Scientists at Gladstone Institutes now say they have identified the reason. Their research shows that in low oxygen environments, red blood cells begin absorbing large amounts of glucose from the bloodstream. In effect, the cells act like sugar sponges under conditions similar to those found on the world’s tallest mountains.

In findings published in Cell Metabolism, the team demonstrated that red blood cells can alter their metabolism when oxygen levels drop. This shift allows the cells to deliver oxygen to tissues more efficiently at high altitude. At the same time, it lowers circulating blood sugar, offering a potential explanation for reduced diabetes risk.

According to senior author Isha Jain, PhD, a Gladstone Investigator, core investigator at Arc Institute, and professor of biochemistry at UC San Francisco, the study resolves a longstanding question in physiology.

“Red blood cells represent a hidden compartment of glucose metabolism that has not been appreciated until now,” Jain says. “This discovery could open up entirely new ways to think about controlling blood sugar.”

Red Blood Cells Identified as a Glucose Sink

Jain’s lab has spent years studying hypoxia, the term for reduced oxygen levels in the blood, and its effects on metabolism. In earlier experiments, her team noticed that mice exposed to low oxygen air had dramatically lower blood glucose levels. The animals rapidly cleared sugar from their bloodstream after eating, which is typically linked to lower diabetes risk. However, when researchers examined major organs to determine where the glucose was being used, they found no clear answer.

“When we gave sugar to the mice in hypoxia, it disappeared from their bloodstream almost instantly,” says Yolanda Martí-Mateos, PhD, a postdoctoral scholar in Jain’s lab and first author of the new study. “We looked at muscle, brain, liver — all the usual suspects — but nothing in these organs could explain what was happening.”

Using a different imaging method, the researchers discovered that red blood cells were serving as the missing “glucose sink,” meaning they were taking in and using significant amounts of glucose from circulation. This was unexpected because red blood cells have traditionally been viewed as simple oxygen carriers.

Follow up experiments in mice confirmed the finding. Under low oxygen conditions, the animals produced more red blood cells overall, and each individual cell absorbed more glucose compared with cells formed under normal oxygen levels.

To uncover the molecular details behind this shift, Jain’s group partnered with Angelo D’Alessandro, PhD, of the University of Colorado Anschutz Medical Campus, and Allan Doctor, MD, from University of Maryland, who has long studied red blood cell biology.

Their work showed that when oxygen is limited, red blood cells use glucose to generate a molecule that helps release oxygen to tissues. This process becomes especially important when oxygen is in short supply.

“What surprised me most was the magnitude of the effect,” D’Alessandro says. “Red blood cells are usually thought of as passive oxygen carriers. Yet, we found that they can account for a substantial fraction of whole-body glucose consumption, especially under hypoxia.”

Implications for Diabetes Treatment

The researchers also found that the metabolic benefits of prolonged hypoxia lasted for weeks to months after mice were returned to normal oxygen levels.

They then evaluated HypoxyStat, a drug recently developed in Jain’s lab that mimics low oxygen exposure. HypoxyStat is taken as a pill and works by causing hemoglobin in red blood cells to bind oxygen more tightly, limiting the amount delivered to tissues. In mouse models of diabetes, the medication completely reversed high blood sugar and outperformed existing treatments.

“This is one of the first use of HypoxyStat beyond mitochondrial disease,” Jain says. “It opens the door to thinking about diabetes treatment in a fundamentally different way — by recruiting red blood cells as glucose sinks.”

The findings may also apply beyond diabetes. D’Alessandro notes potential relevance for exercise physiology and for pathological hypoxia after traumatic injury. Trauma remains a leading cause of death among younger people, and changes in red blood cell production and metabolism could affect glucose availability and muscle performance.

“This is just the beginning,” Jain says. “There’s still so much to learn about how the whole body adapts to changes in oxygen, and how we could leverage these mechanisms to treat a range of conditions.”

