A team of researchers delving into the extensive and complex patterns of mutational patterns found among patients with acute myeloid leukemia (AML) have generated a publicly-accessible dataset that shows the links between the mutational events of the disease and patients’ responses to the drugs used to treat it.
The research team preformed whole-exome sequencing, RNA sequencing and analysis of ex vivo drug sensitivity using a very large patient cohort, as it looked to make sense of a disease with genomic diversity so complicated that Ravindra Majeti, M.D., Ph.D., professor of medicine at Stanford University School of Medicine told Clinical OMICs that “each AML patient is almost like a different case.”
John DiPersio, M.D., Ph.D., chief of the division of oncology, Washington University School of Medicine calls the research a “tour de force.”
One of the two lead authors on the paper, Brian Druker, M.D., the director of the Knight Cancer Institute at Oregon Health & Science University and an investigator with the Howard Hughes Medical Institute, said that for over a decade, his lab “has been trying to integrate functional and genomic data to accelerate our understanding of disease pathogenesis and approaches to therapy” adding that “because of the availability of next-gen sequencing, RNA-seq and improved drug screening technologies, we were able to combine these technologies to analyze a large cohort of samples from patients with AML.”
The research was detailed in a new publication titled “Functional genomic landscape of acute myeloid leukemia,” published on October 17th in Nature, and reported the initial findings from the Beat AML programme on a cohort of 672 tumour specimens collected from 562 patients.
In what DiPersio calls a “heroic effort,” the researchers performed paired-end Illumina HiSeq 2500 exome sequencing on 622 of the tumor samples to identify recurrent mutations, some of which had not been previously described in AML. They dove even more deeply into the genomic information of their samples by analyzing RNA sequences on 451 tumor samples from 411 AML patients to gain gene expression signatures.
However, as DiPersio pointed out, “one of the unique aspects of this work is the identification of drug sensitivity due to combinatorial effects of different mutations.” Testing 122 drugs, they uncovered treatment response patterns that corresponded to specific mutation combinations or gene expression signatures. Dr. Majeti says that, “these data can be studied to find particular mutation groups that confer sensitivity to a particular drug. There is no other way that we could have known that before this work.” He adds that, “we need new ways to test drugs for AML” and that this ex vivo drug testing assay allows to test for “combinations of drugs which would be incredibly challenging to do in mouse models or clinical trials.”
DiPersio added that this testing of various kinds of drugs in a rigorous way has given the field “hypothesis generating information.” However, he notes that there is not strong evidence that the assays recapitulate what would occur in a human. But, he notes that this is a common concern in almost all assays as they are inherently flawed and have significant limitations.
Drucker noted that, “we are just beginning to scratch the surface of analyzing this data, but we are already learning quite a bit about which patients would be predicted to respond to which drugs and why. This is critically important to advancing new and better therapies for patients with AML.”
DiPersio agreed, saying that “this is the first pitch in the game. But, the more pitches that are thrown, the more likely we are to get a strike,” and that the work presented in this paper has developed a framework for going forward.
The team has made their dataset accessible through the Beat AML data viewer (Vizome) so that researchers and clinicians can search what kinds of targeted therapies are most effective against specific subsets of AML cells. However, much more work, including clinical trials, are needed before clinicians can utilize it to make treatment decisions.
Indeed, a clinical trial is in Drucker’s sights. He said that “we are moving this as quickly as possible into biomarker driven clinical trials” and that they “have launched a clinical trial called Beat AML that matches patients with genetically informed treatments. The goals of this study are to treat patients with drugs that will yield significant responses based on pre-clinical data. This is yet another large collaborative effort with 12 universities, 2 genomic providers, 7 pharma companies, and 11 treatment arms that in two years has already enrolled over 400 patients.”
Majeti noted that “AML is undergoing a revolution from a clinical point of view. Up until last year, there were no new drugs on the market for decades. There were 4 or 5 new drugs that gained FDA approval last year and there will be at least that many coming in 2019.”
Mejeti looks forward to the day when these data can be used to stratify patients and to inform which treatments are most appropriate based on combinations of mutations. This work brings the AML field one step closer to achieving the goal of bringing personalized medicine to their patients.