Discrimination of direct and indirect interactions in a network of regulatory effects

The matter of concern are algorithms for the discrimination of direct from indirect regulatory effects from an interaction graph built up by error-prone measurements. Many of these algorithms can be cast as a rule for the removal of a single edge of the graph, such that the remaining graph is still consistent with the data. A set of mild conditions is given under which iterated application of such a rule leads to a unique minimal consistent graph. We show that three of the common methods for direct interactions search fulfill these conditions, thus providing a justification of their use. The main issues a reconstruction algorithm has to deal with, are the noise in the data, the presence of regulatory cycles, and the direction of the regulatory effects. We introduce a novel rule that, in contrast to the previously mentioned methods, simultaneously takes into account all these aspects. An efficient algorithm for the computation of the minimal graph is given, whose time complexity is cubic in the number of vertices of the graph. Finally, we demonstrate the utility of our method in a simulation study.

Marked Gene Transcript Level Alterations Occur Early During Radical Prostatectomy

OBJECTIVES

Gene expression analyses have become an important approach to understand the biology of cancer. However, transcript level patterns and RNA quality could rapidly change in response to ischemic and mechanical stress. Studies have shown that this occurs both perioperatively and after surgical removal of organs.

METHODS

To better understand the relative importance of perioperative and postoperative gene expression changes, we performed quantitative reverse transcription-polymerase chain reactions on the transcripts of 91 cancer-related genes from normal and cancerous prostate tissues from 10 patients at eight different time points during surgical manipulation and after removal of the prostate.

RESULTS

The mRNA levels of 8 (EGR1, p21, KRT17, PIM1, S100P, TNFRSF, WFDC2, and TRIM29) of 91 genes changed significantly with time of surgery in normal and tumor tissue. Remarkably, all eight genes were up-regulated, a reaction that was most prominent during the early intraoperative period. Additional changes occurred but were much less prominent during the first postoperative hour.

CONCLUSIONS

Our results substantially challenge the utility of immediate postoperative tissue sampling. At least for prostate cancer, the data suggest that preoperative tissue collection by core biopsies is optimal for studying molecular changes in normal and neoplastic prostate tissues.

Estimating large-scale signaling networks through nested effect models with intervention effects from microarray data

MOTIVATION

Targeted interventions using RNA interference in combination with the measurement of secondary effects with DNA microarrays can be used to computationally reverse engineer features of upstream non-transcriptional signaling cascades based on the nested structure of effects.

RESULTS

We extend previous work by Markowetz et al., who proposed a statistical framework to score different network hypotheses. Our extensions go in several directions: we show how prior assumptions on the network structure can be incorporated into the scoring scheme by defining appropriate prior distributions on the network structure as well as on hyperparameters. An approach called module networks is introduced to scale up the original approach, which is limited to around 5 genes, to infer large-scale networks of more than 30 genes. Instead of the data discretization step needed in the original framework, we propose the usage of a beta-uniform mixture distribution on the P-value profile, resulting from differential gene expression calculation, to quantify effects. Extensive simulations on artificial data and application of our module network approach to infer the signaling network between 13 genes in the ER-alpha pathway in human MCF-7 breast cancer cells show that our approach gives sensible results. Using a bootstrapping and a jackknife approach, this reconstruction is found to be statistically stable.

AVAILABILITY

The proposed method is available within the Bioconductor R-package nem.

Automated production of recombinant human proteins as resource for proteome research

Background

An arbitrary set of 96 human proteins was selected and tested to set-up a fully automated protein production strategy, covering all steps from DNA preparation to protein purification and analysis. The target proteins are encoded by functionally uncharacterized open reading frames (ORF) identified by the German cDNA consortium. Fusion proteins were produced in E. coli with four different fusion tags and tested in five different purification strategies depending on the respective fusion tag. The automated strategy relies on standard liquid handling and clone picking equipment.

Results

A robust automated strategy for the production of recombinant human proteins in E. coli was established based on a set of four different protein expression vectors resulting in NusA/His, MBP/His, GST and His-tagged proteins. The yield of soluble fusion protein was correlated with the induction temperature and the respective fusion tag. NusA/His and MBP/His fusion proteins are best expressed at low temperature (25°C), whereas the yield of soluble GST fusion proteins was higher when protein expression was induced at elevated temperature. In contrast, the induction of soluble His-tagged fusion proteins was independent of the temperature. Amylose was not found useful for affinity-purification of MBP/His fusion proteins in a high-throughput setting, and metal chelating chromatography is recommended instead.

