George C. Tseng

MicroRNA Expression Profiling of Thyroid Tumors: Biological Significance and Diagnostic Utility

OBJECTIVE MicroRNA (miRNA) expression is deregulated in many types of human cancers. We sought to investigate the expression patterns of miRNA in all major types of thyroid tumors, including tumors carrying distinct oncogenic mutations, and to explore the utility of miRNA profiling for the preoperative diagnosis of thyroid nodules. DESIGN miRNA expression levels were detected in 60 surgically removed thyroid neoplastic and nonneoplastic samples and in 62 fine-needle aspiration (FNA) samples by RT-PCR using TaqMan MicroRNA Panel or individual miRNA sequence-specific primers. miRNA expression levels were calculated relative to normal thyroid tissue. All tumors were genotyped for most common mutations. RESULTS Various histopathological types of thyroid tumors, including those deriving from the same cell type, showed significantly different profiles of miRNA expression. Oncocytic tumors, conventional follicular tumors, papillary carcinomas, and medullary carcinomas formed distinct clusters on the unsupervised hierarchical clustering analysis. Significant correlation between miRNA expression patterns and somatic mutations was observed in papillary carcinomas. A set of seven miRNAs (miR-187, miR-221, miR-222, miR-146b, miR-155, miR-224, and miR-197) that were most differentially overexpressed in thyroid tumors vs. hyperplastic nodules in the surgical samples was validated in the FNA samples, showing high accuracy of thyroid cancer detection. CONCLUSIONS In this study, we demonstrate that various histopathological types of thyroid tumors have distinct miRNA profiles, which further differ within the same tumor type, reflecting specific oncogenic mutations. A limited set of miRNAs can be used diagnostically with high accuracy to detect thyroid cancer in the surgical and preoperative FNA samples.

Comprehensive literature review and statistical considerations for GWAS meta-analysis

With the rapid advances of various high-throughput technologies, generation of ‘-omics’ data is commonplace in almost every biomedical field. Effective data management and analytical approaches are essential to fully decipher the biological knowledge contained in the tremendous amount of experimental data. Meta-analysis, a set of statistical tools for combining multiple studies of a related hypothesis, has become popular in genomic research. Here, we perform a systematic search from PubMed and manual collection to obtain 620 genomic meta-analysis papers, of which 333 microarray meta-analysis papers are summarized as the basis of this paper and the other 249 GWAS meta-analysis papers are discussed in the next companion paper. The review in the present paper focuses on various biological purposes of microarray meta-analysis, databases and software and related statistical procedures. Statistical considerations of such an analysis are further scrutinized and illustrated by a case study. Finally, several open questions are listed and discussed.

Evaluation and comparison of gene clustering methods in microarray analysis

MOTIVATION Microarray technology has been widely applied in biological and clinical studies for simultaneous monitoring of gene expression in thousands of genes. Gene clustering analysis is found useful for discovering groups of correlated genes potentially co-regulated or associated to the disease or conditions under investigation. Many clustering methods including hierarchical clustering, K-means, PAM, SOM, mixture model-based clustering and tight clustering have been widely used in the literature. Yet no comprehensive comparative study has been performed to evaluate the effectiveness of these methods. RESULTS In this paper, six gene clustering methods are evaluated by simulated data from a hierarchical log-normal model with various degrees of perturbation as well as four real datasets. A weighted Rand index is proposed for measuring similarity of two clustering results with possible scattered genes (i.e. a set of noise genes not being clustered). Performance of the methods in the real data is assessed by a predictive accuracy analysis through verified gene annotations. Our results show that tight clustering and model-based clustering consistently outperform other clustering methods both in simulated and real data while hierarchical clustering and SOM perform among the worst. Our analysis provides deep insight to the complicated gene clustering problem of expression profile and serves as a practical guideline for routine microarray cluster analysis.

Molecular evidence for BDNF- and GABA-related dysfunctions in the amygdala of female subjects with Major Depression

Women are twice as likely as men to develop major depressive disorder (MDD) and are more prone to recurring episodes. Hence, we tested the hypothesis that the illness may associate with robust molecular changes in female subjects, and investigated large-scale gene expression in the post-mortem brain of MDD subjects paired with matched controls (n=21 pairs). We focused on the lateral/basolateral/basomedian complex of the amygdala as a neural hub of mood regulation affected in MDD. Among the most robust findings were downregulated transcripts for genes coding for γ-aminobutyric acid (GABA) interneuron-related peptides, including somatostatin (SST), tachykinin, neuropeptide Y (NPY) and cortistatin, in a pattern reminiscent to that previously reported in mice with low brain-derived neurotrophic factor (BDNF). Changes were confirmed by quantitative PCR and not explained by demographic, technical or known clinical parameters. BDNF itself was significantly downregulated at the RNA and protein levels in MDD subjects. Investigating putative mechanisms, we show that this core MDD-related gene profile (including SST, NPY, TAC1, RGS4 and CORT) is recapitulated by complementary patterns in mice with constitutive (BDNF-heterozygous) or activity-dependent (exon IV knockout) decreases in BDNF function, with a common effect on SST and NPY. Together, these results provide both direct (low RNA/protein) and indirect (low BDNF-dependent gene pattern) evidence for reduced BDNF function in the amygdala of female subjects with MDD. Supporting studies in mutant mice models suggest a complex mechanism of low constitutive and activity-dependent BDNF function in MDD, particularly affecting SST/NPY-related GABA neurons, thus linking the neurotrophic and GABA hypotheses of depression.

