Anterior gradient 2 (AGR2) is an endoplasmic reticulum (ER)-resident protein disulfide isomerase (PDI) known to be overexpressed in many human epithelial cancers and is involved in cell migration, cellular transformation, angiogenesis, and metastasis. This protein inhibits the activity of the tumor suppressor p53, and its expression levels can be used to predict cancer patient outcome. However, the precise network of AGR2-interacting partners and clients remains to be fully characterized. Herein, we used label-free quantification and also stable isotope labeling with amino acids in cell culture–based LC–MS/MS analyses to identify proteins interacting with AGR2. Functional annotation confirmed that AGR2 and its interaction partners are associated with processes in the ER that maintain intracellular metabolic homeostasis and participate in the unfolded protein response, including those associated with changes in cellular metabolism, energy, and redox states in response to ER stress. As a proof of concept, the interaction between AGR2 and PDIA3, another ER-resident PDI, was studied in more detail. Pathway analysis revealed that AGR2 and PDIA3 play roles in protein folding in ER, including post-translational modification and in cellular response to stress. We confirmed the AGR2–PDIA3 complex formation in cancer cells, which was enhanced in response to ER stress. Accordingly, molecular docking characterized potential quaternary structure of this complex; however, it remains to be elucidated whether AGR2 rather contributes to PDIA3 maturation in ER, the complex directly acts in cellular signaling, or mediates AGR2 secretion. Our study provides a comprehensive insight into the protein–protein interaction network of AGR2 by identifying functionally relevant proteins and related cellular and biochemical pathways associated with the role of AGR2 in cancer cells.

Renal cell carcinoma (RCC) represents about 2–3% of all cancers with over 400,000 new cases per year. Sunitinib, a vascular endothelial growth factor tyrosine kinase receptor inhibitor, has been used mainly for first-line treatment of metastatic clear-cell RCC with good or intermediate prognosis. However, about one-third of metastatic RCC patients do not respond to sunitinib, leading to disease progression. Here, we aim to find and characterize proteins associated with poor sunitinib response in a pilot proteomics study. Sixteen RCC tumors from patients responding (8) vs. non-responding (8) to sunitinib 3 months after treatment initiation were analyzed using data-independent acquisition mass spectrometry, together with their adjacent non-cancerous tissues. Proteomics analysis quantified 1996 protein groups (FDR = 0.01) and revealed 27 proteins deregulated between tumors non-responding vs. responding to sunitinib, representing a pattern of deregulated proteins potentially contributing to sunitinib resistance. Gene set enrichment analysis showed an up-regulation of epithelial-to-mesenchymal transition with transgelin as one of the most significantly abundant proteins. Transgelin expression was silenced by CRISPR/Cas9 and RNA interference, and the cells with reduced transgelin level exhibited significantly slower proliferation. Our data indicate that transgelin is an essential protein supporting RCC cell proliferation, which could contribute to intrinsic sunitinib resistance.

Cancer treatment is far from optimal also because current classification systems do not reflect the complex molecular status of the tumor and its phenotype in sufficient detail. To construct molecular tumor classifiers, omics tools provide complex molecular data reflecting many aspects from genotype to phenotype. However, the true molecular effectors in the cells are proteins which often serve as potent cancer biomarkers and therapy targets. This review summarizes the method aspects that allowed the data‐independent acquisition (DIA) mass spectrometry (MS) to outperform the traditional, data‐dependent acquisition (DDA) approach in recent years. DIA‐MS studies have already recapitulated molecular classification of colorectal and breast cancer, provided data improving molecular classification of prostate and other cancers, and led to validated diagnostic, prognostic, predictive biomarkers and therapy targets for common solid tumors. Tissue‐specific spectral libraries are important for a deep characterization of tissue proteomes. Further perspectives of current cancer proteomics lie in the fields of single‐cell and spatial proteomics and their integration with clinical data. The importance of functional and clinical validation is highlighted to allow stratified and/or personalized targeted therapy.