An Improved PDE6D Inhibitor Combines with Sildenafil To Inhibit KRAS Mutant Cancer Cell Growth.

The trafficking chaperone PDE6D (or PDEδ) was proposed as a surrogate target for K-Ras, leading to the development of a series of inhibitors that block its prenyl binding pocket. These inhibitors suffered from low solubility and suspected off-target effects, preventing their clinical development. Here, we developed a highly soluble, low nanomolar PDE6D inhibitor (PDE6Di), Deltaflexin3, which has the lowest off-target activity as compared to three prominent reference compounds. Deltaflexin3 reduces Ras signaling and selectively decreases the growth of KRAS mutant and PDE6D-dependent cancer cells. We further show that PKG2-mediated phosphorylation of Ser181 lowers K-Ras binding to PDE6D. Thus, Deltaflexin3 combines with the approved PKG2 activator Sildenafil to more potently inhibit PDE6D/K-Ras binding, cancer cell proliferation, and microtumor growth. As observed previously, inhibition of Ras trafficking, signaling, and cancer cell proliferation remained overall modest. Our results suggest reevaluating PDE6D as a K-Ras surrogate target in cancer.

Comprehensive blood metabolomics profiling of Parkinson's disease reveals coordinated alterations in xanthine metabolism.

Parkinson's disease (PD) is a highly heterogeneous disorder influenced by several environmental and genetic factors. Effective disease-modifying therapies and robust early-stage biomarkers are still lacking, and an improved understanding of the molecular changes in PD could help to reveal new diagnostic markers and pharmaceutical targets. Here, we report results from a cohort-wide blood plasma metabolic profiling of PD patients and controls in the Luxembourg Parkinson's Study to detect disease-associated alterations at the level of systemic cellular process and network alterations. We identified statistically significant changes in both individual metabolite levels and global pathway activities in PD vs. controls and significant correlations with motor impairment scores. As a primary observation when investigating shared molecular sub-network alterations, we detect pronounced and coordinated increased metabolite abundances in xanthine metabolism in de novo patients, which are consistent with previous PD case/control transcriptomics data from an independent cohort in terms of known enzyme-metabolite network relationships. From the integrated metabolomics and transcriptomics network analysis, the enzyme hypoxanthine phosphoribosyltransferase 1 (HPRT1) is determined as a potential key regulator controlling the shared changes in xanthine metabolism and linking them to a mechanism that may contribute to pathological loss of cellular adenosine triphosphate (ATP) in PD. Overall, the investigations revealed significant PD-associated metabolome alterations, including pronounced changes in xanthine metabolism that are mechanistically congruent with alterations observed in independent transcriptomics data. The enzyme HPRT1 may merit further investigation as a main regulator of these network alterations and as a potential therapeutic target to address downstream molecular pathology in PD.

Results and lessons learned from the sbv IMPROVER metagenomics diagnostics for inflammatory bowel disease challenge.

A growing body of evidence links gut microbiota changes with inflammatory bowel disease (IBD), raising the potential benefit of exploiting metagenomics data for non-invasive IBD diagnostics. The sbv IMPROVER metagenomics diagnosis for inflammatory bowel disease challenge investigated computational metagenomics methods for discriminating IBD and nonIBD subjects. Participants in this challenge were given independent training and test metagenomics data from IBD and nonIBD subjects, which could be wither either raw read data (sub-challenge 1, SC1) or processed Taxonomy- and Function-based profiles (sub-challenge 2, SC2). A total of 81 anonymized submissions were received between September 2019 and March 2020. Most participants' predictions performed better than random predictions in classifying IBD versus nonIBD, Ulcerative Colitis (UC) versus nonIBD, and Crohn's Disease (CD) versus nonIBD. However, discrimination between UC and CD remains challenging, with the classification quality similar to the set of random predictions. We analyzed the class prediction accuracy, the metagenomics features by the teams, and computational methods used. These results will be openly shared with the scientific community to help advance IBD research and illustrate the application of a range of computational methodologies for effective metagenomic classification.