Glutathione Peroxidase 3, Deleted or Methylated in Prostate Cancer, Suppresses Prostate Cancer Growth and Metastasis

Glutathione peroxidase 3 is a selenium-dependent enzyme playing a critical role in detoxifying reactive oxidative species and maintaining the genetic integrity of mammalian cells. In this report, we found that the expression of glutathione peroxidase 3 (GPx3) was widely inactivated in prostate cancers. Complete inactivation of GPx3 correlates with a poor clinical outcome. Deletions (hemizygous and homozygous) of GPx3 gene are frequent in prostate cancer samples, occurring in 39% of the samples studied. The rate of methylation of the GPx3 exon 1 region in prostate cancer samples reaches 90%. Overexpression of GPx3 in prostate cancer cell lines induced the suppression of colony formation and anchorage-independent growth of PC3, LNCaP, and Du145 cells. PC3 cells overexpressing GPx3 reduced invasiveness in Matrigel transmigration analysis by an average of 2.7-fold. Xenografted PC3 cells expressing GPx3 showed reduction in tumor volume by 4.8-fold, elimination of metastasis (0/16 versus 7/16), and reduction of animal death (3/16 versus 16/16). The tumor suppressor activity of GPx3 seems to relate to its ability to suppress the expression of c-met. The present findings suggest that GPx3 is a novel tumor suppressor gene.

On ψ-Learning

The concept of large margins have been recognized as an important principle in analyzing learning methodologies, including boosting, neural networks, and support vector machines (SVMs). However, this concept alone is not adequate for learning in nonseparable cases. We propose a learning methodology, called ψ-learning, that is derived from a direct consideration of generalization errors. We provide a theory for ψ-learning and show that it essentially attains the optimal rates of convergence in two learning examples. Finally, results from simulation studies and from breast cancer classification confirm the ability of ψ-learning to outperform SVM in generalization.

MCM7 amplification and overexpression are associated with prostate cancer progression.

The genomic DNA profiles of prostate cancers with aggressive features were compared to the profiles of matched normal DNA to identify genes that are selectively amplified in the cancer cells. One of the identified genes, MCM7, which is a component of the DNA replication licensing complex, has been studied extensively both at the DNA and protein levels in human prostate tissues. Approximately half of the prostate cancer specimens studied showed MCM7 gene amplification, and 60% of the aggressive prostate cancer specimens had increased MCM7 protein expression. Amplification or overexpression of MCM7 was significantly associated with relapse, local invasion and a worse tumor grade. Constitutive expression of MCM7 in a human prostate cancer cell line, DU145, resulted in markedly increased DNA synthesis and cell proliferation compared to vector-only controls, and an increased cell invasion in vitro. Indeed, MCM7 overexpression produced primary tumors 12 times larger than vector-only controls and resulted in a rapid demise of mice bearing those tumors. These studies implicate MCM7, and the DNA replication licensing gene family, in prostate cancer progression, growth and invasion.

Peripheral Blood Mononuclear Cell Gene Expression Profiles Predict Poor Outcome in Idiopathic Pulmonary Fibrosis

