Gene expression arrays in cancer research: methods and applications

The influence of LncRNA H19 polymorphic variants on susceptibility to cancer: A systematic review and updated meta-analysis of 28 case-control studies

OBJECTIVE

Although myriad researches upon the associations between LncRNA H19 polymorphic variants (rs2839698 G>A, rs217727 G>A, rs2107425 C>T, rs2735971 A>G and rs3024270 C>G) and the susceptibility to cancer have been conducted, these results remained contradictory and perplexing. Basing on that, a systematic review and updated meta-analysis was performed to anticipate a fairly precise assessment about such associations.

METHODS

We retrieved the electronic databases EMBASE, PubMed and Web of Science for valuable academic studies before February 28, 2021. Ultimately, 28 of which were encompassed after screening in this meta-analysis, and the available data was extracted and integrated. The pooled odds ratios (ORs) with 95% confidence intervals (CIs) was used to evaluate such associations. For multi-level investigation, subgroup analysis derived from source of controls together with genotypic method was preformed.

RESULTS

Eventually, 28 articles altogether embodying 57 studies were included in this meta-analysis. The results illuminated that LncRNA H19 polymorphisms mentioned above were all irrelevant to cancer susceptibility. Nevertheless, crucial results were found concentrated in population-based control group when subgroup analysis by source of controls were performed in H19 mutation rs2839698 and rs2735971. Meanwhile, in the stratification analysis by genotypic method, apparent cancer risks were discovered by TaqMan method in H19 mutation rs2107425 and rs3024270. Then, trial sequential analysis demonstrated that the results about such associations were firm evidence of effect.

CONCLUSION

Therefore, this meta-analysis indicated that LncRNA H19 polymorphisms were not associated with the susceptibility to human cancer. However, after the stratification analysis, inconsistent results still existed in different genotypic method and source of control. Thus, more high-quality studies on cancer patients of different factors were needed to confirm these findings.

Logistic regression paradigm for training a single-hidden layer feedforward neural network. Application to gene expression datasets for cancer research

OBJECTIVE

The speed of the diagnosis process is vital in pursuing the trial of curing cancer. During the last decade, precision medicine evolved by detecting different types of cancer through microarrays (MA) of deoxyribonucleic acid (DNA) processed by machine learning (ML) algorithms. Personalized diagnosis, followed by personalized treatment, should imply personalized hyperparameters of the ML. The goal of this paper is to propose a novel adaptive ML method that embeds knowledge into the architecture of the algorithm and also filters the features in order to reduce their number, increase computational speed, and decrease computational cost and time.

MATERIALS AND METHODS

fLogSLFN is a novel two-fold theoretically effective ML that can be used in two-class decision problems that embeds the logistic regression in such a manner that the hidden nodes of a single-hidden layer feedforward neural network (SLFN) are problem dependent. A filtering module based on the significance of each attribute is embedded in order to avoid the 'curse of dimensionality' phenomenon. The proposed model has been tested on three publicly available high-dimensional cancer datasets that contain gene expressions provided by complementary DNA (cDNA) array, and DNA microarray. The proposed novel method filtered logistic SLFN (fLogSLFN) has been also compared and statistically benchmarked to four ML algorithms: extreme learning machine (ELM), radial basis function network (RBF), single-hidden layer feedforward neural network trained by the backpropagation algorithm (BPNN), logistic regression with the LASSO penalty, and the adaptive single-hidden layer feedforward network (aSLFN).

MAIN FINDINGS

The experimental results showed that the fLogSLFN is competitive to the other state-of-the-art models, obtaining accuracies between 64.70% and 98.66% depending on the dataset it had been applied on.

CONCLUSIONS

In contrast to other state-of-the-art ML algorithms, the fLogSLFN is capable to embed the knowledge extracted from the data into its architecture, making it problem dependent. The filtering module increases its computational speed, while decreasing computational cost and time. The statistical analysis revealed the fact that by filtering the features the performance is kept, making the algorithm more efficient.

NDRG4 promoter hypermethylation is a mechanistic biomarker associated with metastatic progression in breast cancer patients

The risk of developing metastatic disease in breast cancer patients is traditionally predictable based on the number of positive axillary lymph nodes, complemented with additional clinicopathological factors. However, since lymph node-negative patients have a 20-30% probability of developing metastatic disease, lymph node information alone is insufficient to accurately assess individual risk. Molecular approaches, such as multigene expression panels, analyze a set of cancer-related genes that more accurately predict the early risk of metastasis and the treatment response. Here, we present N-Myc downstream-regulated gene 4 (NDRG4) epigenetic silencing as a mechanistic biomarker of metastasis in ductal invasive breast tumors. While aberrant NDRG4 DNA hypermethylation is significantly associated with the development of metastatic disease, downregulation of NDRG4 transcription and protein expression is functionally associated with enhanced lymph node adhesion and cell mobility. Here, we show that epigenetic silencing of NDRG4 modulates integrin signaling by assembling β1-integrins into large punctate clusters at the leading edge of tumor cells to promote an "adhesive switch," decreasing cell adhesion to fibronectin and increasing cell adhesion and migration towards vitronectin, an important component of human lymph nodes. Taken together, our functional and clinical observations suggest that NDRG4 is a potential mechanistic biomarker in breast cancer that is functionally associated with metastatic disease.

Association between 8q24 Gene Polymorphisms and the Risk of Prostate Cancer: A Systematic Review and Meta-Analysis

Though numerous studies have been conducted to investigate the associations between five 8q24 polymorphisms (rs6983267 T>G, rs1447295 C>A, rs16901979 C>A, rs6983561 A>C and rs10090154 C>T) and prostate cancer (PCa) risk, the available results remained contradictory. Therefore, we performed a comprehensive meta-analysis to derive a precise estimation of such associations. We searched electronic databases PubMed, EMBASE, Web of Science and Wan Fang for the relevant available studies up to February 1st, 2017, and 39 articles were ultimately adopted in this meta-analysis. All data were extracted independently by two investigators and recorded in a unified form. The strength of association between 8q24 polymorphisms and PCa susceptibility was evaluated by the pooled odds ratios (ORs) with 95% confidence intervals (CIs). Subgroup analysis was conducted based on ethnicity, source of controls and genotypic method. Overall, a total of 39 articles containing 80 studies were adopted in this meta-analysis. The results of this meta-analysis indicated that five 8q24 polymorphisms above were all related to PCa susceptibility. Besides, in the subgroup analysis by ethnicity, all selected 8q24 polymorphisms were significantly associated with PCa risk in Asian population. In addition, stratification analysis by source of controls showed that significant results were mostly concentrated in the studies' controls from general population. Moreover, when stratified by genotypic method, significant increased PCa risks were found by TaqMan method. Therefore, this meta-analysis demonstrated that 8q24 polymorphisms (rs6983267 T>G, rs1447295 C>A, rs16901979 C>A, rs6983561 A>C and rs10090154 C>T) were associated with the susceptibility to PCa, which held the potential biomarkers for PCa risk.

