Commonly Used Drug Kills Leukemia Cells Through RNA

HealthK Puspa20 Nov 2026

Nov 20: For more than two decades, azacitidine has been a major treatment for blood cancers such as high-risk myelodysplastic syndromes and acute myeloid leukemia. The drug was thought to work through DNA, by stripping off the chemical tags that keep genes silenced and allowing those genes to switch back on.

Understanding how azacitidine works matters because the drug helps only a third to half of patients, and doctors have no reliable way to predict who will benefit.

In a recent study published in Molecular Cell, researchers at the University of Chicago and UT Health San Antonio report that during the first day of treatment, when the drug kills the most cancer cells, almost none of it happens in DNA, but instead mostly in RNA. There, it strips chemical tags from RNA and shuts down genes that cancer cells need to divide and repair their DNA. The research findings point to NSUN2, the enzyme that installs those tags, as a more precise drug target.

The DNA vs. RNA puzzle

Scientists have long classified azacitidine as a DNA hypomethylating agent because a fraction of the drug ends up in DNA, but that DNA-centric model left major clinical puzzles unsolved. Changes in gene activity after treatment often did not match changes in DNA methylation, a process that adds or removes chemical tags from DNA molecules and changes their activity. Some patients benefited from the drug, even without measurable DNA demethylation.

A sister drug, decitabine, is nearly identical to azacitidine but is built so that only DNA will take it up. If decitabine acts through DNA, then it must be a stronger drug, but in the clinic, azacitidine is far more widely used.

“If DNA were the substrate, then decitabine should work better than azacitidine, but it is the opposite,” said Chuan He, PhD, the study’s senior author and the John T. Wilson Distinguished Service Professor of Chemistry at the University of Chicago. 

How chemical tags control genes

Chemical tags called methyl groups, added to DNA, RNA, and the proteins that package DNA, act as switches that help decide which genes are on or off. In healthy blood cells, this epigenetic regulation keeps growth, DNA repair, and maturation in check.

Blood cancers such as AML hijack this machinery to survive and repair. Earlier work from He’s lab shows that leukemia cells use methyl tags on RNA molecules that sit on the loosely packaged DNA in the cell nucleus to keep growth genes switched on. Because these cancers depend so heavily on distorted gene regulation, disrupting it is a promising way to stop them. The study was also inspired by previous work led by Jason Cheng, MD, PhD, Associate Professor of Pathology at UChicago. 

Catching the drug in the act

To uncover how azacitidine works through RNA, the researchers in the current study focused on the first 8 to 24 hours of treatment. They reasoned that passive DNA demethylation is slow because it often requires cells to divide and azacitidine needs to be modified inside cells before its incorporation into DNA, whereas effects on RNA should appear much faster.

Using high-precision mass spectrometry, the researchers directly measured chemical tags on both DNA and RNA over time. Within just 8 hours, levels of a tag called 5-methylcytosine on chromatin-associated RNA dropped sharply, while DNA methylation remained unchanged for the entire first 24 hours.

To further understand the mechanism, the researchers isolated proteins that had become permanently glued to RNA and DNA after treatment using a specialized biochemical technique. They found that the drug had grabbed directly onto NSUN2, the main enzyme that installs m5C tags on chromatin-associated RNA. This binding peaked at 8 hours, the same moment that RNA methylation dropped and genes began shutting down.

Ruling out DNA

To separate the effects on RNA from those on DNA, the team compared azacitidine with decitabine. Within the first 24 hours, azacitidine rapidly shut down gene activity, disabled DNA repair, and triggered cell death. Decitabine, which acts only through slow changes to DNA, had almost no effect on cell survival in that same window.

“Both are used as hypomethylating agents to treat leukemia, but azacitidine is used in the majority of patients and decitabine in a smaller portion,” He said. “They clearly have different clinical toxicity, yet people somehow think they work the same way, but we know most of the azacitidine goes into RNA.”

To make sure the drug wasn’t acting through an off-target effect, the researchers skipped the drug altogether and used genetic tools to deplete NSUN2 itself. The leukemia cells collapsed in the same way they did under azacitidine: they stopped dividing, lost the ability to repair DNA, and showed the same widespread gene silencing.

By contrast, depleting DNMT1, the main enzyme that adds methyl tags to DNA, produced the opposite effect. Together, these experiments indicate that azacitidine’s rapid toxicity and gene silencing in leukemia cells are driven mainly by its effects on RNA rather than DNA demethylation.

Future directions

The findings may also help explain why azacitidine works so well alongside venetoclax, a drug that pushes cancer cells to self-destruct. Because azacitidine triggers cell death within hours, the researchers found it to be highly synergistic with venetoclax — a combination that is now a standard treatment for many older adults with AML.

The study also offers clues regarding who is most likely to benefit. Leukemia cells with mutations in the genes TET2 or IKZF1 were especially sensitive to the loss of RNA methylation. So, if confirmed in patients, markers like these could help doctors predict who will respond and guide the design of new therapies that target RNA tags.

“Our study suggests that if we can inhibit NSUN2 specifically, we may be able to avoid resistance from inhibiting too many enzymes,” He said. “Our plan is to develop NSUN2 inhibitors that we can test in further studies.”

The study, “5-Azacytosine induces cytotoxicity via 5-methylcytosine depletion on chromatin-associated RNA in leukemia,” was supported by the National Institutes of Health, the Cancer Prevention and Research Institute of Texas, and the Ludwig Center for Metastasis at the University of Chicago.

Additional authors include Boyang Gao from the University of Chicago and Max Planck Institute of Immunobiology and Epigenetics, Germany; Xinyuan Ma, Long Zhao, Zhongyu Zou, Junhong Xiang, and Xiaoyang Dou from the University of Chicago; Ying Li, Juyeong Hong, Feng-Chun Yang and Mingjiang Xu from the University of Texas Health Science Center at San Antonio, Texas.

UChicago Medicine and the Biological Sciences Division continue to be at the forefront of cancer care and research. In April 2027, UChicago Medicine will open the AbbVie Foundation Cancer Pavilion, Chicago’s first freestanding cancer pavilion, to bring advanced diagnostics, innovative treatments, translational discoveries, and comprehensive support to patients and the community.