Identification of a novel gene expression signature associated with overall survival in patients with lung adenocarcinoma: A comprehensive analysis based on TCGA and GEO databases.

OBJECTIVES: Lung adenocarcinoma (LUAD) is the most common subtype of non-small cell lung cancer. Understanding the molecular mechanisms underlying tumor progression is of clinical significance. This study aimed to identify novel molecular markers associated with LUAD prognosis. MATERIALS AND METHODS: RNA sequencing data from the Cancer Genome Atlas (TCGA) database of LUAD tumors and paired normal tissues, and microarray data from the Gene Expression Omnibus (GEO) database were obtained. In the TCGA dataset, differentially expressed (DE) genes were identified by comparing gene expression between early-stage tumors and normal tissue, as well as between advanced-stage and early-stage tumors. A risk score was developed using a weighted linear combination of individual dysregulated protein-coding genes that was associated with overall survival (OS). The prognostic value of the risk score was evaluated using Kaplan-Meier and multivariate Cox analysis. The gene signature was further validated using independent datasets from GEO. RESULTS: Among the 68 identified DE genes, 19 were individually associated with OS in univariate analyses. A risk score was constructed for each patient based on the coefficients in multivariate Cox model and normalized expression levels of these 19 genes. LUAD patients with a low risk score had a significantly better survival than those with a high risk score (log-rank P < 0.0001). After adjusting for age, sex, clinical stage, smoking history, and treatments, the patients with a low risk score had a 81 % decreased risk for death, compared to those with a high risk score (hazard ratio 0.19, 95 % confidence interval 0.097-0.36). The significant association of the risk score with OS in LUAD patients was further validated in three independent GEO datasets. CONCLUSION: A novel 19-gene prognostic signature based on gene expression was identified in LUAD patients. The findings further improve the understanding of LUAD prognostication and have the potential to facilitate risk-stratified disease management.

Machine learning-based genome-wide interrogation of somatic copy number aberrations in circulating tumor DNA for early detection of hepatocellular carcinoma.

BACKGROUND: DNAs released from tumor cells into blood (circulating tumor DNAs, ctDNAs) carry tumor-specific genomic aberrations, providing a non-invasive means for cancer detection. In this study, we aimed to leverage somatic copy number aberration (SCNA) in ctDNA to develop assays to detect early-stage HCCs. METHODS: We conducted low-depth whole-genome sequencing (WGS) to profile SCNAs in 384 plasma samples of hepatitis B virus (HBV)-related HCC and cancer-free HBV patients, using one discovery and two validation cohorts. To fully capture the robust signals of WGS data from the complete genome, we developed a machine learning-based statistical model that is focused on detection accuracy in early-stage HCC. FINDINGS: We built the model using a discovery cohort of 209 patients, achieving an overall area under curve (AUC) of 0.893, with 0.874 for early-stage (Barcelona clinical liver cancer [BCLC] stage 0-A) and 0.933 for advanced-stage (BCLC stage B-D). The performance of the model was then assessed in two validation cohorts (76 and 99 patients) that only consisted of patients with stage 0-A HCC. Our model exhibited a robust predictive performance, with an AUC of 0.920 and 0.812 for the two validation cohorts. Further analyses showed the impact of tumor sample heterogeneity in model training on detecting early-stage tumors, and a refined model addressing the heterogeneity in the discovery cohort significantly increased model performance in validation. INTERPRETATION: We developed an SCNA-based, machine learning-driven model in the non-invasive detection of early-stage HCC in HBV patients and demonstrated its performance through strict independent validations.

Nuclear receptor HNF4α binding sequences are widespread in Alu repeats.

