The Shapeless Cancer Target That May Not Be Undruggable After All

For years, one of cancer biology's most important switches looked like a terrible drug target because it would not hold still. Now a team at the University of British Columbia and BC Cancer says it has built small molecules that can grab that shapeless stretch anyway, a result that matters beyond prostate cancer because much of biology runs on the same kind of floppy, hard-to-target machinery.

In a peer-reviewed paper in Signal Transduction and Targeted Therapy, the researchers report compounds that bind the androgen receptor's transactivation domain, the section that helps switch growth programs on. Several of those compounds reached picomolar to low-nanomolar binding strength, according to the paper, and in mouse xenograft models they outperformed enzalutamide, sold as Xtandi specifically when androgens were present.

This is still a preclinical paper, not a near-term drug launch. Prostate cancer is littered with animal results that did not survive contact with humans. The more interesting claim is narrower and bigger at the same time: a protein region long treated as structurally too messy for conventional drug design may be less untouchable than the field assumed.

The target here is the androgen receptor, the molecular receiver that helps prostate cancer cells respond to male hormones. Existing drugs such as enzalutamide mostly target the receptor's ligand-binding domain, the pocket where hormones dock. That strategy has limits. In a 2023 paper in Nature Structural & Molecular Biology, researchers noted that about 20 percent of people with prostate cancer progress to castration-resistant disease, a lethal form often linked to androgen receptor splice variants that can keep signaling even after the usual docking-pocket drugs stop working.

The new paper goes after a different piece of the receptor: the transactivation domain, a roughly 556-residue region that is predominantly disordered, according to the study. In plain English, this is the spaghetti section. It does important work, but it does not sit in one tidy three-dimensional shape that chemists can target the way they target a classic enzyme pocket.

That problem is not unique to prostate cancer. A 2022 Nature Communications paper cited by the authors says intrinsically disordered proteins make up about 40 percent of the human proteome. Drug discovery has historically preferred proteins with stable structures because they offer cleaner docking sites and simpler screening logic. If disordered regions become tractable, the opportunity is not one new therapy. It is a much larger menu of targets that companies previously crossed off.

The UBC and BC Cancer team did not find one miracle compound on the first try. The paper says the group analyzed more than 560 androgen receptor transactivation-domain inhibitors to optimize potency. It also builds on years of work around this same target class. A 2022 Nature Communications study found that the compound EPI-7170 had 2.5-fold higher affinity for a key region of the transactivation domain than EPI-002, and a 2024 molecular dynamics follow-up hosted on PubMed Central concluded that binding did not force the domain into one rigid folded structure. The molecules appeared to work without taming the disorder into something neat.

That is what makes this paper interesting. The authors report that mass spectrometry identified covalent binding to cysteine 129 in the transactivation domain, giving the field a more concrete physical foothold than vague claims that a molecule interacts somewhere in the mess. They also report that enzalutamide, as expected, had no impact on AR-V7 transcriptional activity, while their transactivation-domain inhibitors retained activity against models built around that resistance mechanism.

There is a backstory here that keeps the result honest. The first small molecule that directly bound an intrinsically disordered region and reached clinical trials was ralaniten acetate, according to the new paper. That Phase I effort ended because of pill burden, the authors write, not because the underlying idea of targeting the disordered domain was disproved. This new generation of compounds is better read as a second attempt to make the concept pharmacologically viable, not as a bolt from nowhere.

The caution flag is obvious. Potent binding, cleaner mechanistic evidence, and stronger xenograft performance do not guarantee a usable drug. What animal models often miss is the long list of things that make medicines fail later: exposure, toxicity, formulation, dosing burden, durability, and the ability to reproduce the effect outside one experimental setup. A field can go from undruggable to druggable in mice and still spend years discovering that the chemistry is not yet clinic-ready.

Still, the paper opens a door the broader industry has wanted open for a long time. If one of the best-known disordered cancer targets can be hit with this level of affinity and in vivo activity, companies working on neurodegeneration, autoimmune disease, and other disorder-heavy areas have reason to revisit targets they once dismissed as too structurally slippery to matter.

That is the real payoff here. Prostate cancer is the proving ground. The larger story is that a whole category of proteins long treated as bad drug real estate may be entering the part of science where impossible has to compete with data.