Predictive molecular signature in lymphoma may direct targeted therapy

Researchers from Dana-Farber Cancer Institute, the Broad Institute and colleagues have found a molecular "signature" in a common form of lymphoma that identifies patients unlikely to respond to standard chemotherapy, and who might benefit instead from treatment with certain experimental targeted drugs.

The discovery came from a massive, fine-grained analysis of DNA structure and gene expression in tumors from patients with diffuse large B cell lymphoma (DLBCL), a cancer of white blood cells. It is the most common form of non-Hodgkin lymphoma. The findings are published in the Sept. 11 issue of Cancer Cell.

"We think that capturing this signature will identify a group of patients whose tumors have a genetic basis for deregulated cell growth," said Margaret Shipp, MD, chief of the Division of Hematologic Neoplasia at Dana-Farber and senior author of the study. "These tumors are less likely to respond completely to standard chemotherapy. Because we now know the basis of this deregulated cell growth, these results suggests ways to target it."

About 60 percent of patients with DLBCL can be cured with current therapy — a combination of a monoclonal antibody and four drugs – while the remaining 40 percent are not, and have an unfavorable prognosis. This study defines genetic and biological mechanisms that underlie the differing responses, said Shipp.

In many cancers, scientists have discovered major mutations that drive tumor growth and whose presence or absence can predict treatment outcome, as well as providing targets for selective drugs. Such dominant mutations are less common in DLCBL, which is "much more genetically complicated," said Shipp, who is also the director of the Lymphoma Program at Dana-Farber. Until recently, oncologists assessed outcomes using a set of clinical features such as age, tumor size and pattern, and some easily measured parameters in blood tests. "But these don't tell you about the basic biology of the disease or how you might treat it more effectively."

Poorer response to chemotherapy has been linked to DLBCL tumors whose cells are proliferating rapidly, Shipp noted, though the reason wasn't known.

In the new study, researchers used a new genomic platform, called high-density SNP arrays, to search for subtle changes on chromosomes known as copy number alterations (CNA). Unlike mutations that disable a gene or cause it to go into overdrive, CNAs change gene dosage by increasing or decreasing gene copy numbers. Advanced genomic methods are needed to detect these alterations.

The analysis provided "much higher resolution and a much more fine-grained map of the alterations" than previous studies," said Shipp. In addition, the researchers measured variations in gene expression across the DLBCL genomes, and combined those results with the CNA map.

Together, these experiments revealed a complex pattern of CNAs and associated gene expression changes in DNA samples from patients who were poor responders to chemotherapy. By contrast, the analysis found "clean" genomes, with few CNAs or gene activity abnormalities, in samples from successfully treated patients.

Together, this pattern of structural alterations and changes in gene expression formed a molecular signature predicting an unfavorable outcome with chemotherapy for DLCBL, the researchers reported. Such a test could be combined with clinical factors to improve prognostic testing, Shipp said. In addition, scientists found that the CNAs in the DNA of poor responders caused disruption of two important molecular pathways often involved in cancer.

First, the CNAs reduced activity of the p53 gene, which protects cells against genetic instability that can lead to cancer. Second, the copy number alterations stimulated genes that control cell division, causing the increased proliferation of cells previously observed in poor-prognosis DLBCL.

For bad-prognosis patients, Shipp added, the good news is that experimental drugs exist that target protein kinases that regulate cell division, and might be successful in blocking excessive cell proliferation. She said one such drug, a "pan-cyclin kinase (CDK) inhibitor" called flavopiridol, showed "nice effectiveness" against lymphoma tumors grafted onto mice.

Clinical trials of panCDK inhibitors in patients are being planned, she said.

The research was funded in part by a National Institutes of Public Health grant (PO1CA092625).

The paper's joint first authors are Stefano Monti, PhD, formerly of the Broad Institute and now at Boston University of Medicine, and Bjoern Chapuy, MD, PhD, of Dana-Farber. The paper's other authors, in addition to Shipp, are from Dana-Farber, the Broad Institute, Brigham and Women's Hospital, and Dana-Farber/Children's Hospital Cancer.