Persistent gene expression and DNA methylation alterations linked to carcinogenic effects of dichloroacetic acid

Background

Mechanistic understanding of transient exposures that lead to adverse health outcomes will enhance our ability to recognize biological signatures of disease. Here, we measured the transcriptomic and epigenomic alterations due to exposure to the metabolic reprogramming agent, dichloroacetic acid (DCA). Previously, we showed that exposure to DCA increased liver tumor incidence in B6C3F1 mice after continuous or early life exposures significantly over background level.

Methods

Using archived formalin-fixed liver samples, we utilized modern methodologies to measure gene expression and DNA methylation levels to link to previously generated phenotypic measures. Gene expression was measured by targeted RNA sequencing (TempO-seq 1500+ toxicity panel: 2754 total genes) in liver samples collected from 10-, 32-, 57-, and 78-week old mice exposed to deionized water (controls), 3.5 g/L DCA continuously in drinking water ("Direct" group), or DCA for 10-, 32-, or 57-weeks followed by deionized water until sample collection ("Stop" groups). Genome-scaled alterations in DNA methylation were measured by Reduced Representation Bisulfite Sequencing (RRBS) in 78-week liver samples for control, Direct, 10-week Stop DCA exposed mice.

Results

Transcriptomic changes were most robust with concurrent or adjacent timepoints after exposure was withdrawn. We observed a similar pattern with DNA methylation alterations where we noted attenuated differentially methylated regions (DMRs) in the 10-week Stop DCA exposure groups compared to the Direct group at 78-weeks. Gene pathway analysis indicated cellular effects linked to increased oxidative metabolism, a primary mechanism of action for DCA, closer to exposure windows especially early in life. Conversely, many gene signatures and pathways reversed patterns later in life and reflected more pro-tumorigenic patterns for both current and prior DCA exposures. DNA methylation patterns correlated to early gene pathway perturbations, such as cellular signaling, regulation and metabolism, suggesting persistence in the epigenome and possible regulatory effects.

Conclusion

Liver metabolic reprogramming effects of DCA interacted with normal age mechanisms, increasing tumor burden with both continuous and prior DCA exposure in the male B6C3F1 rodent model.

Mitochondrial-nuclear epistasis underlying phenotypic variation in breast cancer pathology

The interplay between genes harboring single nucleotide polymorphisms (SNPs) is vital to better understand underlying contributions to the etiology of breast cancer. Much attention has been paid to epistasis between nuclear genes or mutations in the mitochondrial genome. However, there is limited understanding about the epistatic effects of genetic variants in the nuclear and mitochondrial genomes jointly on breast cancer. We tested the interaction of germline SNPs in the mitochondrial (mtSNPs) and nuclear (nuSNPs) genomes of female breast cancer patients in The Cancer Genome Atlas (TCGA) for association with morphological features extracted from hematoxylin and eosin (H&E)-stained pathology images. We identified 115 significant (q-value < 0.05) mito-nuclear interactions that increased nuclei size by as much as 12%. One interaction between nuSNP rs17320521 in an intron of the WSC Domain Containing 2 (WSCD2) gene and mtSNP rs869096886, a synonymous variant mapped to the mitochondrially-encoded NADH dehydrogenase 4 (MT-ND4) gene, was confirmed in an independent breast cancer data set from the Molecular Taxonomy of Breast Cancer International Consortium (METABRIC). None of the 10 mito-nuclear interactions identified from non-diseased female breast tissues from the Genotype-Expression (GTEx) project resulted in an increase in nuclei size. Comparisons of gene expression data from the TCGA breast cancer patients with the genotype homozygous for the minor alleles of the SNPs in WSCD2 and MT-ND4 versus the other genotypes revealed core transcriptional regulator interactions and an association with insulin. Finally, a Cox proportional hazards ratio = 1.7 (C.I. 0.98–2.9, p-value = 0.042) and Kaplan–Meier plot suggest that the TCGA female breast cancer patients with low gene expression of WSCD2 coupled with large nuclei have an increased risk of mortality. The intergenomic dependency between the two variants may constitute an inherent susceptibility of a more severe form of breast cancer and points to genetic targets for further investigation of additional determinants of the disease.

Structural Equation Modeling of In silico Perturbations

Gene expression is controlled by multiple regulators and their interactions. Data from genome-wide gene expression assays can be used to estimate molecular activities of regulators within a model organism and extrapolate them to biological processes in humans. This approach is valuable in studies to better understand complex human biological systems which may be involved in diseases and hence, have potential clinical relevance. In order to achieve this, it is necessary to infer gene interactions that are not directly observed (i.e. latent or hidden) by way of structural equation modeling (SEM) on the expression levels or activities of the downstream targets of regulator genes. Here we developed an R Shiny application, termed “Structural Equation Modeling of In silico Perturbations (SEMIPs)” to compute a two-sided t-statistic (T-score) from analysis of gene expression data, as a surrogate to gene activity in a given human specimen. SEMIPs can be used in either correlational studies between outcome variables of interest or subsequent model fitting on multiple variables. This application implements a 3-node SEM model that consists of two upstream regulators as input variables and one downstream reporter as an outcome variable to examine the significance of interactions among these variables. SEMIPs enables scientists to investigate gene interactions among three variables through computational and mathematical modeling (i.e. in silico). In a case study using SEMIPs, we have shown that putative direct downstream genes of the GATA Binding Protein 2 (GATA2) transcription factor are sufficient to infer its activities in silico for the conserved progesterone receptor (PGR)-GATA2-SRY-box transcription factor 17 (SOX17) genetic network in the human uterine endometrium.

Evaluating the analytical validity of circulating tumor DNA sequencing assays for precision oncology

Circulating tumor DNA (ctDNA) sequencing is being rapidly adopted in precision oncology, but the accuracy, sensitivity and reproducibility of ctDNA assays is poorly understood. Here we report the findings of a multi-site, cross-platform evaluation of the analytical performance of five industry-leading ctDNA assays. We evaluated each stage of the ctDNA sequencing workflow with simulations, synthetic DNA spike-in experiments and proficiency testing on standardized, cell-line-derived reference samples. Above 0.5% variant allele frequency, ctDNA mutations were detected with high sensitivity, precision and reproducibility by all five assays, whereas, below this limit, detection became unreliable and varied widely between assays, especially when input material was limited. Missed mutations (false negatives) were more common than erroneous candidates (false positives), indicating that the reliable sampling of rare ctDNA fragments is the key challenge for ctDNA assays. This comprehensive evaluation of the analytical performance of ctDNA assays serves to inform best practice guidelines and provides a resource for precision oncology.

Cross-oncopanel study reveals high sensitivity and accuracy with overall analytical performance depending on genomic regions

BACKGROUND

Targeted sequencing using oncopanels requires comprehensive assessments of accuracy and detection sensitivity to ensure analytical validity. By employing reference materials characterized by the U.S. Food and Drug Administration-led SEquence Quality Control project phase2 (SEQC2) effort, we perform a cross-platform multi-lab evaluation of eight Pan-Cancer panels to assess best practices for oncopanel sequencing.

RESULTS

All panels demonstrate high sensitivity across targeted high-confidence coding regions and variant types for the variants previously verified to have variant allele frequency (VAF) in the 5-20% range. Sensitivity is reduced by utilizing VAF thresholds due to inherent variability in VAF measurements. Enforcing a VAF threshold for reporting has a positive impact on reducing false positive calls. Importantly, the false positive rate is found to be significantly higher outside the high-confidence coding regions, resulting in lower reproducibility. Thus, region restriction and VAF thresholds lead to low relative technical variability in estimating promising biomarkers and tumor mutational burden.

CONCLUSION

This comprehensive study provides actionable guidelines for oncopanel sequencing and clear evidence that supports a simplified approach to assess the analytical performance of oncopanels. It will facilitate the rapid implementation, validation, and quality control of oncopanels in clinical use.

A verified genomic reference sample for assessing performance of cancer panels detecting small variants of low allele frequency

BACKGROUND

Oncopanel genomic testing, which identifies important somatic variants, is increasingly common in medical practice and especially in clinical trials. Currently, there is a paucity of reliable genomic reference samples having a suitably large number of pre-identified variants for properly assessing oncopanel assay analytical quality and performance. The FDA-led Sequencing and Quality Control Phase 2 (SEQC2) consortium analyze ten diverse cancer cell lines individually and their pool, termed Sample A, to develop a reference sample with suitably large numbers of coding positions with known (variant) positives and negatives for properly evaluating oncopanel analytical performance.

RESULTS

In reference Sample A, we identify more than 40,000 variants down to 1% allele frequency with more than 25,000 variants having less than 20% allele frequency with 1653 variants in COSMIC-related genes. This is 5-100× more than existing commercially available samples. We also identify an unprecedented number of negative positions in coding regions, allowing statistical rigor in assessing limit-of-detection, sensitivity, and precision. Over 300 loci are randomly selected and independently verified via droplet digital PCR with 100% concordance. Agilent normal reference Sample B can be admixed with Sample A to create new samples with a similar number of known variants at much lower allele frequency than what exists in Sample A natively, including known variants having allele frequency of 0.02%, a range suitable for assessing liquid biopsy panels.

