Mapping genetic mutations in skin cancer

This comprehensive characterization of the biological underpinnings of melanoma is the latest work by researchers involved in TCGA, a National Cancer Institute and National Human Genome Research Institute-sponsored effort to create an atlas of genetic and epigenetic changes that drive different cancers. The researchers collected samples from 331 patients and used several molecular methodologies to complete the study, the largest of its kind to-date for cutaneous melanoma. The findings were published in the journal Cell.

"This study explains some long-standing clinical observations that we could not entirely comprehend, in particular for stage III melanoma," said Stergios Moschos, MD, a UNC Lineberger member, a clinical associate professor of medicine at the University of North Carolina School of Medicine Division of Hematology/Oncology and a member of the data analysis and manuscript committee for the TCGA melanoma project. "This work can also serve as a reference map to assist in personalized prognostic and treatment decisions and future clinical trials for patients with cutaneous melanoma."

In the study, the researchers identified four genomic subtypes of melanoma based on mutations. Two subgroups were defined by mutations that have already been shown to be common in melanoma -- mutations in the BRAF and RAS genes. But from a previously heterogeneous group of people who lacked those mutations, one potentially clinically significant subgroup emerged.

That new group was characterized by inactivating mutations in a gene called NF1. Patients in the NF1-mutant group -- which made up about 14 percent of the entire patient population in the study-- were significantly older and their cancers harbored significantly more mutations. Moschos hopes this finding will trigger new research into targeted treatments for patients in that group, and perhaps into the use of treatments that are already FDA-approved.

"For example, BRAF and MEK inhibitor combinations are now used to treat patients with BRAF mutant melanomas, and MEK inhibitor combinations are being explored for RAS-mutant melanomas," said Ian Watson, PhD, an instructor of genomic medicine who was one of three analysis co-chairs for the project from The University of Texas MD Anderson Cancer Center. "Pre-clinical studies have already demonstrated that some NF1-mutant melanoma cell lines respond to MEK inhibitors, but more work is needed to identify responders and non-responders within this new melanoma subtype, as well as to determine strategies to treat (patients in the Triple Wild-type subgroup without BRAF, RAS, and NF1 mutations)."

The fourth subgroup was defined by patients who lacked BRAF, RAS, or NF1 mutations. They didn't have high incidence mutations, but had some low-frequency mutations in genes such as cKIT, and had a significantly higher number of gene copy alterations (gains or losses). Thirty percent of samples in that subgroup had ultraviolent light (UV)-type signature mutations, while more than 90 percent of samples in each of the other groups had UV-type signatures. The high overall incidence of UV signatures points to the important role of sun or tanning bed exposure in this disease, Moschos said. Overall, the researchers hope the subgroup findings can help personalize treatment decisions and guide new targeted treatment strategies.

The study also helped to reveal the importance of the body's immune response in melanoma, Moschos said. In an analysis of RNA expression data led by Katherine Hoadley, PhD, a UNC Lineberger member and research assistant professor of genetics, and Xiaobei Zhao, a postdoctoral research associate at UNC Lineberger, the study found that 51 percent of patients had high expression levels of genes predominately expressed by immune cell subsets. A pathologic analysis of tumors confirmed that genes thought to be immune-related did come from melanoma-infiltrating immune cells, suggesting an augmented immune response in that group. And patients in the immune-high group with regionally metastatic disease had better overall survival than patients of similar stage and without high expression of immune genes. That finding has implications for immunotherapy treatments that work by unlocking the brakes on the body's own immune response to the disease, Moschos said.

"Given that recent studies showed that immune checkpoint inhibitors may be more effective in patients whose tumors are already infiltrated by immune cells, it is not surprising that these immunotherapies are not effective for everybody," he said.

In addition to identifying cells with a higher immune response, the RNA analysis also identified a subgroup of patients with regionally metastatic melanoma who had a worse prognosis. That group exhibited high expression of genes associated with pigmentation and keratins, which are skin-toughening proteins usually found in epithelial skin cells. This keratin-expressing subtype of melanoma had been previously identified by UNC Lineberger Director Norman Sharpless, MD, the Wellcome Distinguished Professor in Cancer Research, and collaborators, but the worse prognosis of this group is a new finding.

Another immune-related finding was that a protein called PD-L1 that's been used to predict responses to immune checkpoint inhibitors can be overexpressed by melanoma cells for genetic reasons. They found amplification of the PD-L1 gene in people with BRAFV600 mutations in particular. The finding helps explain why the test the measures PD-L1 expression in tumors alone is not sufficient to predict responses to PD1/PD-L1 targeted therapies, Moschos said.