Skip to main content
Science Catalogs
Health Sciences

Azacitidine's early anti-leukaemia action centres on RNA methylation, not DNA, study finds

Azacitidine, long used against high-risk myelodysplastic syndromes (MDS) and acute myeloid leukaemia (AML), appears to kill leukaemia cells in the first day of treatment mainly by acting on RNA rather than DNA. A study in Molecular Cell by researchers at the University of Chicago and UT Health San Antonio identifies the RNA-methylating enzyme NSUN2 as the key target. The work challenges the standard classification of the drug as a DNA hypomethylating agent.

A puzzle in the DNA model

Azacitidine has been thought to work by removing methyl tags from DNA, allowing silenced genes to switch back on. Yet the drug benefits only a third to half of patients, and there is no reliable way to predict who will respond. Gene activity after treatment often fails to track DNA methylation changes, and some patients improve without measurable DNA demethylation.

Decitabine, a close chemical relative, is built so that only DNA incorporates it. If DNA were the main route, decitabine should be the stronger drug, but azacitidine is far more widely used. "If DNA were the substrate, then decitabine should work better than azacitidine, but it is the opposite," said senior author Chuan He, professor of chemistry at the University of Chicago.

Looking at the first hours

The team focused on the first eight to 24 hours of exposure. Changes to DNA methylation are slow, partly because they depend on cell division and on the drug being processed inside cells before incorporation into DNA. Effects on RNA were expected to appear sooner.

Using high-precision mass spectrometry, the researchers measured methyl tags on DNA and RNA over time. Within eight hours, levels of 5-methylcytosine (m5C) on chromatin-associated RNA fell sharply, while DNA methylation did not change throughout the first 24 hours. A biochemical method that captures proteins covalently trapped on nucleic acids showed that the drug bound directly to NSUN2, the main enzyme that installs m5C on this class of RNA. Binding peaked at eight hours, coinciding with the fall in RNA methylation and the onset of gene shutdown.

Comparing drugs and depleting the enzyme

In the same 24-hour window, azacitidine rapidly suppressed gene activity, impaired DNA repair and triggered cell death. Decitabine, which acts only through gradual DNA changes, had almost no effect on cell survival.

To rule out off-target effects, the researchers depleted NSUN2 genetically, without any drug. Leukaemia cells stopped dividing, lost DNA repair capacity and showed the same broad gene silencing seen with azacitidine. Depleting DNMT1, the main DNA methyltransferase, had the opposite effect. Together, the results indicate that the rapid toxicity of azacitidine in these cells is driven chiefly by RNA effects.

Implications for combinations and patient selection

The rapid induction of cell death may help explain why azacitidine combines so effectively with venetoclax, which pushes cancer cells towards self-destruction. The pairing is now a standard treatment for many older adults with AML. The researchers found strong synergy between the two drugs.

Leukaemia cells carrying mutations in TET2 or IKZF1 were especially sensitive to loss of RNA methylation. If confirmed in patient samples, such markers could help predict response, although this has not yet been shown clinically.

Limits and next steps

The experiments described were carried out in leukaemia cells, and the source does not report patient outcome data, so clinical relevance remains to be tested. The authors suggest that targeting NSUN2 specifically might avoid the resistance that could arise from inhibiting many enzymes at once, and He said the group plans to develop NSUN2 inhibitors for further study. The work was inspired in part by earlier research led by Jason Cheng, associate professor of pathology at the University of Chicago, and by previous findings from He's laboratory on RNA methylation in leukaemia.

For researchers, the study argues for re-examining other nucleoside analogues through an RNA lens, and for treating RNA modifications as a therapeutic axis in blood cancers. The paper is by Boyang Gao and colleagues (DOI: 10.1016/j.molcel.2026.08.018).

Source: Commonly used drug kills leukemia cells early by acting mainly on RNA, not DNA (Medical Xpress, 21 November 2026).

Image: AI-generated illustration.

Working on this topic? Find a journal for your research or join the discussion on Scholars Post.