CONCLUSION

These new reference samples and their admixtures provide superior capability for performing oncopanel quality control, analytical accuracy, and validation for small to large oncopanels and liquid biopsy assays.

Ancestry-dependent gene expression correlates with reprogramming to pluripotency and multiple dynamic biological processes

Induced pluripotent stem cells (iPSCs) can be derived from differentiated cells, enabling the generation of personalized disease models by differentiating patient-derived iPSCs into disease-relevant cell lines. While genetic variability between different iPSC lines affects differentiation potential, how this variability in somatic cells affects pluripotent potential is less understood. We generated and compared transcriptomic data from 72 dermal fibroblast-iPSC pairs with consistent variation in reprogramming efficiency. By considering equal numbers of samples from self-reported African Americans and White Americans, we identified both ancestry-dependent and ancestry-independent transcripts associated with reprogramming efficiency, suggesting that transcriptomic heterogeneity can substantially affect reprogramming. Moreover, reprogramming efficiency-associated genes are involved in diverse dynamic biological processes, including cancer and wound healing, and are predictive of 5-year breast cancer survival in an independent cohort. Candidate genes may provide insight into mechanisms of ancestry-dependent regulation of cell fate transitions and motivate additional studies for improvement of reprogramming.

Elucidation of Mechanisms of Topotecan-Induced Cell Death in Human Breast MCF-7 Cancer Cells by Gene Expression Analysis

Topotecan is a clinically active anticancer agent for the management of various human tumors. While the principal mechanism of tumor cell killing by topotecan is due to its interactions with topoisomerase I and formation of DNA double-strand breaks, recent studies suggest that mechanisms involving generation of reactive free radicals and induction of oxidative stress may play a significant role in topotecan-dependent tumor cell death. We have shown that topotecan generates a topotecan radical following one-electron oxidation by a peroxidase-hydrogen peroxide system which reacts with reduced glutathione and cysteine, forming the glutathiyl and cysteinyl radicals, respectively. While little is known how these events are involved in topotecan-induced tumor cell death, we have now examined the effects of topotecan short (1 h) and long (24 h) exposure on global gene expression patterns using gene expression microarray analysis in human breast MCF-7 cancer cells, a wild-type p53 containing cell line. We show here that topotecan treatment significantly down-regulated estrogen receptor alpha (ERα/ESR1) and antiapoptotic BCL2 genes in addition to many other p53-regulated genes. Furthermore, 8-oxoguanine DNA glycosylase (OGG1), ferredoxin reductase (FDXR), methionine sulfoxide reductase (MSR), glutathione peroxidases (GPx), and glutathione reductase (GSR) genes were also differentially expressed by topotecan treatment. The differential expression of these genes was observed in a wild-type p53-containing breast ZR-75-1 tumor cell line following topotecan treatment. The involvement of reactive oxygen free radical sensor genes, the oxidative DNA damage (OGG1) repair gene and induction of pro-apoptotic genes suggest that reactive free radical species play a role in topotecan-induced tumor cell death.

Comparison of Normalization Methods for Analysis of TempO-Seq Targeted RNA Sequencing Data

Analysis of bulk RNA sequencing (RNA-Seq) data is a valuable tool to understand transcription at the genome scale. Targeted sequencing of RNA has emerged as a practical means of assessing the majority of the transcriptomic space with less reliance on large resources for consumables and bioinformatics. TempO-Seq is a templated, multiplexed RNA-Seq platform that interrogates a panel of sentinel genes representative of genome-wide transcription. Nuances of the technology require proper preprocessing of the data. Various methods have been proposed and compared for normalizing bulk RNA-Seq data, but there has been little to no investigation of how the methods perform on TempO-Seq data. We simulated count data into two groups (treated vs. untreated) at seven-fold change (FC) levels (including no change) using control samples from human HepaRG cells run on TempO-Seq and normalized the data using seven normalization methods. Upper Quartile (UQ) performed the best with regard to maintaining FC levels as detected by a limma contrast between treated vs. untreated groups. For all FC levels, specificity of the UQ normalization was greater than 0.84 and sensitivity greater than 0.90 except for the no change and +1.5 levels. Furthermore, K-means clustering of the simulated genes normalized by UQ agreed the most with the FC assignments [adjusted Rand index (ARI) = 0.67]. Despite having an assumption of the majority of genes being unchanged, the DESeq2 scaling factors normalization method performed reasonably well as did simple normalization procedures counts per million (CPM) and total counts (TCs). These results suggest that for two class comparisons of TempO-Seq data, UQ, CPM, TC, or DESeq2 normalization should provide reasonably reliable results at absolute FC levels ≥2.0. These findings will help guide researchers to normalize TempO-Seq gene expression data for more reliable results.

The Power of Resolution: Contextualized Understanding of Biological Responses to Liver Injury Chemicals Using High-throughput Transcriptomics and Benchmark Concentration Modeling

Prediction of human response to chemical exposures is a major challenge in both pharmaceutical and toxicological research. Transcriptomics has been a powerful tool to explore chemical-biological interactions, however, limited throughput, high-costs, and complexity of transcriptomic interpretations have yielded numerous studies lacking sufficient experimental context for predictive application. To address these challenges, we have utilized a novel high-throughput transcriptomics (HTT) platform, TempO-Seq, to apply the interpretive power of concentration-response modeling with exposures to 24 reference compounds in both differentiated and non-differentiated human HepaRG cell cultures. Our goals were to (1) explore transcriptomic characteristics distinguishing liver injury compounds, (2) assess impacts of differentiation state of HepaRG cells on baseline and compound-induced responses (eg, metabolically-activated), and (3) identify and resolve reference biological-response pathways through benchmark concentration (BMC) modeling. Study data revealed the predictive utility of this approach to identify human liver injury compounds by their respective BMCs in relation to human internal exposure plasma concentrations, and effectively distinguished drug analogs with varied associations of human liver injury (eg, withdrawn therapeutics trovafloxacin and troglitazone). Impacts of cellular differentiation state (proliferated vs differentiated) were revealed on baseline drug metabolizing enzyme expression, hepatic receptor signaling, and responsiveness to metabolically-activated toxicants (eg, cyclophosphamide, benzo(a)pyrene, and aflatoxin B1). Finally, concentration-response modeling enabled efficient identification and resolution of plausibly-relevant biological-response pathways through their respective pathway-level BMCs. Taken together, these findings revealed HTT paired with differentiated in vitro liver models as an effective tool to model, explore, and interpret toxicological and pharmacological interactions.

RATEmiRs: the rat atlas of tissue-specific and enriched miRNAs for discerning baseline expression exclusivity of candidate biomarkers

MicroRNAs (miRNAs) are small RNAs that regulate mRNA expression and have been targeted as biomarkers of organ damage and disease. To explore the utility of miRNAs to assess injury to specific tissues, a tissue atlas of miRNA abundance was constructed. The Rat Atlas of Tissue-specific and Enriched miRNAs (RATEmiRs) catalogues miRNA sequencing data from 21 and 23 tissues in male and female Sprague-Dawley rats, respectively. RATEmiRs identifies tissue-enriched (TE), tissue-specific (TS), or organ-specific (OS) miRNAs via comparisons of one or more tissue or organ vs others. We provide a brief overview of RATEmiRs and present how to use it to detect miRNA expression abundance of candidate biomarkers as well as to compare the expression of miRNAs between rat and human. The database is available at https://www.niehs.nih.gov/ratemirs/.

Proteasome inhibition creates a chromatin landscape favorable to RNA Pol II processivity

Proteasome activity is required for diverse cellular processes, including transcriptional and epigenetic regulation. However, inhibiting proteasome activity can lead to an increase in transcriptional output that is correlated with enriched levels of trimethyl H3K4 and phosphorylated forms of RNA polymerase (Pol) II at the promoter and gene body. Here, we perform gene expression analysis and ChIP followed by sequencing (ChIP-seq) in MCF-7 breast cancer cells treated with the proteasome inhibitor MG132, and we further explore genome-wide effects of proteasome inhibition on the chromatin state and RNA Pol II transcription. Analysis of gene expression programs and chromatin architecture reveals that chemically inhibiting proteasome activity creates a distinct chromatin state, defined by spreading of the H3K4me3 mark into the gene bodies of differentially-expressed genes. The distinct H3K4me3 chromatin profile and hyperacetylated nucleosomes at transcription start sites establish a chromatin landscape that facilitates recruitment of Ser-5- and Ser-2-phosphorylated RNA Pol II. Subsequent transcriptional events result in diverse gene expression changes. Alterations of H3K36me3 levels in the gene body reflect productive RNA Pol II elongation of transcripts of genes that are induced, underscoring the requirement for proteasome activity at multiple phases of the transcriptional cycle. Finally, by integrating genomics data and pathway analysis, we find that the differential effects of proteasome inhibition on the chromatin state modulate genes that are fundamental for cancer cell survival. Together, our results uncover underappreciated downstream effects of proteasome inhibitors that may underlie targeting of distinct chromatin states and key steps of RNA Pol II-mediated transcription in cancer cells.