Epithelial cells captured from ductal carcinoma in situ reveal a gene expression signature associated with progression to invasive breast cancer

Breast cancer biomarkers that can precisely predict the risk of progression of non-invasive ductal carcinoma in situ (DCIS) lesions to invasive disease are lacking. The identification of molecular alterations that occur during the invasion process is crucial for the discovery of drivers of transition to invasive disease and, consequently, biomarkers with clinical utility. In this study, we explored differences in gene expression in mammary epithelial cells before and after the morphological manifestation of invasion, i.e., early and late stages, respectively. In the early stage, epithelial cells were captured from both pre-invasive lesions with distinct malignant potential [pure DCIS as well as the in situ component that co-exists with invasive breast carcinoma lesions (DCIS-IBC)]; in the late stage, epithelial cells were captured from the two distinct morphological components of the same sample (in situ and invasive components). Candidate genes were identified using cDNA microarray and rapid subtractive hybridization (RaSH) cDNA libraries and validated by RT-qPCR assay using new samples from each group. These analyses revealed 26 genes, including 20 from the early and 6 from the late stage. The expression profile based on the 20 genes, marked by a preferential decrease in expression level towards invasive phenotype, discriminated the majority of DCIS samples. Thus, this study revealed a gene expression signature with the potential to predict DCIS progression and, consequently, provides opportunities to tailor treatments for DCIS patients.

Translating colorectal cancer genetics into clinically useful biomarkers

Colorectal cancer (CRC) is a major health problem worldwide accounting for over a million deaths annually. While many patients with Stage II and III CRC can be cured with combinations of surgery, radiotherapy and chemotherapy, this is morbid costly treatment and a significant proportion will suffer recurrence and eventually die of CRC. Increased understanding of the molecular pathogenesis of CRC has the potential to identify high risk patients and target therapy more appropriately. Despite increased understanding of the molecular events underlying CRC development, established molecular techniques have only produced a limited number of biomarkers suitable for use in routine clinical practice to predict risk, prognosis and response to treatment. Recent rapid technological developments, however, have made genomic sequencing of CRC more economical and efficient, creating potential for the discovery of genetic biomarkers that have greater diagnostic, prognostic and therapeutic capabilities for the management of CRC. This paper reviews the current understanding of the molecular pathogenesis of CRC, and summarizes molecular biomarkers that surgeons will encounter in current clinical use as well as those under development in clinical and preclinical trials. New molecular technologies are reviewed together with their potential impact on the understanding of the molecular pathogenesis of CRC and their potential clinical utility in classification, diagnosis, prognosis and targeting of therapy.

Transcriptional profile of fibroblasts obtained from the primary site, lymph node and bone marrow of breast cancer patients

Cancer-associated fibroblasts (CAF) influence tumor development at primary as well as in metastatic sites, but there have been no direct comparisons of the transcriptional profiles of stromal cells from different tumor sites. In this study, we used customized cDNA microarrays to compare the gene expression profile of stromal cells from primary tumor (CAF, n = 4), lymph node metastasis (N+, n = 3) and bone marrow (BM, n = 4) obtained from breast cancer patients. Biological validation was done in another 16 samples by RT-qPCR. Differences between CAF vs N+, CAF vs BM and N+ vs BM were represented by 20, 235 and 245 genes, respectively (SAM test, FDR < 0.01). Functional analysis revealed that genes related to development and morphogenesis were overrepresented. In a biological validation set, NOTCH2 was confirmed to be more expressed in N+ (vs CAF) and ADCY2, HECTD1, HNMT, LOX, MACF1, SLC1A3 and USP16 more expressed in BM (vs CAF). Only small differences were observed in the transcriptional profiles of fibroblasts from the primary tumor and lymph node of breast cancer patients, whereas greater differences were observed between bone marrow stromal cells and the other two sites. These differences may reflect the activities of distinct differentiation programs.

Estrogen-responsive genes overlap with triiodothyronine-responsive genes in a breast carcinoma cell line

It has been well established that estrogen plays an important role in the progression and treatment of breast cancer. However, the role of triiodothyronine (T₃) remains controversial. We have previously shown its capacity to stimulate the development of positive estrogen receptor breast carcinoma, induce the expression of genes (PR, TGF-alpha) normally stimulated by estradiol (E₂), and suppress genes (TGF-beta) normally inhibited by E₂. Since T₃ regulates growth hormones, metabolism, and differentiation, it is important to verify its action on other genes normally induced by E₂. Therefore, we used DNA microarrays to compare gene expression patterns in MCF-7 breast adenocarcinoma cells treated with E₂ and T₃. Several genes were modulated by both E₂ and T₃ in MCF-7 cells (Student's t-test, P < 0.05). Specifically, we found eight genes that were differentially expressed after treatment with both E₂ and T₃, including amphiregulin, fibulin 1, claudin 6, pericentriolar material 1, premature ovarian failure 1B, factor for adipocyte differentiation-104, sterile alpha motif domain containing 9, and likely ortholog of rat vacuole membrane protein 1 (fold change > 2.0, pFDR < 0.05). We confirmed our microarray results by real-time PCR. Our findings reveal that certain genes in MCF-7 cells can be regulated by both E₂ and T₃.

TWIST1 is a molecular marker for a poor prognosis in oral cancer and represents a potential therapeutic target

BACKGROUND

Locoregional recurrence and distant metastases are ominous events in patients with advanced oral squamous cell carcinoma (OSCC). The objective of this study was to identify functional biomarkers that are predictive of OSCC progression to metastasis.

METHODS

The expression profile of a network of epithelial-mesenchymal transition (EMT) genes was investigated in a large cohort of patients with progressive OSCC using a complimentary DNA microarray platform coupled to quantitative reverse transcriptase-polymerase chain reaction and immunohistochemical analyses. Therapeutic potential was investigated in vitro and in vivo using an orthotopic mouse model of metastatic OSCC growing in the tongue microenvironment.

RESULTS

Among deregulated EMT genes, the Twist-related protein 1 (TWIST1) transcription factor and several of its regulated genes were significantly overexpressed across advanced stages of OSCC. This result was corroborated by the clinical observation that Twist1 up-regulation predicted the occurrence of lymph node and lung metastases as well as poor patient survival. In support of Twist1 as a driver of OSCC progression, the up-regulation of Twist1 was observed in cells isolated from patients with metastatic OSCC. The inhibition of Twist1 in these metastatic cells induced a potent inhibition of cell invasiveness in vitro as well as progression in vivo.

CONCLUSIONS

The current results provide evidence for the prognostic value and therapeutic potential of a network of Twist genes in patients with advanced OSCC.

Contextual data integration in drug discovery

INTRODUCTION

The interpretation of high-throughput profiling data depends on the pathway analysis database. Currently, pathway analysis often has to rely on a set of interactions and pathways measured in every possible human tissue, due to insufficient knowledge about interactions and pathways in the context of the profiling experiment. However, a recent global scale analysis of human tissue proteomes and interactomes reveals significant differences among tissues, suggesting that interaction and pathway data that are used out of biological context are the major source of inaccuracies and noise in the analysis of profiling data.