RvD1n-3 DPA Downregulates the Transcription of Pro-Inflammatory Genes in Oral Epithelial Cells and Reverses Nuclear Translocation of Transcription Factor p65 after TNF-α Stimulation.

Specialized pro-resolving mediators (SPMs) are multifunctional lipid mediators that participate in the resolution of inflammation. We have recently described that oral epithelial cells (OECs) express receptors of the SPM resolvin RvD1n-3 DPA and that cultured OECs respond to RvD1n-3 DPA addition by intracellular calcium release, nuclear receptor translocation and transcription of genes coding for antimicrobial peptides. The aim of the present study was to assess the functional outcome of RvD1n-3 DPA-signaling in OECs under inflammatory conditions. To this end, we performed transcriptomic analyses of TNF-α-stimulated cells that were subsequently treated with RvD1n-3 DPA and found significant downregulation of pro-inflammatory nuclear factor kappa B (NF-κB) target genes. Further bioinformatics analyses showed that RvD1n-3 DPA inhibited the expression of several genes involved in the NF-κB activation pathway. Confocal microscopy revealed that addition of RvD1n-3 DPA to OECs reversed TNF-α-induced nuclear translocation of NF-κB p65. Co-treatment of the cells with the exportin 1 inhibitor leptomycin B indicated that RvD1n-3 DPA increases nuclear export of p65. Taken together, our observations suggest that SPMs also have the potential to be used as a therapeutic aid when inflammation is established.

Expression and function of resolvin RvD1n-3 DPA receptors in oral epithelial cells.

Chronic inflammatory responses can inflict permanent damage to host tissues. Specialized pro-resolving mediators downregulate inflammation but also can have other functions. The aim of this study was to examine whether oral epithelial cells express the receptors FPR2/ALX and DRV1/GPR32, which bind RvD1n-3 DPA , a recently described pro-resolving mediator derived from omega-3 docosapentaenoic acid (DPA), and whether RvD1n-3 DPA exposure induced significant responses in these cells. Gingival biopsies were stained using antibodies to FPR2/ALX and DRV1/GPR32. Expression of FPR2/ALX and DRV1/GPR32 was examined in primary oral epithelial cells by qRT-PCR, flow cytometry, and immunofluorescence. The effect of RvD1n-3 DPA on intracellular calcium mobilization and transcription of beta-defensins 1 and 2, and cathelicidin was evaluated by qRT-PCR. FPR2/ALX and DRV1/GPR32 were expressed by gingival keratinocytes in situ. In cultured oral epithelial cells, FPR2/ALX was detected on the cell surface, whereas FPR2/ALX and DRV1/GPR32 were detected intracellularly. Exposure to RvD1n-3 DPA induced intracellular calcium mobilization, FPR2/ALX internalization, DRV1/GPR32 translocation to the nucleus, and significantly increased expression of genes coding for beta-defensin 1, beta-defensin 2, and cathelicidin. This shows that the signal constituted by RvD1n-3 DPA is recognized by oral keratinocytes and that this can strengthen the antimicrobial and regulatory potential of the oral epithelium.

Biomarker discovery studies for patient stratification using machine learning analysis of omics data: a scoping review.

