Removing intensity effects and identifying significant genes for Affymetrix arrays in macrophage migration inhibitory factor-suppressed neuroblastoma cells

Benchmarking association analyses of continuous exposures with RNA-seq in observational studies

Large datasets of hundreds to thousands of individuals measuring RNA-seq in observational studies are becoming available. Many popular software packages for analysis of RNA-seq data were constructed to study differences in expression signatures in an experimental design with well-defined conditions (exposures). In contrast, observational studies may have varying levels of confounding transcript-exposure associations; further, exposure measures may vary from discrete (exposed, yes/no) to continuous (levels of exposure), with non-normal distributions of exposure. We compare popular software for gene expression-DESeq2, edgeR and limma-as well as linear regression-based analyses for studying the association of continuous exposures with RNA-seq. We developed a computation pipeline that includes transformation, filtering and generation of empirical null distribution of association P-values, and we apply the pipeline to compute empirical P-values with multiple testing correction. We employ a resampling approach that allows for assessment of false positive detection across methods, power comparison and the computation of quantile empirical P-values. The results suggest that linear regression methods are substantially faster with better control of false detections than other methods, even with the resampling method to compute empirical P-values. We provide the proposed pipeline with fast algorithms in an R package Olivia, and implemented it to study the associations of measures of sleep disordered breathing with RNA-seq in peripheral blood mononuclear cells in participants from the Multi-Ethnic Study of Atherosclerosis.

Emerging Role of the Macrophage Migration Inhibitory Factor Family of Cytokines in Neuroblastoma. Pathogenic Effectors and Novel Therapeutic Targets?

Neuroblastoma (NB) is the most frequent extracranial pediatric tumor. Despite the current available multiple therapeutic options, the prognosis for high-risk NB patients remains unsatisfactory and makes the disease a clear unmet medical need. Thus, more tailored therapeutic approaches are warranted to improve both the quality of life and the survival of the patients. Macrophage migration inhibitory factor (MIF) is a pleiotropic cytokine that plays a key role in several diseases, including cancer. Preclinical and clinical studies in NB patients convergently indicate that MIF exerts pro-tumorigenic properties in NB. MIF is upregulated in NB tumor tissues and cell lines and it contributes to NB aggressiveness and immune-escape. To date, there are only a few data about the role of the second member of the MIF family, the MIF homolog d-dopachrome tautomerase (DDT), in NB. Here, we review the preclinical and clinical studies on the role of the MIF family of cytokines in NB and suggest that MIF and possibly DDT inhibitors may be promising novel prognostic and therapeutic targets in NB management.

Microarray data analysis and mining tools

Microarrays are one of the latest breakthroughs in experimental molecular biology that allow monitoring the expression levels of tens of thousands of genes simultaneously. Arrays have been applied to studies in gene expression, genome mapping, SNP discrimination, transcription factor activity, toxicity, pathogen identification and many other applications. In this paper we concentrate on discussing various bioinformatics tools used for microarray data mining tasks with its underlying algorithms, web resources and relevant reference. We emphasize this paper mainly for digital biologists to get an aware about the plethora of tools and programs available for microarray data analysis. First, we report the common data mining applications such as selecting differentially expressed genes, clustering, and classification. Next, we focused on gene expression based knowledge discovery studies such as transcription factor binding site analysis, pathway analysis, protein- protein interaction network analysis and gene enrichment analysis.

Properties of balanced permutations

This paper takes a close look at balanced permutations, a recently developed sample reuse method with applications in bioinformatics. It turns out that balanced permutation reference distributions do not have the correct null behavior, which can be traced to their lack of a group structure. We find that they can give p-values that are too permissive to varying degrees. In particular the observed test statistic can be larger than that of all B balanced permutations of a data set with a probability much higher than 1/(B + 1), even under the null hypothesis.

Use of normalization methods for analysis of microarrays containing a high degree of gene effects

Background

High-throughput microarrays are widely used to study gene expression across tissues and developmental stages. Analysis of gene expression data is challenging in these experiments due to the presence of significant percentages of differentially expressed genes (DEG) observed between tissues and developmental stages. Data normalization methods that are widely used today are not designed for data with a large proportion of tissue or gene effects.