Revealing a human p53 universe

p53 transcriptional networks are well-characterized in many organisms. However, a global understanding of requirements for in vivo p53 interactions with DNA and relationships with transcription across human biological systems in response to various p53 activating situations remains limited. Using a common analysis pipeline, we analyzed 41 data sets from genome-wide ChIP-seq studies of which 16 have associated gene expression data, including our recent primary data with normal human lymphocytes. The resulting extensive analysis, accessible at p53 BAER hub via the UCSC browser, provides a robust platform to characterize p53 binding throughout the human genome including direct influence on gene expression and underlying mechanisms. We establish the impact of spacers and mismatches from consensus on p53 binding in vivo and propose that once bound, neither significantly influences the likelihood of expression. Our rigorous approach revealed a large p53 genome-wide cistrome composed of >900 genes directly targeted by p53. Importantly, we identify a core cistrome signature composed of genes appearing in over half the data sets, and we identify signatures that are treatment- or cell-specific, demonstrating new functions for p53 in cell biology. Our analysis reveals a broad homeostatic role for human p53 that is relevant to both basic and translational studies.

RATEmiRs: the rat atlas of tissue-specific and enriched miRNAs database

Background

MicroRNAs (miRNAs) regulate gene expression and have been targeted as indicators of environmental/toxicologic stressors. Using the data from our deep sequencing of miRNAs in an extensive sampling of rat tissues, we developed a database called RATEmiRs for the Rat Atlas of Tissue-specific and Enriched miRNAs to allow users to dynamically determine mature-, iso- and pre-miR expression abundance, enrichment and specificity in rat tissues and organs.

Results

Illumina sequencing count data from mapped reads and meta data from the miRNA body atlas consisting of 21 and 23 tissues (14 organs) of toxicologic interest from 12 to 13 week old male and female Sprague Dawley rats respectively, were managed in a relational database with a user-friendly query interface. Data-driven pipelines are available to tailor the identification of tissue-enriched (TE) and tissue-specific (TS) miRNAs. Data-driven organ-specific (OS) pipelines reveal miRNAs that are expressed predominately in a given organ. A user-driven approach is also available to assess the tissue expression of user-specified miRNAs. Using one tissue vs other tissues and tissue(s) of an organ vs other organs, we illustrate the utility of RATEmiRs to facilitate the identification of candidate miRNAs. As a use case example, RATEmiRs revealed two TS miRNAs in the liver: rno-miR-122-3p and rno-miR-122-5p. When liver is compared to just the brain tissues for example, rno-miR-192-5p, rno-miR-193-3p, rno-miR-203b-3p, rno-miR-3559-5p, rno-miR-802-3p and rno-miR-802-5p are also detected as abundantly expressed in liver. As another example, 55 miRNAs from the RATEmiRs query of ileum vs brain tissues overlapped with miRNAs identified from the same comparison of tissues in an independent, publicly available dataset of 10 week old male rat microarray data suggesting that these miRNAs are likely not age-specific, platform-specific nor pipeline-dependent. Lastly, we identified 10 miRNAs that have conserved tissue/organ-specific expression between the rat and human species.

Conclusions

RATEmiRs provides a new platform for identification of TE, TS and OS miRNAs in a broad array of rat tissues. RATEmiRs is available at: https://www.niehs.nih.gov/ratemirs

A Comparison of the TempO-Seq S1500+ Platform to RNA-Seq and Microarray Using Rat Liver Mode of Action Samples

The TempO-Seq^TM platform allows for targeted transcriptomic analysis and is currently used by many groups to perform high-throughput gene expression analysis. Herein we performed a comparison of gene expression characteristics measured using 45 purified RNA samples from the livers of rats exposed to chemicals that fall into one of five modes of action (MOAs). These samples have been previously evaluated using Affymetrix^TM rat genome 230 2.0 microarrays and Illumina® whole transcriptome RNA-Seq. Comparison of these data with TempO-Seq analysis using the rat S1500+ beta gene set identified clear differences in the platforms related to signal to noise, root mean squared error, and/or sources of variability. Microarray and TempO-Seq captured the most variability in terms of MOA and chemical treatment whereas RNA-Seq had higher noise and larger differences between samples within a MOA. However, analysis of the data by hierarchical clustering, gene subnetwork connectivity and biological process representation of MOA-varying genes revealed that the samples clearly grouped by treatment as opposed to gene expression platform. Overall these findings demonstrate that the results from the TempO-Seq platform are consistent with findings on other more established approaches for measuring the genome-wide transcriptome.

The nitrone spin trap 5,5-dimethyl-1-pyrroline N-oxide dampens lipopolysaccharide-induced transcriptomic changes in macrophages

OBJECTIVE

M1-like inflammatory phenotype of macrophages plays a critical role in tissue damage in chronic inflammatory diseases. Previously, we found that the nitrone spin trap 5,5-dimethyl-1-pyrroline N-oxide (DMPO) dampens lipopolysaccharide (LPS)-triggered inflammatory priming of RAW 264.7 cells. Herein, we tested whether DMPO by itself can induce changes in macrophage transcriptome, and that these effects may prevent LPS-induced activation of macrophages.

MATERIALS AND METHODS

To test our hypothesis, we performed a transcriptomic and bioinformatics analysis in RAW 264.7 cells incubated with or without LPS, in the presence or in the absence of DMPO.

RESULTS

Functional data analysis showed 79 differentially expressed genes (DEGs) when comparing DMPO vs Control. We used DAVID databases for identifying enriched gene ontology terms and Ingenuity Pathway Analysis for functional analysis. Our data showed that DMPO vs Control comparison of DEGs is related to downregulation immune-system processes among others. Functional analysis indicated that interferon-response factor 7 and toll-like receptor were related (predicted inhibitions) to the observed transcriptomic effects of DMPO. Functional data analyses of the DMPO + LPS vs LPS DEGs were consistent with DMPO-dampening LPS-induced inflammatory transcriptomic profile in RAW 264.7. These changes were confirmed using Nanostring technology.

CONCLUSIONS

Taking together our data, surprisingly, indicate that DMPO by itself affects gene expression related to regulation of immune system and that DMPO dampens LPS-triggered MyD88- and TRIF-dependent signaling pathways. Our research provides critical data for further studies on the possible use of DMPO as a structural platform for the design of novel mechanism-based anti-inflammatory drugs.

Crosstalk between Receptor and Non-receptor Mediated Chemical Modes of Action in Rat Livers Converges through a Dysregulated Gene Expression Network at Tumor Suppressor Tp53

Chemicals, toxicants, and environmental stressors mediate their biologic effect through specific modes of action (MOAs). These encompass key molecular events that lead to changes in the expression of genes within regulatory pathways. Elucidating shared biologic processes and overlapping gene networks will help to better understand the toxicologic effects on biological systems. In this study we used a weighted network analysis of gene expression data from the livers of male Sprague-Dawley rats exposed to chemicals that elicit their effects through receptor-mediated MOAs (aryl hydrocarbon receptor, orphan nuclear hormone receptor, or peroxisome proliferator-activated receptor-α) or non-receptor-mediated MOAs (cytotoxicity or DNA damage). Four gene networks were highly preserved and statistically significant in each of the two MOA classes. Three of the four networks contain genes that enrich for immunity and defense. However, many canonical pathways related to an immune response were activated from exposure to the non-receptor-mediated MOA chemicals and deactivated from exposure to the receptor-mediated MOA chemicals. The top gene network contains a module with 33 genes including tumor suppressor TP53 as the central hub which was slightly up-regulated in gene expression compared to control. Although, there is crosstalk between the two MOA classes of chemicals at the TP53 gene network, more than half of the genes are dysregulated in opposite directions. For example, Thromboxane A Synthase 1 (Tbxas1), a cytochrome P450 protein coding gene regulated by Tp53, is down-regulated by exposure to the receptor-mediated chemicals but up-regulated by the non-receptor-mediated chemicals. The regulation of gene expression by the chemicals within MOA classes was consistent despite varying alanine transaminase and aspartate aminotransferase liver enzyme measurements. These results suggest that overlap in toxicologic pathways by chemicals with different MOAs provides common mechanisms for discordant regulation of gene expression within molecular networks.

Secondhand Smoke-Prevalent Polycyclic Aromatic Hydrocarbon Binary Mixture-Induced Specific Mitogenic and Pro-inflammatory Cell Signaling Events in Lung Epithelial Cells

Low molecular weight polycyclic aromatic hydrocarbons (LMW PAHs; < 206.3 g/mol) are prevalent and ubiquitous environmental contaminants, presenting a human health concern, and have not been as thoroughly studied as the high MW PAHs. LMW PAHs exert their pulmonary effects, in part, through P38-dependent and -independent mechanisms involving cell-cell communication and the production of pro-inflammatory mediators known to contribute to lung disease. Specifically, we determined the effects of two representative LMW PAHs, 1-methylanthracene (1-MeA) and fluoranthene (Flthn), individually and as a binary PAH mixture on the dysregulation of gap junctional intercellular communication (GJIC) and connexin 43 (Cx43), activation of mitogen activated protein kinases (MAPK), and induction of inflammatory mediators in a mouse non-tumorigenic alveolar type II cell line (C10). Both 1-MeA, Flthn, and the binary PAH mixture of 1-MeA and Flthn dysregulated GJIC in a dose and time-dependent manner, reduced Cx43 protein, and activated the following MAPKs: P38, ERK1/2, and JNK. Inhibition of P38 MAPK prevented PAH-induced dysregulation of GJIC, whereas inhibiting ERK and JNK did not prevent these PAHs from dysregulating GJIC indicating a P38-dependent mechanism. A toxicogenomic approach revealed significant P38-dependent and -independent pathways involved in inflammation, steroid synthesis, metabolism, and oxidative responses. Genes in these pathways were significantly altered by the binary PAH mixture when compared with 1-MeA and Flthn alone suggesting interactive effects. Exposure to the binary PAH mixture induced the production and release of cytokines and metalloproteinases from the C10 cells. Our findings with a binary mixture of PAHs suggest that combinations of LMW PAHs may elicit synergistic or additive inflammatory responses which warrant further investigation and confirmation.

goSTAG: gene ontology subtrees to tag and annotate genes within a set

Background

Over-representation analysis (ORA) detects enrichment of genes within biological categories. Gene Ontology (GO) domains are commonly used for gene/gene-product annotation. When ORA is employed, often times there are hundreds of statistically significant GO terms per gene set. Comparing enriched categories between a large number of analyses and identifying the term within the GO hierarchy with the most connections is challenging. Furthermore, ascertaining biological themes representative of the samples can be highly subjective from the interpretation of the enriched categories.