AREAS COVERED

In this review, the major classes of biological context used for experimental detection of molecular interactions and pathways in molecular biology are described. Furthermore, the author reviews methods for predicting biological interactions in order to evaluate the applicability of various contextual interaction data in pathway analysis. Using the results from recent publications that study large-scale tissue composition, the article provides an estimation of the gain in pathway analysis accuracy if only the interactions predicted for the context of a molecular profiling experiment are used, relative to the analysis performed with a context-independent knowledge base.

EXPERT OPINION

It is of the author's opinion that the major source of inaccuracy in pathway analysis is the lack of knowledge about tissue-specific transcriptional regulation. It is therefore suggested that the accuracy of the analysis can be substantially improved if only context-specific interactions and pathways are used for interpretation.

Amyloid-β oligomers induce differential gene expression in adult human brain slices

Cognitive decline in Alzheimer disease (AD) is increasingly attributed to the neuronal impact of soluble oligomers of the amyloid-β peptide (AβOs). Current knowledge on the molecular and cellular mechanisms underlying the toxicity of AβOs stems largely from rodent-derived cell/tissue culture experiments or from transgenic models of AD, which do not necessarily recapitulate the complexity of the human disease. Here, we used DNA microarray and RT-PCR to investigate changes in transcription in adult human cortical slices exposed to sublethal doses of AβOs. The results revealed a set of 27 genes that showed consistent differential expression upon exposure of slices from three different donors to AβOs. Functional classification of differentially expressed genes revealed that AβOs impact pathways important for neuronal physiology and known to be dysregulated in AD, including vesicle trafficking, cell adhesion, actin cytoskeleton dynamics, and insulin signaling. Most genes (70%) were down-regulated by AβO treatment, suggesting a predominantly inhibitory effect on the corresponding pathways. Significantly, AβOs induced down-regulation of synaptophysin, a presynaptic vesicle membrane protein, suggesting a mechanism by which oligomers cause synapse failure. The results provide insight into early mechanisms of pathogenesis of AD and suggest that the neuronal pathways affected by AβOs may be targets for the development of novel diagnostic or therapeutic approaches.

Temporal blastemal cell gene expression analysis in the kidney reveals new Wnt and related signaling pathway genes to be essential for Wilms' tumor onset

Wilms' tumors (WTs) originate from metanephric blastema cells that are unable to complete differentiation, resulting in triphasic tumors composed of epithelial, stromal and blastemal cells, with the latter harboring molecular characteristics similar to those of the earliest kidney development stages. Precise regulation of Wnt and related signaling pathways has been shown to be crucial for correct kidney differentiation. In this study, the gene expression profile of Wnt and related pathways was assessed in laser-microdissected blastemal cells in WTs and differentiated kidneys, in human and in four temporal kidney differentiation stages (i.e. E15.5, E17.5, P1.5 and P7.5) in mice, using an orthologous cDNA microarray platform. A signaling pathway-based gene signature was shared between cells of WT and of earliest kidney differentiation stages, revealing genes involved in the interruption of blastemal cell differentiation in WT. Reverse transcription-quantitative PCR showed high robustness of the microarray data demonstrating 75 and 56% agreement in the initial and independent sample sets, respectively. The protein expression of CRABP2, IGF2, GRK7, TESK1, HDGF, WNT5B, FZD2 and TIMP3 was characterized in WTs and in a panel of human fetal kidneys displaying remarkable aspects of differentiation, which was recapitulated in the tumor. Taken together, this study reveals new genes candidate for triggering WT onset and for therapeutic treatment targets.

Gene network analyses point to the importance of human tissue kallikreins in melanoma progression

Background

A wide variety of high-throughput microarray platforms have been used to identify molecular targets associated with biological and clinical tumor phenotypes by comparing samples representing distinct pathological states.

Methods

The gene expression profiles of human cutaneous melanomas were determined by cDNA microarray analysis. Next, a robust analysis to determine functional classifications and make predictions based on data-oriented hypotheses was performed. Relevant networks that may be implicated in melanoma progression were also considered.

Results

In this study we aimed to analyze coordinated gene expression changes to find molecular pathways involved in melanoma progression. To achieve this goal, ontologically-linked modules with coordinated expression changes in melanoma samples were identified. With this approach, we detected several gene networks related to different modules that were induced or repressed during melanoma progression. Among them we observed high coordinated expression levels of genes involved in a) cell communication (KRT4, VWF and COMP); b) epidermal development (KLK7, LAMA3 and EVPL); and c) functionally related to kallikreins (EVPL, KLK6, KLK7, KLK8, SERPINB13, SERPING1 and SLPI). Our data also indicated that hKLK7 protein expression was significantly associated with good prognosis and survival.

Conclusions

Our findings, derived from a different type of analysis of microarray data, highlight the importance of analyzing coordinated gene expression to find molecular pathways involved in melanoma progression.

Challenges and solutions in proteomics

The accelerated growth of proteomics data presents both opportunities and challenges. Large-scale proteomic profiling of biological samples such as cells, organelles or biological fluids has led to discovery of numerous key and novel proteins involved in many biological/disease processes including cancers, as well as to the identification of novel disease biomarkers and potential therapeutic targets. While proteomic data analysis has been greatly assisted by the many bioinformatics tools developed in recent years, a careful analysis of the major steps and flow of data in a typical highthroughput analysis reveals a few gaps that still need to be filled to fully realize the value of the data. To facilitate functional and pathway discovery for large-scale proteomic data, we have developed an integrated proteomic expression analysis system, iProXpress, which facilitates protein identification using a comprehensive sequence library and functional interpretation using integrated data. With its modular design, iProXpress complements and can be integrated with other software in a proteomic data analysis pipeline. This novel approach to complex biological questions involves the interrogation of multiple data sources, thereby facilitating hypothesis generation and knowledge discovery from the genomic-scale studies and fostering disease diagnosis and drug development.

Gene expression analysis of blastemal component reveals genes associated with relapse mechanism in Wilms tumour

Wilms tumour (WT) is a paediatric kidney tumour, composed of blastemal, epithelial and stromal cells, with a relapse rate of approximately 15%. Long-term survival for patients with relapse remains approximately 50%. Current clinical and molecular research is directed towards identifying prognostic factors to define the minimal and intensive therapy for successful treatment of children with low and high risk of relapse, respectively. Blastemal component presents a high level of aggressiveness and responsiveness to chemotherapy. To identify molecular prognostic markers that are predictive of chemotherapy sensitivity in tumour relapse, blastemal-enriched samples from stage III and IV WT, from patients with relapse or without relapse, were analysed for 4608 human genes immobilised on a customised cDNA platform. These analyses revealed 69 differentially expressed genes, and the top nine genes were further evaluated by qRT-PCR in the initial WT samples. TSPAN3, NCOA6, CDO1, MPP2 and MCM2 were confirmed to be down-regulated in relapse WT, and TSPAN3 and NCOA6 were also validated in an independent sample group. Protein expression of MCM2 and NCOA6 were observed in 38% (13 out of 34) and 28% (9 out of 32), respectively, of independent stage III and IV WT blastema samples, without association with relapse. However, a significant association between MCM2 positive staining and chemotherapy as first treatment suggests the involvement of MCM2 with drug metabolism in WT blastemal cells.

