Effective classification of microRNA precursors using feature mining and AdaBoost algorithms.

MicroRNAs play important roles in most biological processes, including cell proliferation, tissue differentiation, and embryonic development, among others. They originate from precursor transcripts (pre-miRNAs), which contain phylogenetically conserved stem-loop structures. An important bioinformatics problem is to distinguish the pre-miRNAs from pseudo pre-miRNAs that have similar stem-loop structures. We present here a novel method for tackling this bioinformatics problem. Our method, named MirID, accepts an RNA sequence as input, and classifies the RNA sequence either as positive (i.e., a real pre-miRNA) or as negative (i.e., a pseudo pre-miRNA). MirID employs a feature mining algorithm for finding combinations of features suitable for building pre-miRNA classification models. These models are implemented using support vector machines, which are combined to construct a classifier ensemble. The accuracy of the classifier ensemble is further enhanced by the utilization of an AdaBoost algorithm. When compared with two closely related tools on twelve species analyzed with these tools, MirID outperforms the existing tools on the majority of the twelve species. MirID was also tested on nine additional species, and the results showed high accuracies on the nine species. The MirID web server is fully operational and freely accessible at http://bioinformatics.njit.edu/MirID/ . Potential applications of this software in genomics and medicine are also discussed.

Chromatin immunoprecipitation and microarray analysis reveal that TFIIB occupies the SL RNA gene promoter region in Trypanosoma brucei chromosomes.

RNA polymerase II (RNAP-II) synthesizes the m(7)G-capped Spliced Leader (SL) RNA and most protein-coding mRNAs in trypanosomes. RNAP-II recruitment to DNA usually requires a set of transcription factors that make sequence-specific contacts near transcriptional start sites within chromosomes. In trypanosomes, the transcription factor TFIIB is necessary for RNAP-II-dependent SL RNA transcription. However, the trypanosomal TFIIB (tTFIIB) lacks the highly basic DNA binding region normally found in the C-terminal region of TFIIB proteins. To assess the precise pattern of tTFIIB binding within the SL RNA gene locus, as well as within several other loci, we performed chromatin immunoprecipitation/microarray analysis using a tiled gene array with a probe spacing of 10 nucleotides. We found that tTFIIB binds non-randomly within the SL RNA gene locus mainly within a 220-nt long region that straddles the transcription start site. tTFIIB does not bind within the small subunit (SSU) rRNA locus, indicating that trypanosomal TFIIB is not a component of an RNAP-I transcriptional complex. Interestingly, discrete binding sites were observed within the putative promoter regions of two loci on different chromosomes. These data suggest that although trypanosomal TFIIB lacks a highly basic DNA binding region, it nevertheless localizes to discrete regions of chromatin that include the SL RNA gene promoter.

Evaluation of genetic models for response in a randomized clinical trial of duloxetine in major depressive disorder.

In self-identified white patients with major depressive disorder (N=126) treated with open-label duloxetine (60-120 mg/d), a significant association of (P=0.020) of a composite risk score (based on SLC6A2 rs5569 [G1287A] AA, HTR1A rs6295 [C(-1019)G] GG, and COMT rs174697 AA/AG) with 17-item Hamilton Depression Rating Scale total score change from baseline to 12 weeks was observed.

Association of common variations in the norepinephrine transporter gene with response to olanzapine-fluoxetine combination versus continued-fluoxetine treatment in patients with treatment-resistant depression: a candidate gene analysis.

