CoVar: A generalizable machine learning approach to identify the coordinated regulators driving variational gene expression

Network inference is used to model transcriptional, signaling, and metabolic interactions among genes, proteins, and metabolites that identify biological pathways influencing disease pathogenesis. Advances in machine learning (ML)-based inference models exhibit the predictive capabilities of capturing latent patterns in genomic data. Such models are emerging as an alternative to the statistical models identifying causative factors driving complex diseases. We present CoVar, an ML-based framework that builds upon the properties of existing inference models, to find the central genes driving perturbed gene expression across biological states. Unlike differentially expressed genes (DEGs) that capture changes in individual gene expression across conditions, CoVar focuses on identifying variational genes that undergo changes in their expression network interaction profiles, providing insights into changes in the regulatory dynamics, such as in disease pathogenesis. Subsequently, it finds core genes from among the nearest neighbors of these variational genes, which are central to the variational activity and influence the coordinated regulatory processes underlying the observed changes in gene expression. Through the analysis of simulated as well as yeast expression data perturbed by the deletion of the mitochondrial genome, we show that CoVar captures the intrinsic variationality and modularity in the expression data, identifying key driver genes not found through existing differential analysis methodologies.

Aberrant miR-29 is a predictive feature of severe phenotypes in pediatric Crohn's disease

Crohn's disease (CD) is a chronic inflammatory gut disorder. Molecular mechanisms underlying the clinical heterogeneity of CD remain poorly understood. MicroRNAs (miRNAs) are important regulators of gut physiology, and several have been implicated in the pathogenesis of adult CD. However, there is a dearth of large-scale miRNA studies for pediatric CD. We hypothesized that specific miRNAs uniquely mark pediatric CD. We performed small RNA-Seq of patient-matched colon and ileum biopsies from treatment-naive pediatric patients with CD (n = 169) and a control cohort (n = 108). Comprehensive miRNA analysis revealed 58 miRNAs altered in pediatric CD. Notably, multinomial logistic regression analysis revealed that index levels of ileal miR-29 are strongly predictive of severe inflammation and stricturing. Transcriptomic analyses of transgenic mice overexpressing miR-29 show a significant reduction of the tight junction protein gene Pmp22 and classic Paneth cell markers. The dramatic loss of Paneth cells was confirmed by histologic assays. Moreover, we found that pediatric patients with CD with elevated miR-29 exhibit significantly lower Paneth cell counts, increased inflammation scores, and reduced levels of PMP22. These findings strongly indicate that miR-29 upregulation is a distinguishing feature of pediatric CD, highly predictive of severe phenotypes, and associated with inflammation and Paneth cell loss.

Linking gene expression to clinical outcomes in pediatric Crohn's disease using machine learning

Pediatric Crohn's disease (CD) is characterized by a severe disease course with frequent complications. We sought to apply machine learning-based models to predict risk of developing future complications in pediatric CD using ileal and colonic gene expression. Gene expression data was generated from 101 formalin-fixed, paraffin-embedded (FFPE) ileal and colonic biopsies obtained from treatment-naïve CD patients and controls. Clinical outcomes including development of strictures or fistulas and progression to surgery were analyzed using differential expression and modeled using machine learning. Differential expression analysis revealed downregulation of pathways related to inflammation and extra-cellular matrix production in patients with strictures. Machine learning-based models were able to incorporate colonic gene expression and clinical characteristics to predict outcomes with high accuracy. Models showed an area under the receiver operating characteristic curve (AUROC) of 0.84 for strictures, 0.83 for remission, and 0.75 for surgery. Genes with potential prognostic importance for strictures (REG1A, MMP3, and DUOX2) were not identified in single gene differential analysis but were found to have strong contributions to predictive models. Our findings in FFPE tissue support the importance of colonic gene expression and the potential for machine learning-based models in predicting outcomes for pediatric CD.

ROCCO: a robust method for detection of open chromatin via convex optimization

MOTIVATION

Analysis of open chromatin regions across multiple samples from two or more distinct conditions can determine altered gene regulatory patterns associated with biological phenotypes and complex traits. The ATAC-seq assay allows for tractable genome-wide open chromatin profiling of large numbers of samples. Stable, broadly applicable genomic annotations of open chromatin regions are not available. Thus, most studies first identify open regions using peak calling methods for each sample independently. These are then heuristically combined to obtain a consensus peak set. Reconciling sample-specific peak results post hoc from larger cohorts is particularly challenging, and informative spatial features specific to open chromatin signals are not leveraged effectively.

RESULTS

We propose a novel method, ROCCO, that determines consensus open chromatin regions across multiple samples simultaneously. ROCCO employs robust summary statistics and solves a constrained optimization problem formulated to account for both enrichment and spatial dependence of open chromatin signal data. We show this formulation admits attractive theoretical and conceptual properties as well as superior empirical performance compared to current methodology.

AVAILABILITY AND IMPLEMENTATION

Source code, documentation, and usage demos for ROCCO are available on GitHub at: https://github.com/nolan-h-hamilton/ROCCO. ROCCO can also be installed as a stand-alone binary utility using pip/PyPI.

Toxicological and pharmacokinetic properties of sucralose-6-acetate and its parent sucralose: in vitro screening assays

The purpose of this study was to determine the toxicological and pharmacokinetic properties of sucralose-6-acetate, a structural analog of the artificial sweetener sucralose. Sucralose-6-acetate is an intermediate and impurity in the manufacture of sucralose, and recent commercial sucralose samples were found to contain up to 0.67% sucralose-6-acetate. Studies in a rodent model found that sucralose-6-acetate is also present in fecal samples with levels up to 10% relative to sucralose which suggest that sucralose is also acetylated in the intestines. A MultiFlow® assay, a high-throughput genotoxicity screening tool, and a micronucleus (MN) test that detects cytogenetic damage both indicated that sucralose-6-acetate is genotoxic. The mechanism of action was classified as clastogenic (produces DNA strand breaks) using the MultiFlow® assay. The amount of sucralose-6-acetate in a single daily sucralose-sweetened drink might far exceed the threshold of toxicological concern for genotoxicity (TTCgenotox) of 0.15 µg/person/day. The RepliGut® System was employed to expose human intestinal epithelium to sucralose-6-acetate and sucralose, and an RNA-seq analysis was performed to determine gene expression induced by these exposures. Sucralose-6-acetate significantly increased the expression of genes associated with inflammation, oxidative stress, and cancer with greatest expression for the metallothionein 1 G gene (MT1G). Measurements of transepithelial electrical resistance (TEER) and permeability in human transverse colon epithelium indicated that sucralose-6-acetate and sucralose both impaired intestinal barrier integrity. Sucralose-6-acetate also inhibited two members of the cytochrome P450 family (CYP1A2 and CYP2C19). Overall, the toxicological and pharmacokinetic findings for sucralose-6-acetate raise significant health concerns regarding the safety and regulatory status of sucralose itself.

Prdm6 drives ductus arteriosus closure by promoting ductus arteriosus smooth muscle cell identity and contractility

Based upon our demonstration that the smooth muscle cell-selective (SMC-selective) putative methyltransferase, Prdm6, interacts with myocardin-related transcription factor-A, we examined Prdm6's role in SMCs in vivo using cell type-specific knockout mouse models. Although SMC-specific depletion of Prdm6 in adult mice was well tolerated, Prdm6 depletion in Wnt1-expressing cells during development resulted in perinatal lethality and a completely penetrant patent ductus arteriosus (DA) phenotype. Lineage tracing experiments in Wnt1Cre2 Prdm6fl/fl ROSA26LacZ mice revealed normal neural crest-derived SMC investment of the outflow tract. In contrast, myography measurements on DA segments isolated from E18.5 embryos indicated that Prdm6 depletion significantly reduced DA tone and contractility. RNA-Seq analyses on DA and ascending aorta samples at E18.5 identified a DA-enriched gene program that included many SMC-selective contractile associated proteins that was downregulated by Prdm6 depletion. Chromatin immunoprecipitation-sequencing experiments in outflow tract SMCs demonstrated that 50% of the genes Prdm6 depletion altered contained Prdm6 binding sites. Finally, using several genome-wide data sets, we identified an SMC-selective enhancer within the Prdm6 third intron that exhibited allele-specific activity, providing evidence that rs17149944 may be the causal SNP for a cardiovascular disease GWAS locus identified within the human PRDM6 gene.

