Integrative Modeling Reveals Annexin A2-mediated Epigenetic Control of Mesenchymal Glioblastoma

Glioblastomas are characterized by transcriptionally distinct subtypes, but despite possible clinical relevance, their regulation remains poorly understood. The commonly used molecular classification systems for GBM all identify a subtype with high expression of mesenchymal marker transcripts, strongly associated with invasive growth. We used a comprehensive data-driven network modeling technique (augmented sparse inverse covariance selection, aSICS) to define separate genomic, epigenetic, and transcriptional regulators of glioblastoma subtypes. Our model identified Annexin A2 (ANXA2) as a novel methylation-controlled positive regulator of the mesenchymal subtype. Subsequent evaluation in two independent cohorts established ANXA2 expression as a prognostic factor that is dependent on ANXA2 promoter methylation. ANXA2 knockdown in primary glioblastoma stem cell-like cultures suppressed known mesenchymal master regulators, and abrogated cell proliferation and invasion. Our results place ANXA2 at the apex of a regulatory cascade that determines glioblastoma mesenchymal transformation and validate aSICS as a general methodology to uncover regulators of cancer subtypes.

•

Glioblastoma, a form of brain cancer, is characterised by distinct molecular subtypes: proneural, classical and mesenchymal.

•

We used a comprehensive data-driven strategy, aSICS, to elucidate the cellular mechanisms behind the subtypes.

•

Epigenetic control of Annexin A2 (ANXA2) was predicted and confirmed to determine the invasive mesenchymal subtype.

Most cancers have distinct and clinically relevant transcriptional subtypes, but the underlying cellular mechanism behind such subtypes is often hard to resolve. We show that joint analysis across several layers of genomics data can uncover subtype regulators with good accuracy. Our method is applied to the brain cancer glioblastoma multiforme (GBM), revealing that the invasive mesenchymal subtype is driven by epigenetic modulation of the expression of Annexin A2 (ANXA2). Our analysis adds significantly to our understanding of brain cancer subtypes and open for new potential treatment options. The proposed computational technique can be applied to other cancers as well.

Safe and Effective Treatment of Experimental Neuroblastoma and Glioblastoma Using Systemically Delivered Triple MicroRNA-Detargeted Oncolytic Semliki Forest Virus

Background: Glioblastoma multiforme and high-risk neuroblastoma are cancers with poor outcome. Immunotherapy in the form of neurotropic oncolytic viruses is a promising therapeutic approach for these malignancies. Here we evaluate the oncolytic capacity of the neurovirulent and partly IFNβ-resistant Semliki Forest virus (SFV)-4 in glioblastoma multiformes and neuroblastomas. To reduce neurovirulence we constructed SFV4miRT, which is attenuated in normal central nervous system (CNS) cells through insertion of microRNA target sequences for miR124, miR125, miR134.Methods: Oncolytic activity of SFV4miRT was examined in mouse neuroblastoma and glioblastoma multiforme cell lines and in patient-derived human glioblastoma cell cultures (HGCC). In vivo neurovirulence and therapeutic efficacy was evaluated in two syngeneic orthotopic glioma models (CT-2A, GL261) and a syngeneic subcutaneous neuroblastoma model (NXS2). The role of IFNβ in inhibiting therapeutic efficacy was investigated.Results: The introduction of miRNA target sequences reduced neurovirulence of SFV4 in terms of attenuated replication in mouse CNS cells and ability to cause encephalitis when administered intravenously. A single intravenous injection of SFV4miRT prolonged survival and cured four of eight mice (50%) with NXS2 and three of 11 mice (27%) with CT-2A, but not for GL261 tumor-bearing mice. In vivo therapeutic efficacy in different tumor models inversely correlated to secretion of IFNβ by respective cells upon SFV4 infection in vitro Similarly, killing efficacy of HGCC lines inversely correlated to IFNβ response and interferon-α/β receptor-1 expression.Conclusions: SFV4miRT has reduced neurovirulence, while retaining its oncolytic capacity. SFV4miRT is an excellent candidate for treatment of glioblastoma multiforme and neuroblastoma with low IFN-β secretion. Clin Cancer Res; 23(6); 1519-30. ©2016 AACR.

High sensitivity isoelectric focusing to establish a signaling biomarker for the diagnosis of human colorectal cancer

BACKGROUND

The progression of colorectal cancer (CRC) involves recurrent amplifications/mutations in the epidermal growth factor receptor (EGFR) and downstream signal transducers of the Ras pathway, KRAS and BRAF. Whether genetic events predicted to result in increased and constitutive signaling indeed lead to enhanced biological activity is often unclear and, due to technical challenges, unexplored. Here, we investigated proliferative signaling in CRC using a highly sensitive method for protein detection. The aim of the study was to determine whether multiple changes in proliferative signaling in CRC could be combined and exploited as a "complex biomarker" for diagnostic purposes.

METHODS

We used robotized capillary isoelectric focusing as well as conventional immunoblotting for the comprehensive analysis of epidermal growth factor receptor signaling pathways converging on extracellular regulated kinase 1/2 (ERK1/2), AKT, phospholipase Cγ1 (PLCγ1) and c-SRC in normal mucosa compared with CRC stage II and IV. Computational analyses were used to test different activity patterns for the analyzed signal transducers.

