Optimized Data Set and Feature Construction for Substrate Prediction of Membrane Transporters.

α-Helical transmembrane proteins termed membrane transporters mediate the passage of small hydrophilic substrate molecules across biological lipid bilayer membranes. Annotating the specific substrates of the dozens to hundreds of individual transporters of an organism is an important task. In the past, machine learning classifiers have been successfully trained on pan-organism data sets to predict putative substrates of transporters. Here, we critically examine the selection of an optimal data set of protein sequence features for the classification task. We focus on membrane transporters of the three model organisms Escherichia coli, Arabidopsis thaliana, and Saccharomyces cerevisiae, as well as human. We show that organism-specific classifiers can be robustly trained if at least 20 samples are available for each substrate class. If information from position-specific scoring matrices is included, such classifiers have F1 scores between 0.85 and 1.00. For the largest data set (A. thaliana), a 4-class classifier yielded an F-score of 0.97. On a pan-organism data set composed of transporters of all four organisms, amino acid and sugar transporters were predicted with an F1 score of 0.91.

Unspecific CTL Killing Is Enhanced by High Glucose via TNF-Related Apoptosis-Inducing Ligand.

TNF-related apoptosis inducing ligand (TRAIL) is expressed on cytotoxic T lymphocytes (CTLs) and TRAIL is linked to progression of diabetes. However, the impact of high glucose on TRAIL expression and its related killing function in CTLs still remains largely elusive. Here, we report that TRAIL is substantially up-regulated in CTLs in environments with high glucose (HG) both in vitro and in vivo. Non-mitochondrial reactive oxygen species, NFκB and PI3K/Akt are essential in HG-induced TRAIL upregulation in CTLs. TRAILhigh CTLs induce apoptosis of pancreatic beta cell line 1.4E7. Treatment with metformin and vitamin D reduces HG-enhanced expression of TRAIL in CTLs and coherently protects 1.4E7 cells from TRAIL-mediated apoptosis. Our work suggests that HG-induced TRAILhigh CTLs might contribute to the destruction of pancreatic beta cells in a hyperglycemia condition.

Protein Signatures of NK Cell-Mediated Melanoma Killing Predict Response to Immunotherapies.

Despite impressive advances in melanoma-directed immunotherapies, resistance is common and many patients still succumb to metastatic disease. In this context, harnessing natural killer (NK) cells, which have thus far been sidelined in the development of melanoma immunotherapy, could provide therapeutic benefits for cancer treatment. To identify molecular determinants of NK cell-mediated melanoma killing (NKmK), we quantified NK-cell cytotoxicity against a panel of genetically diverse melanoma cell lines and observed highly heterogeneous susceptibility. Melanoma protein microarrays revealed a correlation between NKmK and the abundance and activity of a subset of proteins, including several metabolic factors. Oxidative phoshorylation, measured by oxygen consumption rate, negatively correlated with melanoma cell sensitivity toward NKmK, and proteins involved in mitochondrial metabolism and epithelial-mesenchymal transition were confirmed to regulate NKmK. Two- and three-dimensional killing assays and melanoma xenografts established that the PI3K/AKT/mTOR signaling axis controls NKmK via regulation of NK cell-relevant surface proteins. A "protein-killing-signature" based on the protein analysis predicted NKmK of additional melanoma cell lines and the response of patients with melanoma to anti-PD-1 checkpoint therapy. Collectively, these findings identify novel NK cell-related prognostic biomarkers and may contribute to improved and personalized melanoma-directed immunotherapies. SIGNIFICANCE: NK-cell cytotoxicity assays and protein microarrays reveal novel biomarkers of NK cell-mediated melanoma killing and enable development of signatures to predict melanoma patient responsiveness to immunotherapies.