Genome-scale transcriptomic profiling of peripheral blood mononuclear cells from patients with idiopathic pulmonary fibrosis reveals that decreased expression of CD28, ICOS, LCK, and ITK predicts mortality. Gene Signature Predicts Mortality Idiopathic pulmonary fibrosis (IPF) is a fatal disease that progresses at different rates. Although no therapies exist, giving patients a more accurate prognosis is highly desirable. To this end, Herazo-Maya and colleagues searched the genomes of cells circulating in the blood of IPF patients and found that four genes may be indicators of poor outcome. Patients were recruited into discovery or replication cohorts from two different medical centers in the United States and followed until death or completion of the study. In both groups, genetic material was isolated from the patients’ peripheral blood mononuclear cells (PBMCs) and analyzed for increased or decreased expression. These gene expression profiles were then correlated with transplant-free survival (TFS). In the discovery cohort, Herazo-Maya et al. found that underexpression of the genes CD28, ICOS, LCK, and ITK was associated with decreased TFS. These findings were confirmed in the replication cohort. This “genomic model” incorporating the four genes was combined with the clinical outputs age, gender, and forced vital capacity to create an even stronger predictor of poor outcome. The authors suggest that the decreased expression of these genes might be linked to lower percentages of CD4+CD28+ T cells in the PBMC population, which could contribute to a mechanistic understanding of why some IPF patients progress differently than others. The findings of this study have the potential to affect the care of patients with IPF as well as the understanding of disease mechanism. However, the combined genomic and clinical predictor will need to be validated in additional independent cohorts before translation. We aimed to identify peripheral blood mononuclear cell (PBMC) gene expression profiles predictive of poor outcomes in idiopathic pulmonary fibrosis (IPF) by performing microarray experiments of PBMCs in discovery and replication cohorts of IPF patients. Microarray analyses identified 52 genes associated with transplant-free survival (TFS) in the discovery cohort. Clustering the microarray samples of the replication cohort using the 52-gene outcome-predictive signature distinguished two patient groups with significant differences in TFS. We studied the pathways associated with TFS in each independent microarray cohort and identified decreased expression of “The costimulatory signal during T cell activation” Biocarta pathway and, in particular, the genes CD28, ICOS, LCK, and ITK, results confirmed by quantitative reverse transcription polymerase chain reaction (qRT-PCR). A proportional hazards model, including the qRT-PCR expression of CD28, ICOS, LCK, and ITK along with patient’s age, gender, and percent predicted forced vital capacity (FVC%), demonstrated an area under the receiver operating characteristic curve of 78.5% at 2.4 months for death and lung transplant prediction in the replication cohort. To evaluate the potential cellular source of CD28, ICOS, LCK, and ITK expression, we analyzed and found significant correlation of these genes with the PBMC percentage of CD4+CD28+ T cells in the replication cohort. Our results suggest that CD28, ICOS, LCK, and ITK are potential outcome biomarkers in IPF and should be further evaluated for patient prioritization for lung transplantation and stratification in drug studies.

A combined molecular-pathologic score improves risk stratification of thyroid papillary microcarcinoma.

Thyroid papillary microcarcinoma (TPMC) is an incidentally discovered papillary carcinoma that measures ≤1.0 cm in size. Most TPMCs are indolent, whereas some behave aggressively. The objective of the study was to evaluate whether the combination of v‐raf murine sarcoma viral oncogene homolog B1 (BRAF) mutation and specific histopathologic features allows risk stratification of TPMC.

Mechanisms of hepatocyte growth factor–mediated and epidermal growth factor–mediated signaling in transdifferentiation of rat hepatocytes to biliary epithelium

Previous studies from our laboratory have demonstrated that hepatocytes can transdifferentiate into biliary epithelium (BE) both in vivo and in vitro; however, the mechanisms are unclear. The current study was designed to investigate the mechanisms of hepatocyte transdifferentiation in vitro. Rat hepatocytes were cultured in roller bottles to obtain hepatocyte organoid cultures, which were stimulated with various growth factors (GFs) including hepatocyte growth factor (HGF), epidermal growth factor (EGF), vascular endothelial growth factor (VEGF), platelet‐derived growth factor (PDGF), stem cell factor (SCF), macrophage‐stimulating protein (MSP), fibroblast growth factor‐a (FGF‐a), fibroblast growth factor‐b (FGF‐b), and fibroblast growth factor‐8b (FGF‐8b). Only the cultures treated with HGF, EGF, and their combination exhibited formation of hepatocyte‐derived biliary epithelium (BE) despite the presence and activation of all the pertinent cognate membrane receptors of the rest of the GFs. Microarray analysis of the organoid cultures identified specific up‐regulation of approximately 500 target genes induced by HGF and EGF, including members of the extracellular matrix (ECM) protein family, Wnt/β‐catenin pathway, transforming growth factor beta (TGF‐β)/bone morphogenetic protein (BMP) pathway, and CXC (cysteine‐any amino acid‐cysteine) chemokines. To investigate the downstream signaling involved in hepatocyte to biliary epithelial cell (BEC) transdifferentiation, we investigated expression and activities of mitogen‐activated protein (MAP) kinases [extracellular signal‐regulated kinase (ERK)1/2, p38, and c‐Jun N‐terminal kinase (JNK)/stress‐activated protein kinase (SAPK)] as well as serine/threonine kinase AKT. The analysis indicated that AKT phosphorylation was particularly increased in cultures treated with HGF, EGF, and their combination. Whereas phosphatidylinositol 3‐kinase (PI3K) inhibitor LY294002 completely inhibited biliary epithelium formation, AKT inhibitor could only moderately reduce formation of BE in the organoid cultures treated with HGF+EGF. Most of the HGF+EGF target genes were altered by LY294002. Conclusion: Taken together, these data indicate that hepatocyte to BE transdifferentiation is regulated by HGF and EGF receptors and that PI3 kinase–mediated signaling independent of AKT is a crucial component of the transdifferentiation process. (HEPATOLOGY 2008.)