In vitro characterization of CD133lo cancer stem cells in Retinoblastoma Y79 cell line.

BACKGROUND: Retinoblastoma (Rb), the most common childhood intraocular malignant tumor, is reported to have cancer stem cells (CSCs) similar to other tumors. Our previous investigation in primary tumors identified the small sized cells with low CD133 (Prominin-1) and high CD44 (Hyaluronic acid receptor) expression to be putative Rb CSCs using flow cytometry (FSClo/SSClo/CD133lo/CD44hi). With this preliminary data, we have now utilized a comprehensive approach of in vitro characterization of Y79 Rb cell line following CSC enrichment using CD133 surface marker and subsequent validation to confirm the functional properties of CSCs. METHODS: The cultured Rb Y79 cells were evaluated for surface markers by flow cytometry and CD133 sorted cells (CD133lo/CD133hi) were compared for CSC characteristics by size/percentage, cell cycle assay, colony formation assay, differentiation, Matrigel transwell invasion assay, cytotoxicity assay, gene expression using microarray and validation by semi-quantitative PCR. RESULTS: Rb Y79 cell line shared the profile (CD133, CD90, CXCR4 and ABCB1) of primary tumors except for CD44 expression. The CD133lo cells (16.1 ± 0.2%) were FSClo/SSClo, predominantly within the G0/G1 phase, formed larger and higher number of colonies with ability to differentiate to CD133hi cells, exhibited increased invasive potential in a matrigel transwell assay (p < 0.05) and were resistant to Carboplatin treatment (p < 0.001) as compared to CD133hi cells. The CD133lo cells showed higher expression of several embryonic stem cell genes (HOXB2, HOXA9, SALL1, NANOG, OCT4, LEFTY), stem cells/progenitor genes (MSI2, BMI1, PROX1, ABCB1, ABCB5, ABCG2), and metastasis related gene- MACC1, when compared to the CD133hi cells. CONCLUSIONS: This study validates the observation from our earlier primary tumor study that CSC properties in Rb Y79 cell line are endowed within the CD133lo population, evident by their characteristics- i.e. small sized, dormant in nature, increased colony forming ability, differentiation to CD133hi cells, higher invasiveness potential, drug resistance and primitive gene expression pattern. These findings provide a proof of concept for methodological characterization of the retinoblastoma CSCs with future implications for improved diagnostic and treatment strategies.

UniHI 7: an enhanced database for retrieval and interactive analysis of human molecular interaction networks.

Unified Human Interactome (UniHI) (http://www.unihi.org) is a database for retrieval, analysis and visualization of human molecular interaction networks. Its primary aim is to provide a comprehensive and easy-to-use platform for network-based investigations to a wide community of researchers in biology and medicine. Here, we describe a major update (version 7) of the database previously featured in NAR Database Issue. UniHI 7 currently includes almost 350,000 molecular interactions between genes, proteins and drugs, as well as numerous other types of data such as gene expression and functional annotation. Multiple options for interactive filtering and highlighting of proteins can be employed to obtain more reliable and specific network structures. Expression and other genomic data can be uploaded by the user to examine local network structures. Additional built-in tools enable ready identification of known drug targets, as well as of biological processes, phenotypes and pathways enriched with network proteins. A distinctive feature of UniHI 7 is its user-friendly interface designed to be utilized in an intuitive manner, enabling researchers less acquainted with network analysis to perform state-of-the-art network-based investigations.

Dynamic circadian protein-protein interaction networks predict temporal organization of cellular functions.

Essentially all biological processes depend on protein-protein interactions (PPIs). Timing of such interactions is crucial for regulatory function. Although circadian (~24-hour) clocks constitute fundamental cellular timing mechanisms regulating important physiological processes, PPI dynamics on this timescale are largely unknown. Here, we identified 109 novel PPIs among circadian clock proteins via a yeast-two-hybrid approach. Among them, the interaction of protein phosphatase 1 and CLOCK/BMAL1 was found to result in BMAL1 destabilization. We constructed a dynamic circadian PPI network predicting the PPI timing using circadian expression data. Systematic circadian phenotyping (RNAi and overexpression) suggests a crucial role for components involved in dynamic interactions. Systems analysis of a global dynamic network in liver revealed that interacting proteins are expressed at similar times likely to restrict regulatory interactions to specific phases. Moreover, we predict that circadian PPIs dynamically connect many important cellular processes (signal transduction, cell cycle, etc.) contributing to temporal organization of cellular physiology in an unprecedented manner.