Results

We developed goSTAG for utilizing GO Subtrees to Tag and Annotate Genes that are part of a set. Given gene lists from microarray, RNA sequencing (RNA-Seq) or other genomic high-throughput technologies, goSTAG performs GO enrichment analysis and clusters the GO terms based on the p-values from the significance tests. GO subtrees are constructed for each cluster, and the term that has the most paths to the root within the subtree is used to tag and annotate the cluster as the biological theme. We tested goSTAG on a microarray gene expression data set of samples acquired from the bone marrow of rats exposed to cancer therapeutic drugs to determine whether the combination or the order of administration influenced bone marrow toxicity at the level of gene expression. Several clusters were labeled with GO biological processes (BPs) from the subtrees that are indicative of some of the prominent pathways modulated in bone marrow from animals treated with an oxaliplatin/topotecan combination. In particular, negative regulation of MAP kinase activity was the biological theme exclusively in the cluster associated with enrichment at 6 h after treatment with oxaliplatin followed by control. However, nucleoside triphosphate catabolic process was the GO BP labeled exclusively at 6 h after treatment with topotecan followed by control.

Conclusions

goSTAG converts gene lists from genomic analyses into biological themes by enriching biological categories and constructing GO subtrees from over-represented terms in the clusters. The terms with the most paths to the root in the subtree are used to represent the biological themes. goSTAG is developed in R as a Bioconductor package and is available at https://bioconductor.org/packages/goSTAG

Electronic supplementary material

The online version of this article (doi:10.1186/s13029-017-0066-1) contains supplementary material, which is available to authorized users.

Informatics and Data Analytics to Support Exposome-Based Discovery for Public Health

The complexity of the human exposome-the totality of environmental exposures encountered from birth to death-motivates systematic, high-throughput approaches to discover new environmental determinants of disease. In this review, we describe the state of science in analyzing the human exposome and provide recommendations for the public health community to consider in dealing with analytic challenges of exposome-based biomedical research. We describe extant and novel analytic methods needed to associate the exposome with critical health outcomes and contextualize the data-centered challenges by drawing parallels to other research endeavors such as human genomics research. We discuss efforts for training scientists who can bridge public health, genomics, and biomedicine in informatics and statistics. If an exposome data ecosystem is brought to fruition, it will likely play a role as central as genomic science has had in molding the current and new generations of biomedical researchers, computational scientists, and public health research programs.

The Rat microRNA body atlas; Evaluation of the microRNA content of rat organs through deep sequencing and characterization of pancreas enriched miRNAs as biomarkers of pancreatic toxicity in the rat and dog

Background

MicroRNAs (miRNA) are ~19–25 nucleotide long RNA molecules that fine tune gene expression through the inhibition of translation or degradation of the mRNA through incorporation into the RNA induced silencing complex (RISC). MicroRNAs are stable in the serum and plasma, are detectable in a wide variety of body fluids, are conserved across veterinary species and humans and are expressed in a tissue specific manner. They can be detected at low concentrations in circulation in animals and humans, generating interest in the utilization of miRNAs as serum and/or plasma based biomarkers of tissue injury. MicroRNA tissue profiling in rodents has been published, but sample an insufficient number of organs of toxicologic interest using microarray or qPCR technologies for miRNA detection. Here we impart an improved rat microRNA body atlas consisting of 21 and 23 tissues of toxicologic interest from male and female Sprague Dawley rats respectively, using Illumina miRNA sequencing. Several of the authors created a dog miRNA body atlas and we collaborated to test miRNAs conserved in rat and dog pancreas in caerulein toxicity studies utilizing both species.

Results

A rich data set is presented that more robustly defines the tissue specificity and enrichment profiles of previously published and undiscovered rat miRNAs. We generated 1,927 sequences that mapped to mature miRNAs in rat, mouse and human from miRBase and discovered an additional 1,162 rat miRNAs as compared to the current number of rat miRNAs in miRBase version 21. Tissue specific and enriched miRNAs were identified and a subset of these miRNAs were validated by qPCR for tissue specificity or enrichment. As an example of the power of this approach, we have conducted rat and dog pancreas toxicity studies and examined the levels of some tissue specific and enriched miRNAs conserved between rat and dog in the serum of each species. The studies demonstrate that conserved tissue specific/enriched miRs-216a-5p, 375-3p, 148a-3p, 216b-5p and 141-3p are candidate biomarkers of pancreatic injury in the rat and dog.

Conclusions

A microRNA body atlas for rat and dog was useful in identifying new candidate miRNA biomarkers of organ toxicity in 2 toxicologically relevant species.

Electronic supplementary material

The online version of this article (doi:10.1186/s12864-016-2956-z) contains supplementary material, which is available to authorized users.

Tunable regulation of CREB DNA binding activity couples genotoxic stress response and metabolism

cAMP response element binding protein (CREB) is a key regulator of glucose metabolism and synaptic plasticity that is canonically regulated through recruitment of transcriptional coactivators. Here we show that phosphorylation of CREB on a conserved cluster of Ser residues (the ATM/CK cluster) by the DNA damage-activated protein kinase ataxia-telangiectasia-mutated (ATM) and casein kinase1 (CK1) and casein kinase2 (CK2) positively and negatively regulates CREB-mediated transcription in a signal dependent manner. In response to genotoxic stress, phosphorylation of the ATM/CK cluster inhibited CREB-mediated gene expression, DNA binding activity and chromatin occupancy proportional to the number of modified Ser residues. Paradoxically, substoichiometric, ATM-independent, phosphorylation of the ATM/CK cluster potentiated bursts in CREB-mediated transcription by promoting recruitment of the CREB coactivator, cAMP-regulated transcriptional coactivators (CRTC2). Livers from mice expressing a non-phosphorylatable CREB allele failed to attenuate gluconeogenic genes in response to DNA damage or fully activate the same genes in response to glucagon. We propose that phosphorylation-dependent regulation of DNA binding activity evolved as a tunable mechanism to control CREB transcriptional output and promote metabolic homeostasis in response to rapidly changing environmental conditions.

Smith-Lemli-Opitz syndrome reveals requirement for sterol biosynthesis in the innate immune response

Smith-Lemli-Opitz Syndrome (SLOS) is a rare neurodevelopmental disorder caused by mutation of 7-dehydrocholesterol (7DHC) reductase (DHCR7), a terminal enzyme in cholesterol biosynthesis. SLOS patients have increased incidence of infections, but no mechanism of immunosuppression has been identified. As cholesterol is a key component of membrane raft microdomains and rafts have been implicated in Toll-like receptor (TLR) signaling, we examined whether TLR responsiveness is altered in SLOS. Primary fibroblasts from SLOS patients with varying degrees of deficit in DHCR7 activity were used to assess the TLR4 response. Upon stimulation, significantly lower NF-kB activation and IL-6 and IL-8 production were observed in SLOS cells than in controls. Across patients, a direct relation between residual DHCR7 enzyme activity and IL-6 production was observed, together with an inverse relation between clinical severity scores and IL-8 production. Microarray analysis revealed that several pro-inflammatory chemokines are less responsive to LPS in SLOS. Deficient LPS-induced cytokines were also observed in macrophages from Dhcr7 mutant mice and in RAW264.7 macrophages treated with DHCR7 inhibitors. Supporting cholesterol deficit as the cause of the TLR4 impairment, the LPS response of SLOS macrophages was partially rescued by cholesterol repletion. Mechanistically, confocal microscopy revealed that TLR4 localization to rafts is decreased in SLOS macrophages. Taken together, these findings indicate that the SLOS mutation confers an abnormal TLR4 response. Deficient innate immunity may be a clinically relevant contributor to SLOS pathogenesis, and SLOS highlights the importance of cholesterol biosynthesis to innate immunity.

