Systematic interpretation of microarray data using experiment annotations

Exploring the transcription factor activity in high-throughput gene expression data using RLQ analysis

Background

Interpretation of gene expression microarray data in the light of external information on both columns and rows (experimental variables and gene annotations) facilitates the extraction of pertinent information hidden in these complex data. Biologists classically interpret genes of interest after retrieving functional information from a subset of genes of interest. Transcription factors play an important role in orchestrating the regulation of gene expression. Their activity can be deduced by examining the presence of putative transcription factors binding sites in the gene promoter regions.

Results

In this paper we present the multivariate statistical method RLQ which aims to analyze microarray data where additional information is available on both genes and samples. As an illustrative example, we applied RLQ methodology to analyze transcription factor activity associated with the time-course effect of steroids on the growth of primary human lung fibroblasts. RLQ could successfully predict transcription factor activity, and could integrate various other sources of external information in the main frame of the analysis. The approach was validated by means of alternative statistical methods and biological validation.

Conclusions

RLQ provides an efficient way of extracting and visualizing structures present in a gene expression dataset by directly modeling the link between experimental variables and gene annotations.

Modulation of multidrug resistance in cancer cells by chelidonine and Chelidonium majus alkaloids

Cancer cells often develop multidrug resistance (MDR) which is a multidimensional problem involving several mechanisms and targets. This study demonstrates that chelidonine and an alkaloid extract from Chelidonium majus, which contains protoberberine and benzo[c]phenanthridine alkaloids, has the ability to overcome MDR of different cancer cell lines through interaction with ABC-transporters, CYP3A4 and GST, by induction of apoptosis, and cytotoxic effects. Chelidonine and the alkaloid extract inhibited P-gp/MDR1 activity in a concentration-dependent manner in Caco-2 and CEM/ADR5000 and reversed their doxorubicin resistance. In addition, chelidonine and the alkaloid extract inhibited the activity of the drug modifying enzymes CYP3A4 and GST in a dose-dependent manner. The alkaloids induced apoptosis in MDR cells which was accompanied by an activation of caspase-3, -8,-6/9, and phosphatidyl serine (PS) exposure. cDNA arrays were applied to identify differentially expressed genes after treatment with chelidonine and the alkaloid extract. The expression analysis identified a common set of regulated genes related to apoptosis, cell cycle, and drug metabolism. Treatment of Caco-2 cells with 50 μg/ml alkaloid extract and 50 μM chelidonine for up to 48 h resulted in a significant decrease in mRNA levels of P-gp/MDR1, MRP1, BCRP, CYP3A4, GST, and hPXR and in a significant increase in caspase-3 and caspase-8 mRNA. Thus, chelidonine is a promising model compound for overcoming MDR and for enhancing cytotoxicity of chemotherapeutics, especially against leukaemia cells. Its efficacy needs to be confirmed in animal models.

Robust protein profiling with complex antibody microarrays in a dual-colour mode

Antibody microarrays are a multiplexing technique for the analyses of hundreds of different analytes in parallel from small sample volumes of few microlitres only. With sensitivities in the picomolar to femtomolar range, they are gaining importance in proteomic analyses. These sensitivities can be obtained for complex protein samples without any pre-fractionation or signal amplification. Also, no expensive or elaborate protein depletion steps are needed. As with custom DNA-microarrays, the implementation of a dual-colour assay adds to assay robustness and reproducibility and was therefore a focus of our technical implementation. In order to perform antibody microarray experiments for large sets of samples and analytes in a robust manner, it was essential to optimise the experimental layout, the protein extraction, labelling and incubation as well as data processing steps. Here, we present our current protocol, which is used for the simultaneous analysis of the abundance of more than 800 proteins in plasma, urine, and tissue samples.

Transcript profiling identifies novel key players mediating the growth inhibitory effect of NS-398 on human pancreatic cancer cells

Pancreatic cancer is one of the most aggressive human malignancies with an increasing incidence worldwide. Despite an increase in the number of systemic treatments available for pancreatic cancer, the impact of therapy on the clinical course of the disease has been modest, underscoring an urgent need for new therapeutic options. Although selective cyclooxygenase-2 inhibitors have been demonstrated to have cancer-preventive effects, the mechanism of their effects is not clearly known. Moreover, there have been no unbiased studies to identify novel molecular targets of NS-398 regarding pancreatic cancer. Here we undertook a gene expression profiling study to identify novel molecular targets modulating the growth inhibitory effects of NS-398 on pancreatic cancer cell lines. Our mRNA-based gene expression results showed that the growth inhibitory effect of NS-398 was accompanied with an activation of G1/S and G2/M cell cycle regulation, P53 signalling, apoptotic, aryl hydrocarbon receptor and death receptor signalling pathways. Moreover, we reported, for the first time, that the growth inhibitory effect of NS-398 is mediated by down-regulation of RRM2, CTGF, MCM2 and PCNA and up-regulation of NAG-1 in all cell lines.

