BY Andrea Mongler
Illustration by Abigail Goh
Three billion is such a large number that it’s hard to fathom.
To put it in perspective:
- 3 billion seconds is about 95 years
- If you took 3 billion steps, you could walk around Earth at its equator more than 45 times
- Stacking 3 billion pennies would result in a tower nearly 3,000 miles tall
Three billion also happens to be the number of DNA building blocks, or base pairs, that make up a human genome. And nearly every single cell in our bodies contains two copies of those 3 billion base pairs.
Together, the base pairs that make up our DNA provide the instructions our bodies need to develop, survive and reproduce. They’re like a massive instruction manual packed into a very tiny space. And reading that instruction manual may be the best way to figure out which cancer patients need aggressive treatments to survive.
Oncologists have long known that no two cancer patients are alike. While some patients have fast-moving cancers that will lead to death if they’re not treated quickly and aggressively, others have less severe disease that doesn’t require aggressive treatment.
This is why oncologists tailor treatments to individual patients. There’s no one-size-fits-all approach. And since cancer results from genetic mutations, oncologists’ ability to decide on the best treatment for individual patients often hinges on their ability to accurately identify mutations in a patient’s cancer through genetic testing.
The original form of genetic testing, called karyotyping, involves looking at a patient’s chromosomes under a microscope. It’s still in routine use today, and it’s usually combined with other lab tests, which means oncologists need to interpret and integrate a variety of test results. They use these test results to place patients in risk categories and guide treatment decisions.
“Genetic analysis of cancers often requires multiple testing approaches—some that use technology that’s nearly 50 years old,” says David Spencer, MD, PhD, oncologist at the Alvin J. Siteman Cancer Center at Barnes-Jewish Hospital and Washington University School of Medicine. Spencer also is the medical director of the clinical sequencing facility at Washington University School of Medicine’s Elizabeth H. and James S. McDonnell III Genome Institute. “But if we could quickly obtain a comprehensive genomic profile of a patient up front with one test, it would be a much more effective and efficient way to decide on appropriate treatments.”
Recently, Spencer and colleagues demonstrated that this is possible using a process called whole genome sequencing. Specifically, their research showed that whole genome sequencing is often a more accurate way than conventional testing to stratify patients with acute myeloid leukemia (AML) into risk categories and make treatment decisions accordingly. They also showed that this approach was fast and cost-efficient. Their research holds promise not just for AML patients but also for cancer patients more broadly.
To understand what whole genome sequencing is and why it’s such a promising strategy, it helps to first understand the genome.
Unraveling the genome
From the tiniest bacterium to the largest animal, every living creature has a genome. And while Salmonella has very little in common with, say, a blue whale, they both have DNA and that DNA contains the same basic “building blocks.” A blue whale just happens to have a lot more DNA, and its building blocks are ordered differently than Salmonella’s.
DNA’s building blocks are more formally known as bases, and there are four of them: adenine, thymine, cytosine and guanine. These bases—commonly abbreviated as A, T, C and G—are connected to molecules made up of sugar and phosphate in a continuous strand of DNA.
Each of our cells contains two of these DNA strands that wind together, forming a pattern that’s known as a double helix. It looks like a spiraling ladder, with the sugar and phosphate molecules forming the sides and the bases joining together in pairs to form the rungs. DNA base pairs are particular about their partners: T pairs only with A; C pairs only with G. When 3 billion of these base pairs line up in a certain order, the result is a human genome.
Segments of DNA in the genome are known as genes, and in humans they can vary in length from about 1,000 bases to more than 1 million. Each gene carries a very specific set of instructions, like a chapter out of the body’s massive instruction manual. Every set of instructions is for making a particular protein. Proteins make up everything in our bodies—like bones, teeth, hair and muscles—which means our genes control a lot. Your height, your eye color, your propensity for certain diseases: These are just a few examples of traits controlled by your genes.
On an individual level, everyone’s genome is nearly identical but still unique. There’s variation in 0.1% of our DNA, and that small bit of variation is the reason people aren’t identical. The variation is also part of the reason why some people develop diseases like cancer while others never will.
And in people diagnosed with cancer, identifying acquired changes known as mutations is key to figuring out whether aggressive treatment is needed. But in order to identify all of the relevant mutations, researchers first needed to gain a comprehensive understanding of the entire human genome.