Effects of mannose-binding lectin on pulmonary gene expression and innate immune inflammatory response to ozone

Ozone is a common, potent oxidant pollutant in industrialized nations. Ozone exposure causes airway hyperreactivity, lung hyperpermeability, inflammation, and cell damage in humans and laboratory animals, and exposure to ozone has been associated with exacerbation of asthma, altered lung function, and mortality. The mechanisms of ozone-induced lung injury and differential susceptibility are not fully understood. Ozone-induced lung inflammation is mediated, in part, by the innate immune system. We hypothesized that mannose-binding lectin (MBL), an innate immunity serum protein, contributes to the proinflammatory events caused by ozone-mediated activation of the innate immune system. Wild-type (Mbl(+/+)) and MBL-deficient (Mbl(-/-)) mice were exposed to ozone (0.3 ppm) for up to 72 h, and bronchoalveolar lavage fluid was examined for inflammatory markers. Mean numbers of eosinophils and neutrophils and levels of the neutrophil attractants C-X-C motif chemokines 2 [Cxcl2 (major intrinsic protein 2)] and 5 [Cxcl5 (limb expression, LIX)] in the bronchoalveolar lavage fluid were significantly lower in Mbl(-/-) than Mbl(+/+) mice exposed to ozone. Using genome-wide mRNA microarray analyses, we identified significant differences in transcript response profiles and networks at baseline [e.g., nuclear factor erythroid-related factor 2 (NRF2)-mediated oxidative stress response] and after exposure (e.g., humoral immune response) between Mbl(+/+) and Mbl(-/-) mice. The microarray data were further analyzed to discover several informative differential response patterns and subsequent gene sets, including the antimicrobial response and the inflammatory response. We also used the lists of gene transcripts to search the LINCS L1000CDS(2) data sets to identify agents that are predicted to perturb ozone-induced changes in gene transcripts and inflammation. These novel findings demonstrate that targeted deletion of Mbl caused differential levels of inflammation-related gene sets at baseline and after exposure to ozone and significantly reduced pulmonary inflammation, thus indicating an important innate immunomodulatory role of the gene in this model.

EPIG-Seq: extracting patterns and identifying co-expressed genes from RNA-Seq data

Background

RNA sequencing (RNA-Seq) measures genome-wide gene expression. RNA-Seq data is count-based rendering normal distribution models for analysis inappropriate. Normalization of RNA-Seq data to transform the data has limitations which can adversely impact the analysis. Furthermore, there are a few count-based methods for analysis of RNA-Seq data but they are essentially for pairwise analysis of treatment groups or multiclasses but not pattern-based to identify co-expressed genes.

Results

We adapted our extracting patterns and identifying genes methodology for RNA-Seq (EPIG-Seq) count data. The software uses count-based correlation to measure similarity between genes, quasi-Poisson modelling to estimate dispersion in the data and a location parameter to indicate magnitude of differential expression. EPIG-Seq is different than any other software currently available for pattern analysis of RNA-Seq data in that EPIG-Seq 1) uses count level data and supports cases of inflated zeros, 2) identifies statistically significant clusters of genes that are co-expressed across experimental conditions, 3) takes into account dispersion in the replicate data and 4) provides reliable results even with small sample sizes. EPIG-Seq operates in two steps: 1) extract the pattern profiles from data as seeds for clustering co-expressed genes and 2) cluster the genes to the pattern seeds and compute statistical significance of the pattern of co-expressed genes. EPIG-Seq provides a table of the genes with bootstrapped p-values and profile plots of the patterns of co-expressed genes. In addition, EPIG-Seq provides a heat map and principal component dimension reduction plot of the clustered genes as visual aids. We demonstrate the utility of EPIG-Seq through the analysis of toxicogenomics and cancer data sets to identify biologically relevant co-expressed genes. EPIG-Seq is available at: sourceforge.net/projects/epig-seq.

Conclusions

EPIG-Seq is unlike any other software currently available for pattern analysis of RNA-Seq count level data across experimental groups. Using the EPIG-Seq software to analyze RNA-Seq count data across biological conditions permits the ability to extract biologically meaningful co-expressed genes associated with coordinated regulation.

Electronic supplementary material

The online version of this article (doi:10.1186/s12864-016-2584-7) contains supplementary material, which is available to authorized users.

Blood gene expression profiling of an early acetaminophen response

Acetaminophen can adversely affect the liver especially when overdosed. We used whole blood as a surrogate to identify genes as potential early indicators of an acetaminophen-induced response. In a clinical study, healthy human subjects were dosed daily with 4 g of either acetaminophen or placebo pills for 7 days and evaluated over the course of 14 days. Alanine aminotransferase (ALT) levels for responders to acetaminophen increased between days 4 and 9 after dosing, and 12 genes were detected with expression profiles significantly altered within 24 h. The early responsive genes separated the subjects by class and dose period. In addition, the genes clustered patients who overdosed on acetaminophen apart from controls and also predicted the exposure classifications with 100% accuracy. The responsive genes serve as early indicators of an acetaminophen exposure, and their gene expression profiles can potentially be evaluated as molecular indicators for further consideration.

Blood transcript immune signatures distinguish a subset of people with elevated serum ALT from others given acetaminophen

The diagnosis of drug-induced liver injury is hindered by the limited utility of clinical chemistries. We have shown that hepatotoxicants can produce peripheral blood transcriptome "signatures" (PBTS) in rodents and humans. In this study, 42 adults were treated with acetaminophen (APAP; 1 g every 6 hours) for seven days, followed by three days of placebo. Eleven subjects received only placebo. After five days, 12 subjects (30%) had increases in serum alanine aminotransferase (ALT) levels ("responders"). PBTS of 707 and 760 genes, respectively, could distinguish responders and nonresponders from placebos. Functional analysis of the responder PBTS revealed increased expression of genes involved in TH2-mediated and innate immune responses, whereas the nonresponders demonstrated increased gene expression consistent with a tolerogenic immune response. Taken together, these observations suggest that the clinical subjects with transient increases in serum ALT failed to maintain or intensify a hepatic tolerogenic immune response.

Novel Roles for Notch3 and Notch4 Receptors in Gene Expression and Susceptibility to Ozone-Induced Lung Inflammation in Mice

BACKGROUND

Ozone is a highly toxic air pollutant and global health concern. Mechanisms of genetic susceptibility to ozone-induced lung inflammation are not completely understood. We hypothesized that Notch3 and Notch4 are important determinants of susceptibility to ozone-induced lung inflammation.

METHODS

Wild-type (WT), Notch3 (Notch3-/-), and Notch4 (Notch4-/-) knockout mice were exposed to ozone (0.3 ppm) or filtered air for 6-72 hr.

RESULTS

Relative to air-exposed controls, ozone increased bronchoalveolar lavage fluid (BALF) protein, a marker of lung permeability, in all genotypes, but significantly greater concentrations were found in Notch4-/- compared with WT and Notch3-/- mice. Significantly greater mean numbers of BALF neutrophils were found in Notch3-/- and Notch4-/- mice compared with WT mice after ozone exposure. Expression of whole lung Tnf was significantly increased after ozone in Notch3-/- and Notch4-/- mice, and was significantly greater in Notch3-/- compared with WT mice. Statistical analyses of the transcriptome identified differentially expressed gene networks between WT and knockout mice basally and after ozone, and included Trim30, a member of the inflammasome pathway, and Traf6, an inflammatory signaling member.

CONCLUSIONS

These novel findings are consistent with Notch3 and Notch4 as susceptibility genes for ozone-induced lung injury, and suggest that Notch receptors protect against innate immune inflammation.

Technical guide for applications of gene expression profiling in human health risk assessment of environmental chemicals

Toxicogenomics promises to be an important part of future human health risk assessment of environmental chemicals. The application of gene expression profiles (e.g., for hazard identification, chemical prioritization, chemical grouping, mode of action discovery, and quantitative analysis of response) is growing in the literature, but their use in formal risk assessment by regulatory agencies is relatively infrequent. Although additional validations for specific applications are required, gene expression data can be of immediate use for increasing confidence in chemical evaluations. We believe that a primary reason for the current lack of integration is the limited practical guidance available for risk assessment specialists with limited experience in genomics. The present manuscript provides basic information on gene expression profiling, along with guidance on evaluating the quality of genomic experiments and data, and interpretation of results presented in the form of heat maps, pathway analyses and other common approaches. Moreover, potential ways to integrate information from gene expression experiments into current risk assessment are presented using published studies as examples. The primary objective of this work is to facilitate integration of gene expression data into human health risk assessments of environmental chemicals.

Development of phenotypic and transcriptional biomarkers to evaluate relative activity of potentially estrogenic chemicals in ovariectomized mice

BACKGROUND

Concerns regarding potential endocrine-disrupting chemicals (EDCs) have led to a need for methods to evaluate candidate estrogenic chemicals. Our previous evaluations of two such EDCs revealed a response similar to that of estradiol (E2) at 2 hr, but a less robust response at 24 hr, similar to the short-acting estrogen estriol (E3).

OBJECTIVES

Microarray analysis using tools to recognize patterns of response have been utilized in the cancer field to develop biomarker panels of transcripts for diagnosis and selection of treatments most likely to be effective. Biological effects elicited by long- versus short-acting estrogens greatly affect the risks associated with exposures; therefore, we sought to develop tools to predict the ability of chemicals to maintain estrogenic responses.

METHODS

We used biological end points in uterine tissue and a signature pattern-recognizing tool that identified coexpressed transcripts to develop and test a panel of transcripts in order to classify potentially estrogenic compounds using an in vivo system. The end points used are relevant to uterine tissue, but the resulting classification of the compounds is important for other sensitive tissues and species.

