During the past decade cancer immunotherapy has been revolutionized by the usage of immune checkpoint blockades (ICB): started with the approval of Ipilimumab in 2011, which targets the CTLA-4 protein, and later with the PD-1 and PD-L1 checkpoint inhibitors. These therapies resulted in unprecedented rates of long-lasting tumor responses in patients with a variety of cancers. As a result, these drugs have been FDA-approved for many cancer types. Nevertheless, most patients do not respond to treatment or acquire resistance, reaching response rates of less than 40% in solid tumors with large number of partial responders.
Why some patients fully respond to ICB treatment and others demonstrate partial response or no response at all is still not entirely clear. One factor that has been recently shown to contribute to patient response is the differences in the metabolic adaptation of tumor and immune cells and their competition for resources. In this study, we use single-cell RNA-seq of CD45+ cells taken from 48 melanoma tumor biopsies, pre and post treatment. Using the set of metabolic genes, we identified 10 distinct clusters, with a few of them representing a mixture of different cell types sharing unique metabolic characteristics. 6 of these clusters were found to be significantly associated with patient response, highlighting metabolic genes and pathways that can be targeted to enhance effective immune response. Moreover, we observed specific metabolic patterns that distinguish between responders and non-responders following therapy. Finally, we devised metabolic signature scores that are highly predictive of patient response.