Comprehensive assessment of unfolded protein response and its association with tumor progression in pan-cancer

Disruption of proteostasis in the endoplasmic reticulum leads to the activation of the unfolded protein response (UPR). UPR activation is common in malignant cells and is often considered a contributor to cancer progression. However, the accurate evaluation of UPR status in pan-cancer is still lacking. Furthermore, the UPR-related key characteristics in cancer progression remain to be explored. To fill in these gaps, we first proposed a UPR signature, according to which a UPR score was assigned to each tumor/tumor cell by gene set variation analysis. Based on this, we depicted the UPR landscape across 33 kinds of tumors and 28 kinds of normal tissues, which revealed significant inter- and intra-tumoral heterogeneity in UPR activity across different cancers. Next, by integrating bulk and single-cell transcriptomic data, we unveiled robust correlations of UPR activation with critical genomic and immune features amongst different cancers. In addition, by employing three machine learning algorithms, we identified and validated a promising biomarker derived from the perspective of UPR for breast cancer prognosis. In summary, our study offers an effective approach for evaluating UPR activity in human cancer and underscores the clinical significance of UPR activity in cancer progression.