Poly (A)+ transcriptome assessment of ERBB2-induced alterations in breast cell lines

We report the first quantitative and qualitative analysis of the poly (A)⁺ transcriptome of two human mammary cell lines, differentially expressing (human epidermal growth factor receptor) an oncogene over-expressed in approximately 25% of human breast tumors. Full-length cDNA populations from the two cell lines were digested enzymatically, individually tagged according to a customized method for library construction, and simultaneously sequenced by the use of the Titanium 454-Roche-platform. Comprehensive bioinformatics analysis followed by experimental validation confirmed novel genes, splicing variants, single nucleotide polymorphisms, and gene fusions indicated by RNA-seq data from both samples. Moreover, comparative analysis showed enrichment in alternative events, especially in the exon usage category, in ERBB2 over-expressing cells, data indicating regulation of alternative splicing mediated by the oncogene. Alterations in expression levels of genes, such as LOX, ATP5L, GALNT3, and MME revealed by large-scale sequencing were confirmed between cell lines as well as in tumor specimens with different ERBB2 backgrounds. This approach was shown to be suitable for structural, quantitative, and qualitative assessment of complex transcriptomes and revealed new events mediated by ERBB2 overexpression, in addition to potential molecular targets for breast cancer that are driven by this oncogene.

Transcriptome profiling of whole blood cells identifies PLEK2 and C1QB in human melanoma

Background

Developing analytical methodologies to identify biomarkers in easily accessible body fluids is highly valuable for the early diagnosis and management of cancer patients. Peripheral whole blood is a “nucleic acid-rich” and “inflammatory cell-rich” information reservoir and represents systemic processes altered by the presence of cancer cells.

Methodology/Principal Findings

We conducted transcriptome profiling of whole blood cells from melanoma patients. To overcome challenges associated with blood-based transcriptome analysis, we used a PAXgene™ tube and NuGEN Ovation™ globin reduction system. The combined use of these systems in microarray resulted in the identification of 78 unique genes differentially expressed in the blood of melanoma patients. Of these, 68 genes were further analyzed by quantitative reverse transcriptase PCR using blood samples from 45 newly diagnosed melanoma patients (stage I to IV) and 50 healthy control individuals. Thirty-nine genes were verified to be differentially expressed in blood samples from melanoma patients. A stepwise logit analysis selected eighteen 2-gene signatures that distinguish melanoma from healthy controls. Of these, a 2-gene signature consisting of PLEK2 and C1QB led to the best result that correctly classified 93.3% melanoma patients and 90% healthy controls. Both genes were upregulated in blood samples of melanoma patients from all stages. Further analysis using blood fractionation showed that CD45⁻ and CD45⁺ populations were responsible for the altered expression levels of PLEK2 and C1QB, respectively.

Conclusions/Significance

The current study provides the first analysis of whole blood-based transcriptome biomarkers for malignant melanoma. The expression of PLEK2, the strongest gene to classify melanoma patients, in CD45⁻ subsets illustrates the importance of analyzing whole blood cells for biomarker studies. The study suggests that transcriptome profiling of blood cells could be used for both early detection of melanoma and monitoring of patients for residual disease.

Functional microarray analysis suggests repressed cell-cell signaling and cell survival-related modules inhibit progression of head and neck squamous cell carcinoma

Background

Cancer shows a great diversity in its clinical behavior which cannot be easily predicted using the currently available clinical or pathological markers. The identification of pathways associated with lymph node metastasis (N+) and recurrent head and neck squamous cell carcinoma (HNSCC) may increase our understanding of the complex biology of this disease.

Methods

Tumor samples were obtained from untreated HNSCC patients undergoing surgery. Patients were classified according to pathologic lymph node status (positive or negative) or tumor recurrence (recurrent or non-recurrent tumor) after treatment (surgery with neck dissection followed by radiotherapy). Using microarray gene expression, we screened tumor samples according to modules comprised by genes in the same pathway or functional category.

Results

The most frequent alterations were the repression of modules in negative lymph node (N0) and in non-recurrent tumors rather than induction of modules in N+ or in recurrent tumors. N0 tumors showed repression of modules that contain cell survival genes and in non-recurrent tumors cell-cell signaling and extracellular region modules were repressed.

Conclusions

The repression of modules that contain cell survival genes in N0 tumors reinforces the important role that apoptosis plays in the regulation of metastasis. In addition, because tumor samples used here were not microdissected, tumor gene expression data are represented together with the stroma, which may reveal signaling between the microenvironment and tumor cells. For instance, in non-recurrent tumors, extracellular region module was repressed, indicating that the stroma and tumor cells may have fewer interactions, which disable metastasis development. Finally, the genes highlighted in our analysis can be implicated in more than one pathway or characteristic, suggesting that therapeutic approaches to prevent tumor progression should target more than one gene or pathway, specially apoptosis and interactions between tumor cells and the stroma.

Influence of the interaction between nodal fibroblast and breast cancer cells on gene expression

Our aim was to evaluate the interaction between breast cancer cells and nodal fibroblasts, by means of their gene expression profile. Fibroblast primary cultures were established from negative and positive lymph nodes from breast cancer patients and a similar gene expression pattern was identified, following cell culture. Fibroblasts and breast cancer cells (MDA-MB231, MDA-MB435, and MCF7) were cultured alone or co-cultured separated by a porous membrane (which allows passage of soluble factors) for comparison. Each breast cancer lineage exerted a particular effect on fibroblasts viability and transcriptional profile. However, fibroblasts from positive and negative nodes had a parallel transcriptional behavior when co-cultured with a specific breast cancer cell line. The effects of nodal fibroblasts on breast cancer cells were also investigated. MDA MB-231 cells viability and migration were enhanced by the presence of fibroblasts and accordingly, MDA-MB435 and MCF7 cells viability followed a similar pattern. MDA-MB231 gene expression profile, as evaluated by cDNA microarray, was influenced by the fibroblasts presence, and HNMT, COMT, FN3K, and SOD2 were confirmed downregulated in MDA-MB231 co-cultured cells with fibroblasts from both negative and positive nodes, in a new series of RT-PCR assays. In summary, transcriptional changes induced in breast cancer cells by fibroblasts from positive as well as negative nodes are very much alike in a specific lineage. However, fibroblasts effects are distinct in each one of the breast cancer lineages, suggesting that the inter-relationships between stromal and malignant cells are dependent on the intrinsic subtype of the tumor.

