Targeted therapies are changing cancer treatment. Designed to wield laser-like precision, they focus on overactive proteins produced by mutated genes, destroying cancer cells but sparing normal cells. Most of the cancer drugs created in the last decade follow that model, exploiting knowledge of genetic mutations to stop cancer’s progression rather than aiming a fusillade of chemotherapy agents against all fast-growing cells, normal or cancerous.
But many mutations do not make good targets. Tumor suppressor genes – which normally act like the brakes on runaway cell growth –– are the most commonly mutated genes in cancer, but they are notoriously difficult to target with current techniques. New targets are needed, driving cancer researchers to seek out new therapeutic approaches.
Looking back to a hypothesis first proposed soon after tumor suppressor genes were discovered, a team of scientists from the Broad Institute and Dana-Farber Cancer Institute present a new paradigm for attacking vulnerabilities in cancer cells by targeting genes that don’t necessarily contribute to cancer formation but are altered along the way to becoming cancer.
Nearly 20 years ago Dana-Farber researcher Emil Frei proposed that when tumor suppressor genes were mutated, they weren’t the only parts of the genome being damaged. When mutations disarmed tumor suppressors, they also cut a wider swath of genomic instability. Nearby genes might also be injured, although to a lesser extent. A tumor cell might lose one of its two copies of a nearby gene, for example. That lone copy might be sufficient for cell survival, but additional stress could be more than the cell could tolerate.
In 1993, there was no large-scale way to test this hypothesis. Now scientists led by Broad associate members Rameen Beroukhim and William C. Hahn have harnessed computational strength and biological expertise to answer the question. Writing in the August 15 online edition of Cell, they report proof of principle for a novel strategy that attacks cancer cells by targeting genes partially lost on the path to cancer.
“Cancers delete parts of their genomes when they get rid of tumor suppressor genes, but that process of deletion also makes them susceptible to stresses that otherwise normal cells could tolerate,” said co-senior author Hahn, who is also a medical oncologist at Dana-Farber and an associate professor of medicine at Harvard Medical School. “This discovery opens up a whole new universe of potential targets that seem to distinguish a large fraction of cancer cells from normal cells,” said Beroukhim, co-senior author, a medical oncologist at Dana-Farber, and an assistant professor of medicine at Harvard Medical School.
To systematically identify these genes, the researchers, including co-first authors Deepak Nijhawan and Travis I. Zack, integrated data sets from Project Achilles, which is a comprehensive catalog of genes essential for cancer cell survival, and the Cancer Cell Line Encyclopedia, which includes copy number data. Their analysis yielded 56 genes that had lost one of their two copies. They call them CYCLOPS genes, short for Copy-number alterations Yielding Cancer Liabilities Owing to Partial loSs. Like the mythical one-eyed giant, they depend on that one copy to survive.
The CYCLOPS genes were not randomly assorted throughout the genome. Instead, they largely fell into three complexes important for cell proliferation and survival: the proteosome, the spliceosome, and the ribosome. The scientists narrowed their focus to PSMC2, a gene in the proteasome that has lost one of its gene copies in 10 percent of all cancers. In all, four proteasome CYCLOPS genes were found in a combined 40 percent of cancers.
In normal cells, the proteasome works like a sophisticated garbage disposal to chew up and dispose of proteins that are misfolded or no longer needed. The protein encoded by PSMC2 helps form the cylindrical proteasome’s two caps on its top and bottom. When PSMC2 protein levels are suppressed in a normal cell, the cell calls on a reservoir of PSMC2 protein in the cell. The CYCLOPS version of PSMC2 is unable to supply enough to fill that reservoir in the cell, so when it runs dry, the proteasome stops working and the cell dies.
Using nanoparticle technology to deliver gene-silencing RNA interference, the scientists specifically targeted tumors in mice. When they knocked out PSMC2, tumors were reduced by more than 75 percent and overall survival doubled.
Beyond its cellular importance, the proteasome is also an appealing target because the cancer drug bortezomib, sold as Velcade, is a proteasome inhibitor with tolerable side effects. Its mechanism is not well understood, Hahn said.
“Our paper doesn’t tell us why bortezomib does or doesn’t work in certain cancers, but it does tell us there might be more sophisticated ways to attack the proteasome,” he said. “That might help us understand which patients might benefit.”
Showing that cancer cells are dependent on one CYCLOPS gene in the proteasome opens up new avenues for future work, starting with the proteasome, spliceosome (a splicer of messenger RNA), or ribosome (site of protein synthesis) and possibly beyond, Beroukhim said.
“This opens up a potentially large variety of vulnerabilities to attack in cancer cells,” he said. “There is still work to be done, but if this phenomenon can be taken advantage of therapeutically, it would have a potentially widespread application.”
Other researchers who contributed to this work include Yin Ren, Matthew R. Strickland, Rebecca Lamothe, Steven E. Schumacher, Aviad Tsherniak, Henrike C. Besche, Joseph Rosenbluh, Shyemaa Shehata, Glenn S. Cowley, Barbara A. Weir, Alfred L. Goldberg, Jill P. Mesirov, David E. Root, and Sangeeta Bhatia.
Cancer vulnerabilities unveiled by genomic loss. Nijhawan D et al. Cell, 2012. DOI:10.1016/j.cell.2012.07.023