Microarray Gene Expression Profiling Data Research Articles

Overlapping group screening for binary cancer classification with TCGA high-dimensional genomic data.

Precision medicine has been a global trend of medical development, wherein cancer diagnosis plays an important role. With accurate diagnosis of cancer, we can provide patients with appropriate medical treatments for improving patients' survival. Since disease developments involve complex interplay among multiple factors such as gene-gene interactions, cancer classifications based on microarray gene expression profiling data are expected to be effective, and hence, have attracted extensive attention in computational biology and medicine. However, when using genomic data to build a diagnostic model, there exist several problems to be overcome, including the high-dimensional feature space and feature contamination. In this paper, we propose using the overlapping group screening (OGS) approach to build an accurate cancer diagnosis model and predict the probability of a patient falling into some disease classification category in the logistic regression framework. This new proposal integrates gene pathway information into the procedure for identifying genes and gene-gene interactions associated with the classification of cancer outcome groups. We conduct a series of simulation studies to compare the predictive accuracy of our proposed method for cancer diagnosis with some existing machine learning methods, and find the better performances of the former method. We apply the proposed method to the genomic data of The Cancer Genome Atlas related to lung adenocarcinoma (LUAD), liver hepatocellular carcinoma (LHC), and thyroid carcinoma (THCA), to establish accurate cancer diagnosis models.

Genomic landscape of pancreatic cancer in the Japanese version of the Cancer Genome Atlas.

Pancreatic cancer (PC) is one of the most aggressive cancers worldwide. Although many studies have investigated genomic alterations, the genomic landscape of Japanese PC patients has not been fully elucidated. We used whole-exome sequencing, cancer gene panel deep-sequencing, and microarray gene expression profiling data derived from the Japanese version of the Cancer Genome Atlas (JCGA) in 93 PC cases. Somatic driver mutations were identified in 65.6% of samples in 19 genes. The median tumor mutation burden (TMB) value was 0.24 Muts/Mb (interquartile range, 0.15-0.64 Muts/Mb). The commonly mutated genes were KRAS (58%), TP53 (40%), CDKN2A (10%), SMAD4 (10%), FGFR2 (9%), and PKHD1 (9%). Frequent germline variation genes were BRCA1 (8%), CDH1 (5%), MET (5%), MSH6 (5%), and TEK (5%). Frequent chromosomal arm alterations included copy number gains in 2q (42%), 7q (24%), and 3q (24%), and copy number losses in 19p (62%), 19q (47%), 12q (34%), and 7q (30%). A prognostic analysis according to the presence of driver mutations showed that overall survival (OS) in the driver mutation-positive group was significantly worse in comparison to that of the driver mutation-negative group (median, 23.1 vs 46.7 mo; P=.010). A Cox proportional hazards analysis for OS identified driver mutation (hazard ratio [HR], 1.89; P=.025) and lymph node metastasis (HR, 3.27; P=.002) as independent prognostic factors. The present results from the JCGA dataset constitute a fundamental resource for genomic medicine for PC patients, especially in Japan.

Validation of ZMYND8 as a new treatment target in hepatocellular carcinoma.

ZMYND8 (Zinc finger MYND (Myeloid, Nervy and DEAF-1)-type containing 8) has been known to play an important role in tumor regulation in various types of cancer. However, the results of ZMYND8 expression and their clinical significance in hepatocellular carcinoma (HCC) have not yet been published. In the present study, we investigate the expression of ZMYND8 protein and mRNA in HCC and elucidate its prognostic significance. ZMYND8 protein and mRNA expression in 283 and 234 HCCs were investigated using immunohistochemistry and microarraygene expression profiling data. The relationships between ZMYND8 expression with clinicopathologic features and prognosis of HCC patients were evaluated. Furthermore, we performed the invasion, migration, apoptosis, soft agar formation assay and sphere formation assay in HCC cell lines, and evaluated tumorigenicity in a nude mouse model, after ZMYND8 knockdown. Overexpression of ZMYND8 protein and mRNA was observed in 20.5% and 26.9% of HCC cases, respectively. High ZMYND8 expression showed significant correlations with microvascular invasion, high Edmondson grade, advanced American Joint Committee on Cancer, and increased alpha-fetoprotein level. ZMYND8 mRNA overexpression was an independent prognostic factor for predicting early recurrence as well as short recurrence-free survival (RFS). Downregulation of ZMYND8 reduced migration and invasion of HCC cells, and promoted apoptosis of HCC cells in an in vitro model. In a xenograft nude mouse model, knockdown of ZMYND8 significantly reduced tumor growth. ZMYND8 mRNA overexpression could be a prognostic marker of shorter RFS in HCC patients after curative resection. ZMYND8 might play an important role in the proliferation and progression of HCC and could be a promising candidate for targeted therapy.

