Structural templates predict novel protein interactions and targets from pancreas tumour gene expression data

Biological Networks for Cancer Candidate Biomarkers Discovery

Due to its extraordinary heterogeneity and complexity, cancer is often proposed as a model case of a systems biology disease or network disease. There is a critical need of effective biomarkers for cancer diagnosis and/or outcome prediction from system level analyses. Methods based on integrating omics data into networks have the potential to revolutionize the identification of cancer biomarkers. Deciphering the biological networks underlying cancer is undoubtedly important for understanding the molecular mechanisms of the disease and identifying effective biomarkers. In this review, the networks constructed for cancer biomarker discovery based on different omics level data are described and illustrated from recent advances in the field.

Weighted gene co-expression network analysis reveals key genes involved in pancreatic ductal adenocarcinoma development

PURPOSE

Pancreatic ductal adenocarcinoma (PDAC) is a highly aggressive malignancy. Up till now, the patient's prognosis remains poor which, among others, is due to the paucity of reliable early diagnostic biomarkers. In the past, candidate diagnostic biomarkers and therapeutic targets have been delineated from genes that were found to be differentially expressed in normal versus tumour samples. Recently, new systems biology approaches have been developed to analyse gene expression data, which may yield new biomarkers. As of yet, the weighted gene co-expression network analysis (WGCNA) tool has not been applied to PDAC microarray-based gene expression data.

METHODS

PDAC microarray-based gene expression datasets, listed in the Gene Expression Omnibus (GEO) database, were analysed. After pre-processing of the data, we built two final datasets, Normal and PDAC, encompassing 104 and 129 patient samples, respectively. Next, we constructed a weighted gene co-expression network and identified modules of co-expressed genes distinguishing normal from disease conditions. Functional annotations of the genes in these modules were carried out to highlight PDAC-associated molecular pathways and common regulatory mechanisms. Finally, overall survival analyses were carried out to assess the suitability of the genes identified as prognostic biomarkers.

RESULTS

Using WGCNA, we identified several key genes that may play important roles in PDAC. These genes are mainly related to either endoplasmic reticulum, mitochondrion or membrane functions, exhibit transferase or hydrolase activities and are involved in biological processes such as lipid metabolism or transmembrane transport. As a validation of the applied method, we found that some of the identified key genes (CEACAM1, MCU, VDAC1, CYCS, C15ORF52, TMEM51, LARP1 and ERLIN2) have previously been reported by others as potential PDAC biomarkers. Using overall survival analyses, we found that several of the newly identified genes may serve as biomarkers to stratify PDAC patients into low- and high-risk groups.

CONCLUSIONS

Using this new systems biology approach, we identified several genes that appear to be critical to PDAC development. As such, they may represent potential diagnostic biomarkers as well as therapeutic targets with clinical utility.

Neutrophil-Derived Proteases in the Microenvironment of Pancreatic Cancer -Active Players in Tumor Progression

A hallmark of pancreatic ductal adenocarcinoma (PDAC) is the fibro-inflammatory microenvironment, consisting of activated pancreatic stellate cells, extracellular matrix proteins, and a variety of inflammatory cells, such as T cells, macrophages, or neutrophils. Tumor-infiltrating immune cells, which are found in nearly all cancers, including PDAC, often fail to eliminate the tumor, but conversely can promote its progression by altering the tumor microenvironment. Pancreatic cancer cells are able to attract polymorphonuclear neutrophils (PMN) via tumor secreted chemokines and in human PDAC, PMN infiltrates can be observed in the vicinity of tumor cells and in the desmoplastic tumor stroma, which correlate with undifferentiated tumor growth and poor prognosis. The behavior of tumor-infiltrating neutrophils in the tumor micromilieu is not yet understood at a mechanistic level. It has been shown that PMN have the potential to kill tumor cells, either directly or by antibody-dependent cell-mediated cytotoxicity, but on the other side various adverse effects of PMN, such as promotion of aggressive tumor growth with epithelial-to-mesenchymal transition and increased metastatic potential, have been described. Recent therapeutic approaches for PDAC focus not only the tumor cell itself, but also elements of the tumor microenvironment. Therefore, the role of PMN and their derived products (e.g. cytokines, proteases) as a new vein for a therapeutic target should be critically evaluated in this context. This review summarizes the current understanding of the interplay between proteases of tumor-infiltrating neutrophils and pancreatic tumor cells and elements of the desmoplastic stroma.

