Public Gene Expression Datasets Research Articles

Genotypes of cancer stem cells characterized by epithelial-to-mesenchymal transition and proliferation related functions.

Cancer stem cells (CSCs), or cancer cells with stem cell-like properties, generally exhibit drug resistance and have highly potent cancer inducing capabilities. Genome-wide expression data collected at public repositories over the last few years provide excellent material for studies that can lead to insights concerning the molecular and functional characteristics of CSCs. Here, we conducted functional genomic studies of CSC based on fourteen PCA-screened high quality public CSC whole genome gene expression datasets and, as control, four high quality non-stem-like cancer cell and non-cancerous stem cell datasets from the Gene Expression Omnibus database. A total of 6,002 molecular signatures were taken from the Molecular Signatures Database and used to characterize the datasets, which, under two-way hierarchical clustering, formed three genotypes. Type 1, consisting of mainly glia CSCs, had significantly enhanced proliferation, and significantly suppressed epithelial-mesenchymal transition (EMT), related functions. Type 2, mainly breast CSCs, had significantly enhanced EMT, but not proliferation, related functions. Type 3, composed of ovarian, prostate, and colon CSCs, had significantly suppressed proliferation related functions and mixed expressions on EMT related functions.

Identification of androgen-responsive lncRNAs as diagnostic and prognostic markers for prostate cancer.

Prostate cancer (PCa) is a leading cause of mortality among males. Long non-coding RNAs (lncRNAs) are subclass of noncoding RNAs that may act as biomarkers and therapeutic targets. In this study, we firstly conducted analysis of global lncRNA expression patterns by using our own cohort (GSE73397) and two public available gene expression datasets: The Cancer Genome Atlas (TCGA) and GSE55909. Next, we performed microarray to observe genome-wide lncRNAs' expressions under dihydrotestosterone (DHT) stimulation in LNCaP cells (GSE72866), and overlapped the result with ChIPBase data to predict androgen-responsive lncRNAs with ARE. Combined the two results, a total of 44 androgen-responsive lncRNAs with ARE were found to be over-expressed in PCa samples. Ten lncRNAs were selected for further validation by examining their expressions in LNCaP cells under DHT stimulation, and in PCa samples and cell lines. Among them, RP1-4514.2, LINC01138, SUZ12P1 and KLKP1 were validated as directly AR-targeted lncRNAs by ChIP-PCR. Then we conducted a bioinformatic analysis to identify lncRNAs as putative prognostic and therapeutic targets by using TCGA data. Three androgen-responsive lncRNAs, LINC01138, SUZ12P1 and SNHG1 showed association with gleason score and pT-stage. The biological functions of LINC01138 and SUZ12P1 were also evaluated, both lncRNAs promoted the proliferation and inhibited apoptosis of PCa. These results provide potent information for exploring potential biomarkers and therapeutic targets for prostate cancer, especially for castration-resistant PCa.

A Graphical Model of Smoking-Induced Global Instability in Lung Cancer.

Smoking is the major cause of lung cancer and the leading cause of cancer-related death in the world. The most current view about lung cancer is no longer limited to individual genes being mutated by any carcinogenic insults from smoking. Instead, tumorigenesis is a phenotype conferred by many systematic and global alterations, leading to extensive heterogeneity and variation for both the genotypes and phenotypes of individual cancer cells. Thus, strategically it is foremost important to develop a methodology to capture any consistent and global alterations presumably shared by most of the cancerous cells for a given population. This is particularly true that almost all of the data collected from solid cancers (including lung cancers) are usually distant apart over a large span of temporal or even spatial contexts. Here, we report a multiple non-Gaussian graphical model to reconstruct the gene interaction network using two previously published gene expression datasets. Our graphical model aims to selectively detect gross structural changes at the level of gene interaction networks. Our methodology is extensively validated, demonstrating good robustness, as well as the selectivity and specificity expected based on our biological insights. In summary, gene regulatory networks are still relatively stable during presumably the early stage of neoplastic transformation. But drastic structural differences can be found between lung cancer and its normal control, including the gain of functional modules for cellular proliferations such as EGFR and PDGFRA, as well as the lost of the important IL6 module, supporting their roles as potential drug targets. Interestingly, our method can also detect early modular changes, with the ALDH3A1 and its associated interactions being strongly implicated as a potential early marker, whose activations appear to alter LCN2 module as well as its interactions with the important TP53-MDM2 circuitry. Our strategy using the graphical model to reconstruct gene interaction work with biologically-inspired constraints exemplifies the importance and beauty of biology in developing any bio-computational approach.