CoVar: A generalizable machine learning approach to identify the coordinated regulators driving variational gene expression

Network inference is used to model transcriptional, signaling, and metabolic interactions among genes, proteins, and metabolites that identify biological pathways influencing disease pathogenesis. Advances in machine learning (ML)-based inference models exhibit the predictive capabilities of capturing latent patterns in genomic data. Such models are emerging as an alternative to the statistical models identifying causative factors driving complex diseases. We present CoVar, an inference framework that builds upon the properties of existing inference models, to find the central genes driving perturbed gene expression across biological states. We leverage ML-based network inference to find networks that capture the strength of regulatory interactions. Our model first pinpoints a subset of genes, termed variational, whose expression variabilities typify the differences in network connectivity between the control and perturbed data. Variational genes, by being differentially expressed themselves or possessing differentially expressed neighbor genes, capture gene expression variability. CoVar then creates subnetworks comprising variational genes and their strongly connected neighbor genes and identifies core genes central to these subnetworks that influence the bulk of the variational activity. Through the analysis of yeast expression data perturbed by the deletion of the mitochondrial genome, we show that CoVar identifies key genes not found through independent differential expression analysis.

BET Protein Inhibition Regulates Macrophage Chromatin Accessibility and Microbiota-Dependent Colitis

Introduction

In colitis, macrophage functionality is altered compared to normal homeostatic conditions. Loss of IL-10 signaling results in an inappropriate chronic inflammatory response to bacterial stimulation. It remains unknown if inhibition of bromodomain and extra-terminal domain (BET) proteins alters usage of DNA regulatory elements responsible for driving inflammatory gene expression. We determined if the BET inhibitor, (+)-JQ1, could suppress inflammatory activation of macrophages in Il10^-/- mice.

Methods

We performed ATAC-seq and RNA-seq on Il10^-/- bone marrow-derived macrophages (BMDMs) cultured in the presence and absence of lipopolysaccharide (LPS) with and without treatment with (+)-JQ1 and evaluated changes in chromatin accessibility and gene expression. Germ-free Il10^-/- mice were treated with (+)-JQ1, colonized with fecal slurries and underwent histological and molecular evaluation 14-days post colonization.

Results

Treatment with (+)-JQ1 suppressed LPS-induced changes in chromatin at distal regulatory elements associated with inflammatory genes, particularly in regions that contain motifs for AP-1 and IRF transcription factors. This resulted in attenuation of inflammatory gene expression. Treatment with (+)-JQ1 in vivo resulted in a mild reduction in colitis severity as compared with vehicle-treated mice.

Conclusion

We identified the mechanism of action associated with a new class of compounds that may mitigate aberrant macrophage responses to bacteria in colitis.

Single-Cell Analysis Reveals Unexpected Cellular Changes and Transposon Expression Signatures in the Colonic Epithelium of Treatment-Naïve Adult Crohn's Disease Patients

BACKGROUND & AIMS

The intestinal barrier comprises a monolayer of specialized intestinal epithelial cells (IECs) that are critical in maintaining mucosal homeostasis. Dysfunction within various IEC fractions can alter intestinal permeability in a genetically susceptible host, resulting in a chronic and debilitating condition known as Crohn's disease (CD). Defining the molecular changes in each IEC type in CD will contribute to an improved understanding of the pathogenic processes and the identification of cell type-specific therapeutic targets. We performed, at single-cell resolution, a direct comparison of the colonic epithelial cellular and molecular landscape between treatment-naïve adult CD and non-inflammatory bowel disease control patients.

METHODS

Colonic epithelial-enriched, single-cell sequencing from treatment-naïve adult CD and non-inflammatory bowel disease patients was investigated to identify disease-induced differences in IEC types.

RESULTS

Our analysis showed that in CD patients there is a significant skew in the colonic epithelial cellular distribution away from canonical LGR5+ stem cells, located at the crypt bottom, and toward one specific subtype of mature colonocytes, located at the crypt top. Further analysis showed unique changes to gene expression programs in every major cell type, including a previously undescribed suppression in CD of most enteroendocrine driver genes as well as L-cell markers including GCG. We also dissect an incompletely understood SPIB+ cell cluster, revealing at least 4 subclusters that likely represent different stages of a maturational trajectory. One of these SPIB+ subclusters expresses crypt-top colonocyte markers and is up-regulated significantly in CD, whereas another subcluster strongly expresses and stains positive for lysozyme (albeit no other canonical Paneth cell marker), which surprisingly is greatly reduced in expression in CD. In addition, we also discovered transposable element markers of colonic epithelial cell types as well as transposable element families that are altered significantly in CD in a cell type-specific manner. Finally, through integration with data from genome-wide association studies, we show that genes implicated in CD risk show heretofore unknown cell type-specific patterns of aberrant expression in CD, providing unprecedented insight into the potential biological functions of these genes.

CONCLUSIONS

Single-cell analysis shows a number of unexpected cellular and molecular features, including transposable element expression signatures, in the colonic epithelium of treatment-naïve adult CD.

Subcutaneous adipose tissue splice quantitative trait loci reveal differences in isoform usage associated with cardiometabolic traits

Alternate splicing events can create isoforms that alter gene function, and genetic variants associated with alternate gene isoforms may reveal molecular mechanisms of disease. We used subcutaneous adipose tissue of 426 Finnish men from the METSIM study and identified splice junction quantitative trait loci (sQTLs) for 6,077 splice junctions (FDR < 1%). In the same individuals, we detected expression QTLs (eQTLs) for 59,443 exons and 15,397 genes (FDR < 1%). We identified 595 genes with an sQTL and exon eQTL but no gene eQTL, which could indicate potential isoform differences. Of the significant sQTL signals, 2,114 (39.8%) included at least one proxy variant (linkage disequilibrium r2 > 0.8) located within an intron spanned by the splice junction. We identified 203 sQTLs that colocalized with 141 genome-wide association study (GWAS) signals for cardiometabolic traits, including 25 signals for lipid traits, 24 signals for body mass index (BMI), and 12 signals for waist-hip ratio adjusted for BMI. Among all 141 GWAS signals colocalized with an sQTL, we detected 26 that also colocalized with an exon eQTL for an exon skipped by the sQTL splice junction. At a GWAS signal for high-density lipoprotein cholesterol colocalized with an NR1H3 sQTL splice junction, we show that the alternative splice product encodes an NR1H3 transcription factor that lacks a DNA binding domain and fails to activate transcription. Together, these results detect splicing events and candidate mechanisms that may contribute to gene function at GWAS loci.

Integrative analysis reveals mouse strain-dependent responses to acute ozone exposure associated with airway macrophage transcriptional activity

Acute ozone (O3) exposure is associated with multiple adverse cardiorespiratory outcomes, the severity of which varies across individuals in human populations and inbred mouse strains. However, molecular determinants of response, including susceptibility biomarkers that distinguish who will develop severe injury and inflammation, are not well characterized. We and others have demonstrated that airway macrophages (AMs) are an important resident immune cell type that are functionally and transcriptionally responsive to O3 inhalation. Here, we sought to explore influences of strain, exposure, and strain-by-O3 exposure interactions on AM gene expression and identify transcriptional correlates of O3-induced inflammation and injury across six mouse strains, including five Collaborative Cross (CC) strains. We exposed adult mice of both sexes to filtered air (FA) or 2 ppm O3 for 3 h and measured inflammatory and injury parameters 21 h later. Mice exposed to O3 developed airway neutrophilia and lung injury with strain-dependent severity. In AMs, we identified a common core O3 transcriptional response signature across all strains, as well as a set of genes exhibiting strain-by-O3 exposure interactions. In particular, a prominent gene expression contrast emerged between a low- (CC017/Unc) and high-responding (CC003/Unc) strain, as reflected by cellular inflammation and injury. Further inspection indicated that differences in their baseline gene expression and chromatin accessibility profiles likely contribute to their divergent post-O3 exposure transcriptional responses. Together, these results suggest that aspects of O3-induced respiratory responses are mediated through altered AM transcriptional signatures and further confirm the importance of gene-environment interactions in mediating differential responsiveness to environmental agents.

