Drugs like cisplatin that break DNA are some of the strongest weapons we have against breast, ovarian and other cancers. The problem, common to every form of chemotherapy, is that cisplatin doesn’t work for everyone. Given the potential side effects that go along with the drug—including vomiting, hearing loss and muscle cramps, just to name a few—the decision to give it to a patient becomes something of a gamble: Does the benefit outweigh the risk?
There are tests that examine individual genes and which can give doctors a limited view as to which tumors might respond best to cisplatin. But a multicenter team co-led by Zoltan Szallasi, MD, of Boston Children’s Hospital’s Informatics Program (CHIP), thinks they may have a solution that looks beyond individual genes to see which tumors might succumb to cisplatin and other drugs like it.
Currently, doctors can look for mutations in two genes, BRCA1 and BRCA2, for hints about a breast or ovarian tumor’s sensitivity to cisplatin. These mutations break some of the molecular machinery cells use to repair DNA damage. Cells with these mutations respond well to cisplatin, which exploits and exaggerates DNA damage to kill cancer cells.
But what about tumors that do not harbor BRCA1 or 2 mutations? Might some still be vulnerable to cisplatin? To answer this question, Szallasi and his collaborators decided to take a more all-inclusive approach: Instead of looking for individual gene markers, they would use the whole genome as a marker.
They turned to a measure of genetic damage called allelic imbalance, or AI. All of our healthy cells contain two alleles, or copies, of every gene, one from Mom and one from Dad. In tumor cells, though, the balance of Mom alleles and Dad alleles is frequently out of whack: in any given chromosome there might be extra Mom copies of some genes, or extra Dad copies, or no copies from one parent at all.
The magnitude of that imbalance reflects how much damage a cell’s DNA has already suffered. This, in the team’s thinking, could reveal whether a tumor cell’s DNA repair machinery is working and the likelihood that the cell would die when attacked with cisplatin.
“Because cells employ several genes in addition to BRCA1 and BRCA2 in DNA repair, any method that looks at only one or two genes is bound to fail to predict the response to treatment in a significant portion of cases,” Szallasi explains. “If we instead measure the effect of those mutations on DNA repair as whole, then we may be able to attain a fuller, more accurate prediction of response.”
To see whether their logic played out, the team measured the levels of allelic imbalance in tumor cells from women with “triple-negative” breast cancer (a hard-to-treat breast cancer that doesn’t respond to hormone-based or targeted treatments) who had been treated with cisplatin. Most of these tumors lacked BRCA1 or 2 mutations, ruling out these genes as drug sensitivity markers.
The team also looked at genomic and clinical data from ovarian tumors collected through The Cancer Genome Atlas, a federal effort to catalog genetic features associated with several cancer types.
Analysis of all these data revealed that tumor cells from women who benefited from cisplatin treatment indeed had higher levels of allelic imbalance, suggesting that their tumors had the biggest problems with DNA repair—making them more vulnerable to the drug.
What was most striking was the fact that the tumors that reacted most strongly to cisplatin—meaning the women gained the most benefit from the drug—contained cells with imbalances in their DNA that stretched all the way to the tips of their chromosomes. These tips, called the telomeres, have long been recognized for their roles in helping hold chromosomes together and mitigating DNA damage.
The team reported their findings in the April issue of Cancer Discovery, and will present them live today at the annual meeting of the American Association of Cancer Research.
“With these results, we now have a scientific basis for trying to use breakdowns in DNA repair to fine-tune therapy with drugs like cisplatin that cause DNA damage,” says Szallasi. “We need to translate this method from an experimental one to something that could be used in routine clinical practice.”