Abstract 1066: Synthetic miR-22 inhibits alternative non homologous end-joining DNA repair and increases sensitivity to PARP-inhibition in multiple myeloma cells

Abstract Multiple Myeloma (MM) is an incurable hematologic malignancy characterized by deep genomic instability. Alternative non-homologous end-joining (A-NHEJ) DNA repair is up-regulated in a significant fraction of cancers and it is more error-prone than the major LIG4-dependent NHEJ repair pathway (classic-NHEJ). Principal components of A-NHEJ (PARP1, POLQ and LIG3) represent therefore a new class of cancer cell-specific therapeutic targets. There is Increasing evidence that microRNAs take actively part in the regulation of the DNA damage/repair network. Based on this rationale, we aimed to elucidate the role of miRNAs in the regulation of A-NHEJ-driven genomic instability in MM. We first examined the prognostic role of DNA Ligases involved in NHEJ pathways interrogating the public MM gene expression dataset GSE9782, which revealed higher expression of LIG3 in patients with shorter survival; conversely LIG4 expression was not associated with worse prognosis, thus supporting the idea that LIG3 has a pivotal role in genomic instability and progression of MM. Next, we aimed to identify LIG3-targeting miRNAs by an integrated approach that predicted miR-22 as negative regulator of LIG3 in MM patients. On this basis we evaluated miR-22 and LIG3 expression in a panel of MM cell lines and primary samples compared to plasma cells from healthy donors. We detected significant down-regulation of miR-22 and up-regulation of LIG3, strictly correlated to disease progression. Subsequently, we induced miR-22 overexpression (by transduction or transfection) in a panel of MM cell lines and we found cell growth inhibition, apoptosis induction and cell cycle arrest. Importantly, transfection of miR-22 mimics reduced the viability of primary MM cells but not of HD-PBMCs. We next examined the effects of enforced overexpression of miR-22 on DNA-damage response and we observed an increase of the DNA damage marker yH2AX, reduction of LIG3 at mRNA and protein level, and decrease of LIG3-3’UTR luciferase activity, indicating specific 3’ UTR targeting. Importantly, we observed inhibition of A-NHEJ repair by a functional assay based on the plasmid EJ2-GFP. Finally, miR-22 mimics potentiated cytotoxic effects of PARP-inhibitor Olaparib (Selleckchem), both in vitro and in vivo, confirming that co-inhibition of LIG3 and PARP-1 have a synergistic effect on A-NHEJ dependent MM cells. Taken together, our preliminary findings indicate that miR-22 inhibits A-NHEJ by specific targeting of LIG-3 and modulates sensitivity to PARP inhibition. Further investigations could shed new light on A-NHEJ in MM pathogenesis and define a role for miR-22 as therapeutic agent in this still incurable disease. Acknowledgement: This work has been supported by Italian Association for Cancer Research (AIRC). PI: PT. “Special Program Molecular Clinical Oncology - 5 per mille” n.9980, 2010/15 Citation Format: Daniele Caracciolo, Eugenio Morelli, Maria Teresa Di Martino, Daniela Talarico, Cirino Botta, Nicola Amodio, Cinzia Federico, Emanuela Altomare, Lavinia Biamonte, Maria Angelica Stamato, Pierosandro Tagliaferri, Pierfrancesco Tassone. Synthetic miR-22 inhibits alternative non homologous end-joining DNA repair and increases sensitivity to PARP-inhibition in multiple myeloma cells. [abstract]. In: Proceedings of the 107th Annual Meeting of the American Association for Cancer Research; 2016 Apr 16-20; New Orleans, LA. Philadelphia (PA): AACR; Cancer Res 2016;76(14 Suppl):Abstract nr 1066.

Common Marker Genes Identified from Various Sample Types for Systemic Lupus Erythematosus.