RESULTS

We evaluated biological and transcriptional end points with proven short- and long-acting estrogens and verified the use of our approach using a phytoestrogen. With our model, we were able to classify the diarylheptanoid D3 as a short-acting estrogen.

CONCLUSIONS

We have developed a panel of transcripts as biomarkers which, together with biological end points, might be used to screen and evaluate potentially estrogenic chemicals and infer mode of activity.

Toxicogenomics profiling of bone marrow from rats treated with topotecan in combination with oxaliplatin: a mechanistic strategy to inform combination toxicity

Combinations of anticancer agents may have synergistic anti-tumor effects, but enhanced hematological toxicity often limit their clinical use. We examined whether "microarray profiles" could be used to compare early molecular responses following a single dose of agents administered individually with that of the agents administered in a combination. We compared the mRNA responses within bone marrow of Sprague-Dawley rats after a single 30 min treatment with topotecan at 4.7 mg/kg or oxaliplatin at 15 mg/kg alone to that of sequentially administered combination therapy or vehicle control for 1, 6, and 24 h. We also examined the histopathology of the bone marrow following all treatments. Drug-related histopathological lesions were limited to bone marrow hypocellularity for animals dosed with either agent alone or in combination. Lesions had an earlier onset and higher incidence for animals given topotecan alone or in combination with oxaliplatin. Severity increased from mild to moderate when topotecan was administered prior to oxaliplatin compared with administering oxaliplatin first. Notably, six patterns of co-expressed genes were detected at the 1 h time point that indicate regulatory expression of genes that are dependent on the order of the administration. These results suggest alterations in histone biology, chromatin remodeling, DNA repair, bone regeneration, and respiratory and oxidative phosphorylation are among the prominent pathways modulated in bone marrow from animals treated with an oxaliplatin/topotecan combination. These data also demonstrate the potential for early mRNA patterns derived from target organs of toxicity to inform toxicological risk and molecular mechanisms for agents given in combination.

p53 and NF-κB coregulate proinflammatory gene responses in human macrophages

Macrophages are sentinel immune cells that survey the tissue microenvironment, releasing cytokines in response to both exogenous insults and endogenous events such as tumorigenesis. Macrophages mediate tumor surveillance and therapy-induced tumor regression; however, tumor-associated macrophages (TAM) and their products may also promote tumor progression. Whereas NF-κB is prominent in macrophage-initiated inflammatory responses, little is known about the role of p53 in macrophage responses to environmental challenge, including chemotherapy or in TAMs. Here, we report that NF-κB and p53, which generally have opposing effects in cancer cells, coregulate induction of proinflammatory genes in primary human monocytes and macrophages. Using Nutlin-3 as a tool, we demonstrate that p53 and NF-κB rapidly and highly induce interleukin (IL)-6 by binding to its promoter. Transcriptome analysis revealed global p53/NF-κB co-regulation of immune response genes, including several chemokines, which effectively induced human neutrophil migration. In addition, we show that p53, activated by tumor cell paracrine factors, induces high basal levels of macrophage IL-6 in a TAM model system [tumor-conditioned macrophages (TCM)]. Compared with normal macrophages, TCMs exhibited higher p53 levels, enhanced p53 binding to the IL-6 promoter, and reduced IL-6 levels upon p53 inhibition. Taken together, we describe a mechanism by which human macrophages integrate signals through p53 and NF-κB to drive proinflammatory cytokine induction. Our results implicate a novel role for macrophage p53 in conditioning the tumor microenvironment and suggest a potential mechanism by which p53-activating chemotherapeutics, acting upon p53-sufficient macrophages and precursor monocytes, may indirectly impact tumors lacking functional p53.

Decision tree-based method for integrating gene expression, demographic, and clinical data to determine disease endotypes

Background

Complex diseases are often difficult to diagnose, treat and study due to the multi-factorial nature of the underlying etiology. Large data sets are now widely available that can be used to define novel, mechanistically distinct disease subtypes (endotypes) in a completely data-driven manner. However, significant challenges exist with regard to how to segregate individuals into suitable subtypes of the disease and understand the distinct biological mechanisms of each when the goal is to maximize the discovery potential of these data sets.

Results

A multi-step decision tree-based method is described for defining endotypes based on gene expression, clinical covariates, and disease indicators using childhood asthma as a case study. We attempted to use alternative approaches such as the Student’s t-test, single data domain clustering and the Modk-prototypes algorithm, which incorporates multiple data domains into a single analysis and none performed as well as the novel multi-step decision tree method. This new method gave the best segregation of asthmatics and non-asthmatics, and it provides easy access to all genes and clinical covariates that distinguish the groups.

Conclusions

The multi-step decision tree method described here will lead to better understanding of complex disease in general by allowing purely data-driven disease endotypes to facilitate the discovery of new mechanisms underlying these diseases. This application should be considered a complement to ongoing efforts to better define and diagnose known endotypes. When coupled with existing methods developed to determine the genetics of gene expression, these methods provide a mechanism for linking genetics and exposomics data and thereby accounting for both major determinants of disease.

Gene expression signatures but not cell cycle checkpoint functions distinguish AT carriers from normal individuals

Ataxia telangiectasia (AT) is a rare autosomal recessive disease caused by mutations in the ataxia telangiectasia-mutated gene (ATM). AT carriers with one mutant ATM allele are usually not severely affected although they carry an increased risk of developing cancer. There has not been an easy and reliable diagnostic method to identify AT carriers. Cell cycle checkpoint functions upon ionizing radiation (IR)-induced DNA damage and gene expression signatures were analyzed in the current study to test for differential responses in human lymphoblastoid cell lines with different ATM genotypes. While both dose- and time-dependent G1 and G2 checkpoint functions were highly attenuated in ATM-/- cell lines, these functions were preserved in ATM+/- cell lines equivalent to ATM+/+ cell lines. However, gene expression signatures at both baseline (consisting of 203 probes) and post-IR treatment (consisting of 126 probes) were able to distinguish ATM+/- cell lines from ATM+/+ and ATM-/- cell lines. Gene ontology (GO) and pathway analysis of the genes in the baseline signature indicate that ATM function-related categories, DNA metabolism, cell cycle, cell death control, and the p53 signaling pathway, were overrepresented. The same analyses of the genes in the IR-responsive signature revealed that biological categories including response to DNA damage stimulus, p53 signaling, and cell cycle pathways were overrepresented, which again confirmed involvement of ATM functions. The results indicate that AT carriers who have unaffected G1 and G2 checkpoint functions can be distinguished from normal individuals and AT patients by expression signatures of genes related to ATM functions.

Improved Sparse Multi-Class SVM and Its Application for Gene Selection in Cancer Classification

BACKGROUND

Microarray techniques provide promising tools for cancer diagnosis using gene expression profiles. However, molecular diagnosis based on high-throughput platforms presents great challenges due to the overwhelming number of variables versus the small sample size and the complex nature of multi-type tumors. Support vector machines (SVMs) have shown superior performance in cancer classification due to their ability to handle high dimensional low sample size data. The multi-class SVM algorithm of Crammer and Singer provides a natural framework for multi-class learning. Despite its effective performance, the procedure utilizes all variables without selection. In this paper, we propose to improve the procedure by imposing shrinkage penalties in learning to enforce solution sparsity.

RESULTS

The original multi-class SVM of Crammer and Singer is effective for multi-class classification but does not conduct variable selection. We improved the method by introducing soft-thresholding type penalties to incorporate variable selection into multi-class classification for high dimensional data. The new methods were applied to simulated data and two cancer gene expression data sets. The results demonstrate that the new methods can select a small number of genes for building accurate multi-class classification rules. Furthermore, the important genes selected by the methods overlap significantly, suggesting general agreement among different variable selection schemes.

CONCLUSIONS

High accuracy and sparsity make the new methods attractive for cancer diagnostics with gene expression data and defining targets of therapeutic intervention.

AVAILABILITY

The source MATLAB code are available from http://math.arizona.edu/~hzhang/software.html.

Dynamic expression of 3' UTRs revealed by Poisson hidden Markov modeling of RNA-Seq: implications in gene expression profiling

RNA sequencing (RNA-Seq) allows for the identification of novel exon-exon junctions and quantification of gene expression levels. We show that from RNA-Seq data one may also detect utilization of alternative polyadenylation (APA) in 3' untranslated regions (3' UTRs) known to play a critical role in the regulation of mRNA stability, cellular localization and translation efficiency. Given the dynamic nature of APA, it is desirable to examine the APA on a sample by sample basis. We used a Poisson hidden Markov model (PHMM) of RNA-Seq data to identify potential APA in human liver and brain cortex tissues leading to shortened 3' UTRs. Over three hundred transcripts with shortened 3' UTRs were detected with sensitivity >75% and specificity >60%. Tissue-specific 3' UTR shortening was observed for 32 genes with a q-value ≤ 0.1. When compared to alternative isoforms detected by Cufflinks or MISO, our PHMM method agreed on over 100 transcripts with shortened 3' UTRs. Given the increasing usage of RNA-Seq for gene expression profiling, using PHMM to investigate sample-specific 3' UTR shortening could be an added benefit from this emerging technology.