Results

In our current study, we describe a novel two-dimensional nonparametric normalization method for analyzing microarray data which functions well in the absence or presence of large numbers of gene effects. Rather than relying on an assumption of low variability among most genes, the method implements a unique peak selection strategy to distinguish DEG from genes that are invariant in expression, prior to nonlinear curve fitting. We compared the method under simulated and experimental conditions with five alternative nonlinear normalization approaches: quantile, lowess, robust lowess, invariant set, and cross-correlation (Xcorr). Simulations included various percentages of simulated DEG and the experimental data used is from publicly available datasets known to be difficult to analyze due to the presence of approximately 34% DEG.

Conclusion

We have demonstrated that the new method provides considerable improvement in the accuracy of data normalization when large proportions of gene effects are present. The performance improvement is mostly attributed to its variable selection component, which is designed to separate expression invariant genes from DEG. Adding this key component of the new method to alternative normalization approaches rescues the most of the sensitivity of these methods to gene effects. The results indicate that our method may be used without prior knowledge of or assumptions about housekeeping genes to normalize microarrays that are quite different.

Cooperative regulation of non-small cell lung carcinoma angiogenic potential by macrophage migration inhibitory factor and its homolog, D-dopachrome tautomerase

Tumor-derived growth factors and cytokines stimulate neoangiogenesis from surrounding capillaries to support tumor growth. Recent studies have revealed that macrophage migration inhibitory factor (MIF) expression is increased in lung cancer, particularly non-small cell lung carcinomas (NSCLC). Because MIF has important autocrine effects on normal and transformed cells, we investigated whether autocrine MIF and its only known family member, D-dopachrome tautomerase (D-DT), promote the expression of proangiogenic factors CXCL8 and vascular endothelial growth factor in NSCLC cells. Our results demonstrate that the expression of CXCL8 and vascular endothelial growth factor are strongly reliant upon both the individual and cooperative activities of the two family members. CXCL8 transcriptional regulation by MIF and D-DT appears to involve a signaling pathway that includes the activation of JNK, c-jun phosphorylation, and subsequent AP-1 transcription factor activity. Importantly, HUVEC migration and tube formation induced by supernatants from lung adenocarcinoma cells lacking either or both MIF and D-DT are substantially reduced when compared with normal supernatants. Finally, we demonstrate that the cognate MIF receptor, CD74, is necessary for both MIF- and D-DT-induced JNK activation and CXCL8 expression, suggesting its potential involvement in angiogenic growth factor expression. This is the first demonstration of a biological role for D-DT, and its synergism with MIF suggests that the combined therapeutic targeting of both family members may enhance current anti-MIF-based therapies.

Mu-opioid receptor activation induces transcriptional plasticity in the central extended amygdala

Addiction develops from the gradual adaptation of the brain to chronic drug exposure, and involves genetic reprogramming of neuronal function. The central extended amygdala (EAc) is a network formed by the central amygdala and the bed nucleus of the stria terminalis. This key site controls drug craving and seeking behaviors, and has not been investigated at the gene regulation level. We used Affymetrix microarrays to analyze transcriptional activity in the murine EAc, with a focus on mu-opioid receptor-associated events because these receptors mediate drug reward and dependence. We identified 132 genes whose expression is regulated by a chronic escalating morphine regimen in the EAc from wild-type but not mu-opioid receptor knockout mice. These modifications are mostly EAc-specific. Gene ontology analysis reveals an overrepresentation of neurogenesis, cell growth and signaling protein categories. A separate quantitative PCR analysis of genes in the last of these groups confirms the dysregulation of both orphan (Gpr88) and known (DrD1A, Adora2A, Cnr1, Grm5, Gpr6) G protein-coupled receptors, scaffolding (PSD95, Homer1) and signaling (Sgk, Cap1) proteins, and neuropeptides (CCK, galanin). These transcriptional modifications do not occur following a single morphine injection, and hence result from long-term adaptation to excessive mu receptor activation. Proteins encoded by these genes are classically associated with spine modules function in other brain areas, and therefore our data suggest a remodeling of EAc circuits at sites where glutamatergic and monoaminergic afferences interact. Together, mu receptor-dependent genes identified in this study potentially contribute to drug-induced neural plasticity, and provide a unique molecular repertoire towards understanding drug craving and relapse.