TRIB2 as a biomarker for diagnosis and progression of melanoma.

Malignant melanoma is the most deadly form of skin cancer. There is a critical need to identify the patients that could be successfully treated by surgery alone and those that require adjuvant treatment. In this study, we demonstrate that the expression of tribbles2 (TRIB2) strongly correlates with both the presence and progression of melanocyte-derived malignancies. We examined the expression of TRIB2 in addition to 12 previously described melanoma biomarkers across three independent full genome microarray studies. TRIB2 expression was consistently and significantly increased in benign nevi and melanoma, and was highest in samples from patients with metastatic melanoma. The expression profiles for the 12 biomarkers were poorly conserved throughout these studies with only TYR, S100B and SPP1 showing consistently elevated expression in metastatic melanoma versus normal skin. Strikingly we confirmed these findings in 20 freshly obtained primary melanoma tissue samples from metastatic lesions where the expression of these biomarkers were evaluated revealing that TRIB2 expression correlated with disease stage and clinical prognosis. Our results suggest that TRIB2 is a meaningful biomarker reflecting diagnosis and progression of melanoma, as well as predicting clinical response to chemotherapy.

A novel phosphatidylinositol 3-kinase (PI3K) inhibitor directs a potent FOXO-dependent, p53-independent cell cycle arrest phenotype characterized by the differential induction of a subset of FOXO-regulated genes.

INTRODUCTION: The activation of the phosphoinositide 3-kinase (PI3K)/AKT signalling pathway is one the most frequent genetic events in breast cancer, consequently the development of PI3K inhibitors has attracted much attention. Here we evaluate the effect of PI3K inhibition on global gene expression in breast cancer cells. METHODS: We used a range of methodologies that include in silico compound analysis, in vitro kinase assays, cell invasion assays, proliferation assays, genome-wide transcription studies (Agilent Technologies full genome arrays), gene set enrichment analysis, quantitative real-time PCR, immunoblotting in addition to chromatin immunoprecipitation. RESULTS: We defined the physico-chemical and the biological properties of ETP-45658, a novel potent PI3K inhibitor. We demonstrated that ETP-45658 potently inhibited cell proliferation within a broad range of human cancer cells, most potently suppressing the growth of breast cancer cells via inhibiting cell cycle. We show that this response is Forkhead box O (FOXO) protein dependent and p53 independent. Our genome-wide microarray analysis revealed that the cell cycle was the most affected biological process after exposure to ETP-45658 (or our control PI3K inhibitor PI-103), that despite the multiple transcription factors that are regulated by the PI3K/AKT signalling cascade, only the binding sites for FOXO transcription factors were significantly enriched and only a subset of all FOXO-dependent genes were induced. This disparity in gene transcription was not due to differential FOXO promoter recruitment. CONCLUSIONS: The constitutive activation of PI3Ks and thus the exclusion of FOXO transcription factors from the nucleus is a key feature of breast cancer. Our results presented here highlight that PI3K inhibition activates specific FOXO-dependent genes that mediate cell cycle arrest in breast cancer cells.

The homeobox gene CHX10/VSX2 regulates RdCVF promoter activity in the inner retina.

Rod-derived Cone Viability Factor (RdCVF) is a trophic factor with therapeutic potential for the treatment of retinitis pigmentosa, a retinal disease that commonly results in blindness. RdCVF is encoded by Nucleoredoxin-like 1 (Nxnl1), a gene homologous with the family of thioredoxins that participate in the defense against oxidative stress. RdCVF expression is lost after rod degeneration in the first phase of retinitis pigmentosa, and this loss has been implicated in the more clinically significant secondary cone degeneration that often occurs. Here, we describe a study of the Nxnl1 promoter using an approach that combines promoter and transcriptomic analysis. By transfection of selected candidate transcription factors, chosen based upon their expression pattern, we identified the homeodomain proteins CHX10/VSX2, VSX1 and PAX4, as well as the zinc finger protein SP3, as factors that can stimulate both the mouse and human Nxnl1 promoter. In addition, CHX10/VSX2 binds to the Nxnl1 promoter in vivo. Since CHX10/VSX2 is expressed predominantly in the inner retina, this finding motivated us to demonstrate that RdCVF is expressed in the inner as well as the outer retina. Interestingly, the loss of rods in the rd1 mouse, a model of retinitis pigmentosa, is associated with decreased expression of RdCVF by inner retinal cells as well as by rods. Based upon these results, we propose an alternative therapeutic strategy aimed at recapitulating RdCVF expression in the inner retina, where cell loss is not significant, to prevent secondary cone death and central vision loss in patients suffering from retinitis pigmentosa.