RESULTS

Signaling pathways implicated in cell proliferation were differently dysregulated in CRC and, unexpectedly, several were downregulated in disease. Thus, levels of activated ERK1 (pERK1), but not pERK2, decreased in stage II and IV while total ERK1/2 expression remained unaffected. In addition, c-SRC expression was lower in CRC compared with normal tissues and phosphorylation on the activating residue Y418 was not detected. In contrast, PLCγ1 and AKT expression levels were elevated in disease. Immunoblotting of the different signal transducers, run in parallel to capillary isoelectric focusing, showed higher variability and lower sensitivity and resolution. Computational analyses showed that, while individual signaling changes lacked predictive power, using the combination of changes in three signaling components to create a "complex biomarker" allowed with very high accuracy, the correct diagnosis of tissues as either normal or cancerous.

CONCLUSIONS

We present techniques that allow rapid and sensitive determination of cancer signaling that can be used to differentiate colorectal cancer from normal tissue.

Glioma-derived plasminogen activator inhibitor-1 (PAI-1) regulates the recruitment of LRP1 positive mast cells

Glioblastoma (GBM) is a high-grade glioma with a complex microenvironment, including various inflammatory cells and mast cells (MCs) as one of them. Previously we had identified glioma grade-dependent MC recruitment. In the present study we investigated the role of plasminogen activator inhibitor 1 (PAI-1) in MC recruitment.

PAI-1, a primary regulator in the fibrinolytic cascade is capable of forming a complex with fibrinolytic system proteins together with low-density lipoprotein receptor-related protein 1 (LRP1). We found that neutralizing PAI-1 attenuated infiltration of MCs. To address the potential implication of LRP1 in this process, we used a LRP1 antagonist, receptor-associated protein (RAP), and demonstrated the attenuation of MC migration. Moreover, a positive correlation between the number of MCs and the level of PAI-1 in a large cohort of human glioma samples was observed. Our study demonstrated the expression of LRP1 in human MC line LAD2 and in MCs in human high-grade glioma. The activation of potential PAI-1/LRP1 axis with purified PAI-1 promoted increased phosphorylation of STAT3 and subsequently exocytosis in MCs.

These findings indicate the influence of the PAI-1/LRP1 axis on the recruitment of MCs in glioma. The connection between high-grade glioma and MC infiltration could contribute to patient tailored therapy and improve patient stratification in future therapeutic trials.

ABCG2 regulates self-renewal and stem cell marker expression but not tumorigenicity or radiation resistance of glioma cells

Glioma cells with stem cell traits are thought to be responsible for tumor maintenance and therapeutic failure. Such cells can be enriched based on their inherent drug efflux capability mediated by the ABC transporter ABCG2 using the side population assay, and their characteristics include increased self-renewal, high stem cell marker expression and high tumorigenic capacity in vivo. Here, we show that ABCG2 can actively drive expression of stem cell markers and self-renewal in glioma cells. Stem cell markers and self-renewal was enriched in cells with high ABCG2 activity, and could be specifically inhibited by pharmacological and genetic ABCG2 inhibition. Importantly, despite regulating these key characteristics of stem-like tumor cells, ABCG2 activity did not affect radiation resistance or tumorigenicity in vivo. ABCG2 effects were Notch-independent and mediated by diverse mechanisms including the transcription factor Mef. Our data demonstrate that characteristics of tumor stem cells are separable, and highlight ABCG2 as a potential driver of glioma stemness.

Genome-wide association study identifies multiple susceptibility loci for multiple myeloma

Multiple myeloma (MM) is a plasma cell malignancy with a significant heritable basis. Genome-wide association studies have transformed our understanding of MM predisposition, but individual studies have had limited power to discover risk loci. Here we perform a meta-analysis of these GWAS, add a new GWAS and perform replication analyses resulting in 9,866 cases and 239,188 controls. We confirm all nine known risk loci and discover eight new loci at 6p22.3 (rs34229995, P=1.31 × 10(-8)), 6q21 (rs9372120, P=9.09 × 10(-15)), 7q36.1 (rs7781265, P=9.71 × 10(-9)), 8q24.21 (rs1948915, P=4.20 × 10(-11)), 9p21.3 (rs2811710, P=1.72 × 10(-13)), 10p12.1 (rs2790457, P=1.77 × 10(-8)), 16q23.1 (rs7193541, P=5.00 × 10(-12)) and 20q13.13 (rs6066835, P=1.36 × 10(-13)), which localize in or near to JARID2, ATG5, SMARCD3, CCAT1, CDKN2A, WAC, RFWD3 and PREX1. These findings provide additional support for a polygenic model of MM and insight into the biological basis of tumour development.

Simultaneous Multiplexed Measurement of RNA and Proteins in Single Cells

Significant advances have been made in methods to analyze genomes and transcriptomes of single cells, but to fully define cell states, proteins must also be accessed as central actors defining a cell's phenotype. Methods currently used to analyze endogenous protein expression in single cells are limited in specificity, throughput, or multiplex capability. Here, we present an approach to simultaneously and specifically interrogate large sets of protein and RNA targets in lysates from individual cells, enabling investigations of cell functions and responses. We applied our method to investigate the effects of BMP4, an experimental therapeutic agent, on early-passage glioblastoma cell cultures. We uncovered significant heterogeneity in responses to treatment at levels of RNA and protein, with a subset of cells reacting in a distinct manner to BMP4. Moreover, we found overall poor correlation between protein and RNA at the level of single cells, with proteins more accurately defining responses to treatment.