ObjectiveSystemic lupus erythematosus (SLE) is a complex auto-immune disease. Gene expression studies have been conducted to identify SLE-related genes in various types of samples. It is unknown whether there are common marker genes significant for SLE but independent of sample types, which may have potentials for follow-up translational research. The aim of this study is to identify common marker genes across various sample types for SLE.MethodsBased on four public microarray gene expression datasets for SLE covering three representative types of blood-born samples (monocyte; peripheral blood mononuclear cell, PBMC; whole blood), we utilized three statistics (fold-change, FC; t-test p value; false discovery rate adjusted p value) to scrutinize genes simultaneously regulated with SLE across various sample types. For common marker genes, we conducted the Gene Ontology enrichment analysis and Protein-Protein Interaction analysis to gain insights into their functions.ResultsWe identified 10 common marker genes associated with SLE (IFI6, IFI27, IFI44L, OAS1, OAS2, EIF2AK2, PLSCR1, STAT1, RNASE2, and GSTO1). Significant up-regulation of IFI6, IFI27, and IFI44L with SLE was observed in all the studied sample types, though the FC was most striking in monocyte, compared with PBMC and whole blood (8.82–251.66 vs. 3.73–74.05 vs. 1.19–1.87). Eight of the above 10 genes, except RNASE2 and GSTO1, interact with each other and with known SLE susceptibility genes, participate in immune response, RNA and protein catabolism, and cell death.ConclusionOur data suggest that there exist common marker genes across various sample types for SLE. The 10 common marker genes, identified herein, deserve follow-up studies to dissert their potentials as diagnostic or therapeutic markers to predict SLE or treatment response.

HG-05DIFFERENTIAL EXPRESSION OF PARP1/2 AS A BIOMARKER OF DIPG MALIGNANT PROGRESSION: RELATIONSHIP TO TGF-β/BMP SIGNALING AND EZH2 GLOBAL METHYLATION

HG-05. DIFFERENTIAL EXPRESSION OF PARP1/2 AS A BIOMARKER OF DIPG MALIGNANT PROGRESSION: RELATIONSHIP TO TGF-b/BMP SIGNALING AND EZH2 GLOBAL METHYLATION Uday Bhanu Maachani1, Uma Shankavaram2, Melanie Schweitzer1, Yue Linda Wu1, Kevin Camphausen2, and Mark Souweidane1; Weill Cornell Medicine, New York, NY, USA; National Cancer Institute-NIH, Bethesda, MD, USA Diffuse intrinsic pontine gliomas (DIPG) which occur in the brain stem are the most malignant inoperable and hard to treat brain tumors in children. Many reports indicate DIPG as being unique in terms of the molecular process that drives their malignancy. Recently, Poly (ADP-ribose) polymerase 1 & 2 (PARP1/2) have gained much attention as potential promising therapeutic targets for the treatment of solid tumors including DIPG. In this study, we defined molecular landscape of DIPG tumor samples by using a large public retrospective gene expression data set (n 1⁄4 53) and performed detailed bio-informatics analysis to generate hypotheses regarding potential mechanisms associated with PARP expression. Different associations between PARP and TGF-b/BMP, EZH2 signaling pathway genes were observed (low vs high). We found a negative co-relation between PARP expression and TGBRII, BMP2 and BMP2K genes, with much stronger correlation in patients with H3F3A mutation (present in 65-75% of tumors) suggesting negative role of PARP on TGF-b/BMP signaling. We noticed significant positive correlation of PARP with EZH2 and UHRF1BP1 genes knownto regulate global methylation and their increased expression is recognized to inhibit genes responsible for suppressing tumor development. Further patients with highPARPexpression along with high EZH2 and lowBMP signaling were correlated with poor prognosis to survival. In conclusion, these results reveal potential biological role of PARP in regulating TGF-b/BMP signaling and global methylation driving DIPG tumorigenesis. Moreover, these findings delineate pathways of susceptibility that could be disrupted to confer strong sensitivity to PARP inhibitors. Neuro-Oncology 18:iii48–iii77, 2016. doi:10.1093/neuonc/now073.4 #The Author(s) 2016. Published by Oxford University Press on behalf of the Society for Neuro-Oncology. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com.

Expression of nuclear factor, erythroid 2-like 2-mediated genes differentiates tuberculosis

During infection and host defense, nuclear factor, erythroid 2-like 2 (Nrf2) dependent signaling is an efficient antioxidant defensive mechanism used by host cells to control the destructive effects of reactive oxygen species. This allows for effective defense responses against microbes while minimizing oxidative injury to the host cell itself. As a central regulator of antioxidant genes, Nrf2 has gained great attention in its pivotal role in infection, especially in tuberculosis (TB), the top infectious disease killer worldwide. To elucidate the genes potentially regulated by Nrf2 in TB, we conducted a meta-analysis on published gene expression datasets. Firstly, we compared the global gene expression profiles between control and Nrf2-deficient human cells. The differentially expressed genes were deemed as "Nrf2-mediated genes". Next, the whole blood gene expression pattern of TB patients was compared with that of healthy controls, pneumonia patients, and lung cancer patients. We found that the genes deregulated in TB significantly overlap with the Nrf2-mediated genes. Based on the intersection of Nrf2-mediated and TB-regulated genes, we identified an Nrf2-mediated 17-gene signature, which reflects a cluster of gene ontology terms highly related to TB physiology. We demonstrated that the 17-gene signature can be used to distinguish TB patients from healthy controls and patients with latent TB infection, pneumonia, or lung cancer. Also, the Nrf2-mediated gene signature can be used as an indicator of the anti-TB therapeutic response. More importantly, we confirmed that the predictive power of the Nrf2-mediated 17-gene signature is significantly better than the random gene sets selected from the human transcriptome. Also, the 17-gene signature performs even better than the random gene signatures selected from TB-associated genes. Our study confirms the central role of Nrf2 in TB pathogenesis and provides a novel and useful diagnostic method to differentiate TB patients from other human subjects.