Novel therapeutic targets for pancreatic cancer

Pancreatic cancer has become the fourth leading cause of cancer death in the last two decades. Only 3%-15% of patients diagnosed with pancreatic cancer had 5 year survival rate. Drug resistance, high metastasis, poor prognosis and tumour relapse contributed to the malignancies and difficulties in treating pancreatic cancer. The current standard chemotherapy for pancreatic cancer is gemcitabine, however its efficacy is far from satisfactory, one of the reasons is due to the complex tumour microenvironment which decreases effective drug delivery to target cancer cell. Studies of the molecular pathology of pancreatic cancer have revealed that activation of KRAS, overexpression of cyclooxygenase-2, inactivation of p16^INK4A and loss of p53 activities occurred in pancreatic cancer. Co-administration of gemcitabine and targeting the molecular pathological events happened in pancreatic cancer has brought an enhanced therapeutic effectiveness of gemcitabine. Therefore, studies looking for novel targets in hindering pancreatic tumour growth are emerging rapidly. In order to give a better understanding of the current findings and to seek the direction in future pancreatic cancer research; in this review we will focus on targets suppressing tumour metastatsis and progression, KRAS activated downstream effectors, the relationship of Notch signaling and Nodal/Activin signaling with pancreatic cancer cells, the current findings of non-coding RNAs in inhibiting pancreatic cancer cell proliferation, brief discussion in transcription remodeling by epigenetic modifiers (e.g., HDAC, BMI1, EZH2) and the plausible therapeutic applications of cancer stem cell and hyaluronan in tumour environment.

Cancer networks and beyond: interpreting mutations using the human interactome and protein structure

Over recent years, with the advances in next-generation sequencing, a large number of cancer mutations have been identified and accumulated in public repositories. Coupled to this is our increased ability to generate detailed interactome maps that help to enrich our knowledge of the biological implications of cancer mutations. As a result, network analysis approaches have become an invaluable tool to predict and interpret mutations that are associated with tumour survival and progression. Our understanding of cancer mechanisms is further enhanced by mapping protein structure information to such networks. Here we review the current methodologies for annotating the functional impacts of cancer mutations, which range from analysis of protein structures to protein-protein interaction network studies.

Defining the molecular signature of chemotherapy-mediated lung tumor phenotype modulation and increased susceptibility to T-cell killing

Chemotherapy with platinum doublets, including cisplatin plus vinorelbine, is standard of care for non-small-cell lung cancer. Sublethal exposure to certain chemotherapeutic agents has been demonstrated to alter the phenotype or biology of human tumor cells, rendering them more susceptible to cytotoxic T lymphocyte (CTL)-mediated lysis. The effects of cisplatin/vinorelbine on tumor sensitivity to T-cell cytotoxicity and its molecular mechanisms, however, have not been fully elucidated. We examined the effect of this chemotherapy on growth, cell-surface phenotype, and CTL-mediated lysis of five distinct human lung carcinoma cell lines in vitro and examined the molecular mechanisms associated with enhanced CTL sensitivity. These studies demonstrate that sublethal exposure of human lung tumor cells to the platinum doublet modulates tumor cell phenotype and increases sensitivity to major histocompatibility complex-restricted perforin/granzyme-mediated CTL killing. These studies also demonstrate that exposure to chemotherapy markedly decreased the protein secretion ratio of transforming growth factor-β/interleukin (IL)-8. We examined the gene expression profile of two lung tumor cell lines to identify a shared gene signature in response to sublethal cisplatin/vinorelbine and found coordinate expression of only 16 transcripts, including those for cytokine/chemokine expression and apoptosis such as tumor necrosis factor-α, IL8, CXCL5, and B cell lymphoma-2-like genes (BCL-2). Overall, these results suggest that sublethal exposure to cisplatin/vinorelbine increases sensitivity to perforin/granzyme-mediated CTL killing by modulation of (a) tumor phenotype, (b) cytokine/chemokine milieu, and (c) the proapoptotic/antiapoptotic gene ratio. The data presented here propose a complex mechanism that is distinct from and complementary to that of immunogenic cell death. This molecular signature may be useful in predicting responses to immunotherapy as well as provide the rationale for the potential clinical benefit of the combined use of vaccine with cisplatin/vinorelbine regimens.

