The Cancer Target That Fought Back

The mechanism made biological sense. These tumor types were already struggling with DNA damage. Hit them with a drug that blocks DHX9, and you pile even more replicative stress onto cells that can't handle it. In the lab, that meant dead cancer cells. Lots of them.

The FDA was intrigued enough to grant ATX-559 Fast Track designation for MSI-H colorectal cancer patients who'd already failed checkpoint immunotherapy. That's a real stamp of regulatory interest. Accent started dosing patients in its first-in-human trial in late 2024 and presented early study details at the 2025 ASCO Annual Meeting.

Everything was moving in the right direction. Until it wasn't.

When Biology Bites Back

The fundamental problem with targeting DHX9 is the same thing that made it attractive: the protein does everything. When your drug target has its fingers in transcription, translation, RNA processing, DNA replication, and antiviral defense, shutting it down isn't like snipping a single wire. It's more like cutting the power to an entire building.

Scientists had early warnings. Mice engineered to lack DHX9 entirely die before birth, with massive cell death in embryonic tissue. That's about as loud as a biological alarm bell gets. It told researchers that complete DHX9 inhibition is incompatible with normal life, at least during development.

The hope, of course, was that a drug wouldn't need to shut down DHX9 completely. Partial inhibition might selectively cripple cancer cells while leaving healthy tissue mostly intact. It's a bet the entire field of targeted oncology makes constantly. Sometimes it pays off beautifully, and sometimes biology reminds you who's in charge.

Accent hasn't disclosed what adverse events derailed the trial. But the fact that they terminated the program entirely rather than adjusting the dose or tweaking the patient population tells you the safety signal was deeply concerning. Companies don't abandon a first-in-class asset with Fast Track status unless the data is truly ugly.

A Broader Cautionary Tale

Accent's stumble isn't just about one company or one drug. It raises uncomfortable questions about the entire class of RNA helicase inhibitors.

No drug targeting a DDX or DHX family helicase has ever reached late-stage clinical trials. Not one. Researchers have been chasing these enzymes for years (in cancer, in HIV, in hepatitis C). The targets keep looking promising in the lab, but the drugs keep running into walls. The active sites of these proteins are notoriously hard to drug: flat, featureless surfaces that don't give small molecules much to grab onto. And the enzymes are so essential to normal cell function that the window between killing cancer and harming the patient can be razor thin.

ATX-559's failure doesn't prove DHX9 is undruggable forever. But it does suggest that the therapeutic window (that sweet spot between effective and toxic) may be narrower than preclinical models predicted. Lab-dish experiments and mouse studies can model a lot of things, but they can't fully replicate the complexity of a human body running on DHX9 in dozens of overlapping biological pathways.

Where Accent Goes From Here

To the company's credit, Accent isn't shutting down. It's redirecting its resources toward ATX-295, a KIF18A inhibitor currently in Phase 1/2 trials for advanced ovarian cancer and other solid tumors. KIF18A is a motor protein involved in cell division; a different mechanism entirely, and one that also carries Fast Track designation from the FDA.

Accent also maintains partnerships with AstraZeneca on RNA-modifying protein therapeutics. So while losing the DHX9 program stings, it's not an existential crisis.

But the broader question lingers for everyone working on helicase biology: if a well-funded company with strong preclinical data, FDA Fast Track status, and a biologically rational target can't get past Phase 1, what does that tell us about the difficulty of drugging this protein family?

Maybe it tells us we need better tools: allosteric inhibitors that only partially dial down activity, or drugs that target DHX9's interactions with specific partner proteins rather than its core engine. Maybe it tells us we need better biomarkers to pick patients whose tumors are truly dependent on DHX9 while their normal tissue isn't.

Or maybe it tells us something simpler: Some targets look perfect on paper but carry hidden costs that only reveal themselves when you put a drug into a living, breathing human being. The lab can only take you so far. After that, biology has the final word.

A Chemical Company Just Dropped $920M on an Antiviral Biotech. Wait, What?

A Japanese chemicals-and-housing conglomerate just dropped $920 million on a German antiviral biotech. It sounds absurd — until you look at what Asahi Kasei has been quietly building for the last 50 years.