Conclusion

Soluble fusion proteins can be produced in E. coli in sufficient qualities and μg/ml culture quantities for downstream applications like microarray-based assays, and studies on protein-protein interactions employing a fully automated protein expression and purification strategy. Future applications might include the optimization of experimental conditions for the large-scale production of soluble recombinant proteins from libraries of open reading frames.

Extending pathways based on gene lists using InterPro domain signatures

Background

High-throughput technologies like functional screens and gene expression analysis produce extended lists of candidate genes. Gene-Set Enrichment Analysis is a commonly used and well established technique to test for the statistically significant over-representation of particular pathways. A shortcoming of this method is however, that most genes that are investigated in the experiments have very sparse functional or pathway annotation and therefore cannot be the target of such an analysis. The approach presented here aims to assign lists of genes with limited annotation to previously described functional gene collections or pathways. This works by comparing InterPro domain signatures of the candidate gene lists with domain signatures of gene sets derived from known classifications, e.g. KEGG pathways.

Results

In order to validate our approach, we designed a simulation study. Based on all pathways available in the KEGG database, we create test gene lists by randomly selecting pathway genes, removing these genes from the known pathways and adding variable amounts of noise in the form of genes not annotated to the pathway. We show that we can recover pathway memberships based on the simulated gene lists with high accuracy. We further demonstrate the applicability of our approach on a biological example.

Conclusion

Results based on simulation and data analysis show that domain based pathway enrichment analysis is a very sensitive method to test for enrichment of pathways in sparsely annotated lists of genes. An R based software package domainsignatures, to routinely perform this analysis on the results of high-throughput screening, is available via Bioconductor.

Subtractive gene expression profiling of articular cartilage and mesenchymal stem cells: serpins as cartilage-relevant differentiation markers

OBJECTIVE

Mesenchymal stem cells (MSCs) are a population of cells broadly discussed to support cartilage repair. The differentiation of MSCs into articular chondrocytes is, however, still poorly understood on the molecular level. The aim of this study was to perform an almost genome-wide screen for genes differentially expressed between cartilage and MSCs and to extract new markers useful to define chondrocyte differentiation stages.

METHODS

Gene expression profiles of MSCs (n=8) and articular cartilage from OA patients (n=7) were compared on a 30,000 cDNA-fragment array and differentially expressed genes were extracted by subtraction. Expression of selected genes was assessed during in vitro chondrogenic differentiation of MSCs and during dedifferentiation of expanded chondrocytes using quantitative and semi-quantitative reverse transcriptase-polymerase chain reaction (RT-PCR). Protein secretion was measured by enzyme-linked immunosorbent assay.

RESULTS

Eighty-seven genes were differentially expressed between MSCs and cartilage with a more than three-fold difference. Sixty-seven of them were higher expressed in cartilage and among them 15 genes were previously not detected in cartilage. Differential expression was confirmed for 69% of 26 reanalysed genes by RT-PCR. The profiles of three unknown transcripts and six protease-related molecules were characterised during differentiation. SERPINA1 and SERPINA3 mRNA expression correlated with chondrogenic differentiation of MSCs and dedifferentiation of chondrocytes, and SERPINA1 protein levels in culture supernatants could be correlated alike.

CONCLUSIONS

cDNA-array analysis identified SERPINA1 and A3 as new differentiation-relevant genes for cartilage. Since SERPINA1 secretion correlated with both chondrogenesis of MSCs and dedifferentiation during chondrocyte expansion, it represents an attractive marker for refinement of chondrocyte differentiation.