Gene expression in uterine leiomyoma from tumors likely to be growing (from black women over 35) and tumors likely to be non-growing (from white women over 35)

The study of uterine leiomyomata (fibroids) provides a unique opportunity to investigate the physiological and molecular determinants of hormone dependent tumor growth and spontaneous tumor regression. We conducted a longitudinal clinical study of premenopausal women with leiomyoma that showed significantly different growth rates between white and black women depending on their age. Growth rates for leiomyoma were on average much higher from older black women than for older white women, and we now report gene expression pattern differences in tumors from these two groups of study participants. Total RNA from 52 leiomyoma and 8 myometrial samples were analyzed using Affymetrix Gene Chip expression arrays. Gene expression data was first compared between all leiomyoma and normal myometrium and then between leiomyoma from older black women (age 35 or older) and from older white women. Genes that were found significant in pairwise comparisons were further analyzed for canonical pathways, networks and biological functions using the Ingenuity Pathway Analysis (IPA) software. Whereas our comparison of leiomyoma to myometrium produced a very large list of genes highly similar to numerous previous studies, distinct sets of genes and signaling pathways were identified in comparisons of older black and white women whose tumors were likely to be growing and non-growing, respectively. Key among these were genes associated with regulation of apoptosis. To our knowledge, this is the first study to compare two groups of tumors that are likely to have different growth rates in order to reveal molecular signals likely to be influential in tumor growth.

Widespread Exonization of Transposable Elements in Human Coding Sequences is Associated with Epigenetic Regulation of Transcription

Background

Transposable Elements (TEs) have long been regarded as selfish or junk DNA having little or no role in the regulation or functioning of the human genome. However, over the past several years this view came to be challenged as several studies provided anecdotal as well as global evidence for the contribution of TEs to the regulatory and coding needs of human genes. In this study, we explored the incorporation and epigenetic regulation of coding sequences donated by TEs using gene expression and other ancillary genomics data from two human hematopoietic cell-lines: GM12878 (a lymphoblastoid cell line) and K562 (a Chronic Myelogenous Leukemia cell line). In each cell line, we found several thousand instances of TEs donating coding sequences to human genes. We compared the transcriptome assembly of the RNA sequencing (RNA-Seq) reads with and without the aid of a reference transcriptome and found that the percentage of genes that incorporate TEs in their coding sequences is significantly greater than that obtained from the reference transcriptome assemblies using Refseq and Gencode gene models. We also used histone modifications chromatin immunoprecipitation sequencing (ChIP-Seq) data, Cap Analysis of Gene Expression (CAGE) data and DNAseI Hypersensitivity Site (DHS) data to demonstrate the epigenetic regulation of the TE derived coding sequences. Our results suggest that TEs form a significantly higher percentage of coding sequences than represented in gene annotation databases and these TE derived sequences are epigenetically regulated in accordance with their expression in the two cell types.

Transcriptional responses to loss of RNase H2 in Saccharomyces cerevisiae

We report here the transcriptional responses in Saccharomyces cerevisiae to deletion of the RNH201 gene encoding the catalytic subunit of RNase H2. Deleting RNH201 alters RNA expression of 349 genes by ≥1.5-fold (q-value <0.01), of which 123 are upregulated and 226 are downregulated. Differentially expressed genes (DEGs) include those involved in stress responses and genome maintenance, consistent with a role for RNase H2 in removing ribonucleotides incorporated into DNA during replication. Upregulated genes include several that encode subunits of RNA polymerases I and III, and genes involved in ribosomal RNA processing, ribosomal biogenesis and tRNA modification and processing, supporting a role for RNase H2 in resolving R-loops formed during transcription of rRNA and tRNA genes. A role in R-loop resolution is further suggested by a higher average GC-content proximal to the transcription start site of downregulated as compared to upregulated genes. Several DEGs are involved in telomere maintenance, supporting a role for RNase H2 in resolving RNA-DNA hybrids formed at telomeres. A large number of DEGs encode nucleases, helicases and genes involved in response to dsRNA viruses, observations that could be relevant to the nucleic acid species that elicit an innate immune response in RNase H2-defective humans.

Identification of Identical Transcript Changes in Liver and Whole Blood during Acetaminophen Toxicity

The ability to identify mechanisms underlying drug-induced liver injury (DILI) in man has been hampered by the difficulty in obtaining liver tissue from patients. It has recently been proposed that whole blood toxicogenomics may provide a non-invasive means for mechanistic studies of human DILI. However, it remains unclear to what extent changes in whole blood transcriptome mirror those in liver mechanistically linked to hepatotoxicity. To address this question, we applied the program Extracting Patterns and Identifying co-expressed Genes (EPIG) to publically available toxicogenomic data obtained from rats treated with both toxic and subtoxic doses of acetaminophen (APAP). In a training set of animals, we identified genes (760 at 6 h and 185 at 24 h post dose) with similar patterns of expression in blood and liver during APAP-induced hepatotoxicity. The pathways represented in the coordinately regulated genes largely involved mitochondrial and immune functions. The identified expression signatures were then evaluated in a separate set of animals for discernment of APAP exposure level or APAP-induced hepatotoxicity. At 6 h, the gene sets from liver and blood had equally sufficient classification of APAP exposure levels. At 24 h when toxicity was evident, the gene sets did not perform well in evaluating APAP exposure doses, but provided accurate classification of dose-independent liver injury that was evaluated by serum ALT elevation in the blood. Only 38 genes were common to both the 6 and 24-h gene sets, but these genes had the same capability as the parent gene sets to discern the exposure level and degree of liver injury. Some of the parallel transcript changes reflect pathways that are relevant to APAP hepatotoxicity, including mitochondria and immune functions. However, the extent to which these changes reflect similar mechanisms of action in both tissues remains to be determined.

Differential transcriptomic analysis of spontaneous lung tumors in B6C3F1 mice: comparison to human non-small cell lung cancer

Lung cancer is the leading cause of cancer-related death in people and is mainly due to environmental factors such as smoking and radon. The National Toxicology Program (NTP) tests various chemicals and mixtures for their carcinogenic hazard potential. In the NTP chronic bioassay using B6C3F1 mice, the incidence of lung tumors in treated and control animals is second only to the liver tumors. In order to study the molecular mechanisms of chemically induced lung tumors, an understanding of the genetic changes that occur in spontaneous lung (SL) tumors from untreated control animals is needed. The authors have evaluated the differential transcriptomic changes within SL tumors compared to normal lungs from untreated age-matched animals. Within SL tumors, several canonical pathways associated with cancer (eukaryotic initiation factor 2 signaling, RhoA signaling, PTEN signaling, and mammalian target of rapamycin signaling), metabolism (Inositol phosphate metabolism, mitochondrial dysfunction, and purine and pyramidine metabolism), and immune responses (FcγR-mediated phagocytosis, clathrin-mediated endocytosis, interleukin 8 signaling, and CXCR4 signaling) were altered. Meta-analysis of murine SL tumors and human non-small cell lung cancer transcriptomic data sets revealed a high concordance. These data provide important information on the differential transcriptomic changes in murine SL tumors that will be critical to our understanding of chemically induced lung tumors and will aid in hazard analysis in the NTP 2-year carcinogenicity bioassays.

Population differences in transcript-regulator expression quantitative trait loci

Gene expression quantitative trait loci (eQTL) are useful for identifying single nucleotide polymorphisms (SNPs) associated with diseases. At times, a genetic variant may be associated with a master regulator involved in the manifestation of a disease. The downstream target genes of the master regulator are typically co-expressed and share biological function. Therefore, it is practical to screen for eQTLs by identifying SNPs associated with the targets of a transcript-regulator (TR). We used a multivariate regression with the gene expression of known targets of TRs and SNPs to identify TReQTLs in European (CEU) and African (YRI) HapMap populations. A nominal p-value of <1×10(-6) revealed 234 SNPs in CEU and 154 in YRI as TReQTLs. These represent 36 independent (tag) SNPs in CEU and 39 in YRI affecting the downstream targets of 25 and 36 TRs respectively. At a false discovery rate (FDR) = 45%, one cis-acting tag SNP (within 1 kb of a gene) in each population was identified as a TReQTL. In CEU, the SNP (rs16858621) in Pcnxl2 was found to be associated with the genes regulated by CREM whereas in YRI, the SNP (rs16909324) was linked to the targets of miRNA hsa-miR-125a. To infer the pathways that regulate expression, we ranked TReQTLs by connectivity within the structure of biological process subtrees. One TReQTL SNP (rs3790904) in CEU maps to Lphn2 and is associated (nominal p-value = 8.1×10(-7)) with the targets of the X-linked breast cancer suppressor Foxp3. The structure of the biological process subtree and a gene interaction network of the TReQTL revealed that tumor necrosis factor, NF-kappaB and variants in G-protein coupled receptors signaling may play a central role as communicators in Foxp3 functional regulation. The potential pleiotropic effect of the Foxp3 TReQTLs was gleaned from integrating mRNA-Seq data and SNP-set enrichment into the analysis.

