Acute myeloid leukemia is the deadliest leukemia among children. Only half of patients will achieve long-term remission. But scientists see change on the horizon. Taking advantage of today’s genomic sequencing technologies, researchers are discovering that on the molecular level, AML in children is dramatically different from the disease in adults. At the same time, researchers are clarifying how unique each child’s disease may be, with implications for more tailored treatment, individualized prognosis and monitoring, and the opportunity to develop therapies that target childhood AML’s particular characteristics.
Studies presented at this year’s conference of the American Society of Hematology exemplify the widening scope, and burgeoning hope, of childhood AML research.
“In general, the field [of pediatric AML] has been rather stagnant until recently,” said Dr. Soheil Meshinchi, a pediatric oncologist and investigator of childhood AML at Fred Hutchinson Cancer Research Center. Patients receive the same combination of chemotherapy because their AML “looks the same under the microscope,” said Meshinchi, “but we’re learning more and more that all AML is not the same entity.”
Most AML occurs in adults, especially older adults, and the treatments devised for these patients “trickles down to pediatrics,” said Dr. Katherine Tarlock, a pediatric AML investigator at Fred Hutch who also treats children with AML at Seattle Children’s Hospital. “But the more we’re now learning about pediatric AML [the more we realize that] it’s really drastically different from adult AML … Which has big implications for treatment strategies, especially as we move forward into targeted treatment strategies.”
As researchers begin to expand their investigations into targeted therapies, including immunotherapies, the potential targets are likely to be very different in childhood AML than in adult AML, Tarlock said.
Though pediatric and adult AML cells may look the same under a microscope, the DNA changes driving normal myeloid blood cells to turn cancerous are very different in children and adults. Tarlock, in work that was presented Dec. 5 at ASH, used a commercially available clinical genetic sequencing platform to assess not only sequence changes in adult and pediatric disease, but also large structural changes, in which whole chunks of chromosomes may be lost, duplicated or change position.
Examining samples from 179 children and 755 adults, she found catastrophic chromosomal alterations in 60 percent of AML from children under 18, but in only 38 percent of adult samples. Several of the specific structural alterations Tarlock identified in pediatric AML can be used by oncologists to better predict patient outcome or match them to optimal treatment.
Because the structural changes can have such strong effects, oftentimes a child whose myeloid cell’s genome has been so drastically altered only needs one (or sometimes no) other mutation to tip normal cells into cancer, Tarlock said. Adults, in contrast, usually seem to accumulate many mutations before their cells begin to run amok.
The sequencing platform Tarlock used “is a clinical platform available right now,” said Meshinchi, noting that Seattle Children’s pediatric oncologists already use this same sequencing platform on all AML patients. “The amount of information that is gathered is enormous” and has both clinical and research applications, he said.
Tarlock’s findings echo those of the National Cancer Institute's TARGET (short for Therapeutically Applicable Research to Generate Effective Treatments) AML Initiative, also presented Dec. 5 at ASH by Dr. Jason Farrar, a pediatric hematologist-oncologist at the University of Arkansas for Medical Sciences. Meshinchi leads the TARGET AML Initiative, which seeks to chart the molecular landscape of childhood AML.
The TARGET AML Initiative compared samples from a large cohort of children treated on clinical trials through Children’s Oncology Group with data generated in its adult counterpart, The Cancer Genome Atlas, or TCGA, AML initiative. Again, the team found that children tended to have large genomic alterations compared to the many smaller mutations found in adults. Individualizing even more, the group found that certain genes are more likely to be affected in children than adults, and that some genes are mutated within specific age groups.
“In the majority of younger children AML is caused by large structural chromosomal alterations that lead to a very rapid evolution of the disease,” Meshinchi said. “In adults it is the combination of simple sequence mutations — which people accumulate over years — that drives the disease.”
Common DNA changes also separate AML in the youngest patients from AML in older children, according to work presented by Rhonda Ries, a member of Meshinchi’s lab. In a sample of 164 patients under 2 years of age, Ries found that more than 90 percent had certain types of large structural changes in their DNA. In the genomes of AML cells from this group, Ries saw that either a section of one chromosome had jumped to a new chromosome, or that large segments of a chromosome had been duplicated or lost. And unlike genetic change seen in older AML patients, these leukemia-driving alterations didn’t accumulate with time — they were always present at the time of the child’s birth, Meshinchi said.
The highlight of the study, he said, “is how unique these younger children are. [Therapeutic] targets that are being developed even for older pediatric patients may not really reach this group of very young patients unless we are devising treatments and therapies that are specifically for them."
