Researchers have figured out why a popular therapy for treating colon cancer often eventually stops working: The tumors naturally carry genetic changes that allow some of their cells to evade the drug and keep growing. The good news is that physicians may be able to detect these changes in patients' blood and then halt the cancer with a different drug before it can grow again to a dangerous size.
Many people with advanced colon cancer are treated with proteins called antibodies that target a tumor weak spot, a growth-stimulating protein on the cancer cells known as EGFR. The antibodies often shrink tumors, but the cancer typically comes back within 18 months. To study how this resistance developed, a team led by cancer geneticist Alberto Bardelli of the Institute for Cancer Research and Treatment in Candiolo, Italy, grew colon cancer cells in dishes containing a growth medium laced with an anti-EGFR antibody called cetuximab. The group now reports in Nature that the cells became resistant to the drug because of changes in KRAS, a gene coding for a protein that can reactivate the EGFR growth pathway when EGFR itself is blocked. The antibody-resistant cells showed mutations in KRAS and sometimes even contained extra copies of the gene. The researchers also found these changes in biopsies from six patients with tumors that were resistant to cetuximab or a similar drug called panitumumab.
The blood of a cancer patient typically contains traces of tumor DNA. The Italian team and a separate group led by oncologist Luis Diaz of Johns Hopkins University in Baltimore, Maryland, have now independently found that the KRAS mutations are detectable in blood samples from patients whose tumors became drug resistant, a technique Bardelli calls a "liquid biopsy." The telltale signs in KRAS appeared up to 10 months before tumors became large enough to detect with various standard imaging techniques such as x-rays. That suggests that clinicians could monitor patients' blood for resistance mutations and when they appear, give the patients a second drug that also blocks the EGFR growth pathway, but in a way that the KRAS mutation can't override. Combining one such drug, called a MEK inhibitor, with cetuximab killed colon cancer cells in a lab dish that were resistant to cetuximab alone, Bardelli's group reports.
"The concept of liquid biopsy, which both groups used, is truly a step forward, and I am sure it will be widely used in the clinic," says Bardelli, whose work appears online  today in Nature along with the Hopkins group's study .
Another question was how the resistance mutations got there. Were they present in a few rare cells before the patient was treated with the drug, or did they arise after treatment? To find out, the Hopkins team and collaborators at Harvard University developed a mathematical model of the tumor's genetic evolution that fit their data on KRAS mutations in blood. They concluded that resistance mutations resulting from spontaneous changes in DNA as cells divide are present in a few tumor cells even before treatment begins. The model shows that resistance is "a fait accompli," Diaz and his colleagues write. It is also probably inevitable for other so-called targeted therapies, such as those for melanoma and lung cancer, which also usually work for less than a year, Diaz says.
"The lessons here are that resistance mutations occur spontaneously" within a tumor, even before a therapy is tried, says Diaz. In rare cases, a patient on a targeted therapy lives for years without developing resistance. But such tumors might be very young when the patient is first treated or unusually slow-growing, Diaz says.
"These results are illuminating and sobering" because they show that resistance to targeted colon cancer drugs is inevitable, says oncologist Neal Meropol of Case Western Reserve University in Cleveland, Ohio. Treatments will have to target more than one molecular pathway, he says.
But evolutionary biologist Dominik Wodarz of the University of California, Irvine, also finds the Hopkins study "encouraging" because the same model seems to describe drug resistance in cancer and diseases such as HIV that are "very different from each other biologically." Although cancer is complex, knowing that its dynamics are "relatively simple" will help researchers craft better treatments, he says.