By the 1890s, just as the medical world was beginning to understand that invisible microbes caused infectious diseases, the idea that a missing nutrient could kill you was even harder to fathom.
Then the evidence started piling up. Dutch physician Christiaan Eijkman noticed chickens fed polished rice developed beriberi-like paralysis, cured by adding back the husks. It was the first clue that what we now call vitamin B1 was essential. Naval ships were losing a third of their sailors to scurvy on long voyages. Give them citrus fruit, rich in vitamin C, and deaths dropped to zero.
Over the following decades, researchers began searching for other essential nutrients that the body can't produce on its own. They called them vitamins, and the 'vitamin hunters' who identified them saved millions of lives, earning over a dozen Nobel Prizes along the way. By 1948, all 13 classical vitamins were known.
With readily available vitamins, much of our food started getting supplemented. While this had the obvious benefit of reducing dietary deficiencies, one downside was that people stopped paying attention. Vitamins became something you could buy at the grocery store without a clear rationale of how much, why, or the underlying mechanism for their benefits. There are case reports of physicians giving patients vitamin cocktails for various diseases, and sometimes it works and sometimes it doesn't. The literature is messy, but there are enough glimmers of hope that there's a real signal in there.
Our lab wants to tackle vitamins more systematically. In our new paper in Cell , we flip the traditional approach to finding cures. Instead of starting with one disease and spending a decade searching for a treatment, we start with a treatment that's already safe and accessible and ask: what are all the genetic diseases it can treat?
A genome-wide screen for vitamin-treatable diseases
We started with vitamins B2 and B3 because they're common components of cell culture media that can easily be omitted or added. Taking advantage of this, we performed genome-wide CRISPR screens in K562 cancer cells in media with or without one of these B vitamins. We then consulted the OMIM database of known monogenic diseases to nominate candidates from the genes that grew well with the vitamin but poorly without it. This yielded a list of dozens of disease genes that could be rescued by high vitamin B2 or B3 levels.
For B2, the top hit was SLC52A2, a riboflavin transporter whose mutations cause Brown-Vialetto-Van Laere syndrome, a disorder already known to be treatable with high-dose riboflavin. Another top hit, FLAD1, also corresponds to a riboflavin-responsive condition. Two known positive controls sitting right at the top of our list told us the framework works.
But the screen also picked up GPX4, an enzyme that prevents a type of cell death called ferroptosis, which was unexpected. We validated the connection in cell culture and then in a mouse model, where GPX4-deficient mice on a B2-deficient diet showed accelerated motor decline. This represented a brand new vitamin-disease interaction.
NAXD and vitamin B3
From the vitamin B3 screen, the top hit was the NAXD gene. Mutations in it cause a very rare disease, not a lot of labs study NAXD , and it was completely new to us at the time. What we learned is pretty fascinating, and it led us on a years-long, often challenging, investigation.
The NAXD protein is a metabolic proofreading enzyme. Hundreds of enzymes rely on NAD/NADH (carriers of electrons for oxidation-reduction reactions) for everything from basic energy production to DNA repair. But sometimes, whether through mistakes by enzymes like GAPDH, or due to lower pH and increased temperature, NADH gets erroneously hydrated into something called NADHX. If that's not corrected, it can cyclize into a dead-end product called cyclic NADHX. These rogue metabolites can competitively block the enzymes that normally use NAD and NADH.
NAXD's job is to convert NADHX back to usable NADH, correcting the errors before they cause damage. If people lack NAXD function, they develop severe neurodevelopmental disease that is often fatal in early childhood, though the timeline varies depending on the mutation.
We validated the screen hit in cell culture first, finding that NAXD knockout cells grew poorly in B3-deficient media, and this phenotype was rescued by high B3. But cancer cells in a dish aren't representative of what happens in a living body. We needed an animal model, and no mouse model of NAXD disease existed. That had been a major roadblock for the field.