Already scientists are discovering genetic signatures that hold the potential to improve care and match pediatric AML patients to treatments most likely to eliminate their cancer. Dr. Dalia Selim, also in Meshinchi’s lab, presented at ASH the results of an analysis examining how subtle differences in mutations in the gene nucleophosmin, or NPM1, can alter outcomes.
Selim found that in adults, a particular mutation in the NPM1 gene, known as a Type A variant, predominated. In children, in contrast, she found that it was a slightly different mutation in NPM1, a non-Type-A mutation, that was most common. Selim found that patients with non-Type-A mutations fared better. If validated, these are clinical practice-changing results, Meshinchi said.
Genes can be turned on and off without any alterations in the DNA code, by “epigenetic” alterations in molecules that attach to DNA or its packaging. Epigenetic alterations can combine with mutations to cause AML. As part of the TARGET AML initiative, Dr. Timothy Triche Jr. of the University of Southern California presented work showing the association of specific epigenetic signatures with AML variants arising from different cell types. Triche and his group also found that different epigenetic signatures may reveal vulnerabilities to various drugs.
“We hope to one day use genetic and epigenetic signatures for therapy and risk stratification, or to predict outcome,” Meshinchi said.
Dr. Emilia Lim of the British Columbia Cancer Agency presented the results regarding another epigenetic biomarker, a study in which researchers generated a predictive model based on epi-genetic molecules known as microRNAs. We make many different types of microRNAs, and Lim and her team identified 16 microRNAs that increased and 20 that decreased in pediatric AML patients who experienced an event such as death or relapse within five years of AML treatment. The group was able to use the microRNA model to divide patients groups at low, moderate or high risk of poor outcomes.
Lim’s strategy “is a completely new concept in risk determination, using a cluster of microRNA expression,” Meshinchi said. “It is a completely new way of looking at risk stratification … [which allows the researchers to] much more accurately define risk groups."
As researchers strive to create ever-more individualized therapies, they seek targets in pediatric AML. Tarlock presented evidence that a protein known as mesothelin may be an attractive target for precision medicine strategies in pediatric AML.
Mesothelin is a protein expressed only rarely in certain healthy tissues. For reasons that remain unclear, many solid tumors, including pancreatic and lung cancer, express mesothelin at high levels. Its rarity in healthy tissue but abundance in cancer makes mesothelin a potential goldmine for researchers seeking targets that make cancer cells, but not healthy cells, vulnerable to new treatments such as immunotherapies.
AML cells in some pediatric patients also appear to express the mesothelin gene at high levels — but whether mesothelin is a good, stable target for new therapies remained uncertain. In her study, part of the NCI’s TARGET AML Initiative, Tarlock showed that in one quarter to one third of pediatric patients’ AML, mesothelin appears to be a potential target at both diagnosis and relapse.
“The striking thing is that mesothelin expression correlated with a few common structural variants in pediatric AML,” Tarlock said. She and her team are now looking at whether mesothelin may be a biomarker for these patients, possibly signaling prognosis for treatment response or relapse risk.
Additionally, pediatric oncologists can take advantage of the fact that various immunotherapies against mesothelin are already in the works for solid tumors.
As scientists identify more genetic changes in pediatric AML and uncover their significance, treatment will continue to improve. Even though it may take time to develop new therapies, knowing a child’s AML harbors a genetic variant associated with poor prognosis can help oncologists choose better care.
“At the very least, these are the patients that should go to transplant,” said Meshinchi. “And if there’s any obvious therapeutic target available, we can use it [to tailor treatment].”
And simply knowing the genetic variants in a child’s AML can make it possible to keep an eye out for relapse long before it manifests clinically, he said. Knowing a child’s mutation allows oncologists to finally “see” their disease. Fred Hutch and Children's Oncology Group are working with commercial partners to create the world’s first lab that will be able to design a monitoring test for any child with any mutation, Meshinchi said.
“A deeper understanding of AML is now coming out,” he said. “We can choose treatment more appropriate [to a child’s] AML.”
Now the question is, “How do we do it for everyone?”
Sabrina Richards is a staff writer at Fred Hutchinson Cancer Research Center. She has written about scientific research and the environment for The Scientist and OnEarth Magazine. She has a Ph.D. in immunology from the University of Washington, an M.A. in journalism and an advanced certificate from the Science, Health and Environmental Reporting Program at New York University. Reach her at firstname.lastname@example.org.