Study Details and Funding

The study, titled “Red Blood Cells Serve as a Primary Glucose Sink to Improve Glucose Tolerance at Altitude,” appeared in Cell Metabolism on February 19, 2026. The authors include Yolanda Martí-Mateos, Ayush D. Midha, Will R. Flanigan, Tej Joshi, Helen Huynh, Brandon R. Desousa, Skyler Y. Blume, Alan H. Baik, and Isha Jain of Gladstone; Zohreh Safari, Stephen Rogers, and Allan Doctor of University of Maryland; and Shaun Bevers, Aaron V. Issaian, and Angelo D’Alessandro of University of Colorado Anschutz.

Funding was provided by the National Institutes of Health (DP5 DP5OD026398, R01 HL161071, R01 HL173540, R01HL146442, R01HL149714, DP5OD026398), the California Institute for Regenerative Medicine, Dave Wentz, the Hillblom Foundation, and the W.M. Keck Foundation.

Earlier research found that ultramarathon runners often experience a breakdown of healthy red blood cells during races, which can potentially lead to anemia. However, scientists have not fully understood why this happens. The new study found that after prolonged races, red blood cells become less flexible. Because these cells must bend to pass through tiny blood vessels while delivering oxygen and removing waste, reduced flexibility may limit their efficiency. The team also created the most detailed molecular profile to date showing how endurance races alter red blood cells.

“Participating in events like these can cause general inflammation in the body and damage red blood cells,” said the study’s lead author, Travis Nemkov, PhD, associate professor in the department of biochemistry and molecular genetics at the University of Colorado Anschutz. “Based on these data, we don’t have guidance as to whether people should or should not participate in these types of events; what we can say is, when they do, that persistent stress is damaging the most abundant cell in the body.”

Inside the Study of Ultramarathon Runners

To examine these effects, researchers measured indicators of red blood cell health before and after athletes competed in two demanding races: the Martigny-Combes à Chamonix race (40 kilometers or about 25 miles long) and the Ultra Trail de Mont Blanc race (171 kilometers or 106 miles long). Red blood cells are responsible for carrying oxygen and transporting waste products throughout the body, and their ability to flex is critical for moving through narrow blood vessels.

The team collected blood samples from 23 runners immediately before and after their races. They analyzed thousands of proteins, lipids, metabolites, and trace elements in both plasma and red blood cells. The results consistently showed signs of injury driven by both mechanical (physical) and molecular factors. Mechanical stress likely resulted from shifts in fluid pressure as blood circulates during intense running. Molecular damage appeared linked to inflammation and oxidative stress (when the body has low levels of antioxidants, which fight off molecules that damage DNA and other components within cells).

Longer Races, Greater Cellular Stress

Evidence of accelerated aging and increased breakdown of red blood cells was visible after the 40 kilometer race and was even more pronounced among athletes who completed the 171 kilometer event. Based on these findings, researchers suggest that longer races may lead to greater loss of red blood cells and more damage to those that remain in circulation.

“At some point between marathon and ultra-marathon distances, the damage really starts to take hold,” said Dr. Nemkov. “We’ve observed this damage happening, but we don’t know how long it takes for the body to repair that damage, if that damage has a long-term impact, and whether that impact is good or bad.”

Implications for Performance and Blood Storage

With additional research, the team believes these findings could help guide personalized training, nutrition, and recovery strategies aimed at improving performance while limiting potential harm from extreme endurance exercise. The work may also have broader medical relevance. Stored blood used for transfusions begins to deteriorate after several weeks and must be discarded after six weeks under U.S. Food and Drug Administration regulations. Understanding how intense physical stress affects red blood cells could provide insight into improving blood storage practices.

“Red blood cells are remarkably resilient, but they are also exquisitely sensitive to mechanical and oxidative stress,” said study co-author, Angelo D’Alessandro, PhD, professor at the University of Colorado Anschutz and member of the Hall of Fame of the Association for the Advancement of Blood and Biotherapies. “This study shows that extreme endurance exercise pushes red blood cells toward accelerated aging through mechanisms that mirror what we observe during blood storage. Understanding these shared pathways gives us a unique opportunity to learn how to better protect blood cell function both in athletes and in transfusion medicine.”

