A bacterial compound originally studied as a natural antibiotic can force cancer cells into a fatal spiral by corrupting their ability to process vitamin B2. That is the central finding of a study published in Nature Cell Biology in 2025, which shows that roseoflavin, a molecule made by the soil microbe Streptomyces davawensis, slips into the vitamin B2 metabolic pathway of tumor cells, generates nonfunctional cofactors, and destabilizes a key survival protein called FSP1. Without FSP1, cancer cells lose their shield against a destructive form of cell death known as ferroptosis, in which iron-driven chain reactions shred cell membranes from the inside out.
The discovery matters because FSP1 has emerged over the past several years as one of the main reasons certain cancers resist therapies designed to trigger ferroptosis. Researchers have been hunting for ways to disable it. Roseoflavin, cheap to produce and already well-characterized in microbiology, now offers a proof-of-concept that this can be done with a small molecule targeting the vitamin supply chain rather than the protein itself.
Why FSP1 matters in cancer
Cells protect their membranes from oxidative damage through two main systems. The first, built around an enzyme called GPX4, uses glutathione to neutralize dangerous lipid peroxides. The second, identified in landmark 2019 studies by teams at the Broad Institute and Helmholtz Munich, centers on FSP1. Working independently of glutathione, FSP1 regenerates a membrane-embedded antioxidant called ubiquinol (reduced coenzyme Q10) by using NAD(P)H as an electron source. Ubiquinol then intercepts lipid radicals before they can propagate.
For cancer cells, this backup system is a lifeline. Tumors that overexpress FSP1 can survive even when GPX4 is knocked out or chemically inhibited, a pattern that has frustrated efforts to weaponize ferroptosis in the clinic. Researchers developing FSP1 inhibitors such as the tool compound iFSP1 have shown that blocking the protein sensitizes resistant cancer cells to ferroptotic death. But designing a drug that directly binds and disables FSP1 with the right selectivity has proven difficult. The new study takes a different approach: instead of attacking FSP1 directly, it cuts off the cofactor FSP1 needs to function.
How roseoflavin poisons the pathway
FSP1 requires flavin adenine dinucleotide (FAD) to maintain its structure and catalytic activity. FAD is the end product of a two-step conversion that begins with dietary vitamin B2, also called riboflavin. First, an enzyme called riboflavin kinase (RFK) converts riboflavin into flavin mononucleotide (FMN). Then a second enzyme, FAD synthetase (encoded by the gene FLAD1), converts FMN into FAD.
The Nature Cell Biology team, whose work is indexed in PubMed, systematically knocked out three genes in this supply chain: FLAD1, RFK, and SLC52A2, which encodes the transporter that brings riboflavin into cells. In each case, FSP1 protein levels dropped, lipid peroxidation climbed, and cancer cells became vulnerable to ferroptosis. The experiments established a direct, causal link between vitamin B2 metabolism and the FSP1 defense system.
Roseoflavin exploits this dependency because it is a structural mimic of riboflavin. Earlier enzymology work had already shown that human RFK converts roseoflavin into an analog called RoFMN, and human FAD synthetase then converts RoFMN into RoFAD. These counterfeit cofactors compete with genuine FAD for binding sites on flavoproteins but cannot perform the same chemistry. When roseoflavin floods the pathway, FSP1 ends up loaded with a cofactor that does not work. The protein destabilizes, ubiquinol regeneration stalls, and lipid peroxides accumulate until the membrane collapses.
The researchers confirmed that roseoflavin-treated cells displayed the classic hallmarks of ferroptosis: elevated lipid peroxidation, iron dependence, and rescue by lipophilic antioxidants such as ferrostatin-1. Genetic and pharmacologic disruption of the same pathway produced matching results, reinforcing the conclusion that roseoflavin kills through FSP1 destabilization rather than some unrelated toxic effect.
What the study does not yet show
The findings so far come from controlled cell-based experiments. That is the right starting point for mapping a mechanism, but it leaves major questions unanswered before anyone should think about roseoflavin as a cancer drug.