Huntington's disease and its therapeutic target genes: a global functional profile based on the HD Research Crossroads database.

BACKGROUND: Huntington's disease (HD) is a fatal progressive neurodegenerative disorder caused by the expansion of the polyglutamine repeat region in the huntingtin gene. Although the disease is triggered by the mutation of a single gene, intensive research has linked numerous other genes to its pathogenesis. To obtain a systematic overview of these genes, which may serve as therapeutic targets, CHDI Foundation has recently established the HD Research Crossroads database. With currently over 800 cataloged genes, this web-based resource constitutes the most extensive curation of genes relevant to HD. It provides us with an unprecedented opportunity to survey molecular mechanisms involved in HD in a holistic manner. METHODS: To gain a synoptic view of therapeutic targets for HD, we have carried out a variety of bioinformatical and statistical analyses to scrutinize the functional association of genes curated in the HD Research Crossroads database. In particular, enrichment analyses were performed with respect to Gene Ontology categories, KEGG signaling pathways, and Pfam protein families. For selected processes, we also analyzed differential expression, using published microarray data. Additionally, we generated a candidate set of novel genetic modifiers of HD by combining information from the HD Research Crossroads database with previous genome-wide linkage studies. RESULTS: Our analyses led to a comprehensive identification of molecular mechanisms associated with HD. Remarkably, we not only recovered processes and pathways, which have frequently been linked to HD (such as cytotoxicity, apoptosis, and calcium signaling), but also found strong indications for other potentially disease-relevant mechanisms that have been less intensively studied in the context of HD (such as the cell cycle and RNA splicing, as well as Wnt and ErbB signaling). For follow-up studies, we provide a regularly updated compendium of molecular mechanism, that are associated with HD, at http://hdtt.sysbiolab.eu Additionally, we derived a candidate set of 24 novel genetic modifiers, including histone deacetylase 3 (HDAC3), metabotropic glutamate receptor 1 (GRM1), CDK5 regulatory subunit 2 (CDK5R2), and coactivator 1ß of the peroxisome proliferator-activated receptor gamma (PPARGC1B). CONCLUSIONS: The results of our study give us an intriguing picture of the molecular complexity of HD. Our analyses can be seen as a first step towards a comprehensive list of biological processes, molecular functions, and pathways involved in HD, and may provide a basis for the development of more holistic disease models and new therapeutics.

Multi-color clonal tracking reveals intra-stage proliferative heterogeneity during mammary tumor progression.

Despite major progress in breast cancer research, the functional contribution of distinct cancer cell clones to malignant tumor progression and metastasis remains largely elusive. We have assessed clonal heterogeneity within individual primary tumors and metastases and also during the distinct stages of malignant tumor progression using clonal tracking of cancer cells in the MMTV-PyMT mouse model of metastatic breast cancer. Comparative gene expression analysis of clonal subpopulations reveals a substantial level of heterogeneity across and also within the various stages of breast carcinogenesis. The intra-stage heterogeneity is primarily manifested by differences in cell proliferation, also found within invasive carcinomas of luminal A-, luminal B-, and HER2-enriched human breast cancer. Surprisingly, in the mouse model of clonal tracing of cancer cells, chemotherapy mainly targets the slow-proliferative clonal populations and fails to efficiently repress the fast-proliferative populations. These insights may have considerable impact on therapy selection and response in breast cancer patients.

Author Correction: PyMT-1099, a versatile murine cell model for EMT in breast cancer.

An amendment to this paper has been published and can be accessed via a link at the top of the paper.

Gain Fat-Lose Metastasis: Converting Invasive Breast Cancer Cells into Adipocytes Inhibits Cancer Metastasis.

Cancer cell plasticity facilitates the development of therapy resistance and malignant progression. De-differentiation processes, such as an epithelial-mesenchymal transition (EMT), are known to enhance cellular plasticity. Here, we demonstrate that cancer cell plasticity can be exploited therapeutically by forcing the trans-differentiation of EMT-derived breast cancer cells into post-mitotic and functional adipocytes. Delineation of the molecular pathways underlying such trans-differentiation has motivated a combination therapy with MEK inhibitors and the anti-diabetic drug Rosiglitazone in various mouse models of murine and human breast cancer in vivo. This combination therapy provokes the conversion of invasive and disseminating cancer cells into post-mitotic adipocytes leading to the repression of primary tumor invasion and metastasis formation.