OBJECTIVE: To determine whether single-nucleotide polymorphisms (SNPs) in candidate genes are associated with response to olanzapine-fluoxetine combination. METHOD: A post hoc analysis of a priori-selected SNPs used data from a clinical trial (dates: April 2002-July 2005) of olanzapine-fluoxetine combination, fluoxetine, and olanzapine in patients with major depressive disorder (DSM-IV criteria) and with nonresponse to prestudy antidepressant treatment and nonresponse to fluoxetine treatment during the study. Patients received open-label treatment with fluoxetine for 8 weeks (2 weeks, 25 mg/d; then 6 weeks, 50 mg/d), at the end of which nonresponders (< 25% decline in the 17-item Hamilton Depression Rating Scale score) were randomized to receive double-blind, monotherapy treatment with olanzapine-fluoxetine combination (6/50-18/50 mg/d, n = 71), fluoxetine (50 mg/d, n = 78), or olanzapine (6-18 mg/d, n = 56) for 8 weeks. Statistical significance was assessed at P < .05. The primary efficacy measure for within-study treatment was improvement on the Montgomery-Asberg Depression Rating Scale (MADRS). RESULTS: Rs36024, an intronic SNP in the norepinephrine transporter (SLC6A2), as well as 3 SNPs in melanocortin 3 receptor (MC3R) and 2 SNPs in tryptophan hydroxylase 2 (TPH2), were associated with MADRS-defined response to treatment with olanzapine-fluoxetine combination (adjusted Li-Nyholt P < .05). Except for 1 SNP in TPH2, identified SNPs were not significantly associated with response to continued-fluoxetine or olanzapine treatments. CONCLUSIONS: Our findings further support the hypothesis that the synergistic effect of olanzapine and fluoxetine on prefrontal cortical levels of norepinephrine and dopamine might be an underlying mechanism for the efficacy of olanzapine-fluoxetine combination in the treatment of treatment-resistant depression and, if replicated, may form a basis on which response to olanzapine-fluoxetine combination versus continued fluoxetine can be predicted based on variants in SLC6A2. TRIAL REGISTRATION: Parent study registered at ClinicalTrials.gov identifier: NCT00035321.

Global gene expression profiles of ischemic preconditioning in deceased donor liver transplantation.

The benefits of ischemic preconditioning (IPC) in reducing ischemia/reperfusion injury (IRI) remain indistinct in human liver transplantation (LT). To further understand mechanistic aspects of IPC, we performed microarray analyses as a nested substudy in a randomized trial of 10-minute IPC in 101 deceased donor LTs. Liver biopsies were performed after cold storage and at 90 minutes postreperfusion in 40 of 101 subjects. Global gene expression profiles in 6 biopsy pairs in IPC and work standard organ recovery groups at both time points were compared using the Affymetrix GeneChip Human Gene 1.0 ST array. Transcripts with >1.5-fold change and P < 0.05 were considered significant. IPC altered expression of 82 transcripts in antioxidant, immunological, lipid biosynthesis, cell development and growth, and other groups. Real-time polymerase chain reaction and immunoblotting validated our microarray data. IPC-induced overexpression of glutathione S-transferase mu transcripts (GSTM1, GSTM3, GSTM4, and GSTM5) was accompanied by increased protein expression and may contribute to a decrease in oxidative stress. However, the increased expression of fatty acid synthase may increase oxidative stress, and tumor necrosis factor ligand superfamily member 10 may promote apoptosis. These changes, in combination with decreased expression of heparin-binding epidermal growth factor-like growth factor and insulin-like growth factor binding protein-1, both of which inhibit apoptosis, may increase IRI. In our study of deceased donor LT, IPC induces changes in gene expression, some of which are potentially beneficial but some which are potentially injurious. Thus, our findings of changes in gene expression mirror the outcomes in our clinical trial.

Identification of direct transcriptional targets of the Kaposi's sarcoma-associated herpesvirus Rta lytic switch protein by conditional nuclear localization.

