Modeling signal transduction from protein phosphorylation to gene expression

A Novel Bayesian Framework Infers Driver Activation States and Reveals Pathway-Oriented Molecular Subtypes in Head and Neck Cancer

Head and neck squamous cell cancer (HNSCC) is an aggressive cancer resulting from heterogeneous causes. To reveal the underlying drivers and signaling mechanisms of different HNSCC tumors, we developed a novel Bayesian framework to identify drivers of individual tumors and infer the states of driver proteins in cellular signaling system in HNSCC tumors. First, we systematically identify causal relationships between somatic genome alterations (SGAs) and differentially expressed genes (DEGs) for each TCGA HNSCC tumor using the tumor-specific causal inference (TCI) model. Then, we generalize the most statistically significant driver SGAs and their regulated DEGs in TCGA HNSCC cohort. Finally, we develop machine learning models that combine genomic and transcriptomic data to infer the protein functional activation states of driver SGAs in tumors, which enable us to represent a tumor in the space of cellular signaling systems. We discovered four mechanism-oriented subtypes of HNSCC, which show distinguished patterns of activation state of HNSCC driver proteins, and importantly, this subtyping is orthogonal to previously reported transcriptomic-based molecular subtyping of HNSCC. Further, our analysis revealed driver proteins that are likely involved in oncogenic processes induced by HPV infection, even though they are not perturbed by genomic alterations in HPV+ tumors.

Small peptide substrates and high resolution peptide gels for the analysis of site-specific protein phosphorylation and dephosphorylation

Protein phosphorylation and dephosphorylation reactions play key regulatory roles in many fundamental cellular processes. Due to the large number of kinases and phosphatases in the genome, the identification of the specific enzymes responsible for a given site in a given protein is immensely challenging. However, because protein kinases and phosphatases recognize local specificity determinants within proteins, it is possible to use small peptides to study the characteristics of site-specific phosphorylation. In addition, phosphorylation usually causes retardation in gel mobility, providing an opportunity to investigate peptide phosphorylation and dephosphorylation by monitoring migration on high resolution peptide gels. In this study, we demonstrate the utility of such a technique using small peptides corresponding to cyclin-dependent kinase-1 (Cdk1)/cyclin B1 sites in two important apoptotic regulatory proteins, Bcl-xL and caspase-9. We show that the mobility of the peptides is retarded following Cdk1-mediated phosphorylation, and that peptide dephosphorylation, catalyzed either by purified phosphatase or by crude cell extracts, is readily observable by increased peptide gel mobility. Furthermore, the procedure can be conducted without the use of radioactive adenosine triphosphate (ATP), and does not require any specialized reagents or apparatus. The method can be used to identify and characterize specific kinase and phosphatases responsible for phosphorylation and dephosphorylation of specific sites in any protein of interest.

Interleukin-26 Production in Human Primary Bronchial Epithelial Cells in Response to Viral Stimulation: Modulation by Th17 cytokines

Interleukin (IL)-26 is abundant in human airways and this cytokine is involved in the local immune response to a bacterial stimulus in vivo. Specifically, local exposure to the toll-like receptor (TLR) 4 agonist endotoxin does increase IL-26 in human airways and this cytokine potentiates chemotactic responses in human neutrophils. In addition to T-helper (Th) 17 cells, alveolar macrophages can produce IL-26, but it remains unknown whether this cytokine can also be produced in the airway mucosa per se in response to a viral stimulus. Here, we evaluated whether this is the case using primary bronchial epithelial cells from the airway epithelium in vitro, and exploring the signaling mechanisms involved, including the modulatory effects of additional Th17 cytokines. Finally, we assessed IL-26 and its archetype signaling responses in healthy human airways in vivo. We found increased transcription and release of IL-26 protein after stimulation with the viral-related double stranded (ds) RNA polyinosinic-polycytidylic acid (poly-IC) and showed that this IL-26 release involved mitogen-activated protein (MAP) kinases and nuclear factor kappa-light-chain-enhancer of activated B cells (NF-κB). The release of IL-26 in response to a viral stimulus was modulated by additional Th17 cytokines. Moreover, there was transcription of IL26 mRNA and expression of the protein in epithelial cells of bronchial brush and tissue biopsies respectively after harvest in vivo. In addition, the extracellular IL-26 protein concentrations in bronchoalveolar lavage (BAL) samples did correlate with increased epithelial cell transcription of an archetype intracellular signaling molecule downstream of the IL-26-receptor complex, STAT1, in the bronchial brush biopsies. Thus, our study suggests that viral stimulation causes the production of IL-26 in lining epithelial cells of human airway structural cells that constitute a critical immune barrier and that this production is modulated by Th17 cytokines.

Knowledge-guided fuzzy logic modeling to infer cellular signaling networks from proteomic data

Modeling of signaling pathways is crucial for understanding and predicting cellular responses to drug treatments. However, canonical signaling pathways curated from literature are seldom context-specific and thus can hardly predict cell type-specific response to external perturbations; purely data-driven methods also have drawbacks such as limited biological interpretability. Therefore, hybrid methods that can integrate prior knowledge and real data for network inference are highly desirable. In this paper, we propose a knowledge-guided fuzzy logic network model to infer signaling pathways by exploiting both prior knowledge and time-series data. In particular, the dynamic time warping algorithm is employed to measure the goodness of fit between experimental and predicted data, so that our method can model temporally-ordered experimental observations. We evaluated the proposed method on a synthetic dataset and two real phosphoproteomic datasets. The experimental results demonstrate that our model can uncover drug-induced alterations in signaling pathways in cancer cells. Compared with existing hybrid models, our method can model feedback loops so that the dynamical mechanisms of signaling networks can be uncovered from time-series data. By calibrating generic models of signaling pathways against real data, our method supports precise predictions of context-specific anticancer drug effects, which is an important step towards precision medicine.