Photo courtesy of
University of Michigan
Newly identified molecular signatures may allow us to predict which patients with chronic myelomonocytic leukemia (CMML) will respond to treatment, according to a study published in The Journal of Clinical Investigation.
Finding effective biomarkers is particularly crucial for CMML, the investigators said, because current treatment is slow-acting. Patients must often undergo as much as 6 months of treatment before there are signs of response.
“The slow kinetics is what gets us,” said study author Maria E. Figueroa, MD, of the University of Michigan Medical School in Ann Arbor.
“It’s not just one week or one dose to see signs of response. A good biomarker test could potentially prevent patients who are unlikely to respond from receiving prolonged, unwarranted treatments.”
With this in mind, Dr Figueroa and her colleagues used next-generation sequencing techniques to analyze 40 CMML samples from patients treated with the DNA methyltransferase inhibitor decitabine.
The researchers found 167 differentially methylated regions of DNA at baseline that distinguished patients who responded to decitabine from those who did not. There was a methylation difference of 25% or more between responders and nonresponders.
The investigators speculated that the baseline differences in DNA methylation could be used to predict treatment response at diagnosis.
So they used the percentage of cytosine methylation at each genomic location among the 40 patients as potential predictors and applied a machine-learning approach to build an epigenetic classifier.
The investigators tested the classifier in 28 additional CMML samples and found it was 87% accurate in predicting a patient’s response to decitabine.
The researchers also investigated why nonresponders were resistant to decitabine and found that 2 proteins, CXCL4 and CXCL7, were overexpressed in nonresponders. In cells exposed to high levels of these chemokines, the effects of decitabine were blocked.
“We are pursuing this to understand why these proteins block the effect of the drug and whether we can develop a new compound that could be used along with decitabine to turn nonresponders into responders,” Dr Figueroa said.
The researchers are also working to refine and translate their findings into a viable biomarker test that could be used in the clinic.