Research resource: whole-genome estrogen receptor α binding in mouse uterine tissue revealed by ChIP-seq

To advance understanding of mechanisms leading to biological and transcriptional endpoints related to estrogen action in the mouse uterus, we have mapped ERα and RNA polymerase II (PolII) binding sites using chromatin immunoprecipitation followed by sequencing of enriched chromatin fragments. In the absence of hormone, 5184 ERα-binding sites were apparent in the vehicle-treated ovariectomized uterine chromatin, whereas 17,240 were seen 1 h after estradiol (E₂) treatment, indicating that some sites are occupied by unliganded ERα, and that ERα binding is increased by E₂. Approximately 15% of the uterine ERα-binding sites were adjacent to (<10 kb) annotated transcription start sites, and many sites are found within genes or are found more than 100 kb distal from mapped genes; however, the density (sites per base pair) of ERα-binding sites is significantly greater adjacent to promoters. An increase in quantity of sites but no significant positional differences were seen between vehicle and E₂-treated samples in the overall locations of ERα-binding sites either distal from, adjacent to, or within genes. Analysis of the PolII data revealed the presence of poised promoter-proximal PolII on some highly up-regulated genes. Additionally, corecruitment of PolII and ERα to some distal enhancer regions was observed. A de novo motif analysis of sequences in the ERα-bound chromatin confirmed that estrogen response elements were significantly enriched. Interestingly, in areas of ERα binding without predicted estrogen response element motifs, homeodomain transcription factor-binding motifs were significantly enriched. The integration of the ERα- and PolII-binding sites from our uterine sequencing of enriched chromatin fragments data with transcriptional responses revealed in our uterine microarrays has the potential to greatly enhance our understanding of mechanisms governing estrogen response in uterine and other estrogen target tissues.

The impact of classification of interest on predictive toxicogenomics

The era of toxicogenomics has introduced a new way of monitoring the effect of environmental stressors and toxicants on biological systems via quantification of changes in gene expression. Because the liver is one of the major organs for synthesis and secretion of substances which metabolize endogenous and exogenous materials, there has been a great deal of interest in elucidating predictive and mechanistic genomic markers of hepatotoxicity. This mini-review will bring context to a limited number of toxicogenomics studies which used genomics to evaluate the transcriptional changes in blood and liver in response to acetaminophen (APAP) or other liver toxicants, but differed according to the classification of interest (COI), i.e., the partitioning of the samples a priori according to a common toxicological characteristic. The toxicogenomics studies highlighted are characterized by a classification of either no/low vs. high APAP dose exposure, none vs. observed necrosis, and severity of necrosis. The overlap or lack thereof between the gene classifiers and the modulated biological processes that are elucidated will be discussed to enhance the understanding of the effect of the particular COI model and experimental design used for prediction.

Sources of variance in baseline gene expression in the rodent liver

The use of gene expression profiling in both clinical and laboratory settings would be enhanced by better characterization of variation due to individual, environmental, and technical factors. Analysis of microarray data from untreated or vehicle-treated animals within the control arm of toxicogenomics studies has yielded useful information on baseline fluctuations in liver gene expression in the rodent. Here, studies which highlight contributions of different factors to gene expression variability in the rodent liver are discussed including a large meta-analysis of rat liver, which identified genes that vary in control animals in the absence of chemical treatment. Genes and their pathways that are the most and least variable were identified in a number of these studies. Life stage, fasting, sex, diet, circadian rhythm and liver lobe source can profoundly influence gene expression in the liver. Recognition of biological and technical factors that contribute to variability of background gene expression can help the investigator in the design of an experiment that maximizes sensitivity and reduces the influence of confounders that may lead to misinterpretation of genomic changes. The factors that contribute to variability in liver gene expression in rodents are likely analogous to those contributing to human interindividual variability in drug response and chemical toxicity. Identification of batteries of genes that are altered in a variety of background conditions could be used to predict responses to drugs and chemicals in appropriate models of the human liver.

Maximizing biomarker discovery by minimizing gene signatures

Background

The use of gene signatures can potentially be of considerable value in the field of clinical diagnosis. However, gene signatures defined with different methods can be quite various even when applied the same disease and the same endpoint. Previous studies have shown that the correct selection of subsets of genes from microarray data is key for the accurate classification of disease phenotypes, and a number of methods have been proposed for the purpose. However, these methods refine the subsets by only considering each single feature, and they do not confirm the association between the genes identified in each gene signature and the phenotype of the disease. We proposed an innovative new method termed Minimize Feature's Size (MFS) based on multiple level similarity analyses and association between the genes and disease for breast cancer endpoints by comparing classifier models generated from the second phase of MicroArray Quality Control (MAQC-II), trying to develop effective meta-analysis strategies to transform the MAQC-II signatures into a robust and reliable set of biomarker for clinical applications.

Results

We analyzed the similarity of the multiple gene signatures in an endpoint and between the two endpoints of breast cancer at probe and gene levels, the results indicate that disease-related genes can be preferably selected as the components of gene signature, and that the gene signatures for the two endpoints could be interchangeable. The minimized signatures were built at probe level by using MFS for each endpoint. By applying the approach, we generated a much smaller set of gene signature with the similar predictive power compared with those gene signatures from MAQC-II.

Conclusions

Our results indicate that gene signatures of both large and small sizes could perform equally well in clinical applications. Besides, consistency and biological significances can be detected among different gene signatures, reflecting the studying endpoints. New classifiers built with MFS exhibit improved performance with both internal and external validation, suggesting that MFS method generally reduces redundancies for features within gene signatures and improves the performance of the model. Consequently, our strategy will be beneficial for the microarray-based clinical applications.

Global gene profiling of spontaneous hepatocellular carcinoma in B6C3F1 mice: similarities in the molecular landscape with human liver cancer

Hepatocellular carcinoma (HCC) is an important cause of morbidity and mortality worldwide. Although the risk factors of human HCC are well known, the molecular pathogenesis of this disease is complex, and in general, treatment options remain poor. The use of rodent models to study human cancer has been extensively pursued, both through genetically engineered rodents and rodent models used in carcinogenicity and toxicology studies. In particular, the B6C3F1 mouse used in the National Toxicology Program (NTP) two-year bioassay has been used to evaluate the carcinogenic effects of environmental and occupational chemicals, and other compounds. The high incidence of spontaneous HCC in the B6C3F1 mouse has challenged its use as a model for chemically induced HCC in terms of relevance to the human disease. Using global gene expression profiling, we identify the dysregulation of several mediators similarly altered in human HCC, including re-expression of fetal oncogenes, upregulation of protooncogenes, downregulation of tumor suppressor genes, and abnormal expression of cell cycle mediators, growth factors, apoptosis regulators, and angiogenesis and extracellular matrix remodeling factors. Although major differences in etiology and pathogenesis remain between human and mouse HCC, there are important similarities in global gene expression and molecular pathways dysregulated in mouse and human HCC. These data provide further support for the use of this model in hazard identification of compounds with potential human carcinogenicity risk, and may help in better understanding the mechanisms of tumorigenesis resulting from chemical exposure in the NTP two-year carcinogenicity bioassay.

Principal component analysis-based filtering improves detection for Affymetrix gene expression arrays

Gene expression array technology has reached the stage of being routinely used to study clinical samples in search of diagnostic and prognostic biomarkers. Due to the nature of array experiments, which examine the expression of tens of thousands of genes simultaneously, the number of null hypotheses is large. Hence, multiple testing correction is often necessary to control the number of false positives. However, multiple testing correction can lead to low statistical power in detecting genes that are truly differentially expressed. Filtering out non-informative genes allows for reduction in the number of null hypotheses. While several filtering methods have been suggested, the appropriate way to perform filtering is still debatable. We propose a new filtering strategy for Affymetrix GeneChips®, based on principal component analysis of probe-level gene expression data. Using a wholly defined spike-in data set and one from a diabetes study, we show that filtering by the proportion of variation accounted for by the first principal component (PVAC) provides increased sensitivity in detecting truly differentially expressed genes while controlling false discoveries. We demonstrate that PVAC exhibits equal or better performance than several widely used filtering methods. Furthermore, a data-driven approach that guides the selection of the filtering threshold value is also proposed.

Platforms for biomarker analysis using high-throughput approaches in genomics, transcriptomics, proteomics, metabolomics, and bioinformatics

Global biological responses that reflect disease or exposure biology are kinetic and highly dynamic phenomena. While high-throughput DNA sequencing continues to drive genomics, the possibility of more broadly measuring changes in gene expression has been a recent development manifested by a diversity of technical platforms. Such technologies measure transcripts, proteins and small biological molecules, or metabolites, and respectively define the fields of transcriptomics, proteomics and metabolomics that can be performed at a cell-, tissue-, or organism-wide basis. Bioinformatics is the discipline that derives knowledge from the large quantity and diversity of biological, genetic, genomic and gene expression data by integrating computer science, mathematics, statistics and graphic arts. Gene, protein and metabolite expression profiles can be thought of as snapshots of the current, poorly-mapped molecular landscape. The ultimate aim of genomic platforms is to fully map this landscape to more completely describe all of the biological interactions within a living system, during disease and toxicity, and define the behaviour and relationships of all the components of a biological system. The development of databases and knowledge bases will support the integration of data from multiple domains, as well as computational modelling. This chapter will describe the technical platform methods involving DNA sequencing, mass spectrometry, nuclear magnetic resonance combined with separation systems, and bioinformatics to derive genomic and gene expression data and include the relevant bioinformatic tools for analysis. These genomic, or omics platforms should have wide application to epidemiological studies.