Mapping the Evolution of Cancer Cells in Leukemia
Drs. Timothy Graubert (left) and Matthew Walter mapped the evolution of cancer cells in seven patients with myelodysplastic syndromes who later died of leukemia.
Using whole-genome sequencing, researchers have uncovered new clues to the genetic alterations underlying acute myeloid leukemia (AML) that arises in patients who were previously diagnosed with a myelodysplastic syndrome. Myelodysplastic syndromes develop when the bone marrow stops producing enough healthy cells, and some cases progress to what is known as secondary AML.
To identify genetic mutations driving this type of AML, researchers at the Washington University School of Medicine in St. Louis sequenced the genomes of abnormal cells (obtained from bone marrow) and normal cells (obtained from skin biopsies) from seven patients whose myelodysplastic syndrome had progressed to AML. Each patient had provided samples when they were diagnosed with a myelodysplastic syndrome and later when they were diagnosed with secondary AML.
By comparing the alterations in each sample, the researchers could track the genetic evolution of the cancer cells over time. They could also identify the cluster of cells that initially caused each cancer to develop—known as the founding clone. Then, they observed new clusters of cells, or daughter clones, that emerged.
The findings, published online March 14 in the New England Journal of Medicine, suggest that the secondary AMLs were derived from bone marrow cells carrying mutations that were present during the myelodysplastic phase of disease.
The collection of proteins within a tumor may be just as diverse as the collection of altered genes. A recent clinical study of samples from patient tumors found "a surprising degree of heterogeneity in protein expression both across the different patient samples and even within the same tumor."
"The reality is that tumors within a single patient can have different genetic changes, and we have observed this heterogeneity at the protein level as well," said Dr. Cesar Castro of Massachusetts General Hospital, who co-led the study. His team profiled the cells using a technology developed to diagnose cancer rapidly.
To better understand the heterogeneity in cancer cells, Dr. Castro continued, researchers will need minimally invasive techniques that can collect serial specimens, as well as new ways to extract as much information as possible from these samples.
"These diseases are combinations of multiple clones, and the clones have a relationship," said Dr. Timothy Graubert, the study's senior author. "In every case, the daughter cells of the founding clone carried forward mutations from the cluster of cells that caused the cancer in the first place."
Implications for Treatment
The finding that suspicious mutations in secondary AMLs were always derived from clusters of cells in myelodysplastic syndromes suggests that targeting these early mutations may be the most effective way to prevent the progression of the mutant cells, the researchers predicted. Drugs that target mutations found only in later-evolving cancer cells may be effective only against those cells, they noted.
To develop the most promising targets, "researchers need to determine whether a particular mutation is part of the founding structure of the [cancer] or whether it occurred at a later time," Dr. Graubert explained.
The researchers identified 11 mutations in the patients' cancer cells that were later found in other patients with AML, which is an indication that they may play a role in the disease. Four of the mutations had not been implicated previously in either myelodysplastic syndromes or leukemia.
In addition to sequencing DNA, the researchers profiled changes in the numbers of copies of genes, as well as in patterns of gene expression. Each cancer cell had hundreds of mutations, but the researchers predict that only a small percentage of these may drive the disease.
The additional mutations, however, provided the statistical power the researchers needed to track the evolution of clones over time.
"The deep sequencing allowed us to confirm that a mutation was present, and it gave us a sense of how many cells in a given sample carry the mutation," Dr. Graubert said. "We could then describe the heterogeneity of the [cancers] and construct a picture of clonal evolution."
A Model of Cancer
The study provides evidence to support a model of cancer that has been hard to test experimentally, noted Dr. Lucy Godley of the University of Chicago Medical Center in an accompanying editorial.
The model describes cancer as a series of acquired mutations and epigenetic alterations that start with a single mutated cell and accumulate in a progressive manner. As the process unfolds, "subclones" of cells acquire new properties that give the cells advantages, such as the ability to resist chemotherapy or to metastasize.
This theory has been around for years, Dr. Godley said in an interview. "But new technologies allow you to ask questions that you couldn't ask before. That's what's so exciting about these kinds of genome studies."
Another recent sequencing study—this one led by Dr. Charles Swanton of the London Research Institute—described the genetic variability within different regions of the same kidney tumor. Instead of sequencing the entire genome, the researchers analyzed the protein-coding regions, or exomes.
"We know that cancer can be heterogeneous, and there's probably an evolutionary relationship between different parts of a tumor mass or different cancer cells in leukemia," said Dr. Elaine Mardis, co-director of the Genome Institute at Washington University and a co-author of the AML study.
The new findings on tumor heterogeneity are consistent with previous work in breast cancer, she noted. For instance, her group conducted an analysis of four DNA samples from the same patient whereas another group compared tumors collected from a patient 9 years apart. Similarly, her group recently reported that, in de novo AML, when patients relapse after receiving chemotherapy, the clone that emerges is derived from a clone that was present at initial diagnosis.
Sequencing the genomes of individual cancer cells could reveal additional layers of genetic complexity, as Dr. Graubert and his colleagues noted in their study. This technology is in its infancy, but a recent pilot study suggests how this could work.
New technologies allow you to ask questions that you couldn't ask before. That's what's so exciting about these kinds of genome studies.
—Dr. Lucy Godley
Researchers at BGI (formerly the Beijing Genomics Institute) sequenced the exomes of 25 cells donated by an Asian man with a form of kidney cancer. Among other findings, the researchers discovered that the man did not have alterations in a gene closely associated with that disease in Western populations.
This study, published in Cell, produced interesting findings, but single-cell sequencing is not ready for clinical studies, noted Dr. Michael Dean, an investigator in NCI's Center for Cancer Research and a co-author of the study.
He added: "It is clear that, for at least some [cancer] types, we will need to use multiple samples and multiple methods of genomic analysis to develop a complete picture of a cancer that can be used to guide therapies."
Dr. Mardis agreed. "These studies are all very exciting because we're now getting to the point of being able to approach [cancer] in different ways."