Dosage Compensation and the Distribution of Sex-Biased Gene Expression in Drosophila: Considerations and Genomic Constraints.

Several studies in Drosophila have shown a paucity of male-biased genes (i.e., genes that express higher in males than in females) on the X chromosome. Dosage compensation (DC) is a regulatory mechanism of gene expression triggered in males that hypertranscribes the X-linked genes to the level of transcription in females. There are currently two different hypotheses about the effects of DC on the distribution of male-biased genes: (1) it might limit male-expression level, or (2) it might interfere with the male upregulation of gene expression. Here, we used previously published gene expression datasets to reevaluate both hypotheses and introduce a mutually exclusive prediction that helped us to reject the hypothesis that the paucity of male-biased genes in the X chromosome is due to a limit in the male-expression level. Our analysis also uncovers unanticipated details about how DC interferes with the genomic distribution of both, male-biased and female-biased genes. We suggest that DC actually interferes with female downregulation of gene expression and not male upregulation, as previously suggested.

Integrative meta-analysis identifies microRNA-regulated networks in infantile hemangioma.

BackgroundHemangioma is a common benign tumor in the childhood; however our knowledge about the molecular mechanisms of hemangioma development and progression are still limited. Currently, microRNAs (miRNAs) have been shown as gene expression regulators with an important role in disease pathogenesis. Our goals were to identify miRNA-mRNA expression networks associated with infantile hemangioma.MethodsWe performed a meta-analysis of previously published gene expression datasets including 98 hemangioma samples. Deregulated genes were further used to identify microRNAs as potential regulators of gene expression in infantile hemangioma. Data were integrated using bioinformatics methods, and genes were mapped in proteins, which were then used to construct protein-protein interaction networks.ResultsDeregulated genes play roles in cell growth and differentiation, cell signaling, angiogenesis and vasculogenesis. Regulatory networks identified included microRNAs miR-9, miR-939 and let-7 family; these microRNAs showed the most number of interactions with deregulated genes in infantile hemangioma, suggesting that they may have an important role in the molecular mechanisms of disease. Additionally, results were used to identify drug-gene interactions and druggable gene categories using Drug-Gene Interaction Database. We show that microRNAs and microRNA-target genes may be useful biomarkers for the development of novel therapeutic strategies for patients with infantile hemangioma.ConclusionsmicroRNA-regulated pathways may play a role in infantile hemangioma development and progression and may be potentially useful for future development of novel therapeutic strategies for patients with infantile hemangioma.

Simvastatin prevents triple-negative breast cancer metastasis in pre-clinical models through regulation of FOXO3a.

We previously reported using statins was correlated with improved metastasis-free survival in aggressive breast cancer. The purpose of this study was to examine the effect of statins on metastatic colonization by triple-negative breast cancer (TNBC) cells. TNBC cell lines were treated with simvastatin and then studied for cell cycle progression and proliferation in vitro, and metastasis formation in vivo, following injection of statin-treated cells. Reverse-phase protein assay (RPPA) analysis was performed on statin-treated and control breast cancer cells. RNA interference targeting FOXO3a was used to measure the impact of simvastatin on FOXO3a-expressing cells. The prognostic value of FOXO3a mRNA expression was examined in eight public breast cancer gene expression datasets including 1479 patients. Simvastatin increased G1/S-phase arrest of the cell cycle and inhibited both proliferation and migration of TNBC cells in vitro. In vitro pre-treatment and in vivo treatment with simvastatin reduced metastases. Phosphorylated FOXO3a was downregulated after simvastatin treatment in (RPPA) analysis. Ectopic expression of FOXO3a enhanced mammosphere formation and migratory capacity in vitro. Knockdown of FOXO3a attenuated the effect of simvastatin on mammosphere formation and migration. Analysis of public gene expression data demonstrates FOXO3a mRNA downregulation was independently associated with shorter metastasis-free survival in all breast cancers, as well as in TNBC breast cancers. Simvastatin inhibits in vitro endpoints associated with metastasis through a FOXO3a mechanism and reduced metastasis formation in vivo. FOXO3a expression is prognostic for metastasis formation in patient data. Further investigation of simvastatin as a cancer therapy is warranted.