Respiratory Deleted in Malignant Brain Tumours 1 (DMBT1) levels increase during lung maturation and infection

Deleted in Malignant Brain Tumours 1 (DMBT1) is a secreted scavenger receptor cysteine-rich protein that binds and aggregates various bacteria and viruses in vitro. Studies in adults have shown that DMBT1 is expressed mainly by mucosal epithelia and glands, in particular within the respiratory tract, and plays a role in innate immune defence. We hypothesized that respiratory DMBT1 levels may be influenced by various developmental and clinical factors such as maturity, age and bacterial infection. DMBT1 levels were studied in 205 tracheal aspirate samples of 82 ventilated preterm and full-term infants by enzyme-linked immunosorbent assay. Possible effects of various clinical parameters were tested by multiple regression analysis. DMBT1 levels increased significantly with lung maturity (P < 0.0001 for both gestational and postnatal age) and in small-for-gestational-age infants (P = 0.0179). An increase of respiratory DMBT1 levels was detected in neonatal infections (P < 0.0001). These results were supported by Western blotting. Immunohistochemical analyses of archived newborn lung sections (n = 17) demonstrated high concentrations of DMBT1 in lungs of neonates with bacterial infections. Our data show that preterm infants are able to up-regulate DMBT1 in infection as an unspecific immune reaction.

DMBT1 confers mucosal protection in vivo and a deletion variant is associated with Crohn's disease

BACKGROUND & AIMS

Impaired mucosal defense plays an important role in the pathogenesis of Crohn's disease (CD), one of the main subtypes of inflammatory bowel disease (IBD). Deleted in malignant brain tumors 1 (DMBT1) is a secreted scavenger receptor cysteine-rich protein with predominant expression in the intestine and has been proposed to exert possible functions in regenerative processes and pathogen defense. Here, we aimed at analyzing the role of DMBT1 in IBD.

METHODS

We studied DMBT1 expression in IBD and normal tissues by quantitative reverse transcription-polymerase chain reaction, immunohistochemistry, and mRNA in situ hybridization. Genetic polymorphisms within DMBT1 were analyzed in an Italian IBD case-control sample. Dmbt1(-/-) mice were generated, characterized, and analyzed for their susceptibility to dextran sulfate sodium-induced colitis.

RESULTS

DMBT1 levels correlate with disease activity in inflamed IBD tissues. A highly significant fraction of the patients with IBD displayed up-regulation of DMBT1 specifically in the intestinal epithelial surface cells and Paneth cells. A deletion allele of DMBT1 with a reduced number of scavenger receptor cysteine-rich domain coding exons is associated with an increased risk of CD (P = .00056; odds ratio, 1.75) but not for ulcerative colitis. Dmbt1(-/-) mice display enhanced susceptibility to dextran sulfate sodium-induced colitis and elevated Tnf, Il6, and Nod2 expression levels during inflammation.

CONCLUSIONS

DMBT1 may play a role in intestinal mucosal protection and prevention of inflammation. Impaired DMBT1 function may contribute to the pathogenesis of CD.

The full-ORF clone resource of the German cDNA Consortium

Background

With the completion of the human genome sequence the functional analysis and characterization of the encoded proteins has become the next urging challenge in the post-genome era. The lack of comprehensive ORFeome resources has thus far hampered systematic applications by protein gain-of-function analysis. Gene and ORF coverage with full-length ORF clones thus needs to be extended. In combination with a unique and versatile cloning system, these will provide the tools for genome-wide systematic functional analyses, to achieve a deeper insight into complex biological processes.

Results

Here we describe the generation of a full-ORF clone resource of human genes applying the Gateway cloning technology (Invitrogen). A pipeline for efficient cloning and sequencing was developed and a sample tracking database was implemented to streamline the clone production process targeting more than 2,200 different ORFs. In addition, a robust cloning strategy was established, permitting the simultaneous generation of two clone variants that contain a particular ORF with as well as without a stop codon by the implementation of only one additional working step into the cloning procedure. Up to 92 % of the targeted ORFs were successfully amplified by PCR and more than 93 % of the amplicons successfully cloned.

Conclusion

The German cDNA Consortium ORFeome resource currently consists of more than 3,800 sequence-verified entry clones representing ORFs, cloned with and without stop codon, for about 1,700 different gene loci. 177 splice variants were cloned representing 121 of these genes. The entry clones have been used to generate over 5,000 different expression constructs, providing the basis for functional profiling applications. As a member of the recently formed international ORFeome collaboration we substantially contribute to generating and providing a whole genome human ORFeome collection in a unique cloning system that is made freely available in the community.

