The treatment of cancer by harnessing immune responses has long been pursued. Efforts to turn on the immune system against cancers with inactivated tumor vaccines or intratumor injections of bacterial products to induce local inflammation and recruit an antitumor immune response have led to anecdotal successes. Increasing knowledge about how the immune system is activated, coupled with advances in recombinant DNA technology, has allowed the clinical testing of immune-stimulating cytokines such as interferons and interleukins. These trials have led to a low frequency of durable tumor responses in selected cancers such as melanoma and renal-cell carcinoma at the expense of serious toxic effects. The finding that dendritic cells play a central role in orchestrating a T-cell response to cancer has resulted in multiple clinical trials of dendritic-cell–based vaccines. These studies again provided evidence of occasional tumor responses in a minority of patients. 1

A major limitation of the various approaches to turning on an immune response to cancer is that the immune system exerts a major effort to avoid immune overactivation, which could harm healthy tissues. Cancer takes advantage of this ability to hide from the immune system by exploiting a series of immune escape mechanisms that were developed to avoid autoimmunity. Among these mechanisms are the hijacking of immune-cell–intrinsic checkpoints that are induced on T-cell activation. 2 Blockade of one of these checkpoints, cytotoxic T-lymphocyte–associated antigen 4 (CTLA-4), provided the first evidence of improvement in overall survival for the treatment of patients with metastatic melanoma. 3, 4 The coinhibitory receptor CTLA-4 predominantly regulates T cells at the