A cancer-causing gene called MYC has proven difficult to target with drugs, so oncology researchers have been looking for ways to combat it indirectly. Researchers at the University of Pennsylvania have found a new method for attacking MYC by targeting a protein that they believe is the oncogene’s Achilles’ heel.

Blocking the protein, called ATF4, causes cancer cells to make too much of another type of protein family called 4E-BP. That causes the cancer cells so much stress that they die. The researchers believe that the discovery, published in the journal Nature Cell Biology, could spark new ideas for cancer therapeutics, because experimental compounds that block the synthesis of ATF4 have already been developed.

In normal cells, MYC controls growth. But mutations can cause the gene to malfunction, producing the uncontrollable cell growth that drives cancer. The lead author of the new study, Constantinos Koumenis, Ph.D., Penn professor of radiation oncology, and colleagues previously found that an enzyme called PERK activates ATF4.

But in this study, the team found that blocking PERK didn’t stop tumor growth, because MYC could simply embark upon a separate process to escape and resume its cancer-promoting activities. So they decided to go “further downstream to block tumor growth in a way that cancer cells can’t easily escape,” said Koumenis in a statement.

Using mouse models of lymphoma and colorectal cancer, the Penn researchers removed ATF4 from cells. The tumor cells responded by overproducing 4E-BP proteins. That caused tumors to stop growing in the animal models.

In further lab tests using human tumor samples, the researchers found that when cancer is driven by MYC, both ATF4 and 4E-BP are overexpressed—a signal, they believe, that blocking ATF4 could work in people.

The study comes just a couple of months after a team from Cold Spring Harbor Laboratory proposed a different way to target MYC-driven cancers. They found that deleting the protein PHLPP2 in mouse-derived cancer cells reduces levels of MYC. When they knocked out the PHLPP2 gene in mouse models of prostate cancer, they were able to slow the spread of the cancer.

Despite the challenges inherent in directly targeting MYC, some biotech startups are still pursuing the strategy. Aptos Biosciences, for example, is conducting early-stage human trials of its MYC inhibitor, APTO-253. The program hit a major snag in 2015, when the FDA slapped a clinical hold on it due to manufacturing issues. That hold was lifted a year ago.

Koumenis’ team at Penn favors the indirect approach to attacking MYC that they described in the new study, and the researchers are planning to continue to investigate how ATF4 works in MYC-dependent tumors. They are also planning studies designed to determine whether blocking ATF4 could cause off-target side effects in people.

This article originally appeared in FierceBiotech and was written by Arlene Weintraub