BRCA Mutation-Related and Claudin-Low Breast Cancer: Blood Relatives or Stepsisters?

Background: BRCA mutation-associated (BRCAmut) breast cancer represents a heterogeneous group displaying certain molecular features. Claudin-low breast cancers (CLBC) overlap with characteristics of BRCAmut tumors; therefore, we have investigated whether these are identical subtypes. Methods: Using public gene expression data, CLDN, CDH1, 9-cell line claudin-low predictor (9CLCLP) and PAM50 expression was evaluated in BRCAmut and BRCA wild-type (BRCAwt) breast cancer cases focusing on their possible overlap with the CLBC subtype. A separate formalin-fixed, paraffin-embedded (FFPE) cohort of 22 BRCAmut and 19 BRCAwt tumor tissues was used for immunohistochemical examination of AR, CD24, CD44, CK5/6, claudin-1, -3, -4 and -7, E-cadherin, EGFR, estrogen receptor (ER), EZH2, HER2, Ki67, p53, progesterone receptor (PgR) and vimentin expression. Results: In the data sets, CLDN1 (ROC = 0.785, p < 0.001), CDH1 (ROC = 0.785, p < 0.001), CLDN7 (ROC = 0.723, p < 0.001), CLDN3 (ROC = 0.696, p = 0.020) and CLDN4 (ROC = 0.685, p = 0.027) were expressed at higher level in BRCAmut than BRCAwt tumor tissue. The PAM50 subtype differed from the assigned immunohistochemistry (IHC)-based subtype in 30%. Based on accessible 9CLCLP predictor genes, BRCAmut breast cancer does not display the claudin-low phenotype. Utilizing FFPE samples, claudins were evidently expressed in both BRCAmut and BRCAwt cases. However, at the protein level, only claudin-3 expression was higher in BRCAmut tumors, while claudin-1, -4 and -7 and E-cadherin expression was lower compared to BRCAwt cases. A CD24^low/CD44^high phenotype was found in BRCAmut tumors upon comparison with BRCAwt cases (p < 0.001 and p = 0.001, respectively). Conclusions: There is a prominent correlation between the genes under focus herein and BRCA mutation status. BRCAmut tumors bear stem cell characteristics displaying a distinct cell adhesion molecule profile characterized by high expression of CDH1 and CLDN4 according to public gene expression data set analysis, and higher claudin-3 expression as detected by IHC; thus, BRCAmut breast carcinomas are not identical with the previously identified claudin-low subtype of breast cancer.

A Pilot Proteogenomic Study with Data Integration Identifies MCT1 and GLUT1 as Prognostic Markers in Lung Adenocarcinoma.

We performed a pilot proteogenomic study to compare lung adenocarcinoma to lung squamous cell carcinoma using quantitative proteomics (6-plex TMT) combined with a customized Affymetrix GeneChip. Using MaxQuant software, we identified 51,001 unique peptides that mapped to 7,241 unique proteins and from these identified 6,373 genes with matching protein expression for further analysis. We found a minor correlation between gene expression and protein expression; both datasets were able to independently recapitulate known differences between the adenocarcinoma and squamous cell carcinoma subtypes. We found 565 proteins and 629 genes to be differentially expressed between adenocarcinoma and squamous cell carcinoma, with 113 of these consistently differentially expressed at both the gene and protein levels. We then compared our results to published adenocarcinoma versus squamous cell carcinoma proteomic data that we also processed with MaxQuant. We selected two proteins consistently overexpressed in squamous cell carcinoma in all studies, MCT1 (SLC16A1) and GLUT1 (SLC2A1), for further investigation. We found differential expression of these same proteins at the gene level in our study as well as in other public gene expression datasets. These findings combined with survival analysis of public datasets suggest that MCT1 and GLUT1 may be potential prognostic markers in adenocarcinoma and druggable targets in squamous cell carcinoma. Data are available via ProteomeXchange with identifier PXD002622.