Abstract B2-35: Efficient exploration of multi-cancer networks by generalized covariance selection and interactive web content

Statistical network modeling techniques are increasingly important tools to analyze cancer genomics data. However, current tools for network construction and interpretation are not designed to work across multiple diagnoses and technical platforms, thus limiting their applicability to comprehensive pan-cancer datasets such as the Cancer Genome Atlas (TCGA). Here, we describe a novel strategy to construct and analyze integrative network models heterogeneous data from multiple cancers. First, we introduce a generalization of sparse inverse covariance selection (SICS) designed to integrate genetic, epigenetic and transcriptional data from multiple cancers into a comparative network. The algorithm is shown to be statistically robust, effective at detecting direct pathway links in data from The Cancer Genome Atlas (TCGA), and uses a new strategy involving non-informative priors to balance different cancers and data types. Second, we propose to rationalize the interpretation of the derived networks by a new and publicly accessible tool (cancerlandscapes.org), in which derived models are explored as interactive web content, linked to several pathway and pharmacological databases. To evaluate the performance of the method, we constructed a model of genetic, epigenetic and transcriptional data for eight TCGA cancers, using data from 3900 patients. The derived model rediscovered known mechanisms and contained interesting predictions. Possible applications include the prediction of regulatory relationships between genes in particular cancers, comparison of network modules in across multiple forms of cancer, and identification of drug targets in relation to network structure.

Citation Format: Teresia Kling, Patrik Johansson, Jose Sanchez, Voichita D. Marinescu, Rebecka Jornsten, Sven Nelander. Efficient exploration of multi-cancer networks by generalized covariance selection and interactive web content. [abstract]. In: Proceedings of the AACR Special Conference on Computational and Systems Biology of Cancer; Feb 8-11 2015; San Francisco, CA. Philadelphia (PA): AACR; Cancer Res 2015;75(22 Suppl 2):Abstract nr B2-35.

High levels of WNT-5A in human glioma correlate with increased presence of tumor-associated microglia/monocytes

Malignant gliomas are among the most severe types of cancer, and the most common primary brain tumors. Treatment options are limited and the prognosis is poor. WNT-5A, a member of the WNT family of lipoglycoproteins, plays a role in oncogenesis and tumor progression in various cancers, whereas the role of WNT-5A in glioma remains obscure. Based on the role of WNT-5A as an oncogene, its potential to regulate microglia cells and the glioma-promoting capacities of microglia cells, we hypothesize that WNT-5A has a role in regulation of immune functions in glioma. We investigated WNT-5A expression by in silico analysis of the cancer genome atlas (TCGA) transcript profiling of human glioblastoma samples and immunohistochemistry experiments of human glioma tissue microarrays (TMA). Our results reveal higher WNT-5A protein levels and mRNA expression in a subgroup of gliomas (WNT-5A(high)) compared to non-malignant control brain tissue. Furthermore, we show a significant correlation between WNT-5A in the tumor and presence of major histocompatibility complex Class II-positive microglia/monocytes. Our data pinpoint a positive correlation between WNT-5A and a proinflammatory signature in glioma. We identify increased presence of microglia/monocytes as an important aspect in the inflammatory transformation suggesting a novel role for WNT-5A in human glioma.

The Human Glioblastoma Cell Culture Resource: Validated Cell Models Representing All Molecular Subtypes

Glioblastoma (GBM) is the most frequent and malignant form of primary brain tumor. GBM is essentially incurable and its resistance to therapy is attributed to a subpopulation of cells called glioma stem cells (GSCs). To meet the present shortage of relevant GBM cell (GC) lines we developed a library of annotated and validated cell lines derived from surgical samples of GBM patients, maintained under conditions to preserve GSC characteristics. This collection, which we call the Human Glioblastoma Cell Culture (HGCC) resource, consists of a biobank of 48 GC lines and an associated database containing high-resolution molecular data. We demonstrate that the HGCC lines are tumorigenic, harbor genomic lesions characteristic of GBMs, and represent all four transcriptional subtypes. The HGCC panel provides an open resource for in vitro and in vivo modeling of a large part of GBM diversity useful to both basic and translational GBM research.