Expression of macrophage migration inhibitory factor by neuroblastoma leads to the inhibition of antitumor T cell reactivity in vivo

Neuroblastomas and many other solid tumors produce high amounts of macrophage migration inhibitory factor (MIF), which appears to play a role in tumor progression. We found that MIF expression in neuroblastoma inhibits T cell proliferation in vitro, raising the possibility that MIF promotes tumorigenesis, in part, by suppressing antitumor immunity. To examine whether tumor-derived MIF leads to suppression of T cell immunity in vivo, we generated MIF-deficient neuroblastoma cell lines using short hairpin small interfering RNAs (siRNA). The MIF knockdown (MIFKD) AGN2a neuroblastoma cells were more effectively rejected in immune-competent mice than control siRNA-transduced or wild-type AGN2a. However, the increased rejection of MIFKD AGN2a was not observed in T cell-depleted mice. MIFKD tumors had increased infiltration of CD8(+) and CD4(+) T cells, as well as increased numbers of macrophages, dendritic cells, and B cells. Immunization with MIFKD AGN2a cells significantly increased protection against tumor challenge as compared with immunization with wild-type AGN2a, and the increased protection correlated with elevated frequencies of tumor-reactive CD8(+) T cells in the lymphoid tissue of treated animals. Increased numbers of infiltrating tumor-reactive CD8(+) T cells were also observed at the site of tumor vaccination. In vitro, treatment of AGN2a-derived culture supernatants with neutralizing MIF-specific Ab failed to reverse T cell suppressive activity, suggesting that MIF is not directly responsible for the immune suppression in vivo. This supports a model whereby MIF expression in neuroblastoma initiates a pathway that leads to the suppression of T cell immunity in vivo.

New roles for mononuclear phagocytes in cancer biology

The primary focus in the pathogenesis and treatment of human malignancies has been the tumor cell. However, the biologic properties of a malignancy are not all intrinsically determined. Interactions between heterogeneous cell populations influence the growth and survival of both normal and malignant cells. Studies defining the origin of endothelial cells involved in tumor angiogenesis first demonstrated the contributions of normal cellular environment. Recently, the mononuclear phagocyte lineage has been found to have biologically and clinically significant tumor enhancing and tumor suppressive effects. This article reviews the multiple roles of mononuclear phagocytes in cancer biology. A companion manuscript (J Pediatr Hematol Oncol. 2008, in press) describes the targeting of these cells for therapeutic benefit. Incorporating these strategies into future childhood cancer protocols could be an innovative approach for improving patient outcome.

Macrophage migration inhibitory factor contributes to the immune escape of ovarian cancer by down-regulating NKG2D

The proinflammatory cytokine macrophage migration inhibitory factor (MIF) stimulates tumor cell proliferation, migration, and metastasis; promotes tumor angiogenesis; suppresses p53-mediated apoptosis; and inhibits antitumor immunity by largely unknown mechanisms. We here describe an overexpression of MIF in ovarian cancer that correlates with malignancy and the presence of ascites. Functionally, we find that MIF may contribute to the immune escape of ovarian carcinoma by transcriptionally down-regulating NKG2D in vitro and in vivo which impairs NK cell cytotoxicity toward tumor cells. Together with the additional tumorigenic properties of MIF, this finding provides a rationale for novel small-molecule inhibitors of MIF to be used for the treatment of MIF-secreting cancers.

VILIP-1 downregulation in non-small cell lung carcinomas: mechanisms and prediction of survival