RETINOBASE: a web database, data mining and analysis platform for gene expression data on retina.

BACKGROUND: The retina is a multi-layered sensory tissue that lines the back of the eye and acts at the interface of input light and visual perception. Its main function is to capture photons and convert them into electrical impulses that travel along the optic nerve to the brain where they are turned into images. It consists of neurons, nourishing blood vessels and different cell types, of which neural cells predominate. Defects in any of these cells can lead to a variety of retinal diseases, including age-related macular degeneration, retinitis pigmentosa, Leber congenital amaurosis and glaucoma. Recent progress in genomics and microarray technology provides extensive opportunities to examine alterations in retinal gene expression profiles during development and diseases. However, there is no specific database that deals with retinal gene expression profiling. In this context we have built RETINOBASE, a dedicated microarray database for retina. DESCRIPTION: RETINOBASE is a microarray relational database, analysis and visualization system that allows simple yet powerful queries to retrieve information about gene expression in retina. It provides access to gene expression meta-data and offers significant insights into gene networks in retina, resulting in better hypothesis framing for biological problems that can subsequently be tested in the laboratory. Public and proprietary data are automatically analyzed with 3 distinct methods, RMA, dChip and MAS5, then clustered using 2 different K-means and 1 mixture models method. Thus, RETINOBASE provides a framework to compare these methods and to optimize the retinal data analysis. RETINOBASE has three different modules, "Gene Information", "Raw Data System Analysis" and "Fold change system Analysis" that are interconnected in a relational schema, allowing efficient retrieval and cross comparison of data. Currently, RETINOBASE contains datasets from 28 different microarray experiments performed in 5 different model systems: drosophila, zebrafish, rat, mouse and human. The database is supported by a platform that is designed to easily integrate new functionalities and is also frequently updated. CONCLUSION: The results obtained from various biological scenarios can be visualized, compared and downloaded. The results of a case study are presented that highlight the utility of RETINOBASE. Overall, RETINOBASE provides efficient access to the global expression profiling of retinal genes from different organisms under various conditions.

PyMT-1099, a versatile murine cell model for EMT in breast cancer.

An epithelial-mesenchymal transition (EMT) has been implicated in cancer metastasis, drug resistance, and in conferring stem cell-like traits to cancer cells. Most studies investigating EMT in cancer have either utilized immortalized or cancer cell lines that are already primed to undergo an EMT and do not adequately represent a fully differentiated epithelial state in the absence of an EMT induction. Hence, model systems are required which recapitulate all stages of EMT in cancer cells. Here, we report the derivation and characterization of epithelial PyMT-1099 cancer cells from the MMTV-PyMT mouse model of breast cancer. We demonstrate that PyMT-1099 cells undergo an EMT upon TGFß treatment, while upon TGFß withdrawal they go through a mesenchymal-epithelial transition (MET), as assessed by changes in cell morphology and marker expression and comparable to normal murine mammary gland NMuMG cells. However, in contrast to NMuMG cells, PyMT-1099 cells show an increase in cell migration and are highly tumorigenic and metastatic when transplanted into immunocompromised mice. Finally, we report cancer cell-specific changes in gene expression during EMT of PyMT-1099 cells not found in non-transformed NMuMG cells. Thus, PyMT-1099 cells are a versatile tool to study breast cancer-associated EMT and MET in vitro and in vivo.

A kinome-wide high-content siRNA screen identifies MEK5-ERK5 signaling as critical for breast cancer cell EMT and metastasis.

An epithelial to mesenchymal transition (EMT) has been correlated to malignant tumor progression and metastasis by promoting cancer cell migration and invasion and chemoresistance. Hence, finding druggable EMT effectors is critical to efficiently interfere with metastasis formation and to overcome therapy resistance. We have employed a high-content microscopy screen in combination with a kinome and phosphatome-wide siRNA library to identify signaling pathways underlying an EMT of murine mammary epithelial cells and breast cancer cells. This screen identified the MEK5-ERK5 axis as a critical player in TGFß-mediated EMT. Suppression of MEK5-ERK5 signaling completely prevented the morphological and molecular changes occurring during a TGFß-induced EMT and, conversely, forced highly metastatic breast cancer cells into a differentiated epithelial state. Inhibition of MEK5-ERK5 signaling also repressed breast cancer cell migration and invasion and substantially reduced lung metastasis without affecting primary tumor growth. The results suggest that the MEK5-ERK5 signaling axis via activation of MEF2B and other transcription factors plays an important role in the induction and maintenance of breast cancer cell migration and invasion and thus represents an exploitable target for the pharmacological inhibition of cancer cell metastasis.