Large scale statistical inference of signaling pathways from RNAi and microarray data

Background

The advent of RNA interference techniques enables the selective silencing of biologically interesting genes in an efficient way. In combination with DNA microarray technology this enables researchers to gain insights into signaling pathways by observing downstream effects of individual knock-downs on gene expression. These secondary effects can be used to computationally reverse engineer features of the upstream signaling pathway.

Results

In this paper we address this challenging problem by extending previous work by Markowetz et al., who proposed a statistical framework to score networks hypotheses in a Bayesian manner. Our extensions go in three directions: First, we introduce a way to omit the data discretization step needed in the original framework via a calculation based on p-values instead. Second, we show how prior assumptions on the network structure can be incorporated into the scoring scheme using regularization techniques. Third and most important, we propose methods to scale up the original approach, which is limited to around 5 genes, to large scale networks.

Conclusion

Comparisons of these methods on artificial data are conducted. Our proposed module network is employed to infer the signaling network between 13 genes in the ER-α pathway in human MCF-7 breast cancer cells. Using a bootstrapping approach this reconstruction can be found with good statistical stability.

The code for the module network inference method is available in the latest version of the R-package nem, which can be obtained from the Bioconductor homepage.

Deleted in Malignant Brain Tumors 1 (DMBT1) is present in hyaline membranes and modulates surface tension of surfactant

Background

Deleted in Malignant Brain Tumors 1 (DMBT1) is a secreted scavenger receptor cysteine-rich protein that binds various bacteria and is thought to participate in innate pulmonary host defense. We hypothesized that pulmonary DMBT1 could contribute to respiratory distress syndrome in neonates by modulating surfactant function.

Methods

DMBT1 expression was studied by immunohistochemistry and mRNA in situ hybridization in post-mortem lungs of preterm and full-term neonates with pulmonary hyaline membranes. The effect of human recombinant DMBT1 on the function of bovine and porcine surfactant was measured by a capillary surfactometer. DMBT1-levels in tracheal aspirates of ventilated preterm and term infants were determined by ELISA.

Results

Pulmonary DMBT1 was localized in hyaline membranes during respiratory distress syndrome. In vitro addition of human recombinant DMBT1 to the surfactants increased surface tension in a dose-dependent manner. The DMBT1-mediated effect was reverted by the addition of calcium depending on the surfactant preparation.

Conclusion

Our data showed pulmonary DMBT1 expression in hyaline membranes during respiratory distress syndrome and demonstrated that DMBT1 increases lung surface tension in vitro. This raises the possibility that DMBT1 could antagonize surfactant supplementation in respiratory distress syndrome and could represent a candidate target molecule for therapeutic intervention in neonatal lung disease.

High expression level of bone degrading proteins as a possible inducer of osteolytic features in pigmented villonodular synovitis

Protein expression of osteopontin (OPN), osteoprotegerin (OPG), bone sialoprotein (BSP), osteocalcin (OC), RANKL and PTHrP was determined by use of immunohistochemical analysis on tissue arrays (48 cases of PVNS, 20 cases of active (a-RA), non-active rheumatoid arthritis (na-RA), and osteoarthritis (OA)). Additionally, gene expression was analysed using complimentary DNA (cDNA) microarrays. All PVNS cases showed a higher level of both protein and gene expression of RANKL, OPN and BSP in comparison with OA cases. Expression of OPG was not significantly different in PVNS compared to OA. The RANKL/OPG expression ratio was significantly higher in PVNS than in OA. High expressions level of proteins involved in bone degradation in PVNS may promote an intra-osseous propagation of the lesion. This evidence suggests that PVNS might respond to treatment using specific inhibitors of RANKL, OPN and BSP.

Reduced expression of vacuole membrane protein 1 affects the invasion capacity of tumor cells

Vacuole membrane protein 1 (Vmp1) is described as a cancer-relevant cell cycle modulator, but the function of this protein and its mode of action in tumor progression are still unknown. In this study, we show that the VMP1 mRNA level is significantly reduced in kidney cancer metastases as compared to primary tumors. Further, VMP1 expression is also decreased in the invasive breast cancer cell lines HCC1954 and MDA-MB-231 as compared to the non-invasive cell lines MCF-12A, T-47D and MCF-7. We show for the first time that Vmp1 is a plasma membrane protein and an essential component of initial cell-cell contacts and tight junction formation. It interacts with the tight junction protein Zonula Occludens-1 and colocalizes in spots between neighboring HEK293 cells. Downregulation of VMP1 by RNAi results in loss of cell adherence, and increases the invasion capacity of the non-invasive kidney cancer cell line Caki-2. In conclusion, our findings establish Vmp1 to be a novel cell-cell adhesion protein and that its expression level determines the invasion and metastatic potential of cancer cells.