A Bioinformatic Approach for the Identification of Molecular Determinants of Resistance/Sensitivity to Cancer Thermotherapy.

Application of heat above 43°C and up to 47°C, the so-called “thermal ablation” range, leads to tumor cell destruction either by apoptosis or by necrosis. However, tumor cells have developed mechanisms of defense that render them thermoresistant. Of importance, the in situ application of heat for the treatment of localized solid tumors can also prime specific antitumor immunity. Herein, a bioinformatic approach was employed for the identification of molecular determinants implicated in thermoresistance and immunogenic cell death (ICD). To this end, both literature-derived (text mining) and microarray gene expression profile data were processed, followed by functional enrichment analysis. Two important functional gene modules were detected in hyperthermia resistance and ICD, the former including members of the heat shock protein (HSP) family of molecular chaperones and the latter including immune-related molecules, respectively. Of note, the molecules HSP90AA1 and HSPA4 were found common between thermoresistance and damage signaling molecules (damage-associated molecular patterns (DAMPs)) and ICD. In addition, the prognostic potential of HSP90AA1 and HSPA4 overexpression for cancer patients' overall survival was investigated. The results of this study could constitute the basis for the strategic development of more efficient and personalized therapeutic strategies against cancer by means of thermotherapy, by taking into consideration the genetic profile of each patient.

Development of predictive models for cervical cancer based on gene expression profiling data

Cervical cancer and the prediction of clinical outcome are among the most important emerging applications of gene expression microarray technology with feature sequencing of microRNA. By using reliable and dependable classification of machine learning algorithms available for microarray gene expression profiling data is the key in order to develop the most suitable and possible predictive model to be used by patient. In this paper, two-machine learning algorithms have been used which are Support Vector Machine (SVM) and Random Forests (RF) for the predictive models of cervical cancer. We identify and evaluate the performance of these two algorithms in order to know which algorithm has better performance. In this study, 714 features and 58 samples are used to develop predictive model for cervical cancer and our computational results show that RF algorithm outperform SVM algorithm with the accuracy of 94.21%. Our data also underline the importance of variables, which give the significant role in predicting the occurrence of cervical cancer.

Reverse Engineering of Genome-wide Gene Regulatory Networks from Gene Expression Data.

Transcriptional regulation plays vital roles in many fundamental biological processes. Reverse engineering of genome-wide regulatory networks from high-throughput transcriptomic data provides a promising way to characterize the global scenario of regulatory relationships between regulators and their targets. In this review, we summarize and categorize the main frameworks and methods currently available for inferring transcriptional regulatory networks from microarray gene expression profiling data. We overview each of strategies and introduce representative methods respectively. Their assumptions, advantages, shortcomings, and possible improvements and extensions are also clarified and commented.

Abstract 667: Prognostic value and functional impact of aberrant GABRE∼miR-452∼miR-224 promoter methylation in prostate cancer.