Apurinic/apyrimidinic endonuclease 1, p53, and thioredoxin are linked in control of aging in C. elegans

Deletions in mitochondrial DNA (mtDNA) accumulate during aging. Expression of the Caenorhabditis elegans apurinic/apyrimidinic endonuclease 1 (APE1) ortholog exo-3, involved in DNA repair, is reduced by 45% (P < 0.05) during aging of C. elegans. Suppression of exo-3 by treatment with RNAi resulted in a threefold increase in mtDNA deletions (P < 0.05), twofold enhanced generation of reactive oxygen species (ROS) (P < 0.01), distortion of the structural integrity of the nervous system, reduction of head motility by 43% (P < 0.01) and whole animal motility by 38% (P < 0.05). Suppression of exo-3 significantly reduced life span: mean life span decreased from 18.5 +/- 0.4 to 15.4 +/- 0.1 days (P < 0.001) and maximum life span from 25.9 +/- 0.4 to 23.2 +/- 0.1 days (P = 0.001). Additional treatment of exo-3-suppressed animals with a mitochondrial uncoupler decreased ROS levels, reduced neuronal damage, and increased motility and life span. Additional suppression of the C. elegans p53 ortholog cep-1 in exo-3 RNAi-treated animals similarly decreased ROS levels, preserved neuronal integrity, and increased motility and life span. In wild-type animals, suppression of cep-1, involved in downregulation of exo-3, increased expression of exo-3 without a significant effect on ROS levels, preserved neuronal integrity, and increased motility and life span. Suppression of the C. elegans thioredoxin orthologs trx-1 and trx-2, involved in the redox chaperone activity of exo-3, overrides the protective effect of cep-1 RNAi treatment on neuronal integrity, neuronal function, mean and maximum life span. These results show that APE1/EXO-3, p53/CEP-1, and thioredoxin affect each other and that these interactions determine aging as well as neuronal structure and function.

Dual-color proteomic profiling of complex samples with a microarray of 810 cancer-related antibodies

Antibody microarrays have the potential to enable comprehensive proteomic analysis of small amounts of sample material. Here, protocols are presented for the production, quality assessment, and reproducible application of antibody microarrays in a two-color mode with an array of 1,800 features, representing 810 antibodies that were directed at 741 cancer-related proteins. In addition to measures of array quality, we implemented indicators for the accuracy and significance of dual-color detection. Dual-color measurements outperform a single-color approach concerning assay reproducibility and discriminative power. In the analysis of serum samples, depletion of high-abundance proteins did not improve technical assay quality. On the contrary, depletion introduced a strong bias in protein representation. In an initial study, we demonstrated the applicability of the protocols to proteins derived from urine samples. We identified differences between urine samples from pancreatic cancer patients and healthy subjects and between sexes. This study demonstrates that biomedically relevant data can be produced. As demonstrated by the thorough quality analysis, the dual-color antibody array approach proved to be competitive with other proteomic techniques and comparable in performance to transcriptional microarray analyses.

Comparative analysis of missing value imputation methods to improve clustering and interpretation of microarray experiments

Background

Microarray technologies produced large amount of data. In a previous study, we have shown the interest of k-Nearest Neighbour approach for restoring the missing gene expression values, and its positive impact of the gene clustering by hierarchical algorithm. Since, numerous replacement methods have been proposed to impute missing values (MVs) for microarray data. In this study, we have evaluated twelve different usable methods, and their influence on the quality of gene clustering. Interestingly we have used several datasets, both kinetic and non kinetic experiments from yeast and human.

Results

We underline the excellent efficiency of approaches proposed and implemented by Bo and co-workers and especially one based on expected maximization (EM_array). These improvements have been observed also on the imputation of extreme values, the most difficult predictable values. We showed that the imputed MVs have still important effects on the stability of the gene clusters. The improvement on the clustering obtained by hierarchical clustering remains limited and, not sufficient to restore completely the correct gene associations. However, a common tendency can be found between the quality of the imputation method and the gene cluster stability. Even if the comparison between clustering algorithms is a complex task, we observed that k-means approach is more efficient to conserve gene associations.

Conclusions

More than 6.000.000 independent simulations have assessed the quality of 12 imputation methods on five very different biological datasets. Important improvements have so been done since our last study. The EM_array approach constitutes one efficient method for restoring the missing expression gene values, with a lower estimation error level. Nonetheless, the presence of MVs even at a low rate is a major factor of gene cluster instability. Our study highlights the need for a systematic assessment of imputation methods and so of dedicated benchmarks. A noticeable point is the specific influence of some biological dataset.

Identification of differentially expressed subnetworks based on multivariate ANOVA

Background

Since high-throughput protein-protein interaction (PPI) data has recently become available for humans, there has been a growing interest in combining PPI data with other genome-wide data. In particular, the identification of phenotype-related PPI subnetworks using gene expression data has been of great concern. Successful integration for the identification of significant subnetworks requires the use of a search algorithm with a proper scoring method. Here we propose a multivariate analysis of variance (MANOVA)-based scoring method with a greedy search for identifying differentially expressed PPI subnetworks.