Lytic reactivation from latency is critical for the pathogenesis of Kaposi's sarcoma-associated herpesvirus (KSHV). We previously demonstrated that the 691-amino-acid (aa) KSHV Rta transcriptional transactivator is necessary and sufficient to reactivate the virus from latency. Viral lytic cycle genes, including those expressing additional transactivators and putative oncogenes, are induced in a cascade fashion following Rta expression. In this study, we sought to define Rta's direct targets during reactivation by generating a conditionally nuclear variant of Rta. Wild-type Rta protein is constitutively localized to cell nuclei and contains two putative nuclear localization signals (NLSs). Only one NLS (NLS2; aa 516 to 530) was required for the nuclear localization of Rta, and it relocalized enhanced green fluorescent protein exclusively to cell nuclei. The results of analyses of Rta NLS mutants demonstrated that proper nuclear localization of Rta was required for transactivation and the stimulation of viral reactivation. RTA with NLS1 and NLS2 deleted was fused to the hormone-binding domain of the murine estrogen receptor to generate an Rta variant whose nuclear localization and ability to transactivate and induce reactivation were tightly controlled posttranslationally by the synthetic hormone tamoxifen. We used this strategy in KSHV-infected cells treated with protein synthesis inhibitors to identify direct transcriptional targets of Rta. Rta activated only eight KSHV genes in the absence of de novo protein synthesis. These direct transcriptional targets of Rta were transactivated to different levels and included the genes nut-1/PAN, ORF57/Mta, ORF56/Primase, K2/viral interleukin-6 (vIL-6), ORF37/SOX, K14/vOX, K9/vIRF1, and ORF52. Our data suggest that the induction of most of the KSHV lytic cycle genes requires additional protein expression after the expression of Rta.

Mycobacterium tuberculosis sigma factor E regulon modulates the host inflammatory response.

Mycobacterium tuberculosis survives in macrophages and usually subverts the bactericidal mechanisms of these phagocytes. The understanding of this host-pathogen interaction is relevant for the development of new treatments for tuberculosis. The adaptation of M. tuberculosis to intracellular life depends on its ability to regulate the expression of its genes. Sigma factors are important bacterial transcription activators that bind to the RNA polymerase and give it promoter specificity. Sigma factor E (SigE) controls the expression of genes that are essential for virulence. We have identified the SigE regulon during infection of macrophages, and we analyzed the impact of this regulon on the transcriptional response of phagocytes. Our results indicate that SigE regulates the expression of genes involved in the maintenance of M. tuberculosis cell envelope integrity and function during macrophage infection. Analysis of the phagocytes' transcriptional response indicates that the SigE regulon is involved in the modulation of the inflammatory response.

Global transcriptional profile of Mycobacterium tuberculosis during THP-1 human macrophage infection.

During lung infection, Mycobacterium tuberculosis resides in macrophages and subverts the bactericidal mechanisms of these professional phagocytes. Comprehension of this host-pathogen relationship is fundamental for the development of new therapies to cure and prevent tuberculosis. In this work, we analyzed the transcriptional profile of M. tuberculosis infecting human macrophage-like THP-1 cells in order to identify putative bacterial pathogenic factors that can be relevant for the intracellular survival of M. tuberculosis. We compared the gene expression profile of M. tuberculosis H37Rv after 4 h and 24 h of infection of human macrophage-like THP-1 cells with the gene expression profile of the strain growing exponentially in broth cultures. We found 585 genes expressed differentially by intracellular M. tuberculosis. An analysis of the gene expression profile of M. tuberculosis inside THP-1 cells suggests the perturbation of the cell envelope as a major intracellular stress inside THP-1 macrophages.

A triclosan-ciprofloxacin cross-resistant mutant strain of Staphylococcus aureus displays an alteration in the expression of several cell membrane structural and functional genes.