VILIP-1, a member of the neuronal Ca++ sensor protein family, acts as a tumor suppressor gene in an experimental animal model by inhibiting cell proliferation, adhesion and invasiveness of squamous cell carcinoma cells. Western Blot analysis of human tumor cells showed that VILIP-1 expression was undetectable in several types of human tumor cells, including 11 out of 12 non-small cell lung carcinoma (NSCLC) cell lines. The down-regulation of VILIP-1 was due to loss of VILIP-1 mRNA transcripts. Rearrangements, large gene deletions or mutations were not found. Hypermethylation of the VILIP-1 promoter played an important role in gene silencing. In most VILIP-1-silent cells the VILIP-1 promoter was methylated. In vitro methylation of the VILIP-1 promoter reduced its activity in a promoter-reporter assay. Transcriptional activity of endogenous VILIP-1 promoter was recovered by treatment with 5′-aza-2′-deoxycytidine (5′-Aza-dC). Trichostatin A (TSA), a histone deacetylase inhibitor, potently induced VILIP-1 expression, indicating that histone deacetylation is an additional mechanism of VILIP-1 silencing. TSA increased histone H3 and H4 acetylation in the region of the VILIP-1 promoter. Furthermore, statistical analysis of expression and promoter methylation (n = 150 primary NSCLC samples) showed a significant relationship between promoter methylation and protein expression downregulation as well as between survival and decreased or absent VILIP-1 expression in lung cancer tissues (p<0.0001). VILIP-1 expression is silenced by promoter hypermethylation and histone deacetylation in aggressive NSCLC cell lines and primary tumors and its clinical evaluation could have a role as a predictor of short-term survival in lung cancer patients.

Profiling of CD4+, CD8+, and CD4+CD25+CD45RO+FoxP3+ T cells in patients with malignant glioma reveals differential expression of the immunologic transcriptome compared with T cells from healthy volunteers

PURPOSE

Analyses of T-cell mRNA expression profiles in glioblastoma multiforme has not been previously reported but may help to define and characterize the immunosuppressed phenotype in patients with this type of cancer.

EXPERIMENTAL DESIGN

We did microarray studies that have shown significant and fundamental differences in the expression profiles of CD4(+) and CD8(+) T cells and immunosuppressive CD4(+)CD25(+)CD45RO(+)FoxP3(+) regulatory T cells (T(reg)) from normal healthy volunteers compared with patients with newly diagnosed glioblastoma multiforme. For these investigations, we isolated total RNA from enriched CD4(+) and CD8(+) T cell or T(reg) cell populations from age-matched individuals and did microarray analyses.

RESULTS

ANOVA and principal components analysis show that the various T cell compartments exhibit consistently similar mRNA expression profiles among individuals within either healthy or brain tumor groups but reflect significant differences between these groups. Compared with healthy volunteers, CD4(+) and CD8(+) T cells from patients with glioblastoma multiforme display coordinate down-regulation of genes involved in T cell receptor ligation, activation, and intracellular signaling. In contrast, T(regs) from patients with glioblastoma multiforme exhibit increased levels of transcripts involved in inhibiting host immunity.

CONCLUSION

Our findings support the notion that key differences between expression profiles in T-cell populations from patients with glioblastoma multiforme results from differential expression of the immunologic transcriptome, such that a limited number of genes are principally important in producing the dysregulated T-cell phenotype.

Impaired DNA damage checkpoint response in MIF-deficient mice

Recent studies demonstrated that proinflammatory migration inhibitory factor(MIF) blocks p53-dependent apoptosis and interferes with the tumor suppressor activity of p53. To explore the mechanism underlying this MIF-p53 relationship, we studied spontaneous tumorigenesis in genetically matched p53-/- and MIF-/-p53-/- mice. We show that the loss of MIF expression aggravates the tumor-prone phenotype of p53-/- mice and predisposes them to a broader tumor spectrum, including B-cell lymphomas and carcinomas. Impaired DNA damage response is at the root of tumor predisposition of MIF-/-p53-/- mice. We provide evidence that MIF plays a role in regulating the activity of Cul1-containing SCF ubiquitin ligases. The loss of MIF expression uncouples Chk1/Chk2-responsive DNA damage checkpoints from SCF-dependent degradation of key cell-cycle regulators such as Cdc25A, E2F1 and DP1, creating conditions for the genetic instability of cells. These MIF effects depend on its association with the Jab1/CSN5 subunit of the COP9/CSN signalosome. Given that CSN plays a central role in the assembly of SCF complexes in vivo, regulation of Jab1/CSN5 by MIF is required to sustain optimal composition and function of the SCF complex.