Regulation of DMBT1 via NOD2 and TLR4 in intestinal epithelial cells modulates bacterial recognition and invasion

Mucosal epithelial cell layers are constantly exposed to a complex resident microflora. Deleted in malignant brain tumors 1 (DMBT1) belongs to the group of secreted scavenger receptor cysteine-rich proteins and is considered to be involved in host defense by pathogen binding. This report describes the regulation and function of DMBT1 in intestinal epithelial cells, which form the primary immunological barrier for invading pathogens. We report that intestinal epithelial cells up-regulate DMBT1 upon proinflammatory stimuli (e.g., TNF-alpha, LPS). We demonstrate that DMBT1 is a target gene for the intracellular pathogen receptor NOD2 via NF-kappaB activation. DMBT1 is strongly up-regulated in the inflamed intestinal mucosa of Crohn's disease patients with wild-type, but not with mutant NOD2. We show that DMBT1 inhibits cytoinvasion of Salmonella enterica and LPS- and muramyl dipeptide-induced NF-kappaB activation and cytokine secretion in vitro. Thus, DMBT1 may play an important role in the first line of mucosal defense conferring immune exclusion of bacterial cell wall components. Dysregulated intestinal DMBT1 expression due to mutations in the NOD2/CARD15 gene may be part of the complex pathophysiology of barrier dysfunction in Crohn's disease.

Genetic mapping in mice identifies DMBT1 as a candidate modifier of mammary tumors and breast cancer risk

Low-penetrance breast cancer susceptibility alleles seem to play a significant role in breast cancer risk but are difficult to identify in human cohorts. A genetic screen of 176 N2 backcross progeny of two Trp53(+/-) strains, BALB/c and C57BL/6, which differ in their susceptibility to mammary tumors, identified a modifier of mammary tumor susceptibility in an approximately 25-Mb interval on mouse chromosome 7 (designated SuprMam1). Relative to heterozygotes, homozygosity for BALB/c alleles of SuprMam1 significantly decreased mammary tumor latency from 70.7 to 61.1 weeks and increased risk twofold (P = 0.002). Dmbt1 (deleted in malignant brain tumors 1) was identified as a candidate modifier gene within the SuprMam1 interval because it was differentially expressed in mammary tissues from BALB/c-Trp53(+/-) and C57BL/6-Trp53(+/-) mice. Dmbt1 mRNA and protein was reduced in mammary glands of the susceptible BALB/c mice. Immunohistochemical staining demonstrated that DMBT1 protein expression was also significantly reduced in normal breast tissue from women with breast cancer (staining score, 1.8; n = 46) compared with cancer-free controls (staining score, 3.9; n = 53; P < 0.0001). These experiments demonstrate the use of Trp53(+/-) mice as a sensitized background to screen for low-penetrance modifiers of cancer. The results identify a novel mammary tumor susceptibility locus in mice and support a role for DMBT1 in suppression of mammary tumors in both mice and women.

Identification of aberrant chromosomal regions from gene expression microarray studies applied to human breast cancer

MOTIVATION

In cancer, chromosomal imbalances like amplifications and deletions, or changes in epigenetic mechanisms like DNA methylation influence the transcriptional activity. These alterations are often not limited to a single gene but affect several genes of the genomic region and may be relevant for the disease status. For example, the ERBB2 amplicon (17q21) in breast cancer is associated with poor patient prognosis. We present a general, unsupervised method for genome-wide gene expression data to systematically detect tumor patients with chromosomal regions of distinct transcriptional activity. The method aims to find expression patterns of adjacent genes with a consistently decreased or increased level of gene expression in tumor samples. Such patterns have been found to be associated with chromosomal aberrations and clinical parameters like tumor grading and thus can be useful for risk stratification or therapy.