Evaluation of quantitative rt-PCR using nonamplified and amplified RNA

Quantitative reverse-transcription polymerase chain reaction (qRT-PCR) is a standard assay in molecular medicine for gene expression analysis. Samples from incisional/needle biopsies, laser-microdissected tumor cells and other biologic sources, normally available in clinical cancer studies, generate very small amounts of RNA that are restrictive for expression analysis. As a consequence, an RNA amplification procedure is required to assess the gene expression levels of such sample types. The reproducibility and accuracy of relative gene expression data produced by sensitive methodology as qRT-PCR when cDNA converted from amplified (A) RNA is used as template has not yet been properly addressed. In this study, to properly evaluate this issue, we performed 1 round of linear RNA amplification in 2 breast cell lines (C5.2 and HB4a) and assessed the relative expression of 34 genes using cDNA converted from both nonamplified (NA) and A RNA. Relative gene expression was obtained from beta actin or glyceraldehyde 3-phosphate dehydrogenase normalized data using different dilutions of cDNA, wherein the variability and fold-change differences in the expression of the 2 methods were compared. Our data showed that 1 round of linear RNA amplification, even with suboptimal-quality RNA, is appropriate to generate reproducible and high-fidelity qRT-PCR relative expression data that have similar confidence levels as those from NA samples. The use of cDNA that is converted from both A and NA RNA in a single qRT-PCR experiment clearly creates bias in relative gene expression data.

Challenges in defining predictive markers for response to endocrine therapy in breast cancer

Endocrine therapy is a major treatment modality for hormone-dependent breast cancer. It has a relatively low morbidity, and there is evidence that antihormonal treatments have had a significant effect in reducing mortality for breast cancer. Despite this, resistance to endocrine therapy, either primary or acquired during treatment, occurs in the majority of patients, and is a major obstacle to optimal clinical management. There is therefore an urgent need to identify, on an individual basis, those tumors that are most likely to respond to endocrine therapy (so sparing patients with resistant tumors the needless side effects of ineffective therapy), and the mechanisms of resistance in tumors that are nonresponsive to treatment (so these can be bypassed). These needs are the focus of this review, which discusses the particular issues encountered when investigating the potential of multigene expression signatures as predictive factors for response to aromatase inhibitors, which have recently become front-line endocrine therapies for postmenopausal patients with breast cancer.

Integrative oncogenomic analysis of microarray data in hematologic malignancies

During the last decade, gene expression microarrays and array-based comparative genomic hybridization (array-CGH) have unraveled the complexity of human tumor genomes more precisely and comprehensively than ever before. More recently, the simultaneous assessment of global changes in messenger RNA (mRNA) expression and in DNA copy number through "integrative oncogenomic" analyses has allowed researchers the access to results uncovered through the analysis of one-dimensional data sets, thus accelerating cancer gene discovery. In this chapter, we discuss the major contributions of DNA microarrays to the study of hematological malignancies, focusing on the integrative oncogenomic approaches that correlate genomic and transcriptomic data. We also present the basic aspects of these methodologies and their present and future application in clinical oncology.

Genomic expression, chemotherapy response, and molecular targets in soft tissue sarcomas of the extremities: promising strategies for treatment selection

Neoadjuvant or adjuvant therapy in soft tissue sarcomas is still controversial, especially in regards to the use of chemotherapy. The identification of predictive factors is crucial to avoid the use of chemotherapy in patients with tumors that carry genetic characteristics associated with resistance. Focusing on gene expression data, we performed a review of the actual state of knowledge in molecular predictive factors for chemotherapy response and new targets of therapy in extremity sarcomas.

Expression of human papillomavirus type 16 E7 oncoprotein alters keratinocytes expression profile in response to tumor necrosis factor-alpha

Acute expression of E7 oncogene from human papillomavirus (HPV) 16 or HPV18 is sufficient to overcome tumor necrosis factor (TNF)-alpha cytostatic effect on primary human keratinocytes. In the present study, we investigated the molecular basis of E7-induced TNF resistance through a comparative analysis of the effect of this cytokine on the proliferation and global gene expression of normal and E7-expressing keratinocytes. Using E7 functional mutants, we show that E7-induced TNF resistance correlates with its ability to mediate pRb degradation and cell transformation. On the other hand, this effect does not depend on E7 sequences required to override DNA damage-induced cell cycle arrest or extend keratinocyte life span. Furthermore, we identified a group of 66 genes whose expression pattern differs between normal and E7-expressing cells upon cytokine treatment. These genes are mainly involved in cell cycle regulation suggesting that their altered expression may contribute to sustained cell proliferation even in the presence of a cytostatic stimulus. Differential expression of TCN1 (transcobalamin I), IFI44 (Interferon-induced protein 44), HMGB2 (high-mobility group box 2) and FUS [Fusion (involved in t(12;16) in malignant liposarcoma)] among other genes were further confirmed by western-blot and/or real-time polymerase chain reaction. Moreover, FUS upregulation was detected in HPV-positive cervical high-grade squamous intraepithelial lesions when compared with normal cervical tissue. Further evaluation of the role of such genes in TNF resistance and HPV-associated disease development is warranted.

Reciprocal changes in gene expression profiles of cocultured breast epithelial cells and primary fibroblasts

The importance of epithelial-stroma interaction in normal breast development and tumor progression has been recognized. To identify genes that were regulated by these reciprocal interactions, we cocultured a nonmalignant (MCF10A) and a breast cancer derived (MDA-MB231) basal cell lines, with fibroblasts isolated from breast benign-disease adjacent tissues (NAF) or with carcinoma-associated fibroblasts (CAF), in a transwell system. Gene expression profiles of each coculture pair were compared with the correspondent monocultures, using a customized microarray. Contrariwise to large alterations in epithelial cells genomic profiles, fibroblasts were less affected. In MDA-MB231 highly represented genes downregulated by CAF derived factors coded for proteins important for the specificity of vectorial transport between ER and golgi, possibly affecting cell polarity whereas the response of MCF10A comprised an induction of genes coding for stress responsive proteins, representing a prosurvival effect. While NAF downregulated genes encoding proteins associated to glycolipid and fatty acid biosynthesis in MDA-MB231, potentially affecting membrane biogenesis, in MCF10A, genes critical for growth control and adhesion were altered. NAFs responded to coculture with MDA-MB231 by a decrease in the expression of genes induced by TGFbeta1 and associated to motility. However, there was little change in NAFs gene expression profile influenced by MCF10A. CAFs responded to the presence of both epithelial cells inducing genes implicated in cell proliferation. Our data indicate that interactions between breast fibroblasts and basal epithelial cells resulted in alterations in the genomic profiles of both cell types which may help to clarify some aspects of this heterotypic signaling.