We used CRISPR to generate knockout mice with frameshift mutations in exon 2 of Naxd . The mice were born indistinguishable from their littermates, but within a day or two, their health deteriorated rapidly.
Understanding and reversing NAXD disease
It was a challenge, but we optimized LC-MS methods to detect all the relevant metabolites — NAD, NADH, R-NADHX, S-NADHX, and cyclic NADHX. In the knockout mice, we saw massive accumulation of every form of NADHX across all tissues, while these metabolites were essentially undetectable in wild-type animals.
We were particularly intrigued to find that NAD was specifically depleted in the brain and skin, which is where the disease shows up in human patients. Other organs like liver, heart, and kidney had the error metabolites piling up but were maintaining their NAD levels. Why some tissues are more vulnerable than others is still something we're working to understand.
Broader metabolomics on knockout brains and MALDI-TOF imaging revealed that the most dramatically depleted compounds beyond NAD itself were serine and phosphoserine. Serine depletion was most severe in cortical regions of the brain. But single-nucleus RNA sequencing revealed that the most affected cell types weren't neurons. They were brain endothelial cells, mural cells, and astrocytes – the brain's vasculature. Endothelial cells showed a strong cell death signature that was completely reversed by B3 treatment.
That brought us back to our original question: can the vitamin actually treat this disease?
First we tested what happens when you remove it entirely. We put pregnant mothers on a B3-deficient diet. On regular chow, about a quarter of pups were knockouts, as expected. On the deficient diet, across about 70 pups, not a single knockout was born alive.
Then we tried the opposite. Getting a vitamin to actually increase in the tissues of a tiny newborn mouse pup is harder than you'd think, and we spent over a year on this problem alone. But when we got it right, with daily intraperitoneal injections of nicotinamide riboside (one form of vitamin B3) starting at birth, the rescue was dramatic. Knockout mice were indistinguishable from their littermates at day 50 and beyond. Eight of nine treated animals continued to survive, more than a 40-fold improvement in lifespan. Every aspect of disease we measured was rescued, including body weight, brain pathology, cell death.
When we delayed treatment by just two days, postnatal day 2 instead of day 0, there was no benefit. That's why we advocate including NAXD in neonatal screening panels.
A new framework for vitamin biology
There are clinical case reports where NAXD patients were given supplements including vitamin B3 and some appeared to improve. Those are anecdotal, but they're consistent with what we've found. Translating the right dosing from mice to humans will require careful work, but we're eager to get these findings to the patient and clinical communities as quickly as possible.
We also hope this work inspires others to revisit vitamins with fresh eyes. We'd love some company doing this. Our screens nominated dozens of additional diseases that may respond to B2 or B3 therapy, and we've only tested two out of more than 50 known micronutrients.
More broadly, we think the field should be thinking more creatively about therapies, not just developing drugs, but asking whether there are things we can change about how we live to treat disease. Whether it's what we breathe or what we eat.
Someday, maybe every newborn's genome gets sequenced at birth, and they receive custom dietary recommendations based on their genetics. We see this as picking up where the original vitamin hunters left off, bringing modern tools to the question of which vitamins can treat which diseases, and the mechanisms behind them.
Garg, A., Blume, S.Y., Huynh, H., Barrios, A.M., Karabulut, O.O., Zhao, Q., Midha, A.M., Turner, A., Resnick, B.V., Chen, X., Agrawal, A., Kim, J., Chen, L., Ran, Q., Ryan, A.M., Larson, R.C., Negahban, M., Nelson, S.C.K., Yang, A.C., Traglia, M., Thomas, R., Sun, R., Paredes, M., Corces, R., Lin, H., & Jain, I.H. (2026). Nutritional Genomics Uncovers Vitamin B3 Therapy as Curative for NAXD Disease. Cell . https://doi.org/10.1016/j.cell.2026.01.022
Authors:
Cell
Experimental study
Animals
Nutritional Genomics Uncovers Vitamin B3 Therapy as Curative for NAXD Disease
25-Feb-2026
Funding information and declarations of interest can be found in the manuscript.