OBJECTIVE: To review biomarker discovery studies using omics data for patient stratification which led to clinically validated FDA-cleared tests or laboratory developed tests, in order to identify common characteristics and derive recommendations for future biomarker projects. DESIGN: Scoping review. METHODS: We searched PubMed, EMBASE and Web of Science to obtain a comprehensive list of articles from the biomedical literature published between January 2000 and July 2021, describing clinically validated biomarker signatures for patient stratification, derived using statistical learning approaches. All documents were screened to retain only peer-reviewed research articles, review articles or opinion articles, covering supervised and unsupervised machine learning applications for omics-based patient stratification. Two reviewers independently confirmed the eligibility. Disagreements were solved by consensus. We focused the final analysis on omics-based biomarkers which achieved the highest level of validation, that is, clinical approval of the developed molecular signature as a laboratory developed test or FDA approved tests. RESULTS: Overall, 352 articles fulfilled the eligibility criteria. The analysis of validated biomarker signatures identified multiple common methodological and practical features that may explain the successful test development and guide future biomarker projects. These include study design choices to ensure sufficient statistical power for model building and external testing, suitable combinations of non-targeted and targeted measurement technologies, the integration of prior biological knowledge, strict filtering and inclusion/exclusion criteria, and the adequacy of statistical and machine learning methods for discovery and validation. CONCLUSIONS: While most clinically validated biomarker models derived from omics data have been developed for personalised oncology, first applications for non-cancer diseases show the potential of multivariate omics biomarker design for other complex disorders. Distinctive characteristics of prior success stories, such as early filtering and robust discovery approaches, continuous improvements in assay design and experimental measurement technology, and rigorous multicohort validation approaches, enable the derivation of specific recommendations for future studies.

Comparative transcriptome analysis of Parkinson's disease and Hutchinson-Gilford progeria syndrome reveals shared susceptible cellular network processes.

BACKGROUND: Parkinson's Disease (PD) and Hutchinson-Gilford Progeria Syndrome (HGPS) are two heterogeneous disorders, which both display molecular and clinical alterations associated with the aging process. However, similarities and differences between molecular changes in these two disorders have not yet been investigated systematically at the level of individual biomolecules and shared molecular network alterations. METHODS: Here, we perform a comparative meta-analysis and network analysis of human transcriptomics data from case-control studies for both diseases to investigate common susceptibility genes and sub-networks in PD and HGPS. Alzheimer's disease (AD) and primary melanoma (PM) were included as controls to confirm that the identified overlapping susceptibility genes for PD and HGPS are non-generic. RESULTS: We find statistically significant, overlapping genes and cellular processes with significant alterations in both diseases. Interestingly, the majority of these shared affected genes display changes with opposite directionality, indicating that shared susceptible cellular processes undergo different mechanistic changes in PD and HGPS. A complementary regulatory network analysis also reveals that the altered genes in PD and HGPS both contain targets controlled by the upstream regulator CDC5L. CONCLUSIONS: Overall, our analyses reveal a significant overlap of affected cellular processes and molecular sub-networks in PD and HGPS, including changes in aging-related processes that may reflect key susceptibility factors associated with age-related risk for PD.

A rare loss-of-function variant of ADAM17 is associated with late-onset familial Alzheimer disease.

Common variants of about 20 genes contributing to AD risk have so far been identified through genome-wide association studies (GWAS). However, there is still a large proportion of heritability that might be explained by rare but functionally important variants. One of the so far identified genes with rare AD causing variants is ADAM10. Using whole-genome sequencing we now identified a single rare nonsynonymous variant (SNV) rs142946965 [p.R215I] in ADAM17 co-segregating with an autosomal-dominant pattern of late-onset AD in one family. Subsequent genotyping and analysis of available whole-exome sequencing data of additional case/control samples from Germany, UK, and USA identified five variant carriers among AD patients only. The mutation inhibits pro-protein cleavage and the formation of the active enzyme, thus leading to loss-of-function of ADAM17 alpha-secretase. Further, we identified a strong negative correlation between ADAM17 and APP gene expression in human brain and present in vitro evidence that ADAM17 negatively controls the expression of APP. As a consequence, p.R215I mutation of ADAM17 leads to elevated Aß formation in vitro. Together our data supports a causative association of the identified ADAM17 variant in the pathogenesis of AD.

Integrated Analyses of Microbiome and Longitudinal Metabolome Data Reveal Microbial-Host Interactions on Sulfur Metabolism in Parkinson's Disease.