Sequencing the first cancer genome
In 2003, researchers finished sequencing the human genome for the first time in a multi-institution effort known as the Human Genome Project. (The Genome Institute contributed 25% of that sequence.) When the work was complete, the project’s researchers had figured out the order of the base pairs that make up the genome. All 3 billion of them.
The project spanned 13 years and cost the equivalent of more than $5 billion today. It was a groundbreaking accomplishment, but the idea of sequencing individual genomes from people with cancer to better understand the genetic basis of the disease wasn’t feasible—at least not then.
Timothy Ley, MD, oncologist at Siteman Cancer Center and associate director of cancer genomics at the Genome Institute, says researchers knew that sequencing technology would become faster and less expensive, and it quickly did. Ley and colleagues at the Genome Institute and Siteman Cancer Center were ready. Even before the Human Genome Project was complete, they began planning to sequence the genome of a person with cancer in an effort to find all of the mutations that contributed to that individual’s disease.
“We were convinced that if we compared the genome from a patient’s tumor cells to the genome from their normal, healthy cells, we would find a relatively small number of changes,” Ley says. “And if we compared the cancer genomes of multiple individuals to each other, we expected that recurrent mutations would be incredibly important clues for what was really relevant.”
In 2008 a team of researchers—Ley; Elaine Mardis, PhD; John DiPersio, MD, PhD; Rick Wilson, PhD, and many other colleagues at the Genome Institute, the Division of Oncology at Washington University and Siteman Cancer Center—were the first group in the world to successfully sequence the complete genome of a person with cancer. This endeavor was made possible by a gift from Alvin J. Siteman. (Mardis and Wilson are now working at the Institute for Genomic Medicine at Nationwide Children’s Hospital.) Specifically, the team sequenced the genome of a person’s AML cells and compared it to her skin cells as a reference, looking for differences between the two. In the patient’s AML tumor cells, they found three recurrent mutations that were relevant for her disease.
Within a year, the same group sequenced the genome of a second person with AML, which revealed additional recurrent mutations that were important for that patient’s AML.
“It was a game-changer because it created a blueprint for genome sequencing in cancer,” Ley says.
He also notes that after the first few AML genomes were sequenced, it became obvious that most of the acquired mutations were random and irrelevant for cancer development; they were simply a consequence of aging. Only a small percentage of the mutations were recurrent in other people with AML, and the most recurrent ones were the most likely to be important for the disease.
This speaks to the large amount of genetic diversity in the disease from one person to another—and the need to identify specific mutations in individuals to make appropriate treatment decisions. And with fast-moving AML, identifying those mutations quickly is critical.
“‘Acute’ is the key word in AML,” Ley says. “People with AML are very sick when they come in, so their treatment can’t wait for weeks while we do an extensive diagnostic workup. Unlike many other cancers, we need an answer in days.”
The first cancer genome cost about $1.6 million to sequence and the second about $500,000. Over the years, companies continued to improve the technology so that sequencing became faster and less expensive.
Ley and colleagues sequenced 22 more AML genomes, and by the time they were finished, the cost had fallen to around $50,000. While that was far lower than the cost to sequence the first cancer genome, it was still too expensive to be practical on a wide scale for individual patients.
The cancer field shifted its attention to a technique called whole exome sequencing, which is a much less expensive process. Instead of sequencing the entire genome, whole exome sequencing only identifies mutations in about 1% of the genome. Specifically, it looks for mutations in exons, which are the parts of genes that provide the instructions for making proteins.
Most known disease-causing mutations occur in exons, which makes whole exome sequencing a reasonable approach for some diseases. But for other diseases, including AML, many relevant mutations occur outside the exons.
“This is why we think whole genome sequencing is the best approach for AML,” Spencer says. “To make sure we’re making the right treatment decisions for people with AML, we need to identify all the mutations in the genome right from the start.”
Demonstrating a new approach
Spencer, Ley and Washington University colleagues—including pathologist and immunologist Eric Duncavage, MD, section head of hematopathology, and Molly Schroeder, PhD, a pathology and immunology scientist—set out to demonstrate that whole genome sequencing was an accurate, comprehensive approach for identifying mutations in people with AML. Just as important, they wanted to show that it could be quick and cost-efficient.
They started by developing an approach that focused on speed and on finding mutations that are already known to play a role in AML. In other words, they zeroed in on specific areas of the genome. Spencer notes that historical knowledge made this possible.
