In a 2023 paper on hypoxia and glucose metabolism , our lab showed how organisms rewire their metabolism to adapt to low oxygen levels—such as those found at high altitudes. One of the most striking observations from that work was a dramatic drop in circulating blood sugar.
That study focused on mice exposed to hypoxia. Looking at epidemiological data from the United States, people living at even modest elevations show the same pattern: lower blood glucose, better glucose tolerance, reduced diabetes risk.
It left us wondering where all that glucose from the blood was going. The most traditional explanation would have been insulin signaling, which tells muscle and fat cells to pull glucose out of the blood. But surprisingly, PET/CT scans showed that even after analyzing all major organs, 70% of the increased glucose clearance in hypoxic mice remained unaccounted for. Something else was going on, which we describe today in Cell Metabolism .
The missing sink
We began to suspect glucose was being consumed by a cell within the blood itself.
Red blood cells (RBCs) seemed like an unlikely–but appealing–answer. They've never been thought of as regulators of glucose homeostasis. They don't have nuclei or mitochondria. They're primarily composed of hemoglobin with relatively few known mechanisms to actively regulate their metabolism.
But they're also the most abundant cell type in the body and in chronic hypoxia their numbers dramatically increase. RBCs rely entirely on glucose for energy, since they can't do oxidative metabolism. And they are invisible in PET scans because they're constantly moving.
To test this provocative idea, we returned to some 'old-school' techniques to see if RBCs were truly the missing piece of the puzzle. First, we prevented the hypoxic RBC boost by repeatedly drawing blood from hypoxic mice, keeping their RBC counts at normal levels. This alone was enough to normalize blood glucose and reverse the hypoglycemia.
We then did the opposite and transfused RBCs into normal mice breathing regular air. Just adding more RBCs was sufficient to drop their blood sugar. At this point, we knew we were onto something, so we became very comfortable with our hypothesis.
Understanding the mechanism
Interestingly, during our earlier experiments, we noticed that individual RBCs from hypoxic mice were taking up significantly more glucose than RBCs from animals raised in normal conditions. A straightforward explanation would be an increase in glucose transporters.
Using flow cytometry, we found that RBCs from hypoxic mice had significantly higher GLUT1 glucose transporter abundance per cell. But this was puzzling because mature red blood cells don't have a nucleus, meaning that they can't transcribe genes or make new proteins. So we wondered how those extra GLUT1 proteins got there.
Red blood cells are constantly being produced from the bone marrow and have a lifespan of a couple of months. We wanted to test whether RBCs born in hypoxic bone marrow were being programmed to make more GLUT1 during their development and then maintain that higher transporter level once matured and in circulation.
This was a particularly fun experiment. We labeled all pre-existing RBCs with biotin for three consecutive days, then moved the mice to hypoxia. From that moment on, no additional biotin labeling happened, so any new red cells maturing in the bone marrow would be unlabeled. After four weeks, we separated old and new RBCs and measured their GLUT1 levels.
Interestingly, only the newly synthesized ones showed upregulated GLUT1. The old cells showed no change at all. This told us that once RBCs mature and enter circulation, they keep whatever glucose uptake capacity they were born with, but hypoxia reprograms the bone marrow to churn out a new, glucose-hungry population of RBCs.
RBCs as oxygen-sensors
We had figured out how glucose entered hypoxic RBCs more rapidly, but not what happened to it inside. To find out, we injected labeled glucose into mice and tracked its conversion in RBCs. The hypoxic RBCs metabolized glucose much faster than normal RBCs, converting it within minutes to 2,3-DPG (2,3-diphosphoglycerate). This molecule binds to hemoglobin and helps it release oxygen to tissues—exactly what the body needs at higher altitude.
Considering how acute this mechanism was, it seemed like mature RBCs were reacting real-time to low oxygen. This is where we reached out to Angelo D'Alessandro (University of Colorado) and Allan Doctor (University of Maryland), both leading experts in red blood cell biology. They were very enthusiastic and wanted to help us go deeper into the mechanism.
Together we figured out something that had been proposed decades ago but hadn't really been studied in this physiological context. Under normal oxygen, key glycolytic enzymes including GAPDH get sequestered at the RBC membrane by binding to a protein called Band 3, putting a brake on glycolysis. But when oxygen levels drop, hemoglobin releases its oxygen and changes shape. That deoxygenated hemoglobin can now compete for the Band 3 binding sites, freeing the glycolytic enzymes and allowing them to build more 2,3-DPG.
We confirmed this mechanism using crosslinking proteomics, STED microscopy, and proximity ligation assays in both mouse and human RBCs. This indicates that this elegant metabolic switch is conserved across species.
Implications for diabetes
Once we understood the mechanism, we tested whether it could work therapeutically. Three approaches reversed hyperglycemia in mouse diabetes models: exposing diabetic mice to hypoxia, transfusing RBCs into diabetic mice at normal oxygen, and treating high-fat diet mice with HypoxyStat —a small molecule our lab developed that causes tissue hypoxia in normal oxygen environments by increasing hemoglobin's oxygen affinity.
Of course, RBC transfusions aren't an ideal long-term therapy. But the findings suggest potential directions such as engineering RBCs to be more glucose-avid or targeting RBC turnover to shift populations toward younger, more metabolically active cells.
Major questions remain. The study made us wonder what the ultimate fate of glucose is in RBCs once 2,3-DPG is produced. And more importantly, if RBCs are capable of consuming glucose at this scale: what other aspects of whole-body physiology have we been overlooking?
This project moved unusually fast, less than one year from hypothesis to submission. That came partly from using older approaches that have fallen out of fashion such as phlebotomy, transfusion and biotin tracing. It also came from recognizing when we needed specialized expertise and finding the right collaborators.
Overall, this journey taught us the courage to chase 'crazy' hypotheses and reminded us that, sometimes, truth is hiding in plain sight.
Martí-Mateos, Y., Safari, Z., Bevers, S., Midha, A.D., Flanigan, W.R., Joshi, T., Huynh, H., Desousa, B.R., Blume, S.Y., Baik, A.H., Rogers, S., Issaian, A.V., Doctor, A., D'Alessandro, A., & Jain, I.H. (2026). Red Blood Cells Serve as a Primary Glucose Sink to Improve Glucose Tolerance at Altitude. Cell Metabolism. https://doi.org/10.1016/j.cmet.2026.01.019
Yolanda Martí-Mateos (X: @yolmarmat ) is a postdoctoral fellow in the Jain Lab at Gladstone Institutes.
Isha Jain (X: @ishahjain ) is an Arc Institute Core Investigator and an Associate Investigator at Gladstone Institutes.
