One study shows how it may be possible to stop acute myeloid leukemia by targeting a hidden population of cells that support the cancer. Another reveals how AML evolves to evade anti-cancer drugs, pointing the way to strategies to make the cancer vulnerable. Both findings illustrate the power of a vast library of data and tissue samples assembled by OHSU researchers for exploring features of tumor cells from more than 500 patients with the aggressive blood cancer.
Figure: Mutations arising in AML patients treated with crenolanib. Each column displays a patient; each row a specific gene.
“A lot of projects are leveraging this data set and these are prime examples,” says Jeffrey Tyner, Ph.D., an associate professor of cell, developmental and cancer biology in the OHSU School of Medicine and senior author of the two studies.
AML is the most common acute leukemia in adults, and one of the most difficult to treat. Less than a third of newly diagnosed patients survive beyond five years. Until recently, the standard treatment remained a drug combination established 40 years ago. The need for more effective therapies is dire.
David Edwards, Ph.D., a postdoctoral fellow in the OHSU Knight Cancer Institute, and his collaborators made use of hundreds of AML patient samples to search for the kind of vulnerabilities that can be targeted with drugs: cell-signaling pathways that leukemia cells need to survive, but normal cells can live without.
They focused on a class of enzymes, tyrosine kinases, known to drive growth in many cancers. One by one, they blocked genes that encode different tyrosine kinases and measured the effect on leukemia cell growth and survival. One gene stood out: CSF1R, a gene for a receptor tyrosine kinase that appears on the surface of immune cells and acts to signal growth and differentiation into monocytes, macrophages and dendritic cells.
Following up on this lead, they tested a number of chemical inhibitors that bind to the CSF1R receptor and block its signaling. Two of the more potent inhibitors triggered cell death in AML cells from patients.
But the findings also presented a mystery. For one thing, the inhibitor killed all of the leukemia cells in some AML patient samples, but in some samples it was not effective. And when Edwards and colleagues looked for the CSF1R receptor in their samples of AML cells, they found levels to be negligible. The protein was not present on the surface of most AML cells.
“The question was why CSF1R?” Edwards says. “It is not mutated in AML, and it is not over-expressed. Why is it significant?”
In healthy people, CSF1R receptors are found only on immune cells called macrophages and their precursor cells committed to becoming macrophages. Experiments revealed that a small population of CSF1R-expressing cells emerge among the AML cells of patients, and they appear to be vital supporters of cancer cell proliferation.
Whether they are macrophage cells or leukemia cells remains unclear. Edwards and colleagues hypothesize that the CSF1R-expressing cells behave something like the infiltrating immune cells recruited by solid tumors to support cancerous growth.
Their research indicates that CSF1R-expressing cells support leukemia cells via the release of growth factors called cytokines. In one experiment, for example, the researchers found that adding cytokines could rescue AML cells that would otherwise be killed by doses of a potent drug that blocks CSF1R.
The study, reported in the journal Blood, makes a strong case that using CSF1R inhibitors to eliminate the supportive cells could be an effective treatment for a subset of people with AML. The findings were enough to convince the Knight Cancer Institute to launch a phase 2 clinical trial. It’s being led by Elie Traer, M.D., an assistant professor in the OHSU School of Medicine, and supported by the Leukemia & Lymphoma Society and Janssen Research and Development. The trial will assess the safety and survival associated with a CSF1R inhibitor developed by Janssen. It is enrolling adults with AML that has not responded to initial treatment or that has relapsed.
Overcoming drug resistance
Targeted therapies, drugs designed to cancel the molecular signals driving cancer growth, have been developed against acute myeloid leukemia. But persistent leukemia cells inevitably develop resistance and within weeks or months the disease restarts its fatal progression.
Tyner and colleagues used their AML data set to explore how cancer cells develop resistance to one of the most potent targeted therapies developed to fight AML. “The hope is to understand drug resistance mechanisms well enough to inform the design of better treatment regimens for each individual patient,” Tyner says.
Mutations in a gene called FLT3 are the most commonly identified driver of AML. And drug companies have developed at least eight drugs that target the FLT3 protein, a tyrosine kinase receptor involved in growth signaling in immature blood cells.
Tyner and colleagues focused on a drug called crenolanib. It’s a second-generation FLT3 inhibitor notable because it maintains activity even when leukemia cells develop mutations in the FLT3 gene that allow them to resist other FLT3 inhibitors. In clinical trials, however, resistance to crenolanib emerges in a majority of cases just as with other FLT3 inhibitors.
With the extensive library of AML patient samples, the OHSU researchers were able to use DNA sequencing of leukemia cells from crenolanib-treated patients to track evolution of drug resistance.
They found that with crenolanib, unlike other FLT3 inhibitors, resistance does not arise from further mutations in the FLT3 gene. Instead, they found three distinct patterns of mutations involving a variety of genes other than FLT3. (The researchers have made it possible for others to explore the data in Vizome, an online data brower.)
Each mutation pattern is associated with a different prognosis, and each likely requires different treatment combinations to avoid or overcome drug resistance. The researchers describe the work in a paper published this week in Nature Communications.
The researchers say the study has implications for patient care: comprehensive sequencing should be carried out on patient samples at the start of and during treatment to identify and preemptively target populations of leukemia cells with different resistance mutations.
“You need to know what’s there with very deep sequencing up front,” Tyner says. “That would help you understand which drug combinations are needed.”
CSF1R inhibitors exhibit anti-tumor activity in acute myeloid leukemia by blocking paracrine signals from support cells by David K. Edwards V, Kevin Watanabe-Smith, Angela Rofelty, Alisa Damnernsawad, Ted Laderas, Adam Lamble, Evan F. Lind, Andy Kaempf, Motomi Mori, Mara Rosenberg, Amanda d’Almeida, Nicola Long, Anupriya Agarwal, David Tyler Sweeney, Marc Loriaux, Shannon K. McWeeney and Jeffrey W. Tyner. Blood (preprint online Nov. 13, 2018)
Clinical resistance to crenolanib in acute myeloid leukemia due to diverse molecular mechanisms by Haijiao Zhang, Samantha Savage, Anna Reister Schultz, Daniel Bottomly, Libbey White, Erik Segerdell, Beth Wilmot, Shannon K. McWeeney, Christopher A. Eide, Tamilla Nechiporuk, Amy Carlos, Rachel Henson, Chenwei Lin, Robert Searles, Hoang Ho, Yee Ling Lam, Richard Sweat, Courtney Follit, Vinay Jain, Evan Lind, Gautam Borthakur, Guillermo Garcia-Manero, Farhad Ravandi, Hagop M. Kantarjian, Jorge Cortes, Robert Collins, Daelynn R. Buelow, Sharyn D. Baker, Brian J. Druker and Jeffrey W. Tyner. Nature Communications (Jan. 16, 2019)
OHSU-led effort results in largest cancer dataset of its kind by Amanda Gibbs. OHSU News (Oct. 17, 2018)