Estimating p-values in small microarray experiments

MOTIVATION

Microarray data typically have small numbers of observations per gene, which can result in low power for statistical tests. Test statistics that borrow information from data across all of the genes can improve power, but these statistics have non-standard distributions, and their significance must be assessed using permutation analysis. When sample sizes are small, the number of distinct permutations can be severely limited, and pooling the permutation-derived test statistics across all genes has been proposed. However, the null distribution of the test statistics under permutation is not the same for equally and differentially expressed genes. This can have a negative impact on both p-value estimation and the power of information borrowing statistics.

RESULTS

We investigate permutation based methods for estimating p-values. One of methods that uses pooling from a selected subset of the data are shown to have the correct type I error rate and to provide accurate estimates of the false discovery rate (FDR). We provide guidelines to select an appropriate subset. We also demonstrate that information borrowing statistics have substantially increased power compared to the t-test in small experiments.

Enhanced Quantile Normalization of Microarray Data to Reduce Loss of Information in Gene Expression Profiles

In microarray experiments, removal of systematic variations resulting from array preparation or sample hybridization conditions is crucial to ensure sensible results from the ensuing data analysis. For example, quantile normalization is routinely used in the treatment of both oligonucleotide and cDNA microarray data, even though there might be some loss of information in the normalization process. We recognize that the ideal normalization, if it ever exists, would aim to keep the maximal amount of gene profile information with the lowest possible noise. With this objective in mind, we propose a valuable enhancement to quantile normalization, and demonstrate through three Affymetrix experiments that the enhanced normalization can result in better performance in detecting and ranking differentially expressed genes across experimental conditions.

Epistemological issues in omics and high-dimensional biology: give the people what they want

Gene expression microarrays have been the vanguard of new analytic approaches in high-dimensional biology. Draft sequences of several genomes coupled with new technologies allow study of the influences and responses of entire genomes rather than isolated genes. This has opened a new realm of highly dimensional biology where questions involve multiplicity at unprecedented scales: thousands of genetic polymorphisms, gene expression levels, protein measurements, genetic sequences, or any combination of these and their interactions. Such situations demand creative approaches to the processes of inference, estimation, prediction, classification, and study design. Although bench scientists intuitively grasp the need for flexibility in the inferential process, the elaboration of formal supporting statistical frameworks is just at the very start. Here, we will discuss some of the unique statistical challenges facing investigators studying high-dimensional biology, describe some approaches being developed by statistical scientists, and offer an epistemological framework for the validation of proffered statistical procedures. A key theme will be the challenge in providing methods that a statistician judges to be sound and a biologist finds informative. The shift from family-wise error rate control to false discovery rate estimation and to assessment of ranking and other forms of stability will be portrayed as illustrative of approaches to this challenge.

Selecting normalization genes for small diagnostic microarrays

BACKGROUND

Normalization of gene expression microarrays carrying thousands of genes is based on assumptions that do not hold for diagnostic microarrays carrying only few genes. Thus, applying standard microarray normalization strategies to diagnostic microarrays causes new normalization problems.

RESULTS

In this paper we point out the differences of normalizing large microarrays and small diagnostic microarrays. We suggest to include additional normalization genes on the small diagnostic microarrays and propose two strategies for selecting them from genomewide microarray studies. The first is a data driven univariate selection of normalization genes. The second is multivariate and based on finding a balanced diagnostic signature. Finally, we compare both methods to standard normalization protocols known from large microarrays.

CONCLUSION

Not including additional genes for normalization on small microarrays leads to a loss of diagnostic information. Using house keeping genes from the literature for normalization fails to work for certain datasets. While a data driven selection of additional normalization genes works well, the best results were obtained using a balanced signature.

Statistical Analysis of DNA Microarray Data in Cancer Research: Fig. 1

Microarray techniques have been widely used to monitor gene expression in many areas of biomedical research. They have been widely used for tumor diagnosis and classification, prediction of prognoses and treatment, and understanding of molecular mechanisms, biochemical pathways, and gene networks. Statistical methods are vital for these scientific endeavors. This article reviews recent developments of statistical methods for analyzing data from microarray experiments. Emphasis has been given to normalization of expression from multiple arrays, selecting significantly differentially expressed genes, tumor classifications, and gene expression pathways and networks.