Expression profile of human gingival fibroblasts induced by interleukin-1beta reveals central role of nuclear factor-kappa B in stabilizing human gingival fibroblasts during inflammation

BACKGROUND

Interleukin (IL)-1beta is a key cytokine in the pathogenesis of periodontitis, and it induces inflammatory mediators in periodontal diseases. We developed immortalized human gingival fibroblasts (HGFs), investigated the effects of IL-1beta on the gene expression using expression arrays containing approximately 40,000 genes, and tested the role of nuclear factor-kappa B (NF-kappaB) in maintaining an activated HGF population.

METHODS

Total RNA was isolated from IL-1beta-induced and mock-induced control cells. Gene expression analyses were performed using expression arrays and confirmed by quantitative real-time polymerase chain reaction. Western blot analysis to show inhibitor of kappa B-alpha (IkappaBalpha) phosphorylation and immunostaining of cells for NF-kappaB nuclear translocation were performed. Apoptosis was confirmed by assay of poly ADP-ribose polymerase (PARP) cleavage.

RESULTS

A total of 382 probe sets corresponding to 254 genes were differentially expressed in IL-1beta-induced cells (P <0.001). A total of 215 genes were upregulated, and 39 genes were downregulated. Most notable NF-kappaB pathway members (NFkappaB1, NFkappaB2, IkappaBalpha, IkappaBepsilon, IkappaBzeta, REL, RELB, and TA-NFKBH) were upregulated. IkappaBalpha was phosphorylated, and NF-kappaB accumulated in the nucleus. An IL-1beta-induced set of 27 genes was downregulated by an NF-kappaB inhibitor, leading to a decreased number of viable cells and suggesting an antiapoptotic role for NF-kappaB.

CONCLUSIONS

IL-1beta leads to a large number of significant expression changes consistent with a pathologic role in periodontitis, including enhancement of inflammatory cytokines, chemokines, transcription factors, matrix metalloproteinases, adhesion molecules, and especially NF-kappaB-dependent antiapoptotic genes. NF-kappaB activation blocks apoptosis, thereby stabilizing the HGF population in inflammation.

No-match ORESTES explored as tumor markers

Sequencing technologies and new bioinformatics tools have led to the complete sequencing of various genomes. However, information regarding the human transcriptome and its annotation is yet to be completed. The Human Cancer Genome Project, using ORESTES (open reading frame EST sequences) methodology, contributed to this objective by generating data from about 1.2 million expressed sequence tags. Approximately 30% of these sequences did not align to ESTs in the public databases and were considered no-match ORESTES. On the basis that a set of these ESTs could represent new transcripts, we constructed a cDNA microarray. This platform was used to hybridize against 12 different normal or tumor tissues. We identified 3421 transcribed regions not associated with annotated transcripts, representing 83.3% of the platform. The total number of differentially expressed sequences was 1007. Also, 28% of analyzed sequences could represent noncoding RNAs. Our data reinforces the knowledge of the human genome being pervasively transcribed, and point out molecular marker candidates for different cancers. To reinforce our data, we confirmed, by real-time PCR, the differential expression of three out of eight potentially tumor markers in prostate tissues. Lists of 1007 differentially expressed sequences, and the 291 potentially noncoding tumor markers were provided.

Evidence that molecular changes in cells occur before morphological alterations during the progression of breast ductal carcinoma

Introduction

Ductal carcinoma in situ (DCIS) of the breast includes a heterogeneous group of preinvasive tumors with uncertain evolution. Definition of the molecular factors necessary for progression to invasive disease is crucial to determining which lesions are likely to become invasive. To obtain insight into the molecular basis of DCIS, we compared the gene expression pattern of cells from the following samples: non-neoplastic, pure DCIS, in situ component of lesions with co-existing invasive ductal carcinoma, and invasive ductal carcinoma.

Methods

Forty-one samples were evaluated: four non-neoplastic, five pure DCIS, 22 in situ component of lesions with co-existing invasive ductal carcinoma, and 10 invasive ductal carcinoma. Pure cell populations were isolated using laser microdissection. Total RNA was purified, DNase treated, and amplified using the T7-based method. Microarray analysis was conducted using a customized cDNA platform. The concept of molecular divergence was applied to classify the sample groups using analysis of variance followed by Tukey's test.

Results

Among the tumor sample groups, cells from pure DCIS exhibited the most divergent molecular profile, consequently identifying cells from in situ component of lesions with co-existing invasive ductal carcinoma as very similar to cells from invasive lesions. Additionally, we identified 147 genes that were differentially expressed between pure DCIS and in situ component of lesions with co-existing invasive ductal carcinoma, which can discriminate samples representative of in situ component of lesions with co-existing invasive ductal carcinoma from 60% of pure DCIS samples. A gene subset was evaluated using quantitative RT-PCR, which confirmed differential expression for 62.5% and 60.0% of them using initial and partial independent sample groups, respectively. Among these genes, LOX and SULF-1 exhibited features that identify them as potential participants in the malignant process of DCIS.

Conclusions

We identified new genes that are potentially involved in the malignant transformation of DCIS, and our findings strongly suggest that cells from the in situ component of lesions with co-existing invasive ductal carcinoma exhibit molecular alterations that enable them to invade the surrounding tissue before morphological changes in the lesion become apparent.

Characterization of global transcription profile of normal and HPV-immortalized keratinocytes and their response to TNF treatment

BACKGROUND

Persistent infection by high risk HPV types (e.g. HPV-16, -18, -31, and -45) is the main risk factor for development of cervical intraepithelial neoplasia and cervical cancer. Tumor necrosis factor (TNF) is a key mediator of epithelial cell inflammatory response and exerts a potent cytostatic effect on normal or HPV16, but not on HPV18 immortalized keratinocytes. Moreover, several cervical carcinoma-derived cell lines are resistant to TNF anti-proliferative effect suggesting that the acquisition of TNF-resistance may constitute an important step in HPV-mediated carcinogenesis. In the present study, we compared the gene expression profiles of normal and HPV16 or 18 immortalized human keratinocytes before and after treatment with TNF for 3 or 60 hours.

METHODS

In this study, we determined the transcriptional changes 3 and 60 hours after TNF treatment of normal, HPV16 and HPV18 immortalized keratinocytes by microarray analysis. The expression pattern of two genes observed by microarray was confirmed by Northern Blot. NF-kappaB activation was also determined by electrophoretic mobility shift assay (EMSA) using specific oligonucleotides and nuclear protein extracts.

RESULTS

We observed the differential expression of a common set of genes in two TNF-sensitive cell lines that differs from those modulated in TNF-resistant ones. This information was used to define genes whose differential expression could be associated with the differential response to TNF, such as: KLK7 (kallikrein 7), SOD2 (superoxide dismutase 2), 100P (S100 calcium binding protein P), PI3 (protease inhibitor 3, skin-derived), CSTA (cystatin A), RARRES1 (retinoic acid receptor responder 1), and LXN (latexin). The differential expression of the KLK7 and SOD2 transcripts was confirmed by Northern blot. Moreover, we observed that SOD2 expression correlates with the differential NF-kappaB activation exhibited by TNF-sensitive and TNF-resistant cells.