Increased colonic expression of ACE2 associates with poor prognosis in Crohn's disease

The host receptor for SARS-CoV-2, angiotensin-converting enzyme 2 (ACE2), is highly expressed in small intestine. Our aim was to study colonic ACE2 expression in Crohn's disease (CD) and non-inflammatory bowel disease (non-IBD) controls. We hypothesized that the colonic expression levels of ACE2 impacts CD course. We examined the expression of colonic ACE2 in 67 adult CD and 14 NIBD control patients using RNA-seq and quantitative (q) RT-PCR. We validated ACE2 protein expression and localization in formalin-fixed, paraffin-embedded matched colon and ileal tissues using immunohistochemistry. The impact of increased ACE2 expression in CD for the risk of surgery was evaluated by a multivariate regression analysis and a Kaplan–Meier estimator. To provide critical support for the generality of our findings, we analyzed previously published RNA-seq data from two large independent cohorts of CD patients. Colonic ACE2 expression was significantly higher in a subset of adult CD patients which was defined as the ACE2-high CD subset. IHC in a sampling of ACE2-high CD patients confirmed high ACE2 protein expression in the colon and ileum compared to ACE2-low CD and NIBD patients. Notably, we found that ACE2-high CD patients are significantly more likely to undergo surgery within 5 years of CD diagnosis, and a Cox regression analysis found that high ACE2 levels is an independent risk factor for surgery (OR 2.17; 95% CI, 1.10–4.26; p = 0.025). Increased intestinal expression of ACE2 is associated with deteriorated clinical outcomes in CD patients. These data point to the need for molecular stratification that can impact CD disease-related outcomes.

Increased Colonic Expression of ACE2 Associates with Poor Prognosis in Crohn's disease

Background and Aims

The host receptor for SARS-CoV-2, angiotensin-converting enzyme 2 (ACE2), is highly expressed in small intestine. Our aim was to study colonic ACE2 expression in Crohn's disease (CD) and non-inflammatory bowel disease (non-IBD) controls. We hypothesized that the colonic expression levels of ACE2 impacts CD course.

Methods

We examined the expression of colon ACE2 using RNA-seq and quantitative (q) RT-PCR from 69 adult CD and 14 NIBD control patients. In a subset of this cohort we validated ACE2 protein expression and localization in formalin-fixed, paraffin-embedded matched colon and ileal tissues using immunohistochemistry. The impact of increased ACE2 expression in CD for the risk of surgery was evaluated by a multivariate regression analysis and a Kaplan-Meier estimator. To provide critical support for the generality of our findings, we analyzed previously published RNA-seq data from two large independent cohorts of CD patients.

Results

Colonic ACE2 expression was significantly higher in a subset of adult CD patients (ACE2-high CD). IHC in a sampling of ACE2-high CD patients confirmed high ACE2 protein expression in the colon and ileum compared to ACE2-low CD and NIBD patients. Notably, we found that ACE2-high CD patients are significantly more likely to undergo surgery within 5 years of diagnosis, with a Cox regression analysis finding that high ACE2 levels is an independent risk factor (OR 2.18; 95%CI, 1.05-4.55; p=0.037).

Conclusion

Increased intestinal expression of ACE2 is associated with deteriorated clinical outcomes in CD patients. These data point to the need for molecular stratification that may impact CD disease-related outcomes.

Decreased Colonic Activin Receptor-Like Kinase 1 Disrupts Epithelial Barrier Integrity in Patients With Crohn's Disease

BACKGROUND & AIMS

Intestinal epithelial cell (IEC) barrier dysfunction is critical to the development of Crohn's disease (CD). However, the mechanism is understudied. We recently reported increased microRNA-31-5p (miR-31-5p) expression in colonic IECs of CD patients, but downstream targets and functional consequences are unknown.

METHODS

microRNA-31-5p target genes were identified by integrative analysis of RNA- and small RNA-sequencing data from colonic mucosa and confirmed by quantitative polymerase chain reaction in colonic IECs. Functional characterization of activin receptor-like kinase 1 (ACVRL1 or ALK1) in IECs was performed ex vivo using 2-dimensional cultured human primary colonic IECs. The impact of altered colonic ALK1 signaling in CD for the risk of surgery and endoscopic relapse was evaluated by a multivariate regression analysis and a Kaplan-Meier estimator.

RESULTS

ALK1 was identified as a target of miR-31-5p in colonic IECs of CD patients and confirmed using a 3'-untranslated region reporter assay. Activation of ALK1 restricted the proliferation of colonic IECs in a 5-ethynyl-2-deoxyuridine proliferation assay and down-regulated the expression of stemness-related genes. Activated ALK1 signaling increased colonic IEC differentiation toward colonocytes. Down-regulated ALK1 signaling was associated with increased stemness and decreased colonocyte-specific marker expression in colonic IECs of CD patients compared with healthy controls. Activation of ALK1 enhanced epithelial barrier integrity in a transepithelial electrical resistance permeability assay. Lower colonic ALK1 expression was identified as an independent risk factor for surgery and was associated with a higher risk of endoscopic relapse in CD patients.

CONCLUSIONS

Decreased colonic ALK1 disrupted colonic IEC barrier integrity and was associated with poor clinical outcomes in CD patients.

Crohn's Disease Differentially Affects Region-Specific Composition and Aerotolerance Profiles of Mucosally Adherent Bacteria

BACKGROUND

The intestinal microbiota play a key role in the onset, progression, and recurrence of Crohn disease (CD). Most microbiome studies assay fecal material, which does not provide region-specific information on mucosally adherent bacteria that directly interact with host systems. Changes in luminal oxygen have been proposed as a contributor to CD dybiosis.

METHODS

The authors generated 16S rRNA data using colonic and ileal mucosal bacteria from patients with CD and without inflammatory bowel disease. We developed profiles reflecting bacterial abundance within defined aerotolerance categories. Bacterial diversity, composition, and aerotolerance profiles were compared across intestinal regions and disease phenotypes.

RESULTS

Bacterial diversity decreased in CD in both the ileum and the colon. Aerotolerance profiles significantly differed between intestinal segments in patients without inflammatory bowel disease, although both were dominated by obligate anaerobes, as expected. In CD, high relative levels of obligate anaerobes were maintained in the colon and increased in the ileum. Relative abundances of similar and distinct taxa were altered in colon and ileum. Notably, several obligate anaerobes, such as Bacteroides fragilis, dramatically increased in CD in one or both intestinal segments, although specific increasing taxa varied across patients. Increased abundance of taxa from the Proteobacteria phylum was found only in the ileum. Bacterial diversity was significantly reduced in resected tissues of patients who developed postoperative disease recurrence across 2 independent cohorts, with common lower abundance of bacteria from the Bacteroides, Streptococcus, and Blautia genera.

CONCLUSIONS

Mucosally adherent bacteria in the colon and ileum show distinct alterations in CD that provide additional insights not revealed in fecal material.

