Cracking Cancer's Code
It seems jarring at first, all that violent imagery we use when we talk about cancer. It was 40 years ago, with the National Cancer Act of 1971, that President Nixon launched the War on Cancer, and since then, we fight with knife and laser and radiation and chemical weapons, we target tumors, we run reconnaissance with scans and tests, and we hunt down wayward cells that sneak away from the original lesion to seed new growths elsewhere.
Until now, all the aggressive posturing was intentional and, some would argue, necessary in order to engage a disease as insidious as this. Once a healthy cell picks up signals to grow, grow, grow, it indeed becomes a biological enemy, one that left unchecked can infiltrate, overtake and ultimately shut down normal tissues and organs.
Yet while we have focused so obsessively on cutting out tumors or poisoning them with toxic drugs, we have forgotten that there is another strategy, one that takes a more sophisticated approach than carpet bombing the enemy and instead requires studying it up close, learning its ways and weaknesses and then letting that knowledge work for you.
This is the strategy luring more and more researchers into the next phase of cancer care, one in which the science of genes is applied to the biology of cancer, allowing us to peer deep into the clockwork of a cancer cell and then jam its gears or pull out its mainspring. It means that treating cancer may be not an all-or-nothing endeavor but rather a long-term crusade between tumor and therapy, similar to the way we handle chronic diseases like HIV and other infections — with combinations of drugs in a constantly evolving fashion. Such an approach could be far more precise than the scalpels-and-toxins strategy, but it requires us to do one very big thing: learn to read cancer's DNA.
Cancer genomics represents a relatively simple but powerful idea: that the tumor until now has been a forgotten player — dismissed, in part because it was destined to be eliminated anyway, but more out of ignorance, because scientists simply couldn't suss out its secrets. Turning cancer's DNA against it requires you to understand the malignant cell's likes and dislikes, the biological and molecular enablers that keep it alive at the expense of healthy cells around it. What's more, targeting tumors isn't as easy as drawing a clear line between friend and foe. By its very nature, cancer is a perverted version of what's normal, a case of our cells going rogue. A malignancy is indeed a biological adversary, but it didn't start out that way. In the beginning, every cancer is nothing more than a group of healthy cells that for a variety of reasons — an inherited genetic anomaly, tobacco exposure, too many ultraviolet rays from the sun — turns against us. Finding out why this happens calls for more knowledge about genetic science than we have had in decades past. And just as important, you have to appreciate that all of those factors can be different for every different kind of cancer — or even every different patient. Only then can you hope to conquer the disease from the inside out.
"We like to call it a disruptive technology," says Dr. Matthew Ellis, professor of medicine at Washington University in St. Louis and a leader in sequencing the DNA of cancer cells. "Once we translate this technology into the clinical setting, it will completely rewrite the textbook on cancer, because we can start to fundamentally understand each patient's cancer genome and design treatments to match that genomic information."
A New Look at an Old Idea
revolutionary as cancer genomics sounds, the goal of decoding a tumor's DNA isn't new. As far back as the 1960s, enterprising scientists identified an oddity in the size of one of the 23 pairs of chromosomes, the ropelike bundles in which DNA packages itself in cells. It is still known as the Philadelphia chromosome, a hallmark for identifying chronic myeloid leukemia. The problem with analyzing a tumor's DNA, however, has always been one of resolution. If the physical structure of a chromosome looks different in tumor cells, then there certainly must be changes in the genes packed inside. But how do you decipher the genetic instructions tucked so deeply away? How do you keep up with the uncanny power those genes give a cancer cell to sidestep drugs and find new ways to flourish?
The answer started to emerge in 2001 with the mapping of the human genome and its more than 30,000 genes. That biological blueprint serves as the key to answering all sorts of questions about the human condition — especially how normal development is supposed to proceed and where disease intrudes to alter that path.
Could the same be true of a tumor's genetic map? Could sequencing a cancer cell's genome expose its very operating system? Although Dr. Francis Collins, co-mapper of the human genome, was convinced that it could, it took until 2007, after refinement of the techniques and lowering of the cost associated with the process, for the National Institutes of Health to finally be able to begin sequencing cancer cells in a serious way. The mission of the Cancer Genome Atlas (TCGA) is to sequence dozens of cancers, with about 500 samples of each, by 2014. With "the Human Genome Project, the aim was to provide a catalog of the genes in the human genome. TCGA is building a catalog of the things that go wrong in cancer," says Paul Spellman of the National Cancer Institute.