Simultaneous gene silencing of Bcl-2, XIAP and Survivin re-sensitizes pancreatic cancer cells towards apoptosis

BACKGROUND

Pancreatic ductal adenocarcinoma shows a distinct apoptosis resistance, which contributes significantly to the aggressive nature of this tumor and constrains the effectiveness of new therapeutic strategies. Apoptosis resistance is determined by the net balance of the cells pro-and anti-apoptotic "control mechanisms". Numerous dysregulated anti-apoptotic genes have been identified in pancreatic cancer and seem to contribute to the high anti-apoptotic buffering capacity. We aimed to compare the benefit of simultaneous gene silencing (SGS) of several candidate genes with conventional gene silencing of single genes.

METHODS

From literature search we identified the anti-apoptotic genes XIAP, Survivin and Bcl-2 as commonly upregulated in pancreatic cancer. We performed SGS and silencing of single candidate genes using siRNA molecules in two pancreatic cancer cell lines. Effectiveness of SGS was assessed by qRT-PCR and western blotting. Apoptosis induction was measured by flow cytometry and caspase activation.

RESULTS

Simultaneous gene silencing reduced expression of the three target genes effectively. Compared to silencing of a single target or control, SGS of these genes resulted in a significant higher induction of apoptosis in pancreatic cancer cells.

CONCLUSIONS

In the present study we performed a subliminal silencing of different anti-apoptotic target genes simultaneously. Compared to silencing of single target genes, SGS had a significant higher impact on apoptosis induction in pancreatic cancer cells. Thereby, we give further evidence for the concept of an anti-apoptotic buffering capacity of pancreatic cancer cells.

Human cancer protein-protein interaction network: a structural perspective

Protein-protein interaction networks provide a global picture of cellular function and biological processes. Some proteins act as hub proteins, highly connected to others, whereas some others have few interactions. The dysfunction of some interactions causes many diseases, including cancer. Proteins interact through their interfaces. Therefore, studying the interface properties of cancer-related proteins will help explain their role in the interaction networks. Similar or overlapping binding sites should be used repeatedly in single interface hub proteins, making them promiscuous. Alternatively, multi-interface hub proteins make use of several distinct binding sites to bind to different partners. We propose a methodology to integrate protein interfaces into cancer interaction networks (ciSPIN, cancer structural protein interface network). The interactions in the human protein interaction network are replaced by interfaces, coming from either known or predicted complexes. We provide a detailed analysis of cancer related human protein-protein interfaces and the topological properties of the cancer network. The results reveal that cancer-related proteins have smaller, more planar, more charged and less hydrophobic binding sites than non-cancer proteins, which may indicate low affinity and high specificity of the cancer-related interactions. We also classified the genes in ciSPIN according to phenotypes. Within phenotypes, for breast cancer, colorectal cancer and leukemia, interface properties were found to be discriminating from non-cancer interfaces with an accuracy of 71%, 67%, 61%, respectively. In addition, cancer-related proteins tend to interact with their partners through distinct interfaces, corresponding mostly to multi-interface hubs, which comprise 56% of cancer-related proteins, and constituting the nodes with higher essentiality in the network (76%). We illustrate the interface related affinity properties of two cancer-related hub proteins: Erbb3, a multi interface, and Raf1, a single interface hub. The results reveal that affinity of interactions of the multi-interface hub tends to be higher than that of the single-interface hub. These findings might be important in obtaining new targets in cancer as well as finding the details of specific binding regions of putative cancer drug candidates.

Protein-protein interaction networks provide a global picture of cellular function and biological processes. The dysfunction of some interactions causes many diseases, including cancer. Proteins interact through their interfaces. Therefore, studying the interface properties of cancer-related proteins will help explain their role in the interaction networks. The structural details of interfaces are immensely useful in efforts to answer some fundamental questions such as: (i) what features of cancer-related protein interfaces make them act as hubs; (ii) how hub protein interfaces can interact with tens of other proteins with varying affinities; and (iii) which interactions can occur simultaneously and which are mutually exclusive. Addressing these questions, we propose a method to characterize interactions in a human protein-protein interaction network using three-dimensional protein structures and interfaces. Protein interface analysis shows that the strength and specificity of the interactions of hub proteins and cancer proteins are different than the interactions of non-hub and non-cancer proteins, respectively. In addition, distinguishing overlapping from non-overlapping interfaces, we illustrate how a fourth dimension, that of the sequence of processes, is integrated into the network with case studies. We believe that such an approach should be useful in structural systems biology.