Triclosan is an antimicrobial agent found in many consumer products. Triclosan inhibits the bacterial fatty acid biosynthetic enzyme, enoyl-ACP reductase (FabI). Decreased susceptibility to triclosan correlates with ciprofloxacin resistance in several bacteria. In these bacteria, resistance to both drugs maps to genes encoding multi-drug efflux pumps. The focus of this study was to determine whether triclosan resistance contributes to ciprofloxacin resistance in Staphylococcus aureus. In S. aureus, triclosan resistance maps to a fabI homolog and ciprofloxacin resistance maps to genes encoding DNA gyrase, topoisomerase IV and to the multi-drug efflux pump, NorA. Using a norA overexpressing mutant, we demonstrated that upregulation of NorA does not lead to triclosan resistance. To further investigate triclosan/ciprofloxacin resistance in S. aureus, we isolated triclosan/ciprofloxacin-resistant mutants. The mutants were screened for mutations in the genes encoding the targets of triclosan and ciprofloxacin. One mutant, JJ5, was wild-type for all sequences analyzed. We next monitored the efflux of triclosan from JJ5 and determined that triclosan resistance in the mutant was not due to active efflux of the drug. Finally, gene expression profiling demonstrated that an alteration in cell membrane structural and functional gene expression is likely responsible for triclosan and ciprofloxacin resistance in JJ5.

MicroRNA let-7a down-regulates MYC and reverts MYC-induced growth in Burkitt lymphoma cells.

Regulation of the MYC oncogene remains unclear. Using 10058-F4, a compound that inhibits MYC-MAX transcription factor, MYC protein and gene expression were down-regulated in Namalwa cells, a Burkitt lymphoma. Compound 10058-F4 decreased MYC mRNA (45%), MYC protein (50%), and cell growth (32%). MYC-MAX transcription factor was disrupted 24 h after treatment, resulting in transcriptional inhibition of target genes. Because microRNAs (miRNA) disrupt mRNA translation, let-7a, let-7b, and mir-98 were selected using bioinformatics for targeting MYC. Inhibition of MYC-MAX transcription factor with 10058-F4 increased levels of members of the let-7 family. In inhibited cells at 24 h, let-7a, let-7b, and mir-98 were induced 4.9-, 1.3-, and 2.4-fold, respectively, whereas mir-17-5p decreased 0.23-fold. These results were duplicated using microRNA multianalyte suspension array technology. Regulation of MYC mRNA by let-7a was confirmed by transfections with pre-let-7a. Overexpression of let-7a (190%) decreased Myc mRNA (70%) and protein (75%). Down-regulation of Myc protein and mRNA using siRNA MYC also elevated let-7a miRNA and decreased Myc gene expression. Inverse coordinate regulation of let-7a and mir-17-5p versus Myc mRNA by 10058-F4, pre-let-7a, or siRNA MYC suggested that both miRNAs are Myc-regulated. This supports previous results in lung and colon cancer where decreased levels of the let-7 family resulted in increased tumorigenicity. Here, pre-let-7a transfections led to down-regulation of expression of MYC and its target genes and antiproliferation in lymphoma cells. These findings with let-7a add to the complexity of MYC regulation and suggest that dysregulation of these miRNAs participates in the genesis and maintenance of the lymphoma phenotype in Burkitt lymphoma cells and other MYC-dysregulated cancers.

Molecular signatures of trauma-hemorrhagic shock-induced lung injury: hemorrhage- and injury-associated genes.

The etiology of trauma-hemorrhagic shock (T/HS)-induced acute lung injury has been difficult to elucidate because of, at least in part, the inability of in vivo studies to separate the noninjurious pulmonary effects of trauma-hemorrhage from the tissue-injurious ones. To circumvent this in vivo limitation, we used a model of T/HS in which T/HS lung injury was abrogated by dividing the mesenteric lymph duct. In this way, it was possible to separate the pulmonary injurious response from the noninjurious systemic response to T/HS by comparing the pulmonary molecular responses of rats subjected to T/HS, which did and did not develop lung injury, with those of nonshocked rats. Using high-density oligonucleotide arrays and treatment group comparisons of whole lung tissue collected at 3 h after the end of the shock or sham-shock period, 139 of 8,799 assessed genes were identified by significant analysis of microarrays. Hemorrhage without the secondary effects of lung injury modulated the expression of 21 genes such as interleukin 1beta, metallothionein-2, and myeloctomatosis oncogene (c-myc). In response to injury, 42 genes were identified to be differentially expressed. Upregulated genes included the L1 retroposon and guanine deaminase, whereas downregulated genes included catalase and superoxide dismutase 1. Real-time polymerase chain reaction confirmed the differential expression for selected genes. PathwayAssist analysis identified interleukin 1beta as a central regulator of two subpathways of stress response-related genes (c-myc and superoxide dismutase 1/catalase) as well as several unrelated genes such as lipoprotein lipase. Our model system provided a unique opportunity to distinguish the molecular changes associated with T/HS-induced acute lung injury from the systemic molecular response to T/HS.

