84 A study led by UCLA Health Jonsson Comprehensive Cancer Centre investigators reveals how melanoma, the deadliest form of skin cancer, evolves to resist immunotherapy and identifies a potential strategy to prevent or reverse that resistance. The team found that relapsing melanoma tumours often acquire genomic DNA copy-number variants, which delete or amplify sections of DNA. These variants frequently affect genes that control the cancer cells’ ability to self-destruct in response to damage caused by immune attack. The cumulative effect of copy-number changes, often involving multiple cell-death genes, allows cancer cells to survive immune attacks, leading to tumours relapsing or regrowing months or years after the initial therapy-induced tumour shrinkage. The findings, published in the journal Immunity, suggest making the tumour cells more prone to self-destruct following immune attacks could help maintain or restore the effectiveness of immune checkpoint inhibitors, a widely used form of cancer immunotherapy. “Studies of how cancer genomes evolve to acquire resistance to therapies often focus on small-scale or point mutations,” said Dr Roger Lo, professor of medicine, dermatology and molecular & medical pharmacology at the David Geffen School of Medicine at UCLA and senior author of the study. “This study highlights how large-scale mutations, including gene copy-number changes, can be an efficient way for cancers to evolve.” Resistance to immunotherapy, especially immune checkpoint inhibitors, remains one of the greatest challenges in cancer treatment. While many patients with melanoma initially respond well to checkpoint inhibitors, 40–60% eventually experience relapse as the cancer adapts and becomes resistant. To understand how tumours become resistant over time, the researchers analysed tumour samples from patients before and after they experienced relapse following treatment with immune checkpoint inhibitors. They also analysed tumours from surgeries that were followed by adjuvant checkpoint therapy and subsequently from biopsies at recurrences. They combined analysis of tumour-derived data from patients with analyses of published screening studies aimed at defining all cancer immunotherapy resistance genes, as well as analyses of cell-culture and mouse models the UCLA team developed. The team then tested whether pharmacologically sensitising tumours to programmed cell death, or apoptosis, would re-sensitise tumours to immunotherapy. In both cell-line and mouse models, lowering the apoptotic threshold of tumour cells reinstated killer immune, or T, cells’ ability to induce apoptosis of cancer cells. In a mouse model of melanoma, the UCLA team was able to prevent tumour relapses if a pro-apoptotic drug was added to treat residual tumours that had responded to immunotherapy. “These findings point to a promising preventive therapeutic strategy,” said Dr Lo, who is a member of the UCLA Health Jonsson Comprehensive Cancer Centre. “If we can intervene early to target incipient resistant cancer cells in residual tumours, we may be able to extend the effective durability of immunotherapy in patients with melanoma.” The study also reveals the importance of tracking tumour evolution at the single-cell level. Using single-cell whole-genome sequencing, the team discovered that some of the resistance-related genetic changes were already present in small subsets of tumour cells before treatment began. In addition, relapsing tumours contain heterogeneous subclones with distinct permutations of critical DNA copy-number variations. The genomic copy-number variations contribute to resistance even before therapy begins and fuel natural selection for resistance evolution after therapy begins. For some patients, a wait-and-see or monitoring approach during remission may not be optimal.