Towards inferring time dimensionality in protein-protein interaction networks by integrating structures: the p53 example

Inspection of protein-protein interaction maps illustrates that a hub protein can interact with a very large number of proteins, reaching tens and even hundreds. Since a single protein cannot interact with such a large number of partners at the same time, this presents a challenge: can we figure out which interactions can occur simultaneously and which are mutually excluded? Addressing this question adds a fourth dimension into interaction maps: that of time. Including the time dimension in structural networks is an immense asset; time dimensionality transforms network node-and-edge maps into cellular processes, assisting in the comprehension of cellular pathways and their regulation. While the time dimensionality can be further enhanced by linking protein complexes to time series of mRNA expression data, current robust, network experimental data are lacking. Here we outline how, using structural data, efficient structural comparison algorithms and appropriate datasets and filters can assist in getting an insight into time dimensionality in interaction networks; in predicting which interactions can and cannot co-exist; and in obtaining concrete predictions consistent with experiment. As an example, we present p53-linked processes.

Inhibition of homologous recombination by treatment with BVDU (brivudin) or by RAD51 silencing increases chromosomal damage induced by bleomycin in mismatch repair-deficient tumour cells

Mismatch repair (MMR) has been shown to control homologous recombination (HR) by aborting strand exchange between divergent sequences. We previously demonstrated that MMR-deficient tumour cells are more resistant to chromosomal damage induced by bleomycin (BLM) during the G(2) phase, likely due to the lack of the MMR inhibitory effect on HR. Aim of this study was to investigate whether inhibition of HR by the nucleoside analogue BVDU [(E)-5(2-bromovinyl)-2'-deoxyuridine, brivudin], or silencing of genes involved in HR function, might affect sensitivity of MMR-deficient tumour cells to DNA damage induced by BLM in G(2). The results indicated that BVDU increased chromatid damage and DNA double strand breaks induced by BLM only in MMR-deficient MT-1, HL-60R, HCT116 cells, which are more resistant to BLM with respect to MMR-proficient TK-6, HL-60S and HCT116/3-6 lines. Silencing of RAD51, a key component of HR, increased sensitivity of MMR-deficient HCT-15 cells to BLM clastogenicity; in this case combined treatment with BVDU had no additional effect. Similarly, treatment with BVDU did not affect BLM clastogenicity in CAPAN-1 cells, characterized by a defective HR due to BRCA2 mutations. Conversely, BVDU increased chromatid breaks induced by BLM in HCT-15 cells transiently silenced for DNA-PK catalytic subunit, which plays a key role in non-homologous end joining. The BVDU-mediated increase of chromatid breaks in MMR-deficient cells did not depend on its previously reported inhibitory effect on poly(ADP-ribose) polymerase (PARP). In fact, it was observed also in cells stably silenced for PARP-1, which is responsible for most of cellular PARP activity. These data support the suggestion that the higher sensitivity of MMR-proficient versus MMR-deficient cells to BLM-induced chromatid breaks in the G(2) phase is a consequence of the inhibition of HR by MMR. In MMR-deficient cells, BVDU attenuates the repair of BLM-induced DSBs and this is likely to occur via inhibition of HR.

Co-expression of KLK6 and KLK10 as prognostic factors for survival in pancreatic ductal adenocarcinoma

Kallikreins play an important role in tumour microenvironment and as cancer biomarkers in different cancer entities. Previous studies suggested an upregulation of KLK10 and KLK6 in pancreatic ductal adenocarcinoma (PDAC). Therefore, we evaluated the clinicopathological role of these kallikreins and their value as biomarkers in PDAC.

Differential expression was validated by DNA-microarrays and immunohistochemistry in normal and malignant pancreatic tissues. Sera concentrations of both kallikreins were evaluated using ELISA. In silico analysis of possible protein interactions and gene silencing of KLK10 in vitro using siRNAs gave further insights in the pathomechanisms.

Gene expression analysis and immunohistochemistry demonstrated a strong expression for KLK10 and KLK6 in PDAC. Statistical analysis showed that co-expression of these kallikreins correlated with an R1-resection status (P=0.017) and worse outcome for overall survival (P=0.031). Multivariate analysis proofed that co-expression is an independent prognostic factor for survival (P=0.043). Importantly, KLK10 knockdown in AsPC-1 cells significantly reduced cell migration, whereas computational analysis suggested interaction of KLK6 with angiogenetic factors as an important mechanism.

Co-expression of KLK10 and KLK6 plays an unfavourable role in PDAC. Our results suggest that this effect is likely mediated by an interaction with the factors of the extracellular matrix and enhancement of cancer cell motility.