Spatial proteomics reveals sirtuin 1 to be a determinant of T-cell infiltration in human melanoma

Abstract Background The tumour microenvironment significantly influences the clinical response of patients to therapeutic immune checkpoint inhibition (ICI), but a comprehensive understanding of the underlying immune-regulatory proteome is still lacking. Objectives To decipher targetable biologic processes that determine tumour-infiltrating lymphocytes (TiLs) as a cellular equivalent of clinical response to ICI. Methods We mapped the spatial distribution of proteins in TiL-enriched vs. TiL-low compartments in melanoma by combining microscopy, matrix-assisted laser desorption mass spectrometry imaging and liquid chromatography–mass spectrometry, as well as computational data mining. Pharmacological modulation of sirtuin 1 (SIRT1) activity in syngeneic mouse models was used to evaluate the efficacy of pharmacological SIRT1 activation in two syngeneic melanoma mouse models, one known to be α-programmed cell death protein 1 (PD-1) sensitive and the other α-PD-1 resistant. Results Spatial proteomics and gene ontology-based enrichment analysis identified > 145 proteins enriched in CD8high tumour compartments, including negative regulators of mammalian target of rapamycin signalling such as SIRT1. Multiplexed immunohistochemistry confirmed that SIRT1 protein was expressed more in CD8high than in CD8low compartments. Further analysis of bulk and single-cell RNA sequencing data from melanoma tissue samples suggested the expression of SIRT1 by different lymphocyte subpopulations (CD8+ T cells, CD4+ T cells and B cells). Furthermore, we showed in vivo that pharmacological SIRT1 activation increased the immunological effect of α-PD-1 ICI against melanoma cells in mice, which was accompanied by an increase in T-cell infiltration and T-cell-related cytokines, including interferon (IFN)-γ, CCL4, CXCL9, CXCL10 and tumour necrosis factor-α. In silico analysis of large transcriptional data cohorts showed that SIRT1 was positively associated with the proinflammatory T-cell chemokines CXCL9, CXCL10 and IFN-γ, and prolonged overall survival of patients with melanoma. Conclusions Our study deciphers the proteomics landscape in human melanoma, providing important information on the tumour microenvironment and identifying SIRT1 as having important prognostic and therapeutic implications.