BACKGROUND: Alu repeats, which account for ~10% of the human genome, were originally considered to be junk DNA. Recent studies, however, suggest that they may contain transcription factor binding sites and hence possibly play a role in regulating gene expression. RESULTS: Here, we show that binding sites for a highly conserved member of the nuclear receptor superfamily of ligand-dependent transcription factors, hepatocyte nuclear factor 4alpha (HNF4α, NR2A1), are highly prevalent in Alu repeats. We employ high throughput protein binding microarrays (PBMs) to show that HNF4α binds > 66 unique sequences in Alu repeats that are present in ~1.2 million locations in the human genome. We use chromatin immunoprecipitation (ChIP) to demonstrate that HNF4α binds Alu elements in the promoters of target genes (ABCC3, APOA4, APOM, ATPIF1, CANX, FEMT1A, GSTM4, IL32, IP6K2, PRLR, PRODH2, SOCS2, TTR) and luciferase assays to show that at least some of those Alu elements can modulate HNF4α-mediated transactivation in vivo (APOM, PRODH2, TTR, APOA4). HNF4α-Alu elements are enriched in promoters of genes involved in RNA processing and a sizeable fraction are in regions of accessible chromatin. Comparative genomics analysis suggests that there may have been a gain in HNF4α binding sites in Alu elements during evolution and that non Alu repeats, such as Tiggers, also contain HNF4α sites. CONCLUSIONS: Our findings suggest that HNF4α, in addition to regulating gene expression via high affinity binding sites, may also modulate transcription via low affinity sites in Alu repeats.

Dynamic chromatin states in human ES cells reveal potential regulatory sequences and genes involved in pluripotency.

Pluripotency, the ability of a cell to differentiate and give rise to all embryonic lineages, defines a small number of mammalian cell types such as embryonic stem (ES) cells. While it has been generally held that pluripotency is the product of a transcriptional regulatory network that activates and maintains the expression of key stem cell genes, accumulating evidence is pointing to a critical role for epigenetic processes in establishing and safeguarding the pluripotency of ES cells, as well as maintaining the identity of differentiated cell types. In order to better understand the role of epigenetic mechanisms in pluripotency, we have examined the dynamics of chromatin modifications genome-wide in human ES cells (hESCs) undergoing differentiation into a mesendodermal lineage. We found that chromatin modifications at promoters remain largely invariant during differentiation, except at a small number of promoters where a dynamic switch between acetylation and methylation at H3K27 marks the transition between activation and silencing of gene expression, suggesting a hierarchy in cell fate commitment over most differentially expressed genes. We also mapped over 50 000 potential enhancers, and observed much greater dynamics in chromatin modifications, especially H3K4me1 and H3K27ac, which correlate with expression of their potential target genes. Further analysis of these enhancers revealed potentially key transcriptional regulators of pluripotency and a chromatin signature indicative of a poised state that may confer developmental competence in hESCs. Our results provide new evidence supporting the role of chromatin modifications in defining enhancers and pluripotency.

Integrated approach for the identification of human hepatocyte nuclear factor 4alpha target genes using protein binding microarrays.

UNLABELLED: Hepatocyte nuclear factor 4 alpha (HNF4alpha), a member of the nuclear receptor superfamily, is essential for liver function and is linked to several diseases including diabetes, hemophilia, atherosclerosis, and hepatitis. Although many DNA response elements and target genes have been identified for HNF4alpha, the complete repertoire of binding sites and target genes in the human genome is unknown. Here, we adapt protein binding microarrays (PBMs) to examine the DNA-binding characteristics of two HNF4alpha species (rat and human) and isoforms (HNF4alpha2 and HNF4alpha8) in a high-throughput fashion. We identified approximately 1400 new binding sequences and used this dataset to successfully train a Support Vector Machine (SVM) model that predicts an additional approximately 10,000 unique HNF4alpha-binding sequences; we also identify new rules for HNF4alpha DNA binding. We performed expression profiling of an HNF4alpha RNA interference knockdown in HepG2 cells and compared the results to a search of the promoters of all human genes with the PBM and SVM models, as well as published genome-wide location analysis. Using this integrated approach, we identified approximately 240 new direct HNF4alpha human target genes, including new functional categories of genes not typically associated with HNF4alpha, such as cell cycle, immune function, apoptosis, stress response, and other cancer-related genes. CONCLUSION: We report the first use of PBMs with a full-length liver-enriched transcription factor and greatly expand the repertoire of HNF4alpha-binding sequences and target genes, thereby identifying new functions for HNF4alpha. We also establish a web-based tool, HNF4 Motif Finder, that can be used to identify potential HNF4alpha-binding sites in any sequence.

MicroRNA 92b controls the G1/S checkpoint gene p57 in human embryonic stem cells.