•

The HGCC resource contains 48 annotated human GBM cell lines and an interactive database

•

The GBM cell lines are propagated in stem cell conditions and display GSC characteristics

•

The HGCC resource provides cell lines of all molecular (TCGA) subtypes

•

All data connected with the HGCC cell lines can be accessed at hgcc.se

Variants in ELL2 influencing immunoglobulin levels associate with multiple myeloma

Multiple myeloma (MM) is characterized by an uninhibited, clonal growth of plasma cells. While first-degree relatives of patients with MM show an increased risk of MM, the genetic basis of inherited MM susceptibility is incompletely understood. Here we report a genome-wide association study in the Nordic region identifying a novel MM risk locus at ELL2 (rs56219066T; odds ratio (OR)=1.25; P=9.6 × 10(-10)). This gene encodes a stoichiometrically limiting component of the super-elongation complex that drives secretory-specific immunoglobulin mRNA production and transcriptional regulation in plasma cells. We find that the MM risk allele harbours a Thr298Ala missense variant in an ELL2 domain required for transcription elongation. Consistent with a hypomorphic effect, we find that the MM risk allele also associates with reduced levels of immunoglobulin A (IgA) and G (IgG) in healthy subjects (P=8.6 × 10(-9) and P=6.4 × 10(-3), respectively) and, potentially, with an increased risk of bacterial meningitis (OR=1.30; P=0.0024).

Efficient exploration of pan-cancer networks by generalized covariance selection and interactive web content

Statistical network modeling techniques are increasingly important tools to analyze cancer genomics data. However, current tools and resources are not designed to work across multiple diagnoses and technical platforms, thus limiting their applicability to comprehensive pan-cancer datasets such as The Cancer Genome Atlas (TCGA). To address this, we describe a new data driven modeling method, based on generalized Sparse Inverse Covariance Selection (SICS). The method integrates genetic, epigenetic and transcriptional data from multiple cancers, to define links that are present in multiple cancers, a subset of cancers, or a single cancer. It is shown to be statistically robust and effective at detecting direct pathway links in data from TCGA. To facilitate interpretation of the results, we introduce a publicly accessible tool (cancerlandscapes.org), in which the derived networks are explored as interactive web content, linked to several pathway and pharmacological databases. To evaluate the performance of the method, we constructed a model for eight TCGA cancers, using data from 3900 patients. The model rediscovered known mechanisms and contained interesting predictions. Possible applications include prediction of regulatory relationships, comparison of network modules across multiple forms of cancer and identification of drug targets.

Avoiding pitfalls in L1-regularised inference of gene networks

L₁ regularisation methods fail to infer the correct network even when the data are so informative that all existing links can be proven to exist.

Selective calcium sensitivity in immature glioma cancer stem cells

Tumor-initiating cells are a subpopulation in aggressive cancers that exhibit traits shared with stem cells, including the ability to self-renew and differentiate, commonly referred to as stemness. In addition, such cells are resistant to chemo- and radiation therapy posing a therapeutic challenge. To uncover stemness-associated functions in glioma-initiating cells (GICs), transcriptome profiles were compared to neural stem cells (NSCs) and gene ontology analysis identified an enrichment of Ca²⁺ signaling genes in NSCs and the more stem-like (NSC-proximal) GICs. Functional analysis in a set of different GIC lines regarding sensitivity to disturbed homeostasis using A23187 and Thapsigargin, revealed that NSC-proximal GICs were more sensitive, corroborating the transcriptome data. Furthermore, Ca²⁺ drug sensitivity was reduced in GICs after differentiation, with most potent effect in the NSC-proximal GIC, supporting a stemness-associated Ca²⁺ sensitivity. NSCs and the NSC-proximal GIC line expressed a larger number of ion channels permeable to potassium, sodium and Ca²⁺. Conversely, a higher number of and higher expression levels of Ca²⁺ binding genes that may buffer Ca²⁺, were expressed in NSC-distal GICs. In particular, expression of the AMPA glutamate receptor subunit GRIA1, was found to associate with Ca²⁺ sensitive NSC-proximal GICs, and decreased as GICs differentiated along with reduced Ca²⁺ drug sensitivity. The correlation between high expression of Ca²⁺ channels (such as GRIA1) and sensitivity to Ca²⁺ drugs was confirmed in an additional nine novel GIC lines. Calcium drug sensitivity also correlated with expression of the NSC markers nestin (NES) and FABP7 (BLBP, brain lipid-binding protein) in this extended analysis. In summary, NSC-associated NES⁺/FABP7⁺/GRIA1⁺ GICs were selectively sensitive to disturbances in Ca²⁺ homeostasis, providing a potential target mechanism for eradication of an immature population of malignant cells.

Functional association networks as priors for gene regulatory network inference

Motivation: Gene regulatory network (GRN) inference reveals the influences genes have on one another in cellular regulatory systems. If the experimental data are inadequate for reliable inference of the network, informative priors have been shown to improve the accuracy of inferences.

Results: This study explores the potential of undirected, confidence-weighted networks, such as those in functional association databases, as a prior source for GRN inference. Such networks often erroneously indicate symmetric interaction between genes and may contain mostly correlation-based interaction information. Despite these drawbacks, our testing on synthetic datasets indicates that even noisy priors reflect some causal information that can improve GRN inference accuracy. Our analysis on yeast data indicates that using the functional association databases FunCoup and STRING as priors can give a small improvement in GRN inference accuracy with biological data.

Contact:matthew.studham@scilifelab.se

Supplementary information:Supplementary data are available at Bioinformatics online.