The FAK inhibitor BI 853520 exerts anti-tumor effects in breast cancer.

Focal adhesion kinase (FAK) is a cytoplasmic tyrosine kinase that regulates a plethora of downstream signaling pathways essential for cell migration, proliferation and death, processes that are exploited by cancer cells during malignant progression. These well-established tumorigenic activities, together with its high expression and activity in different cancer types, highlight FAK as an attractive target for cancer therapy. We have assessed and characterized the therapeutic potential and the biological effects of BI 853520, a novel small chemical inhibitor of FAK, in several preclinical mouse models of breast cancer. Treatment with BI 853520 elicits a significant reduction in primary tumor growth caused by an anti-proliferative activity by BI 853520. In contrast, BI 853520 exerts effects with varying degrees of robustness on the different stages of the metastatic cascade. Together, the data demonstrate that the repression of FAK activity by the specific FAK inhibitor BI 853520 offers a promising anti-proliferative approach for cancer therapy.

Expression profiling of genes regulated by TGF-beta: differential regulation in normal and tumour cells.

BACKGROUND: TGF-beta is one of the key cytokines implicated in various disease processes including cancer. TGF-beta inhibits growth and promotes apoptosis in normal epithelial cells and in contrast, acts as a pro-tumour cytokine by promoting tumour angiogenesis, immune-escape and metastasis. It is not clear if various actions of TGF-beta on normal and tumour cells are due to differential gene regulations. Hence we studied the regulation of gene expression by TGF-beta in normal and cancer cells. RESULTS: Using human 19 K cDNA microarrays, we show that 1757 genes are exclusively regulated by TGF-beta in A549 cells in contrast to 733 genes exclusively regulated in HPL1D cells. In addition, 267 genes are commonly regulated in both the cell-lines. Semi-quantitative and real-time qRT-PCR analysis of some genes agrees with the microarray data. In order to identify the signalling pathways that influence TGF-beta mediated gene regulation, we used specific inhibitors of p38 MAP kinase, ERK kinase, JNK kinase and integrin signalling pathways. The data suggest that regulation of majority of the selected genes is dependent on at least one of these pathways and this dependence is cell-type specific. Interestingly, an integrin pathway inhibitor, RGD peptide, significantly affected TGF-beta regulation of Thrombospondin 1 in A549 cells. CONCLUSION: These data suggest major differences with respect to TGF-beta mediated gene regulation in normal and transformed cells and significant role of non-canonical TGF-beta pathways in the regulation of many genes by TGF-beta.

HDNetDB: A Molecular Interaction Database for Network-Oriented Investigations into Huntington's Disease.

Huntington's disease (HD) is a progressive and fatal neurodegenerative disorder caused by an expanded CAG repeat in the huntingtin gene. Although HD is monogenic, its molecular manifestation appears highly complex and involves multiple cellular processes. The recent application of high throughput platforms such as microarrays and mass-spectrometry has indicated multiple pathogenic routes. The massive data generated by these techniques together with the complexity of the pathogenesis, however, pose considerable challenges to researchers. Network-based methods can provide valuable tools to consolidate newly generated data with existing knowledge, and to decipher the interwoven molecular mechanisms underlying HD. To facilitate research on HD in a network-oriented manner, we have developed HDNetDB, a database that integrates molecular interactions with many HD-relevant datasets. It allows users to obtain, visualize and prioritize molecular interaction networks using HD-relevant gene expression, phenotypic and other types of data obtained from human samples or model organisms. We illustrated several HDNetDB functionalities through a case study and identified proteins that constitute potential cross-talk between HD and the unfolded protein response (UPR). HDNetDB is publicly accessible at http://hdnetdb.sysbiolab.eu .

miR-1199-5p and Zeb1 function in a double-negative feedback loop potentially coordinating EMT and tumour metastasis.

