LONDON — Switching off a single faulty gene can halt the limitless division of cancer cells, according to a study published today that opens up a promising way to combat the disease.

The possibility that natural mechanisms for wiping out cancerous cells could be used to kill off liver, blood, or bone tumors comes from a team at Stanford University School of Medicine.

Cancer cells proliferate uncontrollably because of DNA damage and mutations, and they circumvent the mechanisms that cause normal cells to die when they get old or malfunction. By using drugs to turn off a single gene that can play a central role, this process can be halted.

The gene studied is responsible for making a protein called Myc (pronounced “mick”), which enables cells to divide.

A mutation causes cells to overproduce the protein, prompting perpetual cell division — the hallmark of cancer. Faulty versions of Myc are present in 70% of cancers. By turning off the mutated gene, researchers found that not only did uncontrolled cell division cease, but the cells also reactivated a normal physiological mechanism, called senescence. This allowed the cancer cells to die.

“By shutting off a critical cancer gene, tumor cells can realize that they are broken and restore this physiological fail-safe program,” said Dr. Dean Felsher of Stanford, senior author of the study published in the Proceedings of the National Academy of Sciences. “What was unexpected was just the fact that cancer cells had retained the ability to undergo senescence at all.”

The researchers made their discovery using mice that were genetically engineered to have Myc-triggered cancers of either the liver, blood, or bones. The mice’s Myc gene was designed so it could be switched off with antibiotics.

When the mice ate the drugs, the tumors stopped growing and then diminished. Some disappeared in a few days.

The work was described as exciting yesterday by Alan Clarke of Cardiff University. However, he cautioned that although faulty Myc is common in cancer, far fewer tumors are totally dependent — “addicted to Myc” — which could limit the effects of a treatment based on this find.

He added that the study was highly idealized, with scientists engineering mice to make abnormal amounts of Myc protein to develop tumors, then turning off excess Myc to treat the cancer.

The authors had shown that the tumor regression depended on the presence of other genes and these other genes were frequently mutated.

Therefore, “in these contexts, tumor regression by reducing Myc would not necessarily work,” Mr. Clarke said.