Integrative QTL analysis of gene expression and chromatin accessibility identifies multi-tissue patterns of genetic regulation

Gene transcription profiles across tissues are largely defined by the activity of regulatory elements, most of which correspond to regions of accessible chromatin. Regulatory element activity is in turn modulated by genetic variation, resulting in variable transcription rates across individuals. The interplay of these factors, however, is poorly understood. Here we characterize expression and chromatin state dynamics across three tissues-liver, lung, and kidney-in 47 strains of the Collaborative Cross (CC) mouse population, examining the regulation of these dynamics by expression quantitative trait loci (eQTL) and chromatin QTL (cQTL). QTL whose allelic effects were consistent across tissues were detected for 1,101 genes and 133 chromatin regions. Also detected were eQTL and cQTL whose allelic effects differed across tissues, including local-eQTL for Pik3c2g detected in all three tissues but with distinct allelic effects. Leveraging overlapping measurements of gene expression and chromatin accessibility on the same mice from multiple tissues, we used mediation analysis to identify chromatin and gene expression intermediates of eQTL effects. Based on QTL and mediation analyses over multiple tissues, we propose a causal model for the distal genetic regulation of Akr1e1, a gene involved in glycogen metabolism, through the zinc finger transcription factor Zfp985 and chromatin intermediates. This analysis demonstrates the complexity of transcriptional and chromatin dynamics and their regulation over multiple tissues, as well as the value of the CC and related genetic resource populations for identifying specific regulatory mechanisms within cells and tissues.

Colonic epithelial miR-31 associates with the development of Crohn's phenotypes

BACKGROUND

Crohn's disease (CD) is highly heterogeneous, due in large part to variability in cellular processes that underlie the natural history of CD, thereby confounding effective therapy. There is a critical need to advance understanding of the cellular mechanisms that drive CD heterogeneity.

METHODS

We performed small RNA sequencing of adult colon tissue from CD and NIBD controls. Colonic epithelial cells and immune cells were isolated from colonic tissues, and microRNA-31 (miR-31) expression was measured. miR-31 expression was measured in colonoid cultures generated from controls and patients with CD. We performed small RNA-sequencing of formalin-fixed paraffin-embedded colon and ileum biopsies from treatment-naive pediatric patients with CD and controls and collected data on disease features and outcomes.

RESULTS

Small RNA-sequencing and microRNA profiling in the colon revealed 2 distinct molecular subtypes, each with different clinical associations. Notably, we found that miR-31 expression was a driver of these 2 subtypes and, further, that miR-31 expression was particularly pronounced in epithelial cells. Colonoids revealed that miR-31 expression differences are preserved in this ex vivo system. In adult patients, low colonic miR-31 expression levels at the time of surgery were associated with worse disease outcome as measured by need for an end ileostomy and recurrence of disease in the neoterminal ileum. In pediatric patients, lower miR-31 expression at the time of diagnosis was associated with future development of fibrostenotic ileal CD requiring surgeryCONCLUSIONS. These findings represent an important step forward in designing more effective clinical trials and developing personalized CD therapies.

FUNDING

This work was supported by CCF Career Development Award (SZS), R01-ES024983 from NIEHS (SZS and TSF), 1R01DK104828-01A1 from NIDDK (SZS and TSF), P01-DK094779-01A1 from NIDDK (SZS), P30-DK034987 from NIDDK (SZS), 1-16-ACE-47 ADA Pathway Award (PS), UNC Nutrition Obesity Research Center Pilot & Feasibility Grant P30DK056350 (PS), CCF PRO-KIIDS NETWORK (SZS and PS), UNC CGIBD T32 Training Grant from NIDDK (JBB), T32 Training Grant (5T32GM007092-42) from NIGMS (MH), and SHARE from the Helmsley Trust (SZS). The UNC Translational Pathology Laboratory is supported, in part, by grants from the National Cancer Institute (3P30CA016086) and the UNC University Cancer Research Fund (UCRF) (PS).

Tissue- and strain-specific effects of a genotoxic carcinogen 1,3-butadiene on chromatin and transcription

Epigenetic effects of environmental chemicals are under intense investigation to fill existing knowledge gaps between environmental/occupational exposures and adverse health outcomes. Chromatin accessibility is one prominent mechanism of epigenetic control of transcription, and understanding of the chemical effects on both could inform the causal role of epigenetic alterations in disease mechanisms. In this study, we hypothesized that baseline variability in chromatin organization and transcription profiles among various tissues and mouse strains influence the outcome of exposure to the DNA damaging chemical 1,3-butadiene. To test this hypothesis, we evaluated DNA damage along with comprehensive quantification of RNA transcripts (RNA-seq), identification of accessible chromatin (ATAC-seq), and characterization of regions with histone modifications associated with active transcription (ChIP-seq for acetylation at histone 3 lysine 27, H3K27ac). We collected these data in the lung, liver, and kidney of mice from two genetically divergent strains, C57BL/6J and CAST/EiJ, that were exposed to clean air or to 1,3-butadiene (~600 ppm) for 2 weeks. We found that tissue effects dominate differences in both gene expression and chromatin states, followed by strain effects. At baseline, xenobiotic metabolism was consistently more active in CAST/EiJ, while immune system pathways were more active in C57BL/6J across tissues. Surprisingly, even though all three tissues in both strains harbored butadiene-induced DNA damage, little transcriptional effect of butadiene was observed in liver and kidney. Toxicologically relevant effects of butadiene in the lung were on the pathways of xenobiotic metabolism and inflammation. We also found that variability in chromatin accessibility across individuals (i.e., strains) only partially explains the variability in transcription. This study showed that variation in the basal states of epigenome and transcriptome may be useful indicators for individuals or tissues susceptible to genotoxic environmental chemicals.

Molecular classification of Crohn's disease reveals two clinically relevant subtypes

OBJECTIVE

The clinical presentation and course of Crohn's disease (CD) is highly variable. We sought to better understand the cellular and molecular mechanisms that guide this heterogeneity, and characterise the cellular processes associated with disease phenotypes.

DESIGN

We examined both gene expression and gene regulation (chromatin accessibility) in non-inflamed colon tissue from a cohort of adult patients with CD and control patients. To support the generality of our findings, we analysed previously published expression data from a large cohort of treatment-naïve paediatric CD and control ileum.

RESULTS

We found that adult patients with CD clearly segregated into two classes based on colon tissue gene expression-one that largely resembled the normal colon and one where certain genes showed expression patterns normally specific to the ileum. These classes were supported by changes in gene regulatory profiles observed at the level of chromatin accessibility, reflective of a fundamental shift in underlying molecular phenotypes. Furthermore, gene expression from the ilea of a treatment-naïve cohort of paediatric patients with CD could be similarly subdivided into colon-like and ileum-like classes. Finally, expression patterns within these CD subclasses highlight large-scale differences in the immune response and aspects of cellular metabolism, and were associated with multiple clinical phenotypes describing disease behaviour, including rectal disease and need for colectomy.

CONCLUSIONS

Our results strongly suggest that these molecular signatures define two clinically relevant forms of CD irrespective of tissue sampling location, patient age or treatment status.

Variation in DNA-Damage Responses to an Inhalational Carcinogen (1,3-Butadiene) in Relation to Strain-Specific Differences in Chromatin Accessibility and Gene Transcription Profiles in C57BL/6J and CAST/EiJ Mice

BACKGROUND

The damaging effects of exposure to environmental toxicants differentially affect genetically distinct individuals, but the mechanisms contributing to these differences are poorly understood. Genetic variation affects the establishment of the gene regulatory landscape and thus gene expression, and we hypothesized that this contributes to the observed heterogeneity in individual responses to exogenous cellular insults.

OBJECTIVES

We performed an in vivo study of how genetic variation and chromatin organization may dictate susceptibility to DNA damage, and influence the cellular response to such damage, caused by an environmental toxicant.

MATERIALS AND METHODS

We measured DNA damage, messenger RNA (mRNA) and microRNA (miRNA) expression, and genome-wide chromatin accessibility in lung tissue from two genetically divergent inbred mouse strains, C57BL/6J and CAST/EiJ, both in unexposed mice and in mice exposed to a model DNA-damaging chemical, 1,3-butadiene.

RESULTS

Our results showed that unexposed CAST/EiJ and C57BL/6J mice have very different chromatin organization and transcription profiles in the lung. Importantly, in unexposed CAST/EiJ mice, which acquired relatively less 1,3-butadiene-induced DNA damage, we observed increased transcription and a more accessible chromatin landscape around genes involved in detoxification pathways. Upon chemical exposure, chromatin was significantly remodeled in the lung of C57BL/6J mice, a strain that acquired higher levels of 1,3-butadiene-induced DNA damage, around the same genes, ultimately resembling the molecular profile of CAST/EiJ.