Epithelial tumour cells can gain invasive and metastatic capabilities by undergoing an epithelial-mesenchymal transition. Transcriptional regulators and post-transcriptional effectors like microRNAs orchestrate this process of high cellular plasticity and its malignant consequences. Here, using microRNA sequencing in a time-resolved manner and functional validation, we have identified microRNAs that are critical for the regulation of an epithelial-mesenchymal transition and of mesenchymal tumour cell migration. We report that miR-1199-5p is downregulated in its expression during an epithelial-mesenchymal transition, while its forced expression prevents an epithelial-mesenchymal transition, tumour cell migration and invasion in vitro, and lung metastasis in vivo. Mechanistically, miR-1199-5p acts in a reciprocal double-negative feedback loop with the epithelial-mesenchymal transition transcription factor Zeb1. This function resembles the activities of miR-200 family members, guardians of an epithelial cell phenotype. However, miR-1199-5p and miR-200 family members share only six target genes, indicating that, besides regulating Zeb1 expression, they exert distinct functions during an epithelial-mesenchymal transition.

The unfolded protein response and its potential role in Huntington's disease elucidated by a systems biology approach.

Huntington ´s disease (HD) is a progressive, neurodegenerative disease with a fatal outcome. Although the disease-causing gene (huntingtin) has been known for over 20 years, the exact mechanisms leading to neuronal cell death are still controversial. One potential mechanism contributing to the massive loss of neurons observed in the brain of HD patients could be the unfolded protein response (UPR) activated by accumulation of misfolded proteins in the endoplasmic reticulum (ER). As an adaptive response to counter-balance accumulation of un- or misfolded proteins, the UPR upregulates transcription of chaperones, temporarily attenuates new translation, and activates protein degradation via the proteasome. However, persistent ER stress and an activated UPR can also cause apoptotic cell death. Although different studies have indicated a role for the UPR in HD, the evidence remains inconclusive. Here, we present extensive bioinformatic analyses that revealed UPR activation in different experimental HD models based on transcriptomic data. Accordingly, we have identified 53 genes, including RAB5A, HMGB1, CTNNB1, DNM1, TUBB, TSG101, EEF2, DYNC1H1, SLC12A5, ATG5, AKT1, CASP7 and SYVN1 that provide a potential link between UPR and HD. To further elucidate the potential role of UPR as a disease-relevant process, we examined its connection to apoptosis based on molecular interaction data, and identified a set of 40 genes including ADD1, HSP90B1, IKBKB, IKBKG, RPS3A and LMNB1, which seem to be at the crossroads between these two important cellular processes. Remarkably, we also found strong correlation of UPR gene expression with the length of the polyglutamine tract of Huntingtin, which is a critical determinant of age of disease onset in human HD patients pointing to the UPR as a promising target for therapeutic intervention. The study is complemented by a newly developed web-portal called UPR-HD (http://uprhd.sysbiolab.eu) that enables visualization and interactive analysis of UPR-associated gene expression across various HD models.

A chemogenomic screening identifies CK2 as a target for pro-senescence therapy in PTEN-deficient tumours.

Enhancement of cellular senescence in tumours triggers a stable cell growth arrest and activation of an antitumour immune response that can be exploited for cancer therapy. Currently, there are only a limited number of targeted therapies that act by increasing senescence in cancers, but the majority of them are not selective and also target healthy cells. Here we developed a chemogenomic screening to identify compounds that enhance senescence in PTEN-deficient cells without affecting normal cells. By using this approach, we identified casein kinase 2 (CK2) as a pro-senescent target. Mechanistically, we show that Pten loss increases CK2 levels by activating STAT3. CK2 upregulation in Pten null tumours affects the stability of Pml, an essential regulator of senescence. However, CK2 inhibition stabilizes Pml levels enhancing senescence in Pten null tumours. Taken together, our screening strategy has identified a novel STAT3-CK2-PML network that can be targeted for pro-senescence therapy for cancer.

Multidimensional fitting for multivariate data analysis.

Large multidimensional data matrices are frequent in biology. However, statistical methods often have difficulties dealing with such matrices because they contain very complex data sets. Consequently variable selection and dimensionality reduction methods are often used to reduce matrix complexity, although at the expense of information conservation. A new method derived from multidimensional scaling (MDS) is presented for the case where two matrices are available to describe the same population. The presented method transforms one of the matrices, called the target matrix, with some constraints to make it fit with the second matrix, referred to as the reference matrix. The fitting to the reference matrix is performed on the distances computed for the two matrices, and the transformation depends on the problem at hand. A special feature of this method is that a variable can be only partially modified. The method is applied on the exclusive-or (XOR) problem and then on a biological application with large-scale gene expression data.