RETRACTED: Vulnerability of glioblastoma cells to catastrophic vacuolization and death induced by a small molecule

Glioblastoma multiforme (GBM) is the most aggressive form of brain cancer with marginal life expectancy. Based on the assumption that GBM cells gain functions not necessarily involved in the cancerous process, patient-derived glioblastoma cells (GCs) were screened to identify cellular processes amenable for development of targeted treatments. The quinine-derivative NSC13316 reliably and selectively compromised viability. Synthetic chemical expansion reveals delicate structure-activity relationship and analogs with increased potency, termed Vacquinols. Vacquinols stimulate death by membrane ruffling, cell rounding, massive macropinocytic vacuole accumulation, ATP depletion, and cytoplasmic membrane rupture of GCs. The MAP kinase MKK4, identified by a shRNA screen, represents a critical signaling node. Vacquinol-1 displays excellent in vivo pharmacokinetics and brain exposure, attenuates disease progression, and prolongs survival in a GBM animal model. These results identify a vulnerability to massive vacuolization that can be targeted by small molecules and point to the possible exploitation of this process in the design of anticancer therapies.

Comparative drug pair screening across multiple glioblastoma cell lines reveals novel drug-drug interactions

BACKGROUND

Glioblastoma multiforme (GBM) is the most aggressive brain tumor in adults, and despite state-of-the-art treatment, survival remains poor and novel therapeutics are sorely needed. The aim of the present study was to identify new synergistic drug pairs for GBM. In addition, we aimed to explore differences in drug-drug interactions across multiple GBM-derived cell cultures and predict such differences by use of transcriptional biomarkers.

METHODS

We performed a screen in which we quantified drug-drug interactions for 465 drug pairs in each of the 5 GBM cell lines U87MG, U343MG, U373MG, A172, and T98G. Selected interactions were further tested using isobole-based analysis and validated in 5 glioma-initiating cell cultures. Furthermore, drug interactions were predicted using microarray-based transcriptional profiling in combination with statistical modeling.

RESULTS

Of the 5 × 465 drug pairs, we could define a subset of drug pairs with strong interaction in both standard cell lines and glioma-initiating cell cultures. In particular, a subset of pairs involving the pharmaceutical compounds rimcazole, sertraline, pterostilbene, and gefitinib showed a strong interaction in a majority of the cell cultures tested. Statistical modeling of microarray and interaction data using sparse canonical correlation analysis revealed several predictive biomarkers, which we propose could be of importance in regulating drug pair responses.

CONCLUSION

We identify novel candidate drug pairs for GBM and suggest possibilities to prospectively use transcriptional biomarkers to predict drug interactions in individual cases.

Searching for synergies: matrix algebraic approaches for efficient pair screening

Functionally interacting perturbations, such as synergistic drugs pairs or synthetic lethal gene pairs, are of key interest in both pharmacology and functional genomics. However, to find such pairs by traditional screening methods is both time consuming and costly. We present a novel computational-experimental framework for efficient identification of synergistic target pairs, applicable for screening of systems with sizes on the order of current drug, small RNA or SGA (Synthetic Genetic Array) libraries (>1000 targets). This framework exploits the fact that the response of a drug pair in a given system, or a pair of genes' propensity to interact functionally, can be partly predicted by computational means from (i) a small set of experimentally determined target pairs, and (ii) pre-existing data (e.g. gene ontology, PPI) on the similarities between targets. Predictions are obtained by a novel matrix algebraic technique, based on cyclical projections onto convex sets. We demonstrate the efficiency of the proposed method using drug-drug interaction data from seven cancer cell lines and gene-gene interaction data from yeast SGA screens. Our protocol increases the rate of synergism discovery significantly over traditional screening, by up to 7-fold. Our method is easy to implement and could be applied to accelerate pair screening for both animal and microbial systems.

The impact of phenotypic switching on glioblastoma growth and invasion

The brain tumour glioblastoma is characterised by diffuse and infiltrative growth into surrounding brain tissue. At the macroscopic level, the progression speed of a glioblastoma tumour is determined by two key factors: the cell proliferation rate and the cell migration speed. At the microscopic level, however, proliferation and migration appear to be mutually exclusive phenotypes, as indicated by recent in vivo imaging data. Here, we develop a mathematical model to analyse how the phenotypic switching between proliferative and migratory states of individual cells affects the macroscopic growth of the tumour. For this, we propose an individual-based stochastic model in which glioblastoma cells are either in a proliferative state, where they are stationary and divide, or in motile state in which they are subject to random motion. From the model we derive a continuum approximation in the form of two coupled reaction-diffusion equations, which exhibit travelling wave solutions whose speed of invasion depends on the model parameters. We propose a simple analytical method to predict progression rate from the cell-specific parameters and demonstrate that optimal glioblastoma growth depends on a non-trivial trade-off between the phenotypic switching rates. By linking cellular properties to an in vivo outcome, the model should be applicable to designing relevant cell screens for glioblastoma and cytometry-based patient prognostics.

In this work, we develop a spatial mathematical model in order to analyse the growth behavior of the brain tumour glioblastoma. Tumours of this type have a diffuse boundary, with considerable local invasion of surrounding brain tissue, making surgery difficult. At the cellular level, the progression of a glioblastoma is known to depend on the balance between cell division (proliferation) and cell movement (migration). Based on recent evidence, our model assumes that each cell in a glioblastoma tumour resides in either of two mutually exclusive states: proliferating or migrating. From a probabilistic model of switching between these two phenotypes, we go on to derive equations that link cellular phenotypes to disease progression. The model has several possible applications. For instance, it could be used to predict the rate of disease progression in an individual patient, and to improve screening methods.