Study Limitations and Future Research

The research included a small group of participants and lacked racial diversity. Blood samples were also collected at only two time points. The investigators plan to expand future studies to include more participants, additional blood samples, and more detailed measurements after races. They also intend to further explore ways to extend the shelf life of stored blood.

To compare performance directly, researchers assigned identical tasks to different groups. Some teams relied entirely on human expertise, while others used scientists working with AI tools. The challenge was to predict preterm birth using data from more than 1,000 pregnant women.

Even a junior research pair made up of a UCSF master’s student, Reuben Sarwal, and a high school student, Victor Tarca, successfully developed prediction models with AI support. The system generated functioning computer code in minutes — something that would normally take experienced programmers several hours or even days.

The advantage came from AI’s ability to write analytical code based on short but highly specific prompts. Not every system performed well. Only 4 of the 8 AI chatbots produced usable code. Still, those that succeeded did not require large teams of specialists to guide them.

Because of this speed, the junior researchers were able to complete their experiments, verify their findings, and submit their results to a journal within a few months.

“These AI tools could relieve one of the biggest bottlenecks in data science: building our analysis pipelines,” said Marina Sirota, PhD, a professor of Pediatrics who is the interim director of the Bakar Computational Health Sciences Institute (BCHSI) at UCSF and the principal investigator of the March of Dimes Prematurity Research Center at UCSF. “The speed-up couldn’t come sooner for patients who need help now.”

Sirota is co-senior author of the study, published in Cell Reports Medicine on Feb. 17.

Why Preterm Birth Research Matters

Speeding up data analysis could improve diagnostic tools for preterm birth — the leading cause of newborn death and a major contributor to long term motor and cognitive challenges in children. In the United States, roughly 1,000 babies are born prematurely each day.

Researchers still do not fully understand what causes preterm birth. To investigate possible risk factors, Sirota’s team compiled microbiome data from about 1,200 pregnant women whose outcomes were tracked across nine separate studies.

“This kind of work is only possible with open data sharing, pooling the experiences of many women and the expertise of many researchers,” said Tomiko T. Oskotsky MD, co-director of the March of Dimes Preterm Birth Data Repository, associate professor in UCSF BCHSI, and co-author of the paper.

However, analyzing such a vast and complex dataset proved challenging. To tackle this, the researchers turned to a global crowdsourcing competition called DREAM (Dialogue on Reverse Engineering Assessment and Methods).

Sirota co-led one of three DREAM pregnancy challenges, focusing specifically on vaginal microbiome data. More than 100 teams worldwide participated, developing machine learning models designed to detect patterns linked to preterm birth. Most groups completed their work within the three month competition window. Yet it took nearly two years to consolidate the findings and publish them.

Testing AI on Pregnancy and Microbiome Data

Curious whether generative AI could shorten that timeline, Sirota’s group partnered with researchers led by Adi L. Tarca, PhD, co-senior author and professor in the Center for Molecular Medicine and Genetics at Wayne State University in Detroit, MI. Tarca had led the other two DREAM challenges, which focused on improving methods for estimating pregnancy stage.

Together, the researchers instructed eight AI systems to independently generate algorithms using the same datasets from the three DREAM challenges, without direct human coding.

The AI chatbots received carefully written natural language instructions. Much like ChatGPT, the systems were guided through detailed prompts designed to steer them toward analyzing the health data in ways comparable to the original DREAM participants.

Their objectives mirrored the earlier challenges. The AI systems analyzed vaginal microbiome data to identify signs of preterm birth and examined blood or placental samples to estimate gestational age. Pregnancy dating is almost always an estimate, yet it determines the type of care women receive as pregnancies progress. When estimates are inaccurate, preparing for labor becomes more difficult.

Researchers then ran the AI generated code using the DREAM datasets. Only 4 of the 8 tools produced models that matched the performance of the human teams, although in some cases the AI models performed better. The entire generative AI effort — from inception to submission of a paper — took just six months.