CONCLUSIONS

These results suggest that strain-specific changes in chromatin and transcription in response to chemical exposure lead to a "compensation" for underlying genetic-driven interindividual differences in the baseline chromatin and transcriptional state. This work represents an example of how chemical and environmental exposures can be evaluated to better understand gene-by-environment interactions, and it demonstrates the important role of chromatin response in transcriptomic changes and, potentially, in deleterious effects of exposure. https://doi.org/10.1289/EHP1937.

Genetic and epigenetic determinants of inter-individual variability in responses to toxicants

It is well established that genetic variability has a major impact on susceptibility to common diseases, responses to drugs and toxicants, and influences disease-related outcomes. The appreciation that epigenetic marks also vary across the population is growing with more data becoming available from studies in humans and model organisms. In addition, the links between genetic variability, toxicity outcomes and epigenetics are being actively explored. Recent studies demonstrate that gene-by-environment interactions involve both chromatin states and transcriptional regulation, and that epigenetics provides important mechanistic clues to connect expression-related quantitative trait loci (QTL) and disease outcomes. However, studies of Gene×Environment×Epigenetics further extend the complexity of the experimental designs and create a challenge for selecting the most informative epigenetic readouts that can be feasibly performed to interrogate multiple individuals, exposures, tissue types and toxicity phenotypes. We propose that among the many possible epigenetic experimental methodologies, assessment of chromatin accessibility coupled with total RNA levels provides a cost-effective and comprehensive option to sufficiently characterize the complexity of epigenetic and regulatory activity in the context of understanding the inter-individual variability in responses to toxicants.

Genomic dissection of conserved transcriptional regulation in intestinal epithelial cells

The intestinal epithelium serves critical physiologic functions that are shared among all vertebrates. However, it is unknown how the transcriptional regulatory mechanisms underlying these functions have changed over the course of vertebrate evolution. We generated genome-wide mRNA and accessible chromatin data from adult intestinal epithelial cells (IECs) in zebrafish, stickleback, mouse, and human species to determine if conserved IEC functions are achieved through common transcriptional regulation. We found evidence for substantial common regulation and conservation of gene expression regionally along the length of the intestine from fish to mammals and identified a core set of genes comprising a vertebrate IEC signature. We also identified transcriptional start sites and other putative regulatory regions that are differentially accessible in IECs in all 4 species. Although these sites rarely showed sequence conservation from fish to mammals, surprisingly, they drove highly conserved IEC expression in a zebrafish reporter assay. Common putative transcription factor binding sites (TFBS) found at these sites in multiple species indicate that sequence conservation alone is insufficient to identify much of the functionally conserved IEC regulatory information. Among the rare, highly sequence-conserved, IEC-specific regulatory regions, we discovered an ancient enhancer upstream from her6/HES1 that is active in a distinct population of Notch-positive cells in the intestinal epithelium. Together, these results show how combining accessible chromatin and mRNA datasets with TFBS prediction and in vivo reporter assays can reveal tissue-specific regulatory information conserved across 420 million years of vertebrate evolution. We define an IEC transcriptional regulatory network that is shared between fish and mammals and establish an experimental platform for studying how evolutionarily distilled regulatory information commonly controls IEC development and physiology.

The epithelium lining the intestine is an ancient animal tissue that serves as a primary site of nutrient absorption and interaction with microbiota. Its formation and function require complex patterns of gene transcription that vary along the intestine and in specialized intestinal epithelial cell (IEC) subtypes. However, it is unknown how the underlying transcriptional regulatory mechanisms have changed over the course of vertebrate evolution. Here, we used genome-wide profiling of mRNA levels and chromatin accessibility to identify conserved IEC genes and regulatory regions in 4 vertebrate species (zebrafish, stickleback, mouse, and human) separated from a common ancestor by 420 million years. We identified substantial similarities in genes expressed along the vertebrate intestine. These data disclosed putative conserved transcription factor binding sites (TFBS) enriched in accessible chromatin near IEC genes and in regulatory sites with accessibility restricted to IECs. Fluorescent reporter assays in transparent zebrafish showed that these regions, which frequently lacked sequence conservation, were still capable of driving conserved expression patterns. We also found a highly conserved region near mammalian and fish hes1 sufficient to drive expression in a specific population of IECs with active Notch signaling. These results establish a platform to define the conserved transcriptional networks underlying vertebrate IEC physiology.

DeFCoM: analysis and modeling of transcription factor binding sites using a motif-centric genomic footprinter

Motivation

Identifying the locations of transcription factor binding sites is critical for understanding how gene transcription is regulated across different cell types and conditions. Chromatin accessibility experiments such as DNaseI sequencing (DNase-seq) and Assay for Transposase Accessible Chromatin sequencing (ATAC-seq) produce genome-wide data that include distinct 'footprint' patterns at binding sites. Nearly all existing computational methods to detect footprints from these data assume that footprint signals are highly homogeneous across footprint sites. Additionally, a comprehensive and systematic comparison of footprinting methods for specifically identifying which motif sites for a specific factor are bound has not been performed.

Results

Using DNase-seq data from the ENCODE project, we show that a large degree of previously uncharacterized site-to-site variability exists in footprint signal across motif sites for a transcription factor. To model this heterogeneity in the data, we introduce a novel, supervised learning footprinter called Detecting Footprints Containing Motifs (DeFCoM). We compare DeFCoM to nine existing methods using evaluation sets from four human cell-lines and eighteen transcription factors and show that DeFCoM outperforms current methods in determining bound and unbound motif sites. We also analyze the impact of several biological and technical factors on the quality of footprint predictions to highlight important considerations when conducting footprint analyses and assessing the performance of footprint prediction methods. Finally, we show that DeFCoM can detect footprints using ATAC-seq data with similar accuracy as when using DNase-seq data.

Availability and Implementation

Python code available at https://bitbucket.org/bryancquach/defcom.

Contact

bquach@email.unc.edu or tsfurey@email.unc.edu.

Supplementary information

Supplementary data are available at Bioinformatics online.

Alterations to chromatin in intestinal macrophages link IL-10 deficiency to inappropriate inflammatory responses

Intestinal macrophages (IMs) are uniquely programmed to tolerate exposure to bacteria without mounting potent inflammatory responses. The cytokine IL-10 maintains the macrophage anti-inflammatory response such that loss of IL-10 results in chronic intestinal inflammation. To investigate how IL-10-deficiency alters IM programming and bacterial tolerance, we studied changes in chromatin accessibility in response to bacteria in macrophages from two distinct niches, the intestine and bone-marrow, from both wild-type and IL-10-deficient (Il10(-/-) ) mice. We identified chromatin accessibility changes associated with bacterial exposure and IL-10 deficiency in both bone marrow derived macrophages and IMs. Surprisingly, Il10(-/-) IMs adopted chromatin and gene expression patterns characteristic of an inflammatory response, even in the absence of bacteria. Further, when recombinant IL-10 was added to Il10(-/-) cells, it could not revert the chromatin landscape to a normal state. Our results demonstrate that IL-10 deficiency results in stable chromatin alterations in macrophages, even in the absence of bacteria. This supports a model in which IL-10-deficiency leads to chromatin alterations that contribute to a loss of IM tolerance to bacteria, which is a primary initiating event in chronic intestinal inflammation.

New loci for body fat percentage reveal link between adiposity and cardiometabolic disease risk

To increase our understanding of the genetic basis of adiposity and its links to cardiometabolic disease risk, we conducted a genome-wide association meta-analysis of body fat percentage (BF%) in up to 100,716 individuals. Twelve loci reached genome-wide significance (P<5 × 10(-8)), of which eight were previously associated with increased overall adiposity (BMI, BF%) and four (in or near COBLL1/GRB14, IGF2BP1, PLA2G6, CRTC1) were novel associations with BF%. Seven loci showed a larger effect on BF% than on BMI, suggestive of a primary association with adiposity, while five loci showed larger effects on BMI than on BF%, suggesting association with both fat and lean mass. In particular, the loci more strongly associated with BF% showed distinct cross-phenotype association signatures with a range of cardiometabolic traits revealing new insights in the link between adiposity and disease risk.

