-to-lymphocyte ratio, high-sensitivity CRP, and liver enzymes all moved coherently with the phenotype here and are cheap to measure. Lowering chronic inflammatory tone by conventional means remains the defensible play. [Confidence: Medium for the biomarker framing, High for the caution against self-medicating with COX inhibitors]
Context and Source
- Paywalled Paper: Restored clearance of senescent neutrophils by tissue-resident macrophages limits organ aging
- Authors: Yuting Jessy Tan, Travis E. Conley, Fuwen Yao, Fernando J. Garcia-Marques, Damilola E. Akinyemi, Van Vuong Dinh, Qian Wang, Abel Bermudez, Jieun Kim, Julia A. Belk, Oliver Soehnlein, Sharon J. Pitteri, Katrin I. Andreasson (corresponding)
- Institutions: Stanford University School of Medicine (Neurology, Radiology, Wu Tsai Neurosciences Institute, Knight Initiative for Brain Resilience), USA; Institute for Experimental Pathology, University of Munster, Germany; Chan Zuckerberg Biohub San Francisco, USA
- Country: United States (with German collaboration)
- Journal: Science (AAAS), volume 393, issue 6808, article eaea3075, 16 July 2026.
- Funding: NIH, American Heart Association, Phil and Penny Knight Initiative, Glenn Foundation, Chan Zuckerberg Biohub
-
Competing interest disclosure worth noting: senior author K.I.A. is a cofounder of Willow Neuroscience Inc.
Impact evaluation: The impact score of this journal is 47.3 (Journal Impact Factor, 2026 JCR release; CiteScore 48.4), evaluated against a typical high-end range of 0 to 60+ for top general science, therefore this is an Elite impact journal.