Clinically significant copy number alterations and complex rearrangements of MYB and NFIB in head and neck adenoid cystic carcinoma

Adenoid cystic carcinoma (ACC) of the head and neck is a malignant tumor with poor long-term prognosis. Besides the recently identified MYB-NFIB fusion oncogene generated by a t(6;9) translocation, little is known about other genetic alterations in ACC. Using high-resolution, array-based comparative genomic hybridization, and massively paired-end sequencing, we explored genomic alterations in 40 frozen ACCs. Eighty-six percent of the tumors expressed MYB-NFIB fusion transcripts and 97% overexpressed MYB mRNA, indicating that MYB activation is a hallmark of ACC. Thirty-five recurrent copy number alterations (CNAs) were detected, including losses involving 12q, 6q, 9p, 11q, 14q, 1p, and 5q and gains involving 1q, 9p, and 22q. Grade III tumors had on average a significantly higher number of CNAs/tumor compared to Grade I and II tumors (P = 0.007). Losses of 1p, 6q, and 15q were associated with high-grade tumors, whereas losses of 14q were exclusively seen in Grade I tumors. The t(6;9) rearrangements were associated with a complex pattern of breakpoints, deletions, insertions, inversions, and for 9p also gains. Analyses of fusion-negative ACCs using high-resolution arrays and massively paired-end sequencing revealed that MYB may also be deregulated by other mechanisms in addition to gene fusion. Our studies also identified several down-regulated candidate tumor suppressor genes (CTNNBIP1, CASP9, PRDM2, and SFN) in 1p36.33-p35.3 that may be of clinical significance in high-grade tumors. Further, studies of these and other potential target genes may lead to the identification of novel driver genes in ACC.

Network modeling of the transcriptional effects of copy number aberrations in glioblastoma

DNA copy number aberrations (CNAs) are a hallmark of cancer genomes. However, little is known about how such changes affect global gene expression. We develop a modeling framework, EPoC (Endogenous Perturbation analysis of Cancer), to (1) detect disease-driving CNAs and their effect on target mRNA expression, and to (2) stratify cancer patients into long- and short-term survivors. Our method constructs causal network models of gene expression by combining genome-wide DNA- and RNA-level data. Prognostic scores are obtained from a singular value decomposition of the networks. By applying EPoC to glioblastoma data from The Cancer Genome Atlas consortium, we demonstrate that the resulting network models contain known disease-relevant hub genes, reveal interesting candidate hubs, and uncover predictors of patient survival. Targeted validations in four glioblastoma cell lines support selected predictions, and implicate the p53-interacting protein Necdin in suppressing glioblastoma cell growth. We conclude that large-scale network modeling of the effects of CNAs on gene expression may provide insights into the biology of human cancer. Free software in MATLAB and R is provided.

Off-target effects dominate a large-scale RNAi screen for modulators of the TGF-β pathway and reveal microRNA regulation of TGFBR2

BACKGROUND

RNA interference (RNAi) screens have been used to identify novel components of signal-transduction pathways in a variety of organisms. We performed a small interfering (si)RNA screen for novel members of the transforming growth factor (TGF)-β pathway in a human keratinocyte cell line. The TGF-β pathway is integral to mammalian cell proliferation and survival, and aberrant TGF-β responses have been strongly implicated in cancer.

RESULTS

We assayed how strongly single siRNAs targeting each of 6,000 genes affect the nuclear translocation of a green fluorescent protein (GFP)-SMAD2 reporter fusion protein. Surprisingly, we found no novel TGF-β pathway members, but we did find dominant off-target effects. All siRNA hits, whatever their intended direct target, reduced the mRNA levels of two known upstream pathway components, the TGF-β receptors 1 and 2 (TGFBR1 and TGFBR2), via micro (mi)RNA-like off-target effects. The scale of these off-target effects was remarkable, with at least 1% of the sequences in the unbiased siRNA library having measurable off-target effects on one of these two genes. It seems that relatively minor reductions of message levels via off-target effects can have dominant effects on an assay, if the pathway output is very dose-sensitive to levels of particular pathway components. In search of mechanistic details, we identified multiple miRNA-like sequence characteristics that correlated with the off-target effects. Based on these results, we identified miR-20a, miR-34a and miR-373 as miRNAs that inhibit TGFBR2 expression.

CONCLUSIONS

Our findings point to potential improvements for miRNA/siRNA target prediction methods, and suggest that the type II TGF-β receptor is regulated by multiple miRNAs. We also conclude that the risk of obtaining misleading results in siRNA screens using large libraries with single-assay readout is substantial. Control and rescue experiments are essential in the interpretation of such screens, and improvements to the methods to reduce or predict RNAi off-target effects would be beneficial.