A hidden Markov random field-based Bayesian method for the detection of long-range chromosomal interactions in Hi-C data

MOTIVATION

Advances in chromosome conformation capture and next-generation sequencing technologies are enabling genome-wide investigation of dynamic chromatin interactions. For example, Hi-C experiments generate genome-wide contact frequencies between pairs of loci by sequencing DNA segments ligated from loci in close spatial proximity. One essential task in such studies is peak calling, that is, detecting non-random interactions between loci from the two-dimensional contact frequency matrix. Successful fulfillment of this task has many important implications including identifying long-range interactions that assist interpreting a sizable fraction of the results from genome-wide association studies. The task - distinguishing biologically meaningful chromatin interactions from massive numbers of random interactions - poses great challenges both statistically and computationally. Model-based methods to address this challenge are still lacking. In particular, no statistical model exists that takes the underlying dependency structure into consideration.

RESULTS

In this paper, we propose a hidden Markov random field (HMRF) based Bayesian method to rigorously model interaction probabilities in the two-dimensional space based on the contact frequency matrix. By borrowing information from neighboring loci pairs, our method demonstrates superior reproducibility and statistical power in both simulation studies and real data analysis.

AVAILABILITY AND IMPLEMENTATION

The Source codes can be downloaded at: http://www.unc.edu/∼yunmli/HMRFBayesHiC CONTACT: ming.hu@nyumc.org or yunli@med.unc.edu

SUPPLEMENTARY INFORMATION

Supplementary data are available at Bioinformatics online.

Removing reference mapping biases using limited or no genotype data identifies allelic differences in protein binding at disease-associated loci

Background

Genetic variation can alter transcriptional regulatory activity contributing to variation in complex traits and risk of disease, but identifying individual variants that affect regulatory activity has been challenging. Quantitative sequence-based experiments such as ChIP-seq and DNase-seq can detect sites of allelic imbalance where alleles contribute disproportionately to the overall signal suggesting allelic differences in regulatory activity.

Methods

We created an allelic imbalance detection pipeline, AA-ALIGNER, to remove reference mapping biases influencing allelic imbalance detection and evaluate accuracy of allelic imbalance predictions in the absence of complete genotype data. Using the sequence aligner, GSNAP, and varying amounts of genotype information to remove mapping biases we investigated the accuracy of allelic imbalance detection (binomial test) in CREB1 ChIP-seq reads from the GM12878 cell line. Additionally we thoroughly evaluated the influence of experimental and analytical parameters on imbalance detection.

Results

Compared to imbalances identified using complete genotypes, using imputed partial sample genotypes, AA-ALIGNER detected >95 % of imbalances with >90 % accuracy. AA-ALIGNER performed nearly as well using common variants when genotypes were unknown. In contrast, predicting additional heterozygous sites and imbalances using the sequence data led to >50 % false positive rates. We evaluated effects of experimental data characteristics and key analytical parameter settings on imbalance detection. Overall, total base coverage and signal dispersion across the genome most affected our ability to detect imbalances, while parameters such as imbalance significance, imputation quality thresholds, and alignment mismatches had little effect. To assess the biological relevance of imbalance predictions, we used electrophoretic mobility shift assays to functionally test for predicted allelic differences in CREB1 binding in the GM12878 lymphoblast cell line. Six of nine tested variants exhibited allelic differences in binding. Two of these variants, rs2382818 and rs713875, are located within inflammatory bowel disease-associated loci.

Conclusions

AA-ALIGNER accurately detects allelic imbalance in quantitative sequence data using partial genotypes or common variants filling a critical methodological gap in these analyses, as full genotypes are rarely available. Importantly, we demonstrate how experimental and analytical features impact imbalance detection providing guidance for similar future studies.

Electronic supplementary material

The online version of this article (doi:10.1186/s12920-015-0117-x) contains supplementary material, which is available to authorized users.

Chromatin accessibility mapping identifies mediators of basal transcription and retinoid-induced repression of OTX2 in medulloblastoma

Despite an emerging understanding of the genetic alterations giving rise to various tumors, the mechanisms whereby most oncogenes are overexpressed remain unclear. Here we have utilized an integrated approach of genomewide regulatory element mapping via DNase-seq followed by conventional reporter assays and transcription factor binding site discovery to characterize the transcriptional regulation of the medulloblastoma oncogene Orthodenticle Homeobox 2 (OTX2). Through these studies we have revealed that OTX2 is differentially regulated in medulloblastoma at the level of chromatin accessibility, which is in part mediated by DNA methylation. In cell lines exhibiting chromatin accessibility of OTX2 regulatory regions, we found that autoregulation maintains OTX2 expression. Comparison of medulloblastoma regulatory elements with those of the developing brain reveals that these tumors engage a developmental regulatory program to drive OTX2 transcription. Finally, we have identified a transcriptional regulatory element mediating retinoid-induced OTX2 repression in these tumors. This work characterizes for the first time the mechanisms of OTX2 overexpression in medulloblastoma. Furthermore, this study establishes proof of principle for applying ENCODE datasets towards the characterization of upstream trans-acting factors mediating expression of individual genes.

A genome-wide analysis of open chromatin in human epididymis epithelial cells reveals candidate regulatory elements for genes coordinating epididymal function

The epithelium lining the epididymis has a pivotal role in ensuring a luminal environment that can support normal sperm maturation. Many of the individual genes that encode proteins involved in establishing the epididymal luminal fluid are well characterized. They include ion channels, ion exchangers, transporters, and solute carriers. However, the molecular mechanisms that coordinate expression of these genes and modulate their activities in response to biological stimuli are less well understood. To identify cis-regulatory elements for genes expressed in human epididymis epithelial cells, we generated genome-wide maps of open chromatin by DNase-seq. This analysis identified 33,542 epididymis-selective DNase I hypersensitive sites (DHS), which were not evident in five cell types of different lineages. Identification of genes with epididymis-selective DHS at their promoters revealed gene pathways that are active in immature epididymis epithelial cells. These include processes correlating with epithelial function and also others with specific roles in the epididymis, including retinol metabolism and ascorbate and aldarate metabolism. Peaks of epididymis-selective chromatin were seen in the androgen receptor gene and the cystic fibrosis transmembrane conductance regulator (CFTR) gene, which has a critical role in regulating ion transport across the epididymis epithelium. In silico prediction of transcription factor binding sites that were overrepresented in epididymis-selective DHS identified epithelial transcription factors, including ELF5 and ELF3, the androgen receptor, Pax2, and Sox9, as components of epididymis transcriptional networks. Active genes, which are targets of each transcription factor, reveal important biological processes in the epididymis epithelium.

Variation in chromatin accessibility in human kidney cancer links H3K36 methyltransferase loss with widespread RNA processing defects

Comprehensive sequencing of human cancers has identified recurrent mutations in genes encoding chromatin regulatory proteins. For clear cell renal cell carcinoma (ccRCC), three of the five commonly mutated genes encode the chromatin regulators PBRM1, SETD2, and BAP1. How these mutations alter the chromatin landscape and transcriptional program in ccRCC or other cancers is not understood. Here, we identified alterations in chromatin organization and transcript profiles associated with mutations in chromatin regulators in a large cohort of primary human kidney tumors. By associating variation in chromatin organization with mutations in SETD2, which encodes the enzyme responsible for H3K36 trimethylation, we found that changes in chromatin accessibility occurred primarily within actively transcribed genes. This increase in chromatin accessibility was linked with widespread alterations in RNA processing, including intron retention and aberrant splicing, affecting ∼25% of all expressed genes. Furthermore, decreased nucleosome occupancy proximal to misspliced exons was observed in tumors lacking H3K36me3. These results directly link mutations in SETD2 to chromatin accessibility changes and RNA processing defects in cancer. Detecting the functional consequences of specific mutations in chromatin regulatory proteins in primary human samples could ultimately inform the therapeutic application of an emerging class of chromatin-targeted compounds.