A micro-RNA signature associated with race, tumor size, and target gene activity in human uterine leiomyomas.

Human uterine leiomyomas (ULMs) are the most common neoplasms of women. Many genes are dysregulated in ULMs and some of this dysregulation may be due to abnormal expression of micro-RNAs (miRNAs). In this study, 55 ULMs and matched myometrium were collected from 41 patients for microarray-based global miRNA expression analysis. Of 206 miRNAs examined, 45 miRNAs were significantly up- or down-regulated in ULMs in comparison to the matched myometrium (P < 0.001). The top five dysregulated miRNAs in ULMs are the let-7 family, miR-21, miR-23b, miR-29b, and miR-197. Four polycistronic clusters of miRNAs were either up- or down-regulated, but not in a mixed pattern, indicative of coordinated regulation of these miRNAs. Significance analysis revealed that subsets of miRNAs were strongly associated with tumor sizes and race. By prediction analysis we identified some important tumorigenic genes previously identified in ULMs that may be targeted by the dysregulated miRNAs. HMGA2 was identified as one of target genes of the let-7 family of miRNAs and has been found to be suppressed by let-7 in vitro. This article contains Supplementary material available at http://www.interscience.wiley.com/jpages/1045-2257/suppmat.

Analysis of Epstein-Barr virus reservoirs in paired blood and breast cancer primary biopsy specimens by real time PCR.

INTRODUCTION: Epstein-Barr virus (EBV) is present in over 90% of the world's population. This infection is considered benign, even though in limited cases EBV is associated with infectious and neoplastic conditions. Over the past decade, the EBV association with breast cancer has been constantly debated. Adding to this clinical and biological uncertainty, different techniques gave contradictory results for the presence of EBV in breast carcinoma specimens. In this study, minor groove binding (MGB)-TaqMan real time PCR was used to detect the presence of EBV DNA in both peripheral blood and tumor samples of selected patients. METHODS: Peripheral blood and breast carcinoma specimens from 24 patients were collected. DNA was extracted and then amplified by MGB-TaqMan real time PCR. RESULTS: Of 24 breast tumor specimens, 11 (46%) were positive for EBV DNA. Of these 11 breast tumor specimens, 7 (64%) were also positive for EBV DNA in the peripheral blood, while 4 (36%) were positive for EBV DNA in the tumor, but negative in the blood. CONCLUSION: EBV was found at extremely low levels, with a mean of 0.00004 EBV genomes per cell (range 0.00014 to 0.00001 EBV genomes per cell). Furthermore, our finding of the presence of EBV in the tumor specimens coupled to the absence of detection of EBV genomic DNA in the peripheral blood is consistent with the epithelial nature of the virus. Because of the low levels of viral DNA in tumor tissue, further studies are needed to assess the biological input of EBV in breast cancer.

Prostaglandin J2 alters pro-survival and pro-death gene expression patterns and 26 S proteasome assembly in human neuroblastoma cells.