Parkinson's disease (PD) exhibits systemic effects on the human metabolism, with emerging roles for the gut microbiome. Here, we integrate longitudinal metabolome data from 30 drug-naive, de novo PD patients and 30 matched controls with constraint-based modeling of gut microbial communities derived from an independent, drug-naive PD cohort, and prospective data from the general population. Our key results are (1) longitudinal trajectory of metabolites associated with the interconversion of methionine and cysteine via cystathionine differed between PD patients and controls; (2) dopaminergic medication showed strong lipidomic signatures; (3) taurine-conjugated bile acids correlated with the severity of motor symptoms, while low levels of sulfated taurolithocholate were associated with PD incidence in the general population; and (4) computational modeling predicted changes in sulfur metabolism, driven by A. muciniphila and B. wadsworthia, which is consistent with the changed metabolome. The multi-omics integration reveals PD-specific patterns in microbial-host sulfur co-metabolism that may contribute to PD severity.

Transcriptome profiling data reveals ubiquitin-specific peptidase 9 knockdown effects.

Ubiquitin specific peptidase 9 (USP9) is a deubiquitinase encoded by a sex-linked gene with a Y-chromosomal form (USP9Y) and an X-chromosomal form (USP9X) that escapes X-inactivation. Since USP9 is a key regulatory gene with sex-linked expression in the human brain, the gene may be of interest for researchers studying molecular gender differences and ubiquitin signaling in the brain. To assess the downstream effects of knocking down USP9X and USP9Y on a transcriptome-wide scale, we have conducted microarray profiling experiments using the human DU145 prostate cancer cell culture model, after confirming the robust expression of both USP9X and USP9Y in this model. By designing shRNA constructs for the specific knockdown of USP9X and the joint knockdown of USP9X and USP9Y, we have compared gene expression changes in both knockdowns to control conditions to infer potential shared and X- or Y-form specific alterations. Here, we provide details of the corresponding microarray profiling data, which has been deposited in the Gene Expression Omnibus database (GEO series accession number GSE79376). A biological interpretation of the data in the context of a potential involvement of USP9 in Alzheimer's disease has previously been presented in Köglsberger et al. (2016). To facilitate the re-use and re-analysis of the data for other applications, e.g. the study of ubiquitin signaling and protein turnover control, and the regulation of molecular gender differences in the human brain and brain-related disorders, we provide a more in-depth discussion of the data properties, specifications and possible use cases.

Metformin reverses TRAP1 mutation-associated alterations in mitochondrial function in Parkinson's disease.

The mitochondrial proteins TRAP1 and HTRA2 have previously been shown to be phosphorylated in the presence of the Parkinson's disease kinase PINK1 but the downstream signalling is unknown. HTRA2 and PINK1 loss of function causes parkinsonism in humans and animals. Here, we identified TRAP1 as an interactor of HTRA2 using an unbiased mass spectrometry approach. In our human cell models, TRAP1 overexpression is protective, rescuing HTRA2 and PINK1-associated mitochondrial dysfunction and suggesting that TRAP1 acts downstream of HTRA2 and PINK1. HTRA2 regulates TRAP1 protein levels, but TRAP1 is not a direct target of HTRA2 protease activity. Following genetic screening of Parkinson's disease patients and healthy controls, we also report the first TRAP1 mutation leading to complete loss of functional protein in a patient with late onset Parkinson's disease. Analysis of fibroblasts derived from the patient reveal that oxygen consumption, ATP output and reactive oxygen species are increased compared to healthy individuals. This is coupled with an increased pool of free NADH, increased mitochondrial biogenesis, triggering of the mitochondrial unfolded protein response, loss of mitochondrial membrane potential and sensitivity to mitochondrial removal and apoptosis. These data highlight the role of TRAP1 in the regulation of energy metabolism and mitochondrial quality control. Interestingly, the diabetes drug metformin reverses mutation-associated alterations on energy metabolism, mitochondrial biogenesis and restores mitochondrial membrane potential. In summary, our data show that TRAP1 acts downstream of PINK1 and HTRA2 for mitochondrial fine tuning, whereas TRAP1 loss of function leads to reduced control of energy metabolism, ultimately impacting mitochondrial membrane potential. These findings offer new insight into mitochondrial pathologies in Parkinson's disease and provide new prospects for targeted therapies.