DNA microarray study on gene expression profiles in co-cultured endothelial and smooth muscle cells in response to 4- and 24-h shear stress

Shear stress, a major hemodynamic force acting on the vessel wall, plays an important role in physiological processes such as cell growth, differentiation, remodelling, metabolism, morphology, and gene expression. We investigated the effect of shear stress on gene expression profiles in co-cultured vascular endothelial cells (ECs) and smooth muscle cells (SMCs). Human aortic ECs were cultured as a confluent monolayer on top of confluent human aortic SMCs, and the EC side of the co-culture was exposed to a laminar shear stress of 12 dyn/cm(2) for 4 or 24 h. After shearing, the ECs and SMCs were separated and RNA was extracted from the cells. The RNA samples were labelled and hybridized with cDNA array slides that contained 8694 genes. Statistical analysis showed that shear stress caused the differential expression (p < or = 0.05) of a total of 1151 genes in ECs and SMCs. In the co-cultured ECs, shear stress caused the up-regulation of 403 genes and down-regulation of 470. In the co-cultured SMCs, shear stress caused the up-regulation of 152 genes and down-regulation of 126 genes. These results provide new information on the gene expression profile and its potential functional consequences in co-cultured ECs and SMCs exposed to a physiological level of laminar shear stress. Although the effects of shear stress on gene expression in monocultured and co-cultured EC are generally similar, the response of some genes to shear stress is opposite between these two types of culture (e.g., ICAM-1 is up-regulated in monoculture and down-regulated in co-culture), which strongly indicates that EC-SMC interactions affect EC responses to shear stress.

Smooth Muscle Expression of Lipoma Preferred Partner Is Mediated by an Alternative Intronic Promoter That Is Regulated by Serum Response Factor/Myocardin

Lipoma preferred partner (LPP) was recently recognized as a smooth muscle marker that plays a role in smooth muscle cell migration. In this report, we focus on the transcriptional regulation of the LPP gene. In particular, we investigate whether LPP is directly regulated by serum response factor (SRF). We show that the LPP gene contains 3 evolutionarily conserved CArG boxes and that 1 of these is part of an alternative promoter in intron 2. Quantitative RT-PCR shows that this alternative promoter directs transcription specifically to smooth muscle containing tissues in vivo. By using chromatin immunoprecipitation, we demonstrate that 2 of the CArG boxes, including the promoter-associated CArG box, bind to endogenous SRF in cultured aortic smooth muscle cells. Electrophoretic mobility-shift assays show that the conserved CArG boxes bind SRF in vitro. In reporter experiments, we show that the alternative promoter has transcriptional capacity that is dependent on SRF/myocardin and that the promoter associated CArG box is required for that activity. Finally, we show by quantitative RT-PCR that the alternative promoter is strongly downregulated in SRF-deficient embryonic stem cells and in smooth muscle tissues derived from conditional SRF knockout mice. Collectively, our data demonstrate that expression of LPP in smooth muscle is mediated by an alternative promoter that is regulated by SRF/myocardin.

An improved method for detecting and delineating genomic regions with altered gene expression in cancer

A method is presented for identifying genomic regions with altered gene expression in gene expression maps.

Genomic regions with altered gene expression are a characteristic feature of cancer cells. We present a novel method for identifying such regions in gene expression maps. This method is based on total variation minimization, a classical signal restoration technique. In systematic evaluations, we show that our method combines top-notch detection performance with an ability to delineate relevant regions without excessive over-segmentation, making it a significant advance over existing methods. Software (Rendersome) is provided.

Threshold-free high-power methods for the ontological analysis of genome-wide gene-expression studies

Ontological analysis facilitates the interpretation of microarray data. Here we describe new ontological analysis methods which, unlike existing approaches, are threshold-free and statistically powerful. We perform extensive evaluations and introduce a new concept, detection spectra, to characterize methods. We show that different ontological analysis methods exhibit distinct detection spectra, and that it is critical to account for this diversity. Our results argue strongly against the continued use of existing methods, and provide directions towards an enhanced approach.

Gene expression analysis suggests that EBF-1 and PPARγ2 induce adipogenesis of NIH-3T3 cells with similar efficiency and kinetics

Differentiation of multipotent mesenchymal stem cells into lipid-accumulating adipocytes is a physiological process induced by transcription factors in combination with hormonal stimulation. We have used Affymetrix microarrays to compare the adipogenic differentiation pathways of NIH-3T3 fibroblasts induced to undergo in vitro differentiation by ectopic expression of early B cell factor (EBF)-1 or peroxisome proliferator-activated receptor (PPAR)γ2. These experiments revealed that commitment to the adipogenic pathway in the NIH-3T3 cells was not reflected in gene expression until 4 days after induction of differentiation. Furthermore, gene expression patterns at the earlier time points after stimulation indicated that EBF-1 and PPARγ2 induced different sets of genes, while the similarities increased upon differentiation, and that several genes linked to adipocyte differentiation were also transiently induced in the vector-transduced cells. These data suggest that the initial activation of genes associated with adipocyte development is independent of commitment to the adipogenic pathway and that EBF-1 and PPARγ2 induce adipocyte differentiation with comparable kinetics and efficiency.

