Cancer Cells Tricked into Manufacturing the Drugs That Kill Them

One obstacle stood in the way. Inside living cells, a substance called glutathione rapidly deactivates the palladium at the heart of the catalyst, shutting down the reaction. To protect it, the researchers enclosed the palladacycle within albumin, a common blood protein. Like a courier shielding a fragile parcel, the albumin carrier lets the complex travel through the bloodstream and keeps the catalyst active after it enters a cell.

The logic of selectivity is straightforward. Cancer cells hold excess carbon monoxide; healthy cells do not. The drug-producing reaction should therefore proceed in tumors and stall elsewhere, because only cancer cells supply enough of the gas to drive it.

What Did the Laboratory Tests Show?

The researchers worked through a sequence of experiments:

• They screened different palladacycles to find the one generating the greatest amount of phenanthridinone.
• They exposed cultured human liver, cervical, breast, and lung cancer cells to the albumin-palladacycle complex for 48 hours, alongside a noncancerous cell line.
• After 48 hours, they detected phenanthridinone in the cancer cells but not in the normal cells.

The team then refined the system so the newly formed compound could detach from its carrier and move freely within the cell, and further modifications enabled the reaction to produce a particularly toxic phenanthridinone derivative already used in anticancer medicines. When introduced into cultured liver cancer cells, the modified system generated the drug inside the cells, which subsequently showed signs of damage from the toxic compound.

Why Did the Drug Not Form in Blood?

A safety concern remained. Blood itself carries carbon monoxide, and if the gas triggered the reaction prematurely, the resulting drug could damage blood cells before the compound ever reached a tumor. Laboratory testing showed the toxic drug was not produced in blood. The researchers attribute this to hemoglobin, the oxygen-transporting protein: in the bloodstream, carbon monoxide is largely bound to hemoglobin and appears unavailable to fuel the reaction.

The finding suggests the system can travel through blood without generating harmful substances along the way, although the authors caution that further studies must establish its safety in living organisms. All results so far come from cultured cells, not animals or patients.

"This work represents an important advance in therapeutic in vivo synthetic chemistry," Tanaka said. "In our therapeutic strategy, molecules naturally present or generated within the body — including disease-associated metabolites — are used as endogenous chemical feedstocks to synthesize therapeutic molecules directly at the site of disease."

The team now plans to test the approach in animal models. If those experiments succeed, the technology could seed a new class of treatments that exploit tumor chemistry itself — using the disease's own exhaust to build the drug that destroys it, while healthy tissue stays untouched.

The study, by Masayuki Kawai, Tsung-Che Chang, Ambara R. Pradipta, and Katsunori Tanaka, appears in Nature Communications, volume 17 (DOI: 10.1038/s41467-026-77616-8).