A microfluidics-based in vitro model of the gastrointestinal human-microbe interface.

Changes in the human gastrointestinal microbiome are associated with several diseases. To infer causality, experiments in representative models are essential, but widely used animal models exhibit limitations. Here we present a modular, microfluidics-based model (HuMiX, human-microbial crosstalk), which allows co-culture of human and microbial cells under conditions representative of the gastrointestinal human-microbe interface. We demonstrate the ability of HuMiX to recapitulate in vivo transcriptional, metabolic and immunological responses in human intestinal epithelial cells following their co-culture with the commensal Lactobacillus rhamnosus GG (LGG) grown under anaerobic conditions. In addition, we show that the co-culture of human epithelial cells with the obligate anaerobe Bacteroides caccae and LGG results in a transcriptional response, which is distinct from that of a co-culture solely comprising LGG. HuMiX facilitates investigations of host-microbe molecular interactions and provides insights into a range of fundamental research questions linking the gastrointestinal microbiome to human health and disease.

Transcriptomic analyses of primary astrocytes under TNFα treatment.

Astrocytes, the most abundant glial cell population in the central nervous system, have important functional roles in the brain as blood brain barrier maintenance, synaptic transmission or intercellular communications [1], [2]. Numerous studies suggested that astrocytes exhibit a functional and morphological high degree of plasticity. For example, following any brain injury, astrocytes become reactive and hypertrophic. This phenomenon, also called reactive gliosis, is characterized by a set of progressive gene expression and cellular changes [3]. Interestingly, in this context, astrocytes can re-acquire neurogenic properties. It has been shown that astrocytes can undergo dedifferentiation upon injury and inflammation, and may re-acquire the potentiality of neural progenitors [4], [5], [6], [7]. To assess the effect of inflammation on astrocytes, primary mouse astrocytes were treated with tumor necrosis factor α (TNFα), one of the main pro-inflammatory cytokines. The strength of this study is that pure primary astrocytes were used. As microglia are highly reactive immune cells, we used a magnetic cell sorting separation (MACS) method to further obtain highly pure astrocyte cultures devoid of microglia. Here, we provide details of the microarray data, which have been deposited in the Gene Expression Omnibus (GEO) under the series accession number GSE73022. The analysis and interpretation of these data are included in Gabel et al. (2015). Analysis of gene expression indicated that the NFκB pathway-associated genes were induced after a TNFα treatment. We have shown that primary astrocytes devoid of microglia can respond to a TNFα treatment with the re-expression of genes implicated in the glial cell development.

Using prior knowledge from cellular pathways and molecular networks for diagnostic specimen classification.

For many complex diseases, an earlier and more reliable diagnosis is considered a key prerequisite for developing more effective therapies to prevent or delay disease progression. Classical statistical learning approaches for specimen classification using omics data, however, often cannot provide diagnostic models with sufficient accuracy and robustness for heterogeneous diseases like cancers or neurodegenerative disorders. In recent years, new approaches for building multivariate biomarker models on omics data have been proposed, which exploit prior biological knowledge from molecular networks and cellular pathways to address these limitations. This survey provides an overview of these recent developments and compares pathway- and network-based specimen classification approaches in terms of their utility for improving model robustness, accuracy and biological interpretability. Different routes to translate omics-based multifactorial biomarker models into clinical diagnostic tests are discussed, and a previous study is presented as example.

Inflammation Promotes a Conversion of Astrocytes into Neural Progenitor Cells via NF-κB Activation.