“We’ve sequenced hundreds of genomes in the last 10 years, so we have a very good idea which genes to pay attention to when we’re sequencing,” he says. “That means we know where to look, and when we find a difference in a patient’s genome versus the reference genome, we know whether it’s important.”
With their approach in place, the researchers recently identified a group of people with AML who had been treated at Siteman Cancer Center and who had undergone conventional genetic testing. They picked individuals with specific mutations to see if their method could accurately identify those mutations. They also selected a group of current patients who would undergo conventional tests as well. Then they conducted whole genome sequencing on the participants’ AML samples and compared the results with the conventional test results.
Whole genome sequencing identified all of the clinically relevant mutations revealed by standard testing. It also identified clinically relevant mutations that conventional tests didn’t detect.
The results came fast, too came fast, too—in an average of five days and sometimes as few as three. That’s far faster than turnaround times in the early days of whole genome sequencing.
At less than $2,000, the cost to sequence a participant’s genome in the study was much lower than whole genome sequencing costs a decade ago. It’s also in line with the cost of conventional genetic testing for AML, which is typically $1,000 to $2,000.
In addition, whole genome sequencing identified mutations in participants for whom conventional testing had failed. That’s because conventional testing requires living cells, but sometimes blood samples don’t contain enough living cells for the test to work. In comparison, whole genome sequencing requires only DNA.
The researchers weren’t surprised that their approach worked. In fact, they had expected it to work. Their primary goal was less about proving that whole genome sequencing could accurately identify cancer-causing mutations and more about demonstrating speed and ease of use.
“There is a perception in the field that whole genome sequencing takes a long time, and it’s really complicated,” Spencer says. “We have demonstrated that with today’s technology, it doesn’t have to take a long time, and it doesn’t have to be complicated if you focus on detecting mutations that you already know are clinically relevant.”
“The gold standard’
Whole genome sequencing holds great promise for people with AML who need to start an appropriate treatment regimen quickly for a good chance of success. But how soon the technology can become widely available will likely hinge on how soon insurers begin to cover it. Spencer, Duncavage and colleagues are in the process of pursuing insurance reimbursement and will continue their research in the meantime.
Investigators at Washington University also hope to see whole genome sequencing used for other cancers in the near future, and they’ve already started a pilot study with people who have lung cancer, with other diseases to follow soon.
In addition, they note that the technology will likely provide other information that’s useful. For example, whole genome sequencing has the potential to provide the information needed for HLA typing, the process used to match people with cancer to donors for bone marrow transplants.
“One of our goals is to prove that whole genome sequencing will result in kind of a one-stop shop,” Spencer says. “Oncologists would have all the information they needed in one place, with no delay, to talk to their patients about the therapy that might be right for them.”
And Ley emphasizes that whole genome sequencing has the potential to improve cancer treatment decisions around the world.
“Whole genome sequencing is a gold standard for defining all of the major mutations in a cancer genome, and it could become a global standard because of the nature of the technology,” Ley says. “If this approach is widely adopted, oncologists everywhere will have the same information for their patients and be better equipped to recommend the best possible treatment decisions.”
Learn more about the recent study by Duncavage, Schroeder, Ley, Spencer and colleagues.
In addition to the people named in this article, many others have made significant contributions to the work described here. In the Division of Oncology at Washington University School of Medicine, additional members of the team have included Jack Baty; Matthew Christopher, MD, PhD; Li Ding, PhD; Timothy Graubert, MD, now at Massachusetts General Hospital; Malachi Griffith, PhD; Obi Griffith, PhD; Sharon Heath; Jeffery Klco, MD, PhD, now at St Jude Children’s Hospital; Daniel Link, MD; Chris Miller, PhD; Allegra Petti, PhD; Michael Rettig, PhD; Matthew Walter, MD; Lukas Wartman, MD; John Welch, MD, PhD; Peter Westervelt, MD, PhD. In the Department of Pathology and Immunology, members of the team have included Andrew Hughes, MD, PhD; Julie Neidich, MD; Jacqueline Payton, MD, PhD; Mark Watson, MD, PhD. At Washington University School of Medicine’s Elizabeth J. and James S. McDonnell III Genome Institute, members of the team have included Andrew Bohannon; Matt Cordes; Feiyu Du; Catrina Fronick; Bob Fulton; Lucinda Fulton; John Garza; Scott Kruchowski; David Larson, PhD Sandra MacMillan; Tom Mooney; Michelle O’Laughlin; Jason Walker; Roxanne Wilson.