Those things can range from an overactive gene to a complete swapping of DNA from one chromosome to another. "The cardinal feature of a cancer cell is that it's lost the identity it was born with," says Dr. Ernest Hawk, head of the division of cancer prevention and population sciences at MD Anderson Cancer Center in Houston. "It simply doesn't live a normal life and then die as normal cells do."
Researchers already had some limited experience in isolating and explaining the workings of cancer genes. The specific mutations linked to breast cancer in the BRCA 1 and BRCA 2 genes as well as alterations in a gene called APC, which normally suppresses tumor growth and is linked to colon cancer, are behind anywhere from 5% to a third of these diseases. But these are inherited aberrations, and cancers are triggered not only by the genes we get from our parents but also by corruptions to our genome that we acquire in our daily lives — from smoking, sun and diet as well as simple aging. "What has happened in cancer care over the past 20 years has been very piecemeal and ad hoc," says Dr. Todd Golub, director of the cancer program at the Broad Institute. "We discovered some cancer-causing genes here and there, often by stumbling across them. But the notion of being able to say that we are going to systematically and comprehensively interrogate the cancer genome has never been possible before."
Now, he says, improvements in the speed and precision of gene-sequencing technology are making broader sequencing of tumors more realistic. In 2001 the Human Genome Project produced one entire sequence of the human genome at a cost of $1 billion; today it's possible to map a cancer cell's genome for about $5,000, and it won't be long before that will dip to $1,000. In fact, it may soon cost more to store and analyze the data extracted from tumor genomes than it will to generate the maps.
Any cost may well prove worthwhile, however, since even the first slivers of information from cancer genotyping are proving tantalizing. In the first large-scale sequencing of a cancer, using 38 tumors from patients with multiple myeloma, for example, Golub and his team found that this rare blood cancer shared with melanoma a genetic abnormality that causes cells to grow too fast. And that turned out to be a potentially lifesaving piece of information for multiple myeloma patients, since melanoma is already being treated with a drug that targets this aberrant gene.
That's just the first and easiest example of how clinicians and pharmaceutical companies will start exploiting their newfound genetic knowledge with trials of other existing drugs to treat new cancers. The multiple myeloma genome also highlighted several genes that no scientist had ever even described before in the literature, which could become targets for entirely new classes of drugs.
Ellis and his team have sketched out a similar map of one type of breast cancer — tumors that are positive for receptors of the hormone estrogen. Like Golub and researchers at TCGA, they are beginning to see patterns in the genetic triggers of cancer. They suspect, for example, that cancer is not a disease of blockbuster mutations, in which a majority of patients with a particular cancer share aberrations in one or even a few genes. Instead, it's likely that each type of cancer may have a few "driver" mutations and a host of "passenger" changes that appear at a very low frequency. The good news is that both suites of aberrations tend to funnel into a common molecular trunk, like the branches of a tree, and it's that trunk of processes that can become a powerful target for new treatments.
Eager to put genotyping into practice, doctors at MD Anderson and Massachusetts General Hospital, among others, have already begun using sequencing technology to guide treatment of patients in clinical trials. Even without the full genome map of certain cancers, clinicians are using known mutations linked to cancer to dictate which drugs patients receive. In MD Anderson's program, all lung-cancer patients are offered the chance to join a trial in which their tumors are genetically analyzed for some well-known genetic defects thought to play a role in cancer. About 15% of lung cancers, for example, show mutations in a gene that makes a protein critical for cell growth. Patients with this aberration can enroll in trials in which FDA-approved drugs targeting that mutation are being tested as a first-line therapy, instead of chemotherapy, for treating their disease, giving them a head start in gaining any benefits the drugs might provide. (At the moment, these drugs are approved only for patients with advanced cancer for whom other therapies have failed.)
Cancer experts aren't naive enough to believe that sequencing a tumor just once will reveal all they need to know. Cancer is constantly changing its offensive and defensive plans in response to whatever treatments doctors are using against it. The idea is to rebiopsy patients periodically and allow the dynamic genetic changes in the tumors to educate doctors about how aggressive the cancer is, whether it has developed resistance to drugs and even whether it has spread. "The concept is to let the tumor teach us how to treat patients," says Dr. Waun Ki Hong, head of cancer medicine at MD Anderson.
It's all part of the leap toward personalized cancer care, the therapeutic beacon toward which researchers and doctors have been navigating for a long time. "We fully expect that 10 years from now, each cancer patient is going to want to get a genomic analysis of their cancer and will expect customized therapy based on that information," says Brad Ozenberger, TCGA's program director. Only with more individualized therapies that match the right treatment with the right patient at the right time will the battle ultimately be won.