Brain inflammation, a common feature in neurodegenerative diseases, is a complex series of events, which can be detrimental and even lead to neuronal death. Nonetheless, several studies suggest that inflammatory signals are also positively influencing neural cell proliferation, survival, migration, and differentiation. Recently, correlative studies suggested that astrocytes are able to dedifferentiate upon injury and may thereby re-acquire neural stem cell (NSC) potential. However, the mechanism underlying this dedifferentiation process upon injury remains unclear. Here, we report that during the early response of reactive gliosis, inflammation induces a conversion of mature astrocytes into neural progenitors. A TNF treatment induces the decrease of specific astrocyte markers, such as glial fibrillary acidic protein (GFAP) or genes related to glycogen metabolism, while a subset of these cells re-expresses immaturity markers, such as CD44, Musashi-1, and Oct4. Thus, TNF treatment results in the appearance of cells that exhibit a neural progenitor phenotype and are able to proliferate and differentiate into neurons and/or astrocytes. This dedifferentiation process is maintained as long as TNF is present in the culture medium. In addition, we highlight a role for Oct4 in this process, since the TNF-induced dedifferentiation can be prevented by inhibiting Oct4 expression. Our results show that activation of the NF-κB pathway through TNF plays an important role in the dedifferentiation of astrocytes via the re-expression of Oct4. These findings indicate that the first step of reactive gliosis is in fact a dedifferentiation process of resident astrocytes mediated by the NF-κB pathway.

Functional and phenotypic differences of pure populations of stem cell-derived astrocytes and neuronal precursor cells.

Availability of homogeneous astrocyte populations would facilitate research concerning cell plasticity (metabolic and transcriptional adaptations; innate immune responses) and cell cycle reactivation. Current protocols to prepare astrocyte cultures differ in their final content of immature precursor cells, preactivated cells or entirely different cell types. A new method taking care of all these issues would improve research on astrocyte functions. We found here that the exposure of a defined population of pluripotent stem cell-derived neural stem cells (NSC) to BMP4 results in pure, nonproliferating astrocyte cultures within 24-48 h. These murine astrocytes generated from embryonic stem cells (mAGES) expressed the positive markers GFAP, aquaporin 4 and GLT-1, supported neuronal function, and acquired innate immune functions such as the response to tumor necrosis factor and interleukin 1. The protocol was applicable to several normal or disease-prone pluripotent cell lines, and the corresponding mAGES all exited the cell cycle and lost most of their nestin expression, in contrast to astrocytes generated by serum-addition or obtained as primary cultures. Comparative gene expression analysis of mAGES and NSC allowed quantification of differences between the two cell types and a definition of an improved marker set to define astrocytes. Inclusion of several published data sets in this transcriptome comparison revealed the similarity of mAGES with cortical astrocytes in vivo. Metabolic analysis of homogeneous NSC and astrocyte populations revealed distinct neurochemical features: both cell types synthesized glutamine and citrate, but only mature astrocytes released these metabolites. Thus, the homogeneous cultures allowed an improved definition of NSC and astrocyte features.

EnrichNet: network-based gene set enrichment analysis.

MOTIVATION: Assessing functional associations between an experimentally derived gene or protein set of interest and a database of known gene/protein sets is a common task in the analysis of large-scale functional genomics data. For this purpose, a frequently used approach is to apply an over-representation-based enrichment analysis. However, this approach has four drawbacks: (i) it can only score functional associations of overlapping gene/proteins sets; (ii) it disregards genes with missing annotations; (iii) it does not take into account the network structure of physical interactions between the gene/protein sets of interest and (iv) tissue-specific gene/protein set associations cannot be recognized. RESULTS: To address these limitations, we introduce an integrative analysis approach and web-application called EnrichNet. It combines a novel graph-based statistic with an interactive sub-network visualization to accomplish two complementary goals: improving the prioritization of putative functional gene/protein set associations by exploiting information from molecular interaction networks and tissue-specific gene expression data and enabling a direct biological interpretation of the results. By using the approach to analyse sets of genes with known involvement in human diseases, new pathway associations are identified, reflecting a dense sub-network of interactions between their corresponding proteins. AVAILABILITY: EnrichNet is freely available at http://www.enrichnet.org. CONTACT: Natalio.Krasnogor@nottingham.ac.uk, reinhard.schneider@uni.lu or avalencia@cnio.es SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics Online.