Scientists emphasize that AI still requires careful oversight. These systems can produce misleading results, and human expertise remains essential. However, by rapidly sorting through massive health datasets, generative AI may allow researchers to spend less time troubleshooting code and more time interpreting results and asking meaningful scientific questions.

“Thanks to generative AI, researchers with a limited background in data science won’t always need to form wide collaborations or spend hours debugging code,” Tarca said. “They can focus on answering the right biomedical questions.”

Authors: UCSF authors are Reuben Sarwal; Claire Dubin; Sanchita Bhattacharya, MS; and Atul Butte, MD, PhD. Other authors are Victor Tarca (Huron High School, Ann Arbor, MI); Nikolas Kalavros and Gustavo Stolovitzky, PhD (New York University); Gaurav Bhatti (Wayne State University); and Roberto Romero, MD, D(Med)Sc (National Institute of Child Health and Human Development (NICHD)).

Funding: This work was funded by the March of Dimes Prematurity Research Center at UCSF, and by ImmPort. The data used in this study was generated in part with support from the Pregnancy Research Branch of the NICHD.

The research zeroed in on polymer binders used in the negative electrodes of lithium-ion batteries (anodes). These binders act like a glue that holds the electrode materials together. Even though they account for less than 5% of the electrode’s total weight, they strongly influence mechanical strength, electrical and ionic conductivity, and how long a battery can operate through repeated charge cycles.

Because binders are present in such small amounts and lack clear visual signatures, scientists have had difficulty determining exactly where they are located within the electrode. This has limited efforts to fine tune battery performance, since the way binders are distributed directly affects conductivity, structural stability, and long term durability.

Patent Pending Staining Technique Reveals Hidden Structure

To overcome this obstacle, the researchers designed a patent-pending staining approach that attaches traceable silver and bromine markers to widely used cellulose- and latex-based binders in graphite- and silicon-based anodes. Once labeled, the binders can be detected because they emit characteristic X-rays (measured with energy-dispersive X-ray spectroscopy) or reflect high-energy electrons from the sample surface (measured with energy-selective backscattered electron imaging).

When viewed under an electron microscope, these signals provide detailed maps of where specific elements are located and what the electrode surface looks like. This allows scientists to analyze binder distribution with far greater precision than before.

Lead author Dr. Stanislaw Zankowski (Department of Materials, University of Oxford) said: “This staining technique opens up an entirely new toolbox for understanding how modern binders behave during electrode manufacturing. For the first time, we can accurately see the distribution of these binders not only generally (i.e., their thickness throughout the electrode), but also locally, as nanoscale binder layers and clusters, and correlate them with anode performance.”

The method works with standard graphite electrodes as well as advanced materials such as silicon or SiOx, making it relevant for both current lithium-ion batteries and next-generation designs.

Faster Charging and Longer Battery Life

By applying the new imaging tool, the team discovered that even subtle shifts in binder distribution can significantly change how efficiently a battery charges and how long it lasts. In testing, adjustments to slurry mixing and drying steps reduced the internal ionic resistance of experimental electrodes by as much as 40% — a major barrier to fast charging.

The researchers also captured detailed images of extremely thin layers of carboxymethyl cellulose (CMC) binder that coat graphite particles. The technique enabled clear detection of CMC layers just 10 nm thick and visualized structures spanning four orders of magnitude within a single image. The images revealed that what begins as a uniform CMC coating can break apart into uneven, patchy fragments during electrode processing, which may weaken battery performance and stability.

Co-author Professor Patrick Grant (Department of Materials, University of Oxford) said: “This multidisciplinary effort-spanning chemistry, electron microscopy, electrochemical testing, and modelling- has resulted in an innovative imaging approach that will help us to understand key surface processes that affect battery longevity and performance. This will drive forward advancements across a wide range of battery applications.”

The work was supported by the Faraday Institution’s Nextrode project and has already drawn significant interest from industry, including major battery producers and electric vehicle manufacturers.