Many neurodegenerative disorders are characterized by two pathological hallmarks: progressive loss of neurons and occurrence of inclusion bodies containing ubiquitinated proteins. Inflammation may be critical to neurodegeneration associated with ubiquitin-protein aggregates. We previously showed that prostaglandin J2 (PGJ2), one of the endogenous products of inflammation, induces neuronal death and the accumulation of ubiquitinated proteins into distinct aggregates. We now report that temporal microarray analysis of human neuroblastoma SK-N-SH revealed that PGJ2 triggered a "repair" response including increased expression of heat shock, protein folding, stress response, detoxification and cysteine metabolism genes. PGJ2 also decreased expression of cell growth/maintenance genes and increased expression of apoptotic genes. Over time pro-death responses prevailed over pro-survival responses, leading to cellular demise. Furthermore, PGJ2 increased the expression of proteasome and other ubiquitin-proteasome pathway genes. This increase failed to overcome PGJ2 inhibition of 26 S proteasome activity. Ubiquitinated proteins are degraded by the 26 S proteasome, shown here to be the most active proteasomal form in SK-N-SH cells. We demonstrate that PGJ2 impairs 26 S proteasome assembly, which is an ATP-dependent process. PGJ2 perturbs mitochondrial function, which could be critical to the observed 26 S proteasome disassembly, suggesting a cross-talk between mitochondrial and proteasomal impairment. In conclusion neurotoxic products of inflammation, such as PGJ2, may play a role in neurodegenerative disorders associated with the aggregation of ubiquitinated proteins by impairing 26 S proteasome activity and inducing a chain of events that culminates in neuronal cell death. Temporal characterization of these events is relevant to understanding the underlying mechanisms and to identifying potential early biomarkers.

Noise filtering and nonparametric analysis of microarray data underscores discriminating markers of oral, prostate, lung, ovarian and breast cancer.

BACKGROUND: A major goal of cancer research is to identify discrete biomarkers that specifically characterize a given malignancy. These markers are useful in diagnosis, may identify potential targets for drug development, and can aid in evaluating treatment efficacy and predicting patient outcome. Microarray technology has enabled marker discovery from human cells by permitting measurement of steady-state mRNA levels derived from thousands of genes. However many challenging and unresolved issues regarding the acquisition and analysis of microarray data remain, such as accounting for both experimental and biological noise, transcripts whose expression profiles are not normally distributed, guidelines for statistical assessment of false positive/negative rates and comparing data derived from different research groups. This study addresses these issues using Affymetrix HG-U95A and HG-U133 GeneChip data derived from different research groups. RESULTS: We present here a simple non parametric approach coupled with noise filtering to identify sets of genes differentially expressed between the normal and cancer states in oral, breast, lung, prostate and ovarian tumors. An important feature of this study is the ability to integrate data from different laboratories, improving the analytical power of the individual results. One of the most interesting findings is the down regulation of genes involved in tissue differentiation. CONCLUSIONS: This study presents the development and application of a noise model that suppresses noise, limits false positives in the results, and allows integration of results from individual studies derived from different research groups.

Functional cloning, sorting, and expression profiling of nucleic acid-binding proteins.

A major challenge in the post-sequencing era is to elucidate the activity and biological function of genes that reside in the human genome. An important subset includes genes that encode proteins that regulate gene expression or maintain the structural integrity of the genome. Using a novel oligonucleotide-binding substrate as bait, we show the feasibility of a modified functional expression-cloning strategy to identify human cDNAs that encode a spectrum of nucleic acid-binding proteins (NBPs). Approximately 170 cDNAs were identified from screening phage libraries derived from a human colorectal adenocarcinoma cell line and from noncancerous fetal lung tissue. Sequence analysis confirmed that virtually every clone contained a known DNA- or RNA-binding motif. We also report on a complementary sorting strategy that, in the absence of subcloning and protein purification, can distinguish different classes of NBPs according to their particular binding properties. To extend our functional annotation of NBPs, we have used GeneChip expression profiling of 14 different breast-derived cell lines to examine the relative transcriptional activity of genes identified in our screen and cluster analysis to discover other genes that have similar expression patterns. Finally, we present strategies to analyze the upstream regulatory region of each gene within a cluster group and select unique combinations of transcription factor binding sites that may be responsible for dictating the observed synexpression.