Novel gene sets improve set-level classification of prokaryotic gene expression data

BackgroundSet-level classification of gene expression data has received significant attention recently. In this setting, high-dimensional vectors of features corresponding to genes are converted into lower-dimensional vectors of features corresponding to biologically interpretable gene sets. The dimensionality reduction brings the promise of a decreased risk of overfitting, potentially resulting in improved accuracy of the learned classifiers. However, recent empirical research has not confirmed this expectation. Here we hypothesize that the reported unfavorable classification results in the set-level framework were due to the adoption of unsuitable gene sets defined typically on the basis of the Gene ontology and the KEGG database of metabolic networks. We explore an alternative approach to defining gene sets, based on regulatory interactions, which we expect to collect genes with more correlated expression. We hypothesize that such more correlated gene sets will enable to learn more accurate classifiers.MethodsWe define two families of gene sets using information on regulatory interactions, and evaluate them on phenotype-classification tasks using public prokaryotic gene expression data sets. From each of the two gene-set families, we first select the best-performing subtype. The two selected subtypes are then evaluated on independent (testing) data sets against state-of-the-art gene sets and against the conventional gene-level approach.ResultsThe novel gene sets are indeed more correlated than the conventional ones, and lead to significantly more accurate classifiers. The novel gene sets are indeed more correlated than the conventional ones, and lead to significantly more accurate classifiers.ConclusionNovel gene sets defined on the basis of regulatory interactions improve set-level classification of gene expression data. The experimental scripts and other material needed to reproduce the experiments are available at http://ida.felk.cvut.cz/novelgenesets.tar.gz.Electronic supplementary materialThe online version of this article (doi:10.1186/s12859-015-0786-7) contains supplementary material, which is available to authorized users.

Mitochondrial mass, a new metabolic biomarker for stem-like cancer cells: Understanding WNT/FGF-driven anabolic signaling.

Here, we developed an isogenic cell model of "stemness" to facilitate protein biomarker discovery in breast cancer. For this purpose, we used knowledge gained previously from the study of the mouse mammary tumor virus (MMTV). MMTV initiates mammary tumorigenesis in mice by promoter insertion adjacent to two main integration sites, namely Int-1 (Wnt1) and Int-2 (Fgf3), which ultimately activates Wnt/β-catenin signaling, driving the propagation of mammary cancer stem cells (CSCs). Thus, to develop a humanized model of MMTV signaling, we over-expressed WNT1 and FGF3 in MCF7 cells, an ER(+) human breast cancer cell line. We then validated that MCF7 cells over-expressing both WNT1 and FGF3 show a 3.5-fold increase in mammosphere formation, and that conditioned media from these cells is also sufficient to promote stem cell activity in untransfected parental MCF7 and T47D cells, as WNT1 and FGF3 are secreted factors. Proteomic analysis of this model system revealed the induction of i) EMT markers, ii) mitochondrial proteins, iii) glycolytic enzymes and iv) protein synthesis machinery, consistent with an anabolic CSC phenotype. MitoTracker staining validated the expected WNT1/FGF3-induced increase in mitochondrial mass and activity, which presumably reflects increased mitochondrial biogenesis. Importantly, many of the proteins that were up-regulated by WNT/FGF-signaling in MCF7 cells, were also transcriptionally over-expressed in human breast cancer cells in vivo, based on the bioinformatic analysis of public gene expression datasets of laser-captured patient samples. As such, this isogenic cell model should accelerate the discovery of new biomarkers to predict clinical outcome in breast cancer, facilitating the development of personalized medicine.Finally, we used mitochondrial mass as a surrogate marker for increased mitochondrial biogenesis in untransfected MCF7 cells. As predicted, metabolic fractionation of parental MCF7 cells, via MitoTracker staining, indicated that high mitochondrial mass is a new metabolic biomarker for the enrichment of anabolic CSCs, as functionally assessed by mammosphere-forming activity. This observation has broad implications for understanding the role of mitochondrial biogenesis in the propagation of stem-like cancer cells. Technically, this general metabolic approach could be applied to any cancer type, to identify and target the mitochondrial-rich CSC population.The implications of our work for understanding the role of mitochondrial metabolism in viral oncogenesis driven by random promoter insertions are also discussed, in the context of MMTV and ALV infections.

