Scientists have discovered that directly reducing the production of reactive oxygen species (ROS) at their source in astrocytes, mitochondrial complex III, improves neuronal health and significantly increases lifespan in a mouse model of Alzheimer’s [1].
Dangerous species
Reactive oxygen species (ROS) are short-lived, highly reactive oxygen-containing molecules such as superoxide and hydrogen peroxide that are formed as byproducts of normal metabolism, particularly in mitochondria. In small numbers, ROS can be important signaling molecules, but excessive ROS levels trigger oxidative stress, which damages proteins, lipids, and DNA. Oxidative stress has been found to play a major role in age-related conditions, including dementia [2].
Scientists have proposed various antioxidant strategies, and some antioxidants have become popular supplements. The human body also produces natural antioxidants, such as glutathione. However, those scavengers cannot neutralize ROS immediately at the sites of production, so some damage is inevitable [3]. In a new study published in Nature Metabolism, researchers from Weill Cornell Medical College and other institutions attempted a novel, highly targeted approach to brain ROS.
“Decades of research implicate mitochondrial ROS in neurodegenerative diseases,” said Dr. Adam Orr, an assistant professor of research in neuroscience at the Feil Family Brain and Mind Research Institute at Weill Cornell, who co-led this study. “But most antioxidants tested in clinical studies have failed. That lack of success might be related to the inability of antioxidants to block ROS at their source and do so selectively without altering cell metabolism.”
Patching the leak at the source
The same team previously identified small molecules that can target mitochondrial ROS right where they are generated. They call these molecules site-selective electron-leak suppressors (SELs): S3QELs (pronounced “sequels”) for the complex III site and S1QELs (pronounced “cycles”) for the complex I site.
Trying to understand the role of astrocytes, a type of supportive brain cell, in brain pathologies, the researchers exposed them to disease-relevant cues, such as the inflammatory cytokine IL-1α and oligomeric amyloid-β (Aβ), a hallmark of Alzheimer’s disease. Both of these compounds increased mitochondrial hydrogen peroxide (H₂O₂), a major ROS, indicating stimulus-dependent ROS generation at complex III. S3QELs blunted these increases while preserving ATP production.
Importantly, the researchers mapped specific protein cysteine oxidations, showing that ROS produced at complex III serve as signaling inputs. Under pathological cues, however, this signaling becomes overactive, amplifying disease-associated transcriptional changes in astrocytes.
In neuron-astrocyte co-cultures, astrocytes primed to produce complex-III ROS made nearby neurons fare worse than neurons paired with quiescent astrocytes. Crucially, applying S3QELs to the astrocytes (not the neurons) attenuated neuronal harm. The effect also held in a conditioned medium, meaning that neuronal injury was largely driven by the molecules that astrocytes secrete downstream of complex-III ROS.
Interestingly, results differed when neurons were co-cultured with microglia, the brain’s resident macrophages. Primed microglia could also harm neurons, but dialing down complex-III ROS with S3QELs did not rescue neurons the way it did in the astrocyte–neuron setup. The authors suggest that in microglia, neuronal injury depends less on complex-III ROS and more on other pathways.
Longer lifespan in an Alzheimer’s model
Finally, the researchers moved to an in vivo tauopathy model: PS19 mice expressing human tauP301S. In pharmacokinetics experiments, S3QEL2 crossed the blood-brain barrier and was generally well tolerated.
Mice given a six-week bolus regimen (yielding higher brain exposure) showed significantly lower hippocampal phosphorylated tau and reductions in several inflammation markers compared to controls. With chronic chow dosing, pathology effects were minimal and attributed to much lower brain levels compared with bolus feeding. However, in a lifespan cohort, chow-dosed mice still lived longer than controls: median lifespan increased by about 17%, and the oldest ages reached increased by 19.9%; of course, PS19 mice have markedly shorter lifespans than wild-type mice.
The idea of suppressing ROS production at the source may have broader applications. “I’m really excited about the translational potential of this work,” said Dr. Anna Orr, the Nan and Stephen Swid Associate Professor of Frontotemporal Dementia Research at the Feil Family Brain and Mind Research Institute and member of the Appel Alzheimer’s Disease Research Institute at Weill Cornell, who co-led this research. “We can now target specific mechanisms and go after the exact sites that are relevant for disease.”
Literature
[1] Barnett, D., Zimmer, T.S., Booraem, C. et al. (2025). Mitochondrial complex III-derived ROS amplify immunometabolic changes in astrocytes and promote dementia pathology. Nat Metab.
[2] Perluigi, M., Di Domenico, F., & Butterfield, D. A. (2023). Oxidative damage in neurodegeneration: roles in the pathogenesis and progression of Alzheimer disease. Physiological Reviews.
[3] Watson, M. A., Wong, H. S., & Brand, M. D. (2019). Use of S1QELs and S3QELs to link mitochondrial sites of superoxide and hydrogen peroxide generation to physiological and pathological outcomes. Biochemical Society Transactions, 47(5), 1461-1469.
View the article at lifespan.io