Results

Given the MANOVA-based scoring method, we performed a greedy search to identify the subnetworks with the maximum scores in the PPI network. Our approach was successfully applied to human microarray datasets. Each identified subnetwork was annotated with the Gene Ontology (GO) term, resulting in the phenotype-related functional pathway or complex. We also compared these results with those of other scoring methods such as t statistic- and mutual information-based scoring methods. The MANOVA-based method produced subnetworks with a larger number of proteins than the other methods. Furthermore, the subnetworks identified by the MANOVA-based method tended to consist of highly correlated proteins.

Conclusion

This article proposes a MANOVA-based scoring method to combine PPI data with expression data using a greedy search. This method is recommended for the highly sensitive detection of large subnetworks.

Gene expression profiling identifies novel key players involved in the cytotoxic effect of Artesunate on pancreatic cancer cells

Pancreatic cancer is one of the most aggressive human malignancies, with an extremely poor prognosis. The paucity of curative therapies has translated into an overall 5-year survival rate of less than 5%, underscoring a desperate need for new therapeutic options. Artesunate (ART), clinically used as anti-malarial agent, has recently revealed remarkable anti-tumor activity. However, the mechanisms underlying those activities in pancreatic cancer are not yet known. Here we evaluated the anti-tumor activity of Artesunate and the possible underlying mechanisms in pancreatic cancer. MiaPaCa-2 (poorly differentiated) and BxPC-3 (moderately differentiated) pancreatic cancer cell lines were treated with Artesunate and the effect was monitored by a tetrazolium-based assay (MTS) for evaluating cell viability and by flow cytometry and caspase 3/7 activation for apoptosis evaluation. In addition cDNA arrays were used to identify differentially expressed genes. The microarray data were then validated by RT-PCR and Western blotting. Moreover, pathways associated with these expression changes were identified using the Ingenuity Pathway Analysis. The expression analysis identified a common set of genes that were regulated by Artesunate in pancreatic cancer. Our results provide the first in vitro evidence for the therapeutic utility of Artesunate in pancreatic cancer. Moreover, we identified Artesunate as a novel topoisomerase IIalpha inhibitor that inhibits pancreatic cancer growth through modulation of multiple signaling pathways. The present analysis is a starting point for the generation of hypotheses on candidate genes and for a more detailed dissection of the functional role of individual genes for the activity of Artesunate in tumor cells.

From experimental setup to data analysis in transcriptomics: copper metabolism in the human pathogen Candida albicans

Transcript profiling by microarray analysis offers a great opportunity to reveal unknown effects in a comprehensive context. To be able to interpret the data, some basic issues in experimental setting and design including type and number of replications have to be considered and are discussed in this work. In order to facilitate and automate data interpretation, the experimental data were projected and clustered by Correspondence Analysis, subsequently associated with gene ontology (GO) terms for functional classification. We applied the technology to investigate copper metabolism in the human pathogen Candida albicans. The presented dataset gives an example of how different fluorescent labeling, biological and technical replicas and data analysis strategies for microarray experiments may influence the final outcome of the results.

Identification of malignancy factors by analyzing cystic tumors of the pancreas

AIM

The diversity in the aggressiveness of cystic tumors of the pancreas - ranging from the usually benign serous cystadenoma to lesions of variable degrees of malignancy - was utilized for the identification of molecular factors that are involved in the occurrence of malignancy.

METHODS

We analyzed the transcript profiles of different cystic tumor types. The results were confirmed at the protein level by immunohistochemistry. Also, functional studies with siRNA silencing were performed.

RESULTS

Expression variations at the RNA and protein level were identified that are closely correlated with the degree of malignancy. Besides, all tumors could be classified effectively by this means. Many of the identified factors had not previously been known to be associated with malignant cystic lesions. siRNA silencing of the gene with the most prominent variation - the anti-apoptotic factor FASTK (Fas-activated serine/threonine kinase) - revealed a regulative effect on several genes known to be relevant to the development of tumors.

CONCLUSION

By a molecular analysis of rare types of pancreatic cancer, which are less frequent in terms of disease, variations could be identified that could be critical for the regulation of malignancy and thus relevant to the treatment of also the majority of pancreatic tumors.

Direct functional assessment of the composite phenotype through multivariate projection strategies

We present a novel approach for the analysis of transcriptomics data that integrates functional annotation of gene sets with expression values in a multivariate fashion, and directly assesses the relation of functional features to a multivariate space of response phenotypical variables. Multivariate projection methods are used to obtain new correlated variables for a set of genes that share a given function. These new functional variables are then related to the response variables of interest. The analysis of the principal directions of the multivariate regression allows for the identification of gene function features correlated with the phenotype. Two different transcriptomics studies are used to illustrate the statistical and interpretative aspects of the methodology. We demonstrate the superiority of the proposed method over equivalent approaches.