DNase-seq predicts regions of rotational nucleosome stability across diverse human cell types

DNase-seq is primarily used to identify nucleosome-depleted DNase I hypersensitive (DHS) sites genome-wide that correspond to active regulatory elements. However, ∼40 yr ago it was demonstrated that DNase I also digests with a ∼10-bp periodicity around nucleosomes matching the exposure of the DNA minor groove as it wraps around histones. Here, we use DNase-seq data from 49 samples representing diverse cell types to reveal this digestion pattern at individual loci and predict genomic locations where nucleosome rotational positioning, the orientation of DNA with respect to the histone surface, is stably maintained. We call these regions DNase I annotated regions of nucleosome stability (DARNS). Compared to MNase-seq experiments, we show DARNS correspond well to annotated nucleosomes. Interestingly, many DARNS are positioned over only one side of annotated nucleosomes, suggesting that the periodic digestion pattern attenuates over the nucleosome dyad. DARNS reproduce the arrangement of nucleosomes around transcription start sites and are depleted at ubiquitous DHS sites. We also generated DARNS from multiple lymphoblast cell line (LCL) samples. We found that LCL DARNS were enriched at DHS sites present in most of the original 49 samples but absent in LCLs, while multi-cell-type DARNS were enriched at LCL-specific DHS sites. This indicates that variably open DHS sites are often occupied by rotationally stable nucleosomes in cell types where the DHS site is closed. DARNS provide additional information about precise DNA orientation within individual nucleosomes not available from other nucleosome positioning assays and contribute to understanding the role of chromatin in gene regulation.

Patterns of regulatory activity across diverse human cell types predict tissue identity, transcription factor binding, and long-range interactions

Regulatory elements recruit transcription factors that modulate gene expression distinctly across cell types, but the relationships among these remains elusive. To address this, we analyzed matched DNase-seq and gene expression data for 112 human samples representing 72 cell types. We first defined more than 1800 clusters of DNase I hypersensitive sites (DHSs) with similar tissue specificity of DNase-seq signal patterns. We then used these to uncover distinct associations between DHSs and promoters, CpG islands, conserved elements, and transcription factor motif enrichment. Motif analysis within clusters identified known and novel motifs in cell-type-specific and ubiquitous regulatory elements and supports a role for AP-1 regulating open chromatin. We developed a classifier that accurately predicts cell-type lineage based on only 43 DHSs and evaluated the tissue of origin for cancer cell types. A similar classifier identified three sex-specific loci on the X chromosome, including the XIST lincRNA locus. By correlating DNase I signal and gene expression, we predicted regulated genes for more than 500K DHSs. Finally, we introduce a web resource to enable researchers to use these results to explore these regulatory patterns and better understand how expression is modulated within and across human cell types.

Identifying and characterizing regulatory sequences in the human genome with chromatin accessibility assays

After finishing a human genome reference sequence in 2002, the genomics community has turned to the task of interpreting it. A primary focus is to identify and characterize not only protein-coding genes, but all functional elements in the genome. The effort includes both individual investigators and large-scale projects like the Encyclopedia of DNA Elements (ENCODE) project. As part of the ENCODE project, several groups have identified millions of regulatory elements in hundreds of human cell-types using DNase-seq and FAIRE-seq experiments that detect regions of nucleosome-free open chromatin. ChIP-seq experiments have also been used to discover transcription factor binding sites and map histone modifications. Nearly all identified elements are found in non-coding DNA, hypothesizing a function for previously unannotated sequence. In this review, we provide an overview of the ENCODE effort to define regulatory elements, summarize the main results, and discuss implications of the millions of regulatory elements distributed throughout the genome.

Chromatin accessibility reveals insights into androgen receptor activation and transcriptional specificity

BACKGROUND

Epigenetic mechanisms such as chromatin accessibility impact transcription factor binding to DNA and transcriptional specificity. The androgen receptor (AR), a master regulator of the male phenotype and prostate cancer pathogenesis, acts primarily through ligand-activated transcription of target genes. Although several determinants of AR transcriptional specificity have been elucidated, our understanding of the interplay between chromatin accessibility and AR function remains incomplete.

RESULTS

We used deep sequencing to assess chromatin structure via DNase I hypersensitivity and mRNA abundance, and paired these datasets with three independent AR ChIP-seq datasets. Our analysis revealed qualitative and quantitative differences in chromatin accessibility that corresponded to both AR binding and an enrichment of motifs for potential collaborating factors, one of which was identified as SP1. These quantitative differences were significantly associated with AR-regulated mRNA transcription across the genome. Base-pair resolution of the DNase I cleavage profile revealed three distinct footprinting patterns associated with the AR-DNA interaction, suggesting multiple modes of AR interaction with the genome.

CONCLUSIONS

In contrast with other DNA-binding factors, AR binding to the genome does not only target regions that are accessible to DNase I cleavage prior to hormone induction. AR binding is invariably associated with an increase in chromatin accessibility and, consequently, changes in gene expression. Furthermore, we present the first in vivo evidence that a significant fraction of AR binds only to half of the full AR DNA motif. These findings indicate a dynamic quantitative relationship between chromatin structure and AR-DNA binding that impacts AR transcriptional specificity.

The accessible chromatin landscape of the human genome

DNase I hypersensitive sites (DHSs) are markers of regulatory DNA and have underpinned the discovery of all classes of cis-regulatory elements including enhancers, promoters, insulators, silencers and locus control regions. Here we present the first extensive map of human DHSs identified through genome-wide profiling in 125 diverse cell and tissue types. We identify ∼2.9 million DHSs that encompass virtually all known experimentally validated cis-regulatory sequences and expose a vast trove of novel elements, most with highly cell-selective regulation. Annotating these elements using ENCODE data reveals novel relationships between chromatin accessibility, transcription, DNA methylation and regulatory factor occupancy patterns. We connect ∼580,000 distal DHSs with their target promoters, revealing systematic pairing of different classes of distal DHSs and specific promoter types. Patterning of chromatin accessibility at many regulatory regions is organized with dozens to hundreds of co-activated elements, and the transcellular DNase I sensitivity pattern at a given region can predict cell-type-specific functional behaviours. The DHS landscape shows signatures of recent functional evolutionary constraint. However, the DHS compartment in pluripotent and immortalized cells exhibits higher mutation rates than that in highly differentiated cells, exposing an unexpected link between chromatin accessibility, proliferative potential and patterns of human variation.

A genome-wide analysis of open chromatin in human tracheal epithelial cells reveals novel candidate regulatory elements for lung function

BACKGROUND

Distal cell-type-specific regulatory elements may be located at very large distances from the genes that they control and are often hidden within intergenic regions or in introns of other genes. The development of methods that enable mapping of regions of open chromatin genome wide has greatly advanced the identification and characterisation of these elements.

METHODS

Here we use DNase I hypersensitivity mapping followed by deep sequencing (DNase-seq) to generate a map of open chromatin in primary human tracheal epithelial (HTE) cells and use bioinformatic approaches to characterise the distribution of these sites within the genome and with respect to gene promoters, intronic and intergenic regions.

RESULTS

Genes with HTE-selective open chromatin at their promoters were associated with multiple pathways of epithelial function and differentiation. The data predict novel cell-type-specific regulatory elements for genes involved in HTE cell function, such as structural proteins and ion channels, and the transcription factors that may interact with them to control gene expression. Moreover, the map of open chromatin can identify the location of potentially critical regulatory elements in genome-wide association studies (GWAS) in which the strongest association is with single nucleotide polymorphisms in non-coding regions of the genome. We demonstrate its relevance to a recent GWAS that identifies modifiers of cystic fibrosis lung disease severity.