Abstract Prostate cancer (PC) is a major cause of morbidity and mortality among men in Western countries. Currently available tools for PC diagnosis and prognosis are suboptimal and novel biomarkers are urgently needed. This study investigated the regulation, function, and PC biomarker potential of a genomic locus encompassing miR-224, miR-452, and their host gene gamma-aminobutyric acid (GABA) A receptor family epsilon subunit (GABRE). First, mRNA and miRNA expression profiling data for three independent patient sets, including a total of 80 nonmalignant and 281 PC tissue samples, showed significant downregulation of all three genes in PC. Next, we asked if this locus was silenced in PC by aberrant hypermethylation of its promoter-associated CpG island. Indeed, bisulfite sequencing of 11 nonmalignant and 21 PC tissue samples showed frequent cancer-specific hypermethylation, which was successfully validated in an independent set of 35 nonmalignant and 245 PC tissue samples by methylation-specific qPCR. Notably, GABRE promoter methylation distinguished PC from nonmalignant prostate with 96% sensitivity and 94% specificity (AUC=0.98). Consistent with an epigenetic silencing mechanism, methylation levels correlated inversely with GABRE, miR-224, and miR-452 expression in both clinical samples and prostatic cell lines. By univariate cox regression analysis, high GABRE methylation was significantly (p=0.011, p=0.015, p=0.002) associated with PSA recurrence after radical prostatectomy in three independent PC patient cohorts (n=186, n=129, n=148). Moreover, in multivariate models, GABRE methylation added significant independent prognostic value to established clinicopathological predictors in two cohorts (p=0.031 and p=0.020) and was borderline significant in the smaller third cohort (p=0.073). In line with the significant adverse prognostic value of GABRE promoter methylation, functional studies in vitro suggested that silencing of this locus may lead to increased malignancy of PC cells. Thus, ectopic expression of miR-224 and miR-452 mimics, respectively, inhibited proliferation, migration, and invasion of two PC cell lines (PC3 and DU145) with undetectable endogenous levels of both miRNAs. We are currently analyzing post-transfection microarray gene expression profiling data to identify target genes mediating the tumor suppressive functions of miR-224 and miR-452; results will be presented at the conference. In conclusion, our results showed that the GABRE∼miR-452∼miR-224 locus is frequently silenced in PC by aberrant promoter hypermethylation and, furthermore, is a new promising epigenetic candidate biomarker for PC diagnosis and prognosis. Also, the tumor suppressive functions of miR-224 and miR-452 demonstrated in two PC cell lines may suggest this locus as a potential target for epigenetic therapy of PC. Citation Format: Helle Kristensen, Christa Haldrup, Martin M. Mortensen, Michael Borre, Søren Høyer, Peter J. Wild, Christian Arsov, Wolfgang Goering, Tapio Visakorpi, Torben F. Ørntoft, Karina D. Sørensen. Prognostic value and functional impact of aberrant GABRE∼miR-452∼miR-224 promoter methylation in prostate cancer. [abstract]. In: Proceedings of the 104th Annual Meeting of the American Association for Cancer Research; 2013 Apr 6-10; Washington, DC. Philadelphia (PA): AACR; Cancer Res 2013;73(8 Suppl):Abstract nr 667. doi:10.1158/1538-7445.AM2013-667

Abstract A1: SNAIL inhibitor GN-25 induces MET through Global Reprograming of EMT Network Genes: a Systems and Network Analysis

Abstract Epithelial-to-mesenchymal transition (EMT) is a complex mechanism that arises through a cohesive and intricate biological cross-talk between numerous cancer-related networks. Among them sonic hedgehog directed snail network is one of the major components that drives EMT. This has led to strategies targeting either hedgehog signaling or snail protein using highly specific agents. However, it is needless to say that removal of individual components through single protein-targeted drug (here snail) from robust EMT networks may surprisingly have little functional consequence in cancer. We propose that, to have an impact, interventions within EMT network need to be multiple but highly selective. Further, predicting the functional outcome of interventions, such as those originating from single pathway targeted drug treatment, is not as straightforward as originally thought and computational/integrated tools are needed. The emergence of computational network biology has enhanced our knowledge of multi-pathway interactions in cancer and helped in making sense of drug response signature datasets that collectively decode the complex mechanisms of drug action. These holistic approaches have laid the foundation of the emerging concepts of “network pharmacology” that is solidifying its position in cancer medicine. Unlike earlier reductionist one-drug, one-target approaches, network pharmacology invokes the idea of drug promiscuity often engages a fortuitous combination of appropriate high-value targets in cancer cells to produce treatment success. Network pharmacology is being used in the identification of weak nodes in global cancer networks, predicting drug toxicity, drug repurposing, identifying multiscale mechanisms of drug action and in the rational design of potent anticancer drug combinations. To this end, we had earlier developed a novel snail-p53 interaction inhibitor (GN-25) that was shown to suppress snail function. However, this drug was not investigated for its effects against EMT pathways. Here using network pharmacology drug repurposing principles, we show that our novel SNAIL inhibitor (GN-25), through its promiscuous action induces mesenchymal-to-epithelial transition (MET) in well-recognized HMLE-snail cell line model developed by Dr. Weinberg. Network modeling of microarray gene expression profiling data from GN-25 treated cells showed an orchestrated global reorganization of EMT network genes. Most importantly, the expression signatures were validated both at the mRNA and protein expression levels, and mechanistically studied by RNA interference which provided unequivocal evidence on the role of GN-25 targeting EMT specific genes. This proof-of-concept investigation shows that computational biology and network pharmacology can be very powerful tools for rational design and development of agents that are typically promiscuous like the snail inhibitor as a candidate EMT-targeting drug. Citation Format: Asfar S. Azmi, Ramzi M. Mohammad, Fazlul H. Sarkar. SNAIL inhibitor GN-25 induces MET through Global Reprograming of EMT Network Genes: a Systems and Network Analysis [abstract]. In: Proceedings of the AACR Special Conference on Chemical Systems Biology: Assembling and Interrogating Computational Models of the Cancer Cell by Chemical Perturbations; 2012 Jun 27-30; Boston, MA. Philadelphia (PA): AACR; Cancer Res 2012;72(13 Suppl):Abstract nr A1.