CONCLUSION

This is the first in depth analysis of the differential effect of TNF on normal and HPV16 or HPV18 immortalized keratinocytes. Our findings may be useful for the identification of genes involved in TNF resistance acquisition and candidate genes which deregulated expression may be associated with cervical disease establishment and/or progression.

Genome-wide studies in thyroid neoplasia

There is much interest in the application of genome biology to the field of thyroid neoplasia, despite the relatively low mortality rate associated with thyroid cancer in general. The principal reason for this interest is that the field of thyroid neoplasia stands to benefit from the application of genomic information to address a variety of pathologic and clinical issues. In addition to practical patient care issues, there is an excellent opportunity of expand the basic understanding of thyroid carcinogenesis. In this article, the most relevant genomic work on thyroid tumors performed to date is reviewed along with some general comments about the potential impact of genomic biology on thyroid pathology and the management of patients with thyroid nodules and cancer.

Molecular diagnostics for the detection and characterization of microbial pathogens

New and advanced methods of molecular diagnostics are changing the way we practice clinical microbiology, which affects the practice of medicine. Signal amplification and real-time nucleic acid amplification technologies offer a sensitive and specific result with a more rapid turnaround time than has ever before been possible. Numerous methods of postamplification analysis afford the simultaneous detection and differentiation of numerous microbial pathogens, their mechanisms of resistance, and the construction of disease-specific assays. The technical feasibility of these assays has already been demonstrated. How these new, often more expensive tests will be incorporated into routine practice and the impact they will have on patient care remain to be determined. One of the most attractive uses for such techniques is to achieve a more rapid characterization of the infectious agent so that a narrower-spectrum antimicrobial agent may be used, which should have an impact on resistance patterns.

Proteomic expression profiling of breast cancer

Breast cancer is one of the most common cancers observed in women in industrialized Western countries. The development of novel diagnostic methods and the application of modern systemic therapies have significantly optimized early detection and therapy of breast cancer. However, many patients are currently overtreated. Traditionally, tumours have been categorized on the basis of histopathological criteria. However, staining pattern and intensity of cancer cells are not sufficient to reflect the molecular events driving tumour development and progression. Therefore, new genomic, transcriptomic and proteomic techniques are applied to clinical samples aiming to identify new targets for a therapy tailored for an individual patient. After an introduction to common genomic and transcriptomic profiling technologies and their relevance for clinical use, we will focus on analytical and preanalytical applications for the identification of new therapeutic targets by protein profiling, with a special emphasis on two-dimensional gel-technologies (2D-PAGE), particularly as they apply to the study of breast cancer.

Chemical and biological differentiation of three human breast cancer cell types using time-of-flight secondary ion mass spectrometry

We use time-of-flight secondary ion mass spectrometry (TOF-SIMS) to image and classify individual cells on the basis of their characteristic mass spectra. Using statistical data reduction on the large data sets generated during TOF-SIMS analysis, similar biological materials can be differentiated on the basis of a combination of small changes in protein expression, metabolic activity and cell structure. We apply this powerful technique to image and differentiate three carcinoma-derived human breast cancer cell lines (MCF-7, T47D, and MDA-MB-231). In homogenized cells, we show the ability to differentiate the cell types as well as cellular compartments (cytosol, nuclear, and membrane). These studies illustrate the capacity of TOF-SIMS to characterize individual cells by chemical composition, which could ultimately be applied to detect and identify single aberrant cells within a normal cell population. Ultimately, we anticipate characterizing rare chemical changes that may provide clues to single cell progression within carcinogenic and metastatic pathways.

DNA microarray technology in toxicogenomics of aquatic models: methods and applications

Toxicogenomics represents the merging of toxicology with genomics and bioinformatics to investigate biological functions of genome in response to environmental contaminants. Aquatic species have traditionally been used as models in toxicology to characterize the actions of environmental stresses. Recent completion of the DNA sequencing for several fish species has spurred the development of DNA microarrays allowing investigators access to toxicogenomic approaches. However, since microarray technology is thus far limited to only a few aquatic species and derivation of biological meaning from microarray data is highly dependent on statistical arguments, the full potential of microarray in aquatic species research has yet to be realized. Herein we review some of the issues related to construction, probe design, statistical and bioinformatical data analyses, and current applications of DNA microarrays. As a model a recently developed medaka (Oryzias latipes) oligonucleotide microarray was described to highlight some of the issues related to array technology and its application in aquatic species exposed to hypoxia. Although there are known non-biological variations present in microarray data, it remains unquestionable that array technology will have a great impact on aquatic toxicology. Microarray applications in aquatic toxicogenomics will range from the discovery of diagnostic biomarkers, to establishment of stress-specific signatures and molecular pathways hallmarking the adaptation to new environmental conditions.

Gene stage-specific expression in the microenvironment of pediatric myelodysplastic syndromes

Using cDNA microarray assays we have observed a clear difference in the gene expression pattern between bone marrow stromal cells obtained from healthy children (CT) and from pediatric patients with either myelodysplastic syndromes (MDS) or acute myeloid leukemia (AML) associated with MDS (MDS-AML). The global gene function profiling analysis indicated that in the pediatric MDS microenvironment the disease stages may be characterized mainly by underexpression of genes associated with biological processes such as transport. Furthermore, a subset of downregulated genes related to endocytosis and protein secretion was able to discriminate MDS from MDS-AML.

A simple method to combine multiple molecular biomarkers for dichotomous diagnostic classification

Background

In spite of the recognized diagnostic potential of biomarkers, the quest for squelching noise and wringing in information from a given set of biomarkers continues. Here, we suggest a statistical algorithm that – assuming each molecular biomarker to be a diagnostic test – enriches the diagnostic performance of an optimized set of independent biomarkers employing established statistical techniques. We validated the proposed algorithm using several simulation datasets in addition to four publicly available real datasets that compared i) subjects having cancer with those without; ii) subjects with two different cancers; iii) subjects with two different types of one cancer; and iv) subjects with same cancer resulting in differential time to metastasis.

Results

Our algorithm comprises of three steps: estimating the area under the receiver operating characteristic curve for each biomarker, identifying a subset of biomarkers using linear regression and combining the chosen biomarkers using linear discriminant function analysis. Combining these established statistical methods that are available in most statistical packages, we observed that the diagnostic accuracy of our approach was 100%, 99.94%, 96.67% and 93.92% for the real datasets used in the study. These estimates were comparable to or better than the ones previously reported using alternative methods. In a synthetic dataset, we also observed that all the biomarkers chosen by our algorithm were indeed truly differentially expressed.

Conclusion

The proposed algorithm can be used for accurate diagnosis in the setting of dichotomous classification of disease states.

Microarray technology in obstetrics and gynecology: a guide for clinicians

Microarrays can be constructed with dozens to millions of probes on their surface to allow high-throughput analyses of many biologic processes to be performed simultaneously on the same sample. Microarrays are now widely used for gene expression analysis, deoxyribonucleic acid resequencing, single-nucleotide polymorphism genotyping, and comparative genomic hybridization. Microarray technology is accelerating research in many fields and now microarrays are moving into clinical application. This review discusses the emerging role of microarrays in molecular diagnostics, pathogen detection, oncology, and pharmacogenomics.