RESULTS

Our approach was applied to 12 independent human breast cancer microarray studies comprising 1422 tumor samples. We prioritized chromosomal regions and genes predominantly found across all studies. The result highlighted not only regions which are well known to be amplified like 17q21 and 11q13, but also others like 8q24 (distal to MYC) and 17q24-q25 which may harbor novel putative oncogenes. Since our approach can be applied to any microarray study it may become a valuable tool for the exploration of transcriptional changes in diverse disease types.

AVAILABILITY

The R source codes which implement the method and an exemplary analysis are available at http://www.dkfz.de/mga2/people/buness/CTP/.

SUPPLEMENTARY INFORMATION

Supplementary data are available at Bioinformatics online.

Identification of cellular targets for the human papillomavirus E6 and E7 oncogenes by RNA interference and transcriptome analyses

Specific types of human papillomaviruses (HPVs) cause cervical cancer, the second most common tumor in women worldwide. Both cellular transformation and the maintenance of the oncogenic phenotype of HPV-positive tumor cells are linked to the expression of the viral E6 and E7 oncogenes. To identify downstream cellular target genes for the viral oncogenes, we silenced endogenous E6 and E7 expression in HPV-positive HeLa cells by RNA interference (RNAi). Subsequently, we assessed changes of the cellular transcriptome by genome-wide microarray analysis. We identified 648 genes, which were either downregulated (360 genes) or upregulated (288 genes), upon inhibition of E6/E7 expression. A large fraction of these genes is involved in tumor-relevant processes, such as apoptosis control, cell cycle regulation, or spindle formation. Others may represent novel cellular targets for the HPV oncogenes, such as a large group of C-MYC-associated genes involved in RNA processing and splicing. Comparison with published microarray data revealed a substantial concordance between the genes repressed by RNAi-mediated E6/E7 silencing in HeLa cells and genes reported to be upregulated in HPV-positive cervical cancer biopsies.

DMBT1 is frequently downregulated in well-differentiated gastric carcinoma but more frequently upregulated across various gastric cancer types

Well-differentiated gastric carcinomas are considered to represent a distinct entity emerging via specific molecular changes different from those found in other gastric carcinoma types. The gene deleted in malignant brain tumours 1 (DMBT1) at 10q25.3-q26.1 codes for a protein presumably involved in cell differentiation and protection and has been proposed as a candidate tumour suppressor for brain and epithelial cancer. One study reported a loss of DMBT1 expression in 12.5% (5/40) of gastric cancer samples. Here, we examined in more detail DMBT1 protein and mRNA expression in 78 primary gastric tumour samples and corresponding normal gastric mucosa. DMBT1 was expressed in all non-tumour gastric mucosa tissues. Eleven out of 71 (15%) gastric tumours were negative for the DMBT1 protein in immunohistochemical analyses. Lack of DMBT1 expression was significantly more frequently found in well-differentiated gastric tumours (6/18 well-differentiated tumours vs. 5/53 other subtypes; P=0.025). Quantitative RT-PCR revealed a downregulation of the DMBT1 mRNA for 8/21 (38%) cases, while the remaining 13 cases (62%) displayed a substantial upregulation. Our data suggest that a loss of DMBT1 expression may preferentially take place in well-differentiated gastric carcinoma. However, an upregulation of DMBT1 expression is more frequently found across all gastric cancer types.

GOSim--an R-package for computation of information theoretic GO similarities between terms and gene products

Background

With the increased availability of high throughput data, such as DNA microarray data, researchers are capable of producing large amounts of biological data. During the analysis of such data often there is the need to further explore the similarity of genes not only with respect to their expression, but also with respect to their functional annotation which can be obtained from Gene Ontology (GO).

Results

We present the freely available software package GOSim, which allows to calculate the functional similarity of genes based on various information theoretic similarity concepts for GO terms. GOSim extends existing tools by providing additional lately developed functional similarity measures for genes. These can e.g. be used to cluster genes according to their biological function. Vice versa, they can also be used to evaluate the homogeneity of a given grouping of genes with respect to their GO annotation. GOSim hence provides the researcher with a flexible and powerful tool to combine knowledge stored in GO with experimental data. It can be seen as complementary to other tools that, for instance, search for significantly overrepresented GO terms within a given group of genes.

Conclusion

GOSim is implemented as a package for the statistical computing environment R and is distributed under GPL within the CRAN project.