Interleukin‐6 maintains bone marrow‐derived mesenchymal stem cell stemness by an ERK1/2‐dependent mechanism

Adult human mesenchymal stem cells (MSCs) hold promise for an increasing list of therapeutic uses due to their ease of isolation, expansion, and multi-lineage differentiation potential. To maximize the clinical potential of MSCs, the underlying mechanisms by which MSC functionality is controlled must be understood. We have taken a deconstructive approach to understand the individual components in vitro, namely the role of candidate "stemness" genes. Our recent microarray gene expression profiling data suggest that interleukin-6 (IL-6) may contribute to the maintenance of MSCs in their undifferentiated state. In this study, we showed that IL-6 gene expression is significantly higher in undifferentiated MSCs as compared to their chondrogenic, osteogenic, and adipogenic derivatives. Moreover, we found that MSCs secrete copious amounts of IL-6 protein, which decreases dramatically during osteogenic differentiation. We further evaluated the role of IL-6 for maintenance of MSC "stemness," using a series of functional assays. The data showed that IL-6 is both necessary and sufficient for enhanced MSC proliferation, protects MSCs from apoptosis, inhibits adipogenic and chondrogenic differentiation of MSCs, and increases the rate of in vitro wound healing of MSCs. We further identified ERK1/2 activation as the key pathway through which IL-6 regulates both MSC proliferation and inhibition of differentiation. Taken together, these findings show for the first time that IL-6 maintains the proliferative and undifferentiated state of bone marrow-derived MSCs, an important parameter for the optimization of both in vitro and in vivo manipulation of MSCs.

EVI1 Overexpression in Pediatric Acute Myeloid Leukemia Associated with Unfavorable Subtypes.