Abstract 3756: Identification of prognostic gene signatures of glioblastoma: a study based on TCGA data analysis

Abstract The Cancer Genome Atlas (TCGA) project is a large-scale effort with the goal of identifying novel molecular aberrations in glioblastoma (GBM). Here, we describe an in-depth analysis of gene expression data and copy number aberration (CNA) data to classify GBMs into prognostic groups to determine correlates of subtypes that may be biologically significant. To identify predictive survival models, we searched TCGA in 173 patients and identified 42 probe sets that could be used to divide the tumor samples into 3 groups and showed a significantly improved overall survival. Kaplan-Meier plots showed that the median survival of group 3 was markedly longer (127 weeks) than that of groups 1 and 2 (47 and 52 weeks, respectively). We then validated the 42 probe sets to stratify the patients according to survival in other public GBM gene expression datasets (eg, GSE4290 dataset). An overall analysis of the gene expression and copy number aberration using a multivariate Cox regression model showed that the 42 probe sets had a significant prognostic value independent of other variables. By integrating multidimensional genomic data from TCGA, we identified a specific survival model in a new prognostic group of GBM and suggest that molecular stratification of patients with GBM into homogeneous subgroups may provide opportunities for the development of new treatment modalities. [This work was supported by the National Research Foundation of Korea (NRF) grant funded by the Korean government (MEST) (NRF-2013R1A1A2063115) and the MSIP (Ministry of Science, ICT and Future Planning), Korea, under the “IT Consilience Creative Program” (NIPA-2014-H0201-14-1002) supervised by the NIPA (National IT Industry Promotion Agency)] Citation Format: Yong-Wan Kim. Identification of prognostic gene signatures of glioblastoma: a study based on TCGA data analysis. [abstract]. In: Proceedings of the 106th Annual Meeting of the American Association for Cancer Research; 2015 Apr 18-22; Philadelphia, PA. Philadelphia (PA): AACR; Cancer Res 2015;75(15 Suppl):Abstract nr 3756. doi:10.1158/1538-7445.AM2015-3756

Fuzzy logic selection as a new reliable tool to identify molecular grade signatures in breast cancer--the INNODIAG study.

BackgroundPersonalized medicine has become a priority in breast cancer patient management. In addition to the routinely used clinicopathological characteristics, clinicians will have to face an increasing amount of data derived from tumor molecular profiling. The aims of this study were to develop a new gene selection method based on a fuzzy logic selection and classification algorithm, and to validate the gene signatures obtained on breast cancer patient cohorts.MethodsWe analyzed data from four published gene expression datasets for breast carcinomas. We identified the best discriminating genes by comparing molecular expression profiles between histologic grade 1 and 3 tumors for each of the training datasets. The most pertinent probes were selected and used to define fuzzy molecular grade 1-like (good prognosis) and fuzzy molecular grade 3-like (poor prognosis) profiles. To evaluate the prognostic performance of the fuzzy grade signatures in breast cancer tumors, a Kaplan-Meier analysis was conducted to compare the relapse-free survival deduced from histologic grade and fuzzy molecular grade classification.ResultsWe applied the fuzzy logic selection on breast cancer databases and obtained four new gene signatures. Analysis in the training public sets showed good performance of these gene signatures for grade (sensitivity from 90% to 95%, specificity 67% to 93%). To validate these gene signatures, we designed probes on custom microarrays and tested them on 150 invasive breast carcinomas. Good performance was obtained with an error rate of less than 10%. For one gene signature, among 74 histologic grade 3 and 18 grade 1 tumors, 88 cases (96%) were correctly assigned. Interestingly histologic grade 2 tumors (n = 58) were split in these two molecular grade categories.ConclusionWe confirmed the use of fuzzy logic selection as a new tool to identify gene signatures with good reliability and increased classification power. This method based on artificial intelligence algorithms was successfully applied to breast cancers molecular grade classification allowing histologic grade 2 classification into grade 1 and grade 2 like to improve patients prognosis. It opens the way to further development for identification of new biomarker combinations in other applications such as prediction of treatment response.Electronic supplementary materialThe online version of this article (doi:10.1186/s12920-015-0077-1) contains supplementary material, which is available to authorized users.

The expression of four genes as a prognostic classifier for stage I lung adenocarcinoma in 12 independent cohorts.