Human embryonic stem (ES) cells exhibit a shorter G(1) cell cycle phase than most somatic cells. Here, we examine the role of an abundant, human ES cell-enriched microRNA, miR-92b, in cell cycle distribution. Inhibition of miR-92b in human ES cells results in a greater number of cells in the G(1) phase and a lower number in the S phase. Conversely, overexpression of miR-92b in differentiated cells results in a decreased number of cells in G1 phase and an increased number in S-phase. p57, a gene whose product inhibits G(1) to S-phase progression, is one of the predicted targets of miR-92b. Inhibition of miR-92b in human ES cells increases p57 protein levels, and miR-92b overexpression in differentiated cells decreases p57 protein levels. Furthermore, miR-92b inhibits a luciferase reporter construct that includes part of the 3' untranslated region of the p57 gene containing the predicted target of the miR-92b seed sequence. Thus, we show that the miRNA miR-92b directly downregulates protein levels of the G(1)/S checkpoint gene p57. STEM CELLS 2009;27:1524-1528.

A study of the relationships between oligonucleotide properties and hybridization signal intensities from NimbleGen microarray datasets.

Well-defined relationships between oligonucleotide properties and hybridization signal intensities (HSI) can aid chip design, data normalization and true biological knowledge discovery. We clarify these relationships using the data from two microarray experiments containing over three million probes from 48 high-density chips. We find that melting temperature (T(m)) has the most significant effect on HSI while length for the long oligonucleotides studied has very little effect. Analysis of positional effect using a linear model provides evidence that the protruding ends of probes contribute more than tethered ends to HSI, which is further validated by specifically designed match fragment sliding and extension experiments. The impact of sequence similarity (SeqS) on HSI is not significant in comparison with other oligonucleotide properties. Using regression and regression tree analysis, we prioritize these oligonucleotide properties based on their effects on HSI. The implications of our discoveries for the design of unbiased oligonucleotides are discussed. We propose that isothermal probes designed by varying the length is a viable strategy to reduce sequence bias, though imposing selection constraints on other oligonucleotide properties is also essential.

Prevalence of the initiator over the TATA box in human and yeast genes and identification of DNA motifs enriched in human TATA-less core promoters.

The core promoter of eukaryotic genes is the minimal DNA region that recruits the basal transcription machinery to direct efficient and accurate transcription initiation. The fraction of human and yeast genes that contain specific core promoter elements such as the TATA box and the initiator (INR) remains unclear and core promoter motifs specific for TATA-less genes remain to be identified. Here, we present genome-scale computational analyses indicating that approximately 76% of human core promoters lack TATA-like elements, have a high GC content, and are enriched in Sp1-binding sites. We further identify two motifs - M3 (SCGGAAGY) and M22 (TGCGCANK) - that occur preferentially in human TATA-less core promoters. About 24% of human genes have a TATA-like element and their promoters are generally AT-rich; however, only approximately 10% of these TATA-containing promoters have the canonical TATA box (TATAWAWR). In contrast, approximately 46% of human core promoters contain the consensus INR (YYANWYY) and approximately 30% are INR-containing TATA-less genes. Significantly, approximately 46% of human promoters lack both TATA-like and consensus INR elements. Surprisingly, mammalian-type INR sequences are present - and tend to cluster - in the transcription start site (TSS) region of approximately 40% of yeast core promoters and the frequency of specific core promoter types appears to be conserved in yeast and human genomes. Gene Ontology analyses reveal that TATA-less genes in humans, as in yeast, are frequently involved in basic "housekeeping" processes, while TATA-containing genes are more often highly regulated, such as by biotic or stress stimuli. These results reveal unexpected similarities in the occurrence of specific core promoter types and in their associated biological processes in yeast and humans and point to novel vertebrate-specific DNA motifs that might play a selective role in TATA-independent transcription.

Whole-genome analysis of histone H3 lysine 4 and lysine 27 methylation in human embryonic stem cells.

We mapped Polycomb-associated H3K27 trimethylation (H3K27me3) and Trithorax-associated H3K4 trimethylation (H3K4me3) across the whole genome in human embryonic stem (ES) cells. The vast majority of H3K27me3 colocalized on genes modified with H3K4me3. These commodified genes displayed low expression levels and were enriched in developmental function. Another significant set of genes lacked both modifications and was also expressed at low levels in ES cells but was enriched for gene function in physiological responses rather than development. Commodified genes could change expression levels rapidly during differentiation, but so could a substantial number of genes in other modification categories. SOX2, POU5F1, and NANOG, pluripotency-associated genes, shifted from modification by H3K4me3 alone to colocalization of both modifications as they were repressed during differentiation. Our results demonstrate that H3K27me3 modifications change during early differentiation, both relieving existing repressive domains and imparting new ones, and that colocalization with H3K4me3 is not restricted to pluripotent cells.