The EVI1 (ecotropic virus integration-1) gene plays an important role in hematopoiesis especially in megakaryocyte development. The MDS1 gene is located upstream of EVI1, and its function is currently unknown. Normally the MDS1/EVI1 intergenic splice variant is co-expressed with EVI1. In adult acute myeloid leukemia (AML) overexpression of EVI1 (EVI1+) can be found in patients with chromosome 3q26-rearrangements. Often, these patients do not co-express MDS1/EVI1. Recently high EVI1 expression was also discovered in a separate subgroup of patients that did not have 3q26-rearrangements. Occasionally, they did not show overexpression of MDS1/EVI1. In these patients cryptic inversions of chromosome 3 were identified with fluorescence in situ hybridization (FISH). Of interest, EVI1+ was found to be an independent poor prognostic marker in adult AML (Lugthart et al, Blood 2008). In pediatric AML, 3q26-rearrangements are rare and the role of EVI1 is unknown. In this study, we investigated the frequency and clinical relevance of EVI1+ in pediatric AML. EVI1 expression was analyzed in 233 pediatric AML patients, of whom microarray gene expression profiling data were available. EVI1+ was found in 25 pediatric AML patients (11%), and confirmed with real-time quantitative PCR. This included 13/49 (26%) patients with MLL-rearranged AML: 5/22 (23%) cases with t(9;11); and all (n=4) cases with t(6;11). Moreover, EVI1+ was found in 4/7 (57%) cases with AML M7; in 2/3 (66%) cases with AML M6; in both cases with monosomy 7; in 1/43 (2%) cases with normal cytogenetics; in 2 patients with random cytogenetics, and in 1 patient with a cytogenetic failure. EVI1+ was not found in the t(8;21), inv(16) and t(15;17) subgroups. 3/25 EVI1+ patients lacked the MDS/EVI1 transcript, but no cryptic 3q26-rearrangements were detected with FISH. Molecular analysis showed that one patient had a CEBPα mutation; one patient had an FLT3-ITD; and 3 patients showed a mutation in the RAS oncogene. EVI+ was not correlated with sex or white blood cell count. However, the frequency in children younger than 10 years old was twice as high when compared to older children (14% vs 7%, p=0.12). Survival analysis was restricted to the subset of patients who were treated using uniform DCOG and BFM treatment protocols (n=204). In this cohort, EVI1+ patients had a worse 5-years event-free survival (pEFS) compared to patients without EVI1+ (30 vs. 43%, p=0.02). However, multivariate analysis, including cytogenetics (favorable [t(8;21, inv(16), t(15;17)] vs. other), FLT3-ITD, age and WBC, showed that EVI1+ was not an independent prognostic factor for survival. Moreover, within the unfavorable/normal cytogenetic subgroup, there was no difference in outcome between patients with and without EVI1+. We conclude that EVI1+ is found in ~10% of pediatric AML, and highly correlated with specific unfavorable cytogenetic (MLL-rearrangements) and morphologic (FAB M6/7) subtypes. In contrast to adult AML, no 3q26-rearrangements or cryptic inversions were found, and EVI1+ was not an independent prognostic factor. This difference in prognostic relevance may be due to differences in treatment. Alternatively, these results may indicate that EVI1 plays a different role in disease biology between adult and pediatric AML. This is at least suggested by the lack of 3q26 aberrations in pediatric AML.

A trend pattern assessment approach to microarray gene expression profiling data analysis

We study the problem of how to assess the reliability of a statistical measurement on data set containing unknown quantity of noises, inconsistencies, and outliers. A practical approach that analyzes the dynamical patterns (trends) of the statistical measurements through a sequential extreme-boundary-points (EBP) weed-out process is explored. We categorize the weed-out trend patterns (WOTP) and examine their relation to the reliability of the measurement. The approach is applied to the processes of extracting genes that are predictive to BCL2 translocations and to clinical survival outcomes of diffuse large B-cell lymphoma (DLBCL) from DNA Microarray gene expression profiling data sets. Fisher’s Discriminate Criterion (FDC) is used as a statistical measurement in the processes. It is found that the weed-out trend analysis (WOTA) approach is effective for qualitatively assessing the statistics-based measurements in the experimentations conducted.

Triple-negative breast cancer: therapeutic options

Triple-negative breast cancers are defined by a lack of expression of oestrogen, progesterone, and ERBB2 receptors. This subgroup accounts for 15% of all types of breast cancer and for a higher percentage of breast cancer arising in African and African-American women who are premenopausal. Because of the absence of specific treatment guidelines for this subgroup, triple-negative breast cancers are managed with standard treatment; however, such treatment leaves them associated with a high rate of local and systemic relapse. Histologically, such cancers are poorly differentiated, and most fall into the basal subgroup of breast cancers, characterised by staining for basal markers (ie, cytokeratin 5/6). Analyses of microarray gene-expression profiling data show that they form a homogeneous group (or so-called cluster) in transcriptional terms and, increasingly, research studies are identifying basal cancers on the basis of exhibiting this distinctive transcriptional profile. Histologically and transcriptionally, triple-negative breast cancers have many similarities to BRCA1-associated breast cancers, which suggests that dysfunction in BRCA1 or related pathways occurs in this subset of sporadic cancers. In this review, we discuss the molecular features of triple-negative breast cancers and consider how the use of existing cytotoxic agents can be optimised for this patient group. We discuss the implications of a possible underlying BRCA1-pathway dysfunction in this subgroup in terms of treatment and we also investigate the predominant proliferative signals and the on-going research addressing the suitability of these signals as therapeutic targets.