The most pressing gap is the absence of animal data. No published results from this study describe tumor xenograft experiments, pharmacokinetic measurements, or toxicity profiling in living organisms. Without that information, basic questions remain open: Can roseoflavin reach tumors at concentrations high enough to disable FSP1? How quickly is it cleared from the bloodstream? What is its half-life?
Selectivity is an equally serious concern. Every human cell uses riboflavin, and roseoflavin enters the same metabolic machinery in healthy tissue as it does in tumors. Rapidly dividing cells in bone marrow and the gut lining, which already face significant oxidative stress, could be especially sensitive. The therapeutic window, meaning the gap between a dose that kills cancer cells and one that harms normal tissue, has not been defined.
Specific IC50 values for roseoflavin against different cancer cell lines have not been reported in the available data. Nor has the study identified which tumor types are most vulnerable. Cancers with high FSP1 expression and low GPX4 activity would be the logical candidates, but that prediction awaits experimental validation. Some aggressive cancers, including certain melanomas and therapy-resistant lung adenocarcinomas, are known to express elevated FSP1, making them plausible targets for future testing.
Resistance is another open question. Cancer cells under selective pressure could potentially evolve mutations in RFK or FLAD1 that distinguish riboflavin from roseoflavin, alter SLC52A2 transporter activity to exclude the compound, or shift toward alternative antioxidant defenses. None of these escape routes have been formally tested, but they will need to be addressed in any long-term development program.
Where this fits in ferroptosis research
Ferroptosis has attracted intense interest in oncology over the past decade because many treatment-resistant cancers appear to depend on specific anti-ferroptotic defenses. The GPX4 pathway was the first major target, and several experimental inhibitors (including RSL3 and ML210) have been used in preclinical studies. But tumors that co-express FSP1 can shrug off GPX4 inhibition, creating a need for combination strategies or alternative attack points.
The roseoflavin findings suggest that vitamin B2 metabolism could serve as that alternative node. Rather than trying to design a molecule that fits into FSP1’s active site, researchers could target the upstream supply chain that FSP1 depends on. This approach has a precedent in cancer pharmacology: antifolates like methotrexate work by disrupting vitamin B9 (folate) metabolism, and they remain among the most widely used chemotherapy agents decades after their introduction.
Roseoflavin itself may or may not be the compound that advances to clinical testing. Its value right now is as a probe molecule that reveals a druggable vulnerability. Medicinal chemists could use its structure as a starting point to design analogs with better tumor selectivity, tighter binding to FAD-processing enzymes, or more predictable pharmacokinetics. The fact that roseoflavin is a natural product with a known biosynthetic pathway also simplifies early-stage production and characterization.
For researchers working on ferroptosis-based cancer therapies, the practical implication is that FSP1’s reliance on FAD creates a targetable bottleneck. Combining roseoflavin or a derivative with GPX4 inhibitors could, in theory, shut down both major anti-ferroptotic defenses simultaneously. That dual-strike strategy has not been tested in animals or patients, but the mechanistic logic is sound, and the genetic knockout data from this study support the rationale.
From mechanism to medicine: what comes next
Translating a cell-culture discovery into a viable therapy requires years of work, and most candidates fail along the way. The immediate next steps for roseoflavin would include dose-response studies in animal tumor models, pharmacokinetic profiling to determine how the compound distributes through the body, and toxicity assessments in normal tissues. Researchers will also need to identify biomarkers, likely FSP1 expression levels and riboflavin pathway activity, that could predict which patients’ tumors are most susceptible.
If roseoflavin proves too toxic or too nonselective in animals, the pathway itself remains a valid target. Screening campaigns could search for molecules that inhibit RFK or FLAD1 with greater specificity, or that are preferentially taken up by tumor cells through differences in SLC52A2 expression. The genetic data from this study provide a clear set of targets for such efforts.
What the Nature Cell Biology paper has accomplished, as of its publication, is a rigorous demonstration that FSP1-dependent cancers carry a metabolic vulnerability that can be exploited with a simple, well-understood compound. That is not a cure. It is a map showing where to dig. The digging, through animal studies, safety testing, and eventually human trials, still lies ahead.
More from Morning Overview
*This article was researched with the help of AI, with human editors creating the final content.