Molecular predictors of response and outcome in ovarian cancer

A major problem in clinical management of patients with epithelial ovarian cancer (EOC) is the largely unpredictable response to first-line treatment and the occurrence of relapse after complete initial response, associated with broad cross-resistance to even structurally dissimilar drugs. During tumor development and progression, multiple genic alterations take place that might contribute specifically to the treatment response and eventually impact on disease outcome. One area of intense research is the identification of molecular markers to accurately assess the prognosis of EOC patients and to define innovative therapeutic strategies. A large survey of recent published data indicates the need to revisit traditional molecular markers with respect to their contribution to the assessment of overall survival in selected populations. Furthermore, recent technological developments that enable simultaneous measurement of many parameters ("omic" approaches) hold the promise of identifying new molecular prognostic and predictive markers.

Childhood acute lymphoblastic leukemia in the age of genomics

The recent sequencing of the human genome and technical breakthroughs now make it possible to simultaneously determine mRNA expression levels of almost all of the identified genes in the human genome. DNA "chip" or microarray technology holds great promise for the development of more refined, biologically-based classification systems for childhood ALL, as well as the identification of new targets for novel therapy. To date gene expression profiles have been described that correlate with subtypes of ALL defined by morphology, immunophenotype, cytogenetic alterations, and response to therapy. Mechanistic insights into treatment failure have come from the definition of mRNA signatures that predict in vitro chemoresistance, as well as differences between blasts at relapse and new diagnosis. New bioinformatics tools optimize data mining, but validation of findings is essential since "over-fitting" the data is a common danger. In the future, genomic analysis will be complemented by evaluation of the cancer proteome.

Gene expression profile associated with response to doxorubicin-based therapy in breast cancer

PURPOSE

This study was designed to identify genes that could predict response to doxorubicin-based primary chemotherapy in breast cancer patients.

EXPERIMENTAL DESIGN

Biopsy samples were obtained before primary treatment with doxorubicin and cyclophosphamide. RNA was extracted and amplified and gene expression was analyzed using cDNA microarrays.

RESULTS

Response to chemotherapy was evaluated in 51 patients, and based on Response Evaluation Criteria in Solid Tumors guidelines, 42 patients, who presented at least a partial response (> or =30% reduction in tumor dimension), were classified as responsive. Gene profile of samples, divided into training set (n = 38) and independent validation set (n = 13), were at first analyzed against a cDNA microarray platform containing 692 genes. Unsupervised clustering could not separate responders from nonresponders. A classifier was identified comprising EMILIN1, FAM14B, and PBEF, which however could not correctly classify samples included in the validation set. Our next step was to analyze gene profile in a more comprehensive cDNA microarray platform, containing 4,608 open reading frame expressed sequence tags. Seven samples of the initial training set (all responder patients) could not be analyzed. Unsupervised clustering could correctly group all the resistant samples as well as at least 85% of the sensitive samples. Additionally, a classifier, including PRSS11, MTSS1, and CLPTM1, could correctly distinguish 95.4% of the 44 samples analyzed, with only two misclassifications, one sensitive sample and one resistant tumor. The robustness of this classifier is 2.5 greater than the first one.

CONCLUSION

A trio of genes might potentially distinguish doxorubicin-responsive from nonresponsive tumors, but further validation by a larger number of samples is still needed.

An ensemble approach for phenotype classification based on fuzzy partitioning of gene expression data

We focus on developing a pattern recognition method suitable for performing supervised analysis tasks on molecular data resulting from microarray experiments. Molecular characterization of tissue samples using microarray gene expression profiling is expected to uncover fundamental aspects related to cancer diagnosis and drug discovery. There is therefore a need for reliable, accurate classification methods. With this study we propose a framework for constructing an ensemble of individually trained SVM classifiers, each of them specialized on subsets of the input space. The fuzzy approach used for partitioning the data produces overlapping subsets of the input space that facilitates subsequent classification tasks.

Class distinction between follicular adenomas and follicular carcinomas of the thyroid gland on the basis of their signature expression

BACKGROUND

Nodules of the thyroid gland are observed frequently in patients who undergo ultrasound studies. The majority of these nodules are benign, corresponding to goiters or adenomas, and only a small fraction corresponds to carcinomas. Among thyroid tumors, the diagnosis of follicular adenocarcinomas by preoperative fine-needle aspiration biopsy is a major challenge, because it requires inspection of the entire capsule to differentiate it from adenoma. Consequently, large numbers of patients undergo unnecessary thyroidectomy.

METHODS

Using data from gene expression analysis, the authors applied Fisher linear discriminant analysis and searched for expression signatures of individual samples of adenomas and follicular carcinomas that could be used as molecular classifiers for the precise classification of malignant and nonmalignant lesions.

RESULTS

Fourteen trios of genes were described that fulfilled the criteria for the correct classification of 100% of samples. The robustness of these trios was verified by using leave-1-out cross-validation and bootstrap analyses. The results demonstrated that, by combining trios, better classifiers could be generated that correctly classified >92% of samples.

CONCLUSIONS

The strategy of classifiers based on individual signatures was a useful strategy for distinguishing between samples with very similar expression profiles.

Expression profile of malignant and nonmalignant lesions of esophagus and stomach: differential activity of functional modules related to inflammation and lipid metabolism

Adenocarcinomas of stomach and esophagus are frequently associated with preceding inflammatory alterations of the normal mucosa. Whereas intestinal metaplasia of the gastric mucosa is associated with higher risk of malignization, Barrett's disease is a risk factor for adenocarcinoma of the esophagus. Barrett's disease is characterized by the substitution of the squamous mucosa of the esophagus by a columnar tissue classified histopathologically as intestinal metaplasia. Using cDNA microarrays, we determined the expression profile of normal gastric and esophageal mucosa as well as intestinal metaplasia and adenocarcinomas from both organs. Data were explored to define functional alterations related to the transformation from squamous to columnar epithelium and the malignant transformation from intestinal metaplasia to adenocarcinomas. Based on their expression profile, adenocarcinomas of the esophagus showed stronger correlation with intestinal metaplasia of the stomach than with Barrett's mucosa. Second, we identified two functional modules, lipid metabolism and cytokine, as being altered with higher statistical significance. Whereas the lipid metabolism module is active in samples representing intestinal metaplasia and inactive in adenocarcinomas, the cytokine module is inactive in samples representing normal esophagus and esophagitis. Using the concept of relevance networks, we determined the changes in linear correlation of genes pertaining to these two functional modules. Exploitation of the data presented herein will help in the precise molecular characterization of adenocarcinoma from the distal esophagus, avoiding the topographical and descriptive classification that is currently adopted, and help with the proper management of patients with Barrett's disease.