Gene Regulation Networks Related to Neural Differentiation of hESC

With the unique property of self-renewal and developmental pluripotency, human embryonic stem cells (hESC) provide an opportunity to study molecular aspects of developmental biology. Understanding gene regulation of hESC pluripotency is a critical step toward directing hESC differentiation for regenerative medicine. However, currently little is known about hESC gene regulation of hESC pluripotency. Applying network analysis to microarray gene expression profiling data, we compared gene expression profiles from pluripotent hESC to hESC-derived astrocytes and identified potential gene regulation networks. These gene regulation networks suggest that hECS has stringent control of cell cycle and apoptosis. Our data reveal several potential hESC differentiation biomarkers and suggest that IGF2 and A2M could play a role in hESC pluripotency by altering the availability of cytokines at the local environment of hECS. These findings underscore the importance of network analysis among differentially expressed genes, and should facilitate future study for understanding the gene regulation of hESC pluripotency.

DISCOVERY OF HIGHLY DIFFERENTIATIVE GENE GROUPS FROM MICROARRAY GENE EXPRESSION DATA USING THE GENE CLUB APPROACH

It is commonly believed that suitable analysis of microarray gene expression profile data can lead to better understanding of diseases, and better ways to diagnose and treat diseases. To achieve those goals, it is of interest to discover the gene interaction networks, and perhaps even pathways, underlying given diseases from such data. In this paper, we consider methods for efficiently discovering highly differentiative gene groups (HDGG), which may provide insights on gene interaction networks. HDGGs are groups of genes which completely or nearly completely characterize the diseased or normal tissues. Discovering HDGGs is challenging, due to the high dimensionality of the data. Our methods are based on the novel concept of gene clubs. A gene club consists of a set of genes having high potential to be interactive with each other. The methods can (i) efficiently discover signature HDGGs which completely characterize the diseased and the normal tissues respectively, (ii) find strongest or near strongest HDGGs containing any given gene, and (iii) find much stronger HDGGs than previous methods. As part of the experimental evaluation, the methods are applied to colon, prostate, ovarian, and breast cancer, and leukemia and so on. Some of the genes in the extracted signature HDGGs have known biological functions, and some have attracted little attention in biology and medicine. We hope that appropriate study on them can lead to medical breakthroughs. Some HDGGs for colon and prostate cancers are listed here. The website listed below contains HDGGs for the other cancers. HDGG is implemented in C++ and runs on Unix or Windows platform. The code is available at: http://www.cs.wright.edu/~gdong/hdgg/.

Mapping of candidate tumor suppressor genes on chromosome 12 in adenoid cystic carcinoma

Adenoid cystic carcinoma (ACC) is a common malignancy of salivary glands, for which the underlying genetic mechanisms of tumorigenesis are poorly understood. Prior studies in ACC have identified deletions in chromosome 12. To further characterize these changes, we performed an extensive LOH analysis in 58 ACC using a panel of 28 microsatellite markers. Results show 66% overall genetic loss. Three markers (D12S1713, D12S2196, D12S398) are contiguous and define a 6.84 Mb region of deletion at 12q13.11–q13.13. Two other markers (D12S2078, D12S1628) are also contiguous and define a 4.5 Mb region of deletion at 12q24.32–q24.33. The three remaining markers, D12S1056 at 12q14.1, D12S1051 at 12q23.1 and D12S1636 at 12q23.3 define smaller regions of deletion. An analysis of microarray gene expression profiling data available for ACC shows several genes with significant transcriptional downregulation that map to these areas of genetic deletion. This combined genetic and genomic analysis provides several candidate genes to test for functional tumor suppressor activity in ACC.

Use of gene networks for identifying and validating drug targets.

We propose a new method for identifying and validating drug targets by using gene networks, which are estimated from cDNA microarray gene expression profile data. We created novel gene disruption and drug response microarray gene expression profile data libraries for the purpose of drug target elucidation. We use two types of microarray gene expression profile data for estimating gene networks and then identifying drug targets. The estimated gene networks play an essential role in understanding drug response data and this information is unattainable from clustering methods, which are the standard for gene expression analysis. In the construction of gene networks, we use the Bayesian network model. We use an actual example from analysis of the Saccharomyces cerevisiae gene expression profile data to express a concrete strategy for the application of gene network information to drug discovery.