We previously developed a prognostic classifier using the expression levels of BRCA1, HIF1A, DLC1, and XPO1 that identified stage I lung adenocarcinoma patients with a high risk of relapse. That study evaluated patients in five independent cohorts from various regions of the world. In an attempt to further validate the classifier, we have used a meta-analysis-based approach to study 12 cohorts consisting of 1,069 tumor-node-metastasis stage I lung adenocarcinoma patients from every suitable, publically available dataset. Cohorts were obtained through a systematic search of public gene expression datasets. These data were used to calculate the risk score using the previously published 4-gene risk model. A fixed effect meta-analysis model was used to generate a pooled estimate for all cohorts. The classifier was associated with prognosis in 10 of the 12 cohorts (P < 0.05). This association was highly consistent regardless of the ethnic diversity or microarray platform. The pooled estimate demonstrated that patients classified as high risk had worse overall survival for all stage I [HR, 2.66; 95% confidence interval (CI), 1.93-3.67; P < 0.0001] patients and in stratified analyses of stage IA (HR, 2.69; 95% CI, 1.66-4.35; P < 0.0001) and stage IB (HR, 2.69; 95% CI, 1.74-4.16; P < 0.0001) patients. The 4-gene classifier provides independent prognostic stratification of stage IA and stage IB patients beyond conventional clinical factors. Our results suggest that the 4-gene classifier may assist clinicians in decisions about the postoperative management of early-stage lung adenocarcinoma patients.

Accessibilome of human glioblastoma: collagen-VI-alpha-1 is a new target and a marker of poor outcome.

Functional targeted therapy has unfortunately failed to improve the outcome of glioblastoma patients. Success stories evidenced by the use of antibody-drug conjugates in other tumor types are encouraging, but targets specific to glioblastoma and accessible through the bloodstream remain scarce. In the current work, we have identified and characterized novel and accessible proteins using an innovative proteomic approach on six human glioblastomas; the corresponding data have been deposited in the PRIDE database identifier PXD001398. Among several clusters of uniquely expressed proteins, we highlight collagen-VI-alpha-1 (COL6A1) as a highly expressed tumor biomarker with low levels in most normal tissues. Immunohistochemical analysis of glioma samples from 61 patients demonstrated that COL6A1 is a significant and consistent feature of high-grade glioma. Deposits of COL6A1 were evidenced in the perivascular regions of the tumor-associated vasculature and in glioma cells found in pseudopalisade structures. Retrospective analysis of public gene-expression data sets from over 300 glioma patients demonstrated a significant correlation of poor patient outcome and high COL6A1 expression. In a proof-of-concept study, we use chicken chorioallantoic membrane in vivo model to show that COL6A1 is a reachable target for IV-injected antibodies. The present data warrant further development of human COL6A1 antibodies for assessing the quantitative biodistribution in the preclinical tumor models.

Angiogenesis interactome and time course microarray data reveal the distinct activation patterns in endothelial cells.

Angiogenesis involves stimulation of endothelial cells (EC) by various cytokines and growth factors, but the signaling mechanisms are not completely understood. Combining dynamic gene expression time-course data for stimulated EC with protein-protein interactions associated with angiogenesis (the “angiome”) could reveal how different stimuli result in different patterns of network activation and could implicate signaling intermediates as points for control or intervention. We constructed the protein-protein interaction networks of positive and negative regulation of angiogenesis comprising 367 and 245 proteins, respectively. We used five published gene expression datasets derived from in vitro assays using different types of blood endothelial cells stimulated by VEGFA (vascular endothelial growth factor A). We used the Short Time-series Expression Miner (STEM) to identify significant temporal gene expression profiles. The statistically significant patterns between 2D fibronectin and 3D type I collagen substrates for telomerase-immortalized EC (TIME) show that different substrates could influence the temporal gene activation patterns in the same cell line. We investigated the different activation patterns among 18 transmembrane tyrosine kinase receptors, and experimentally measured the protein level of the tyrosine-kinase receptors VEGFR1, VEGFR2 and VEGFR3 in human umbilical vein EC (HUVEC) and human microvascular EC (MEC). The results show that VEGFR1–VEGFR2 levels are more closely coupled than VEGFR1–VEGFR3 or VEGFR2–VEGFR3 in HUVEC and MEC. This computational methodology can be extended to investigate other molecules or biological processes such as cell cycle.

Mining cancer gene expression databases for latent information on intronic microRNAs

Around 50% of all human microRNAs reside within introns of coding genes and are usually co-transcribed. Gene expression datasets, therefore, should contain a wealth of miRNA-relevant latent information, exploitable for many basic and translational research aims. The present study was undertaken to investigate this possibility. We developed an in silico approach to identify intronic-miRNAs relevant to breast cancer, using public gene expression datasets. This led to the identification of a miRNA signature for aggressive breast cancer, and to the characterization of novel roles of selected miRNAs in cancer-related biological phenotypes. Unexpectedly, in a number of cases, expression regulation of the intronic-miRNA was more relevant than the expression of their host gene. These results provide a proof of principle for the validity of our intronic miRNA mining strategy, which we envision can be applied not only to cancer research, but also to other biological and biomedical fields.