Hepatocyte nuclear factor 4alpha orchestrates expression of cell adhesion proteins during the epithelial transformation of the developing liver.

Epithelial formation is a central facet of organogenesis that relies on intercellular junction assembly to create functionally distinct apical and basal cell surfaces. How this process is regulated during embryonic development remains obscure. Previous studies using conditional knockout mice have shown that loss of hepatocyte nuclear factor 4alpha (HNF4alpha) blocks the epithelial transformation of the fetal liver, suggesting that HNF4alpha is a central regulator of epithelial morphogenesis. Although HNF4alpha-null hepatocytes do not express E-cadherin (also called CDH1), we show here that E-cadherin is dispensable for liver development, implying that HNF4alpha regulates additional aspects of epithelial formation. Microarray and molecular analyses reveal that HNF4alpha regulates the developmental expression of a myriad of proteins required for cell junction assembly and adhesion. Our findings define a fundamental mechanism through which generation of tissue epithelia during development is coordinated with the onset of organ function.

Hepatocyte nuclear factor 4alpha is essential for embryonic development of the mouse colon.

BACKGROUND & AIMS: Hepatocyte nuclear factor 4 alpha (HNF4alpha) is a transcription factor that has been shown to be required for hepatocyte differentiation and development of the liver. It has also been implicated in regulating expression of genes that act in the epithelium of the lower gastrointestinal tract. This implied that HNF4alpha might be required for development of the gut. METHODS: Mouse embryos were generated in which Hnf4a was ablated in the epithelial cells of the fetal colon by using Cre-loxP technology. Embryos were examined by using a combination of histology, immunohistochemistry, DNA microarray, reverse-transcription polymerase chain reaction, electrophoretic mobility shift assays, and chromatin immunoprecipitation analyses to define the consequences of loss of HNF4alpha on colon development. RESULTS: Embryos were recovered at E18.5 that lacked HNF4alpha in their colons. Although early stages of colonic development occurred, HNF4alpha-null colons failed to form normal crypts. In addition, goblet-cell maturation was perturbed and expression of an array of genes that encode proteins with diverse roles in colon function was disrupted. Several genes whose expression in the colon was dependent on HNF4alpha contained HNF4alpha-binding sites within putative transcriptional regulatory regions and a subset of these sites were occupied by HNF4alpha in vivo. CONCLUSIONS: HNF4alpha is a transcription factor that is essential for development of the mammalian colon, regulates goblet-cell maturation, and is required for expression of genes that control normal colon function and epithelial cell differentiation.

Identifying transcription factor binding sites through Markov chain optimization.

Even though every cell in an organism contains the same genetic material, each cell does not express the same cohort of genes. Therefore, one of the major problems facing genomic research today is to determine not only which genes are differentially expressed and under what conditions, but also how the expression of those genes is regulated. The first step in determining differential gene expression is the binding of sequence-specific DNA binding proteins (i.e. transcription factors) to regulatory regions of the genes (i.e. promoters and enhancers). An important aspect to understanding how a given transcription factor functions is to know the entire gamut of binding sites and subsequently potential target genes that the factor may bind/regulate. In this study, we have developed a computer algorithm to scan genomic databases for transcription factor binding sites, based on a novel Markov chain optimization method, and used it to scan the human genome for sites that bind to hepatocyte nuclear factor 4 alpha (HNF4alpha). A list of 71 known HNF4alpha binding sites from the literature were used to train our Markov chain model. By looking at the window of 600 nucleotides around the transcription start site of each confirmed gene on the human genome, we identified 849 sites with varying binding potential and experimentally tested 109 of those sites for binding to HNF4alpha. Our results show that the program was very successful in identifying 77 new HNF4alpha binding sites with varying binding affinities (i.e. a 71% success rate). Therefore, this computational method for searching genomic databases for potential transcription factor binding sites is a powerful tool for investigating mechanisms of differential gene regulation.