Using rule-based machine learning for candidate disease gene prioritization and sample classification of cancer gene expression data.

Microarray data analysis has been shown to provide an effective tool for studying cancer and genetic diseases. Although classical machine learning techniques have successfully been applied to find informative genes and to predict class labels for new samples, common restrictions of microarray analysis such as small sample sizes, a large attribute space and high noise levels still limit its scientific and clinical applications. Increasing the interpretability of prediction models while retaining a high accuracy would help to exploit the information content in microarray data more effectively. For this purpose, we evaluate our rule-based evolutionary machine learning systems, BioHEL and GAssist, on three public microarray cancer datasets, obtaining simple rule-based models for sample classification. A comparison with other benchmark microarray sample classifiers based on three diverse feature selection algorithms suggests that these evolutionary learning techniques can compete with state-of-the-art methods like support vector machines. The obtained models reach accuracies above 90% in two-level external cross-validation, with the added value of facilitating interpretation by using only combinations of simple if-then-else rules. As a further benefit, a literature mining analysis reveals that prioritizations of informative genes extracted from BioHEL's classification rule sets can outperform gene rankings obtained from a conventional ensemble feature selection in terms of the pointwise mutual information between relevant disease terms and the standardized names of top-ranked genes.

PathVar: analysis of gene and protein expression variance in cellular pathways using microarray data.

SUMMARY: Finding significant differences between the expression levels of genes or proteins across diverse biological conditions is one of the primary goals in the analysis of functional genomics data. However, existing methods for identifying differentially expressed genes or sets of genes by comparing measures of the average expression across predefined sample groups do not detect differential variance in the expression levels across genes in cellular pathways. Since corresponding pathway deregulations occur frequently in microarray gene or protein expression data, we present a new dedicated web application, PathVar, to analyze these data sources. The software ranks pathway-representing gene/protein sets in terms of the differences of the variance in the within-pathway expression levels across different biological conditions. Apart from identifying new pathway deregulation patterns, the tool exploits these patterns by combining different machine learning methods to find clusters of similar samples and build sample classification models. AVAILABILITY: freely available at http://pathvar.embl.de CONTACT: enrico.glaab@uni.lu SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

RERG (Ras-like, oestrogen-regulated, growth-inhibitor) expression in breast cancer: a marker of ER-positive luminal-like subtype.

Global gene expression profiling studies have classified breast cancer into a number of distinct biological and molecular classes with clinical relevance. The heterogeneous luminal group, which is largely characterised by oestrogen receptor (ER) expression, appears to contain distinct subgroups with differing behaviour. In this study, we analysed 47,293 gene transcripts in 128 invasive breast carcinomas (BC) using Artificial Neural Networks and a cross-validation analysis in combination with an ensemble sample classification to identify genes that can be used to subclassify ER+ luminal tumours. The results were validated using immunohistochemistry on TMAs containing 1,140 invasive breast cancers. Our results showed that the RERG gene is one of the highest ranked genes to differentiate between ER+ luminal-like and ER- non-luminal cancers based on a 10-fold external cross-validation analysis with an average classification accuracy of 89%. This was confirmed in our protein expression studies that showed RERG positive associations with markers of luminal differentiation including ER, luminal cytokeratins (CK19, CK18 and CK7/8) and FOXA1 (P = 0.004) and other markers of good prognosis in BC including small size, lower histologic grade and positive expression of androgen receptor, nuclear BRCA1, FHIT and cell cycle inhibitors p27 and p21. RERG expression was inversely associated with the proliferation marker MIB1 (P = 0.005) and p53. Strong RERG expression showed an association with longer breast cancer specific survival and distant metastasis free interval in the whole series as well as in the ER+ luminal group and these associations were independent of other prognostic variables. In conclusion, we used novel bioinformatics methods to identify candidate genes to characterise ER+ luminal-like breast cancer. RERG gene is a key marker of the luminal BC class and can be used to separate distinct prognostic subgroups.