Neural cell adhesion molecule-deficient beta-cell tumorigenesis results in diminished extracellular matrix molecule expression and tumour cell-matrix adhesion

To understand by which mechanism neural cell adhesion molecule (N-CAM) limits beta tumour cell disaggregation and dissemination, we searched for potential downstream genes of N-CAM during beta tumour cell progression by gene expression profiling. Here, we show that N-CAM-deficient beta-cell tumorigenesis is associated with changes in the expression of genes involved in cell-matrix adhesion and cytoskeletal dynamics, biological processes known to affect the invasive and metastatic behaviour of tumour cells. The extracellular matrix (ECM) molecules emerged as the primary target, i.e. N-CAM deficiency resulted in down-regulated mRNA expression of a broad range of ECM molecules. Consistent with this result, deficient deposition of major ECM stromal components, such as fibronectin, laminin 1 and collagen IV, was observed. Moreover, N-CAM-deficient tumour cells displayed defective matrix adhesion. These results offer a potential mechanism for tumour cell disaggregation during N-CAM-deficient beta tumour cell progression. Prospective consequences of these findings for the role of N-CAM in beta tumour cell dissemination are discussed.

Predictive screening for regulators of conserved functional gene modules (gene batteries) in mammals

Background

The expression of gene batteries, genomic units of functionally linked genes which are activated by similar sets of cis- and trans-acting regulators, has been proposed as a major determinant of cell specialization in metazoans. We developed a predictive procedure to screen the mouse and human genomes and transcriptomes for cases of gene-battery-like regulation.

Results

In a screen that covered ~40 per cent of all annotated protein-coding genes, we identified 21 co-expressed gene clusters with statistically supported sharing of cis-regulatory sequence elements. 66 predicted cases of over-represented transcription factor binding motifs were validated against the literature and fell into three categories: (i) previously described cases of gene battery-like regulation, (ii) previously unreported cases of gene battery-like regulation with some support in a limited number of genes, and (iii) predicted cases that currently lack experimental support. The novel predictions include for example Sox 17 and RFX transcription factor binding sites that were detected in ~10% of all testis specific genes, and HNF-1 and 4 binding sites that were detected in ~30% of all kidney specific genes respectively. The results are publicly available at .

Conclusion

21 co-expressed gene clusters were enriched for a total of 66 shared cis-regulatory sequence elements. A majority of these predictions represent novel cases of potential co-regulation of functionally coupled proteins. Critical technical parameters were evaluated, and the results and the methods provide a valuable resource for future experimental design.

Identification of human intestinal alkaline sphingomyelinase as a novel ecto-enzyme related to the nucleotide phosphodiesterase family

Alkaline sphingomyelinase (alk-SMase) hydrolyzes dietary sphingomyelin and generates sphingolipid messengers in the gut. In the present study, we purified the enzyme, identified a part of the amino acid sequence, and found a cDNA in the GenBank coding for the protein. The cDNA contains 1841 bp, and the open reading frame encodes 458 amino acids. Transient expression of the cDNA linked to a Myc tag in COS-7 cells increased alk-SMase activity in the cell extract by 689-fold and in the medium by 27-fold. High activity was also identified in the anti-Myc immunoprecipitated proteins and the proteins cross-reacted with anti-human alk-SMase. Northern blotting of human intestinal tissues found high levels of alk-SMase mRNA in the intestine and liver. The amino acid sequence shared no similarity with acid and neutral SMases but was related to the ecto-nucleotide phosphodiesterase (NPP) family with 30-36% identity to human NPPs. Alk-SMase has a predicted signal peptide domain at the N terminus and a signal anchor domain at the C terminus. The ion-binding sites and the catalytic residue of NPPs were conserved, but the substrate specificity domain was modified. Alk-SMase had no detectable nucleotidase activity, but its activity against sphingomyelin could be inhibited by orthovanadate, imidazole, and ATP. In contrast to NPPs, alk-SMase activity was not stimulated by divalent metal ions but inhibited by Zn2+. Differing from NPP2, the alk-SMase cleaved phosphocholine but not choline from lysophosphatidylcholine. Phylogenetic tree indicated that the enzyme is a new branch derived from the NPP family. Two cDNA sequences of mouse and rat that shared 83% identity to human alk-SMase were identified in the GenBank. In conclusion, we identified the amino acid and cDNA sequences of human intestinal alk-SMase, and found that it is a novel ecto-enzyme related to the NPP family with specific features essential for its SMase activity.

Prediction of cell type-specific gene modules: identification and initial characterization of a core set of smooth muscle-specific genes

Genes that are expressed in the same subset of cells potentially constitute a module regulated by shared cis-regulatory elements and a distinct set of transcription factors. Identifying such units is an important entry point to the molecular study of cell differentiation. We developed a general method to classify cell type-specific genes from expressed sequence tag (EST) data, and we optimized it for identification of smooth muscle cell (SMC)-specific genes. Expression profiles were derived from the quantitative distribution of EST data in mouse, and genes were classified based on their profile similarity to known reference genes, in this case smooth muscle myosin heavy chain. A large majority (>90%) of known SMC-specific genes were identified, together with novel candidates. Extensive experimental validation confirmed SMC-specific expression of candidates, for example, lipoma preferred partner (LPP) and a novel SMC-specific putative monoamine oxidase, SMAO. Our method performed considerably better than other computational methods in an objective cross validation comparison. The total number of SMC-specific genes is estimated to be approximately 50.