The compounds in question were originally developed as inhibitors of KDM4, a family of proteins that can drive the growth and spread of cancer cells when they become overactive. KDM4 enzymes belong to a broader class of epigenetic regulators that modify histones, the protein scaffolds around which DNA is wound, and their overactivity has been linked to several aggressive tumour types. Zavondemstat, built on the same chemical principles as the research tool QC6352, progressed into clinical testing on the assumption that blocking KDM4 was the key to its therapeutic effect. The new study upends that assumption, showing that both compounds instead predominantly inhibit DHODH, an enzyme that cancer cells depend on to manufacture the molecular building blocks required for rapid proliferation.

Lead author Professor Lenka Munoz of the University of Sydney School of Medical Sciences and the Charles Perkins Centre offered a vivid analogy for what DHODH inhibition means inside a tumour cell. She described DHODH as a machine producing the bricks needed to build new DNA. If the machine is switched off, the cell starts running out of bricks and can no longer efficiently copy its DNA or keep dividing. In other words, the compounds are not adjusting the epigenetic programming of cancer cells through KDM4; they are starving them of the pyrimidine nucleotides essential for DNA synthesis, a fundamentally different form of metabolic stress.

The discovery emerged, somewhat unexpectedly, from research with a different goal altogether. The team set out to investigate whether zavondemstat and QC6352 could be repurposed for glioblastoma, the most common and most aggressive form of brain cancer, a disease for which effective treatments remain desperately limited. Munoz explained that when the researchers tested other KDM4 inhibitors, those compounds failed to reproduce the anti-cancer effects observed with QC6352. If blocking KDM4 were truly driving the observed activity, other inhibitors of the same target would have been expected to behave similarly. Instead, the results pointed to QC6352 acting through an entirely different mechanism, prompting the team to dig deeper.

Using patient-derived glioblastoma stem cells, tumour models and a series of genetic, mechanistic and molecular experiments, the researchers systematically traced the compounds’ anti-cancer activity to DHODH rather than KDM4. Patient-derived stem cells are considered a particularly demanding model system because they preserve many of the characteristics of the original tumour, including the cellular subpopulations thought to drive recurrence and resistance. The convergence of genetic and pharmacological evidence across these models gave the team confidence that DHODH, not KDM4, is the key target of QC6352 in glioblastoma stem cells, as the paper’s title states.

The implications extend well beyond glioblastoma. Researchers around the world have used QC6352 as a leading tool compound to study KDM4 biology, and Munoz noted that the related drug zavondemstat progressed into clinical trials based on the same understanding. Her team’s study found that much of the anti-cancer activity of both compounds is driven by blocking DHODH rather than KDM4. That means a substantial body of published research may have drawn conclusions about KDM4 function from experiments in which the compound was, in fact, hitting a different target entirely. Scientific conclusions built on such misattributed mechanisms can propagate through the literature, shaping hypotheses, grant funding and drug discovery programmes in ways that are difficult to unwind.

For the clinical programme, the stakes are equally high. Munoz emphasised that getting a drug’s mechanism wrong can lead to poorly designed clinical trials, inappropriate patient selection and years of research focused on the wrong biological target. Clinical trials for targeted cancer therapies typically enrol patients whose tumours show evidence of the presumed target, on the logic that those patients are most likely to benefit. If the true target is different, the biomarker strategy may select patients who are unlikely to respond, while patients who could benefit from DHODH inhibition are excluded. There are well-known historical examples of cancer drugs advancing through large clinical trials before researchers realised they were not working through the mechanism originally proposed, and Munoz stressed that the new study shows this is not merely a historical problem but one still happening today.

Beyond the immediate implications for zavondemstat and QC6352, the findings open a potentially promising new avenue for glioblastoma treatment. DHODH inhibition has attracted growing interest in oncology, and several drugs targeting the enzyme are already being investigated for other cancers. The possibility that such agents could eventually be tested in patients with brain cancer adds a therapeutic silver lining to what is otherwise a cautionary tale. Glioblastoma remains one of the deadliest human cancers, and any mechanism-based strategy that can be pursued with compounds already in clinical development offers a faster route to the clinic than starting from scratch.

The researchers also took a constructive step to repair the damaged research toolchain. Having established that QC6352 confounds KDM4 biology through off-target DHODH inhibition, they developed new compounds that inhibit KDM4 without affecting DHODH. These improved chemical probes should allow scientists to study the genuine role of KDM4 enzymes in cancer with far greater accuracy, replacing a compromised tool with ones whose selectivity has been rigorously characterised. The episode underscores a principle increasingly emphasised in chemical biology: the value of a research compound depends on knowing precisely what it does, and validation of target engagement should be a prerequisite for both laboratory use and clinical translation.

Munoz argued that greater rigour is needed to ensure researchers understand exactly how potential treatments work before they move into clinical testing. Establishing a drug’s true mechanism early, she said, can protect patients, prevent wasted research effort and help ensure that limited funding is directed towards genuinely promising treatments. The study, an experimental investigation conducted in cells and tumour models, was funded by the National Health and Medical Research Council, Tour de Cure and the Tim Trigg Scholarship into Glioblastoma Research, and the authors declared no competing interests. As zavondemstat continues its journey through clinical development, the work stands as a striking reminder that in modern drug discovery, the question of how a drug works can be as consequential as whether it works at all.

Subject of Research: Mechanism of action of the KDM4 inhibitor QC6352 and the clinical drug zavondemstat in glioblastoma stem cells

Article Title: Cancer trial drug may work differently than scientists believed, raising concerns for clinical research

Article References: Cancer trial drug may work differently than scientists believed, raising concerns for clinical research. (n.d.). Original publication

Image Credits: AI Generated

DOI: Not provided

Keywords: zavondemstat, QC6352, KDM4, DHODH, glioblastoma, cancer drug development, drug mechanism, clinical trials, chemical biology, University of Sydney, Nature Chemical Biology, targeted therapy