CONCLUSION

Since HTE cells have many functional similarities with bronchial epithelial cells and other differentiated cells in the respiratory epithelium, these data are of direct relevance to elucidating the molecular basis of normal lung function and lung disease.

Integrating genetic and gene expression evidence into genome-wide association analysis of gene sets

Single variant or single gene analyses generally account for only a small proportion of the phenotypic variation in complex traits. Alternatively, gene set or pathway association analyses are playing an increasingly important role in uncovering genetic architectures of complex traits through the identification of systematic genetic interactions. Two dominant paradigms for gene set analyses are association analyses based on SNP genotypes and those based on gene expression profiles. However, gene-disease association can manifest in many ways, such as alterations of gene expression, genotype, and copy number; thus, an integrative approach combining multiple forms of evidence can more accurately and comprehensively capture pathway associations. We have developed a single statistical framework, Gene Set Association Analysis (GSAA), that simultaneously measures genome-wide patterns of genetic variation and gene expression variation to identify sets of genes enriched for differential expression and/or trait-associated genetic markers. Simulation studies illustrate that joint analyses of genomic data increase the power to detect real associations when compared with gene set methods that use only one genomic data type. The analysis of two human diseases, glioblastoma and Crohn's disease, detected abnormalities in previously identified disease-associated pathways, such as pathways related to PI3K signaling, DNA damage response, and the activation of NFKB. In addition, GSAA predicted novel pathway associations, for example, differential genetic and expression characteristics in genes from the ABC transporter family in glioblastoma and from the HLA system in Crohn's disease. These demonstrate that GSAA can help uncover biological pathways underlying human diseases and complex traits.

Cell-type specific and combinatorial usage of diverse transcription factors revealed by genome-wide binding studies in multiple human cells

Cell-type diversity is governed in part by differential gene expression programs mediated by transcription factor (TF) binding. However, there are few systematic studies of the genomic binding of different types of TFs across a wide range of human cell types, especially in relation to gene expression. In the ENCODE Project, we have identified the genomic binding locations across 11 different human cell types of CTCF, RNA Pol II (RNAPII), and MYC, three TFs with diverse roles. Our data and analysis revealed how these factors bind in relation to genomic features and shape gene expression and cell-type specificity. CTCF bound predominantly in intergenic regions while RNAPII and MYC preferentially bound to core promoter regions. CTCF sites were relatively invariant across diverse cell types, while MYC showed the greatest cell-type specificity. MYC and RNAPII co-localized at many of their binding sites and putative target genes. Cell-type specific binding sites, in particular for MYC and RNAPII, were associated with cell-type specific functions. Patterns of binding in relation to gene features were generally conserved across different cell types. RNAPII occupancy was higher over exons than adjacent introns, likely reflecting a link between transcriptional elongation and splicing. TF binding was positively correlated with the expression levels of their putative target genes, but combinatorial binding, in particular of MYC and RNAPII, was even more strongly associated with higher gene expression. These data illuminate how combinatorial binding of transcription factors in diverse cell types is associated with gene expression and cell-type specific biology.

Dynamics of the epigenetic landscape during erythroid differentiation after GATA1 restoration

Interplays among lineage-specific nuclear proteins, chromatin modifying enzymes, and the basal transcription machinery govern cellular differentiation, but their dynamics of action and coordination with transcriptional control are not fully understood. Alterations in chromatin structure appear to establish a permissive state for gene activation at some loci, but they play an integral role in activation at other loci. To determine the predominant roles of chromatin states and factor occupancy in directing gene regulation during differentiation, we mapped chromatin accessibility, histone modifications, and nuclear factor occupancy genome-wide during mouse erythroid differentiation dependent on the master regulatory transcription factor GATA1. Notably, despite extensive changes in gene expression, the chromatin state profiles (proportions of a gene in a chromatin state dominated by activating or repressive histone modifications) and accessibility remain largely unchanged during GATA1-induced erythroid differentiation. In contrast, gene induction and repression are strongly associated with changes in patterns of transcription factor occupancy. Our results indicate that during erythroid differentiation, the broad features of chromatin states are established at the stage of lineage commitment, largely independently of GATA1. These determine permissiveness for expression, with subsequent induction or repression mediated by distinctive combinations of transcription factors.

High-resolution genome-wide in vivo footprinting of diverse transcription factors in human cells

Regulation of gene transcription in diverse cell types is determined largely by varied sets of cis-elements where transcription factors bind. Here we demonstrate that data from a single high-throughput DNase I hypersensitivity assay can delineate hundreds of thousands of base-pair resolution in vivo footprints in human cells that precisely mark individual transcription factor-DNA interactions. These annotations provide a unique resource for the investigation of cis-regulatory elements. We find that footprints for specific transcription factors correlate with ChIP-seq enrichment and can accurately identify functional versus nonfunctional transcription factor motifs. We also find that footprints reveal a unique evolutionary conservation pattern that differentiates functional footprinted bases from surrounding DNA. Finally, detailed analysis of CTCF footprints suggests multiple modes of binding and a novel DNA binding motif upstream of the primary binding site.

Global epigenomic analysis of primary human pancreatic islets provides insights into type 2 diabetes susceptibility loci

Identifying cis-regulatory elements is important to understanding how human pancreatic islets modulate gene expression in physiologic or pathophysiologic (e.g., diabetic) conditions. We conducted genome-wide analysis of DNase I hypersensitive sites, histone H3 lysine methylation modifications (K4me1, K4me3, K79me2), and CCCTC factor (CTCF) binding in human islets. This identified ∼18,000 putative promoters (several hundred unannotated and islet-active). Surprisingly, active promoter modifications were absent at genes encoding islet-specific hormones, suggesting a distinct regulatory mechanism. Of 34,039 distal (nonpromoter) regulatory elements, 47% are islet unique and 22% are CTCF bound. In the 18 type 2 diabetes (T2D)-associated loci, we identified 118 putative regulatory elements and confirmed enhancer activity for 12 of 33 tested. Among six regulatory elements harboring T2D-associated variants, two exhibit significant allele-specific differences in activity. These findings present a global snapshot of the human islet epigenome and should provide functional context for noncoding variants emerging from genetic studies of T2D and other islet disorders.

Allele-specific and heritable chromatin signatures in humans

Next-generation sequencing-based assays to detect gene regulatory elements are enabling the analysis of individual-to-individual and allele-specific variation of chromatin status and transcription factor binding in humans. Recently, a number of studies have explored this area, using lymphoblastoid cell lines. Around 10% of chromatin sites show either individual-level differences or allele-specific behavior. Future studies are likely to be limited by cell line accessibility, meaning that white-bloodcell-based studies are likely to continue to be the main source of samples. A detailed understanding of the relationship between normal genetic variation and chromatin variation can shed light on how polymorphisms in non-coding regions in the human genome might underlie phenotypic variation and disease.

Heritable individual-specific and allele-specific chromatin signatures in humans

The extent to which variation in chromatin structure and transcription factor binding may influence gene expression, and thus underlie or contribute to variation in phenotype, is unknown. To address this question, we cataloged both individual-to-individual variation and differences between homologous chromosomes within the same individual (allele-specific variation) in chromatin structure and transcription factor binding in lymphoblastoid cells derived from individuals of geographically diverse ancestry. Ten percent of active chromatin sites were individual-specific; a similar proportion were allele-specific. Both individual-specific and allele-specific sites were commonly transmitted from parent to child, which suggests that they are heritable features of the human genome. Our study shows that heritable chromatin status and transcription factor binding differ as a result of genetic variation and may underlie phenotypic variation in humans.

DNaseI hypersensitivity at gene-poor, FSH dystrophy-linked 4q35.2

A subtelomeric region, 4q35.2, is implicated in facioscapulohumeral muscular dystrophy (FSHD), a dominant disease thought to involve local pathogenic changes in chromatin. FSHD patients have too few copies of a tandem 3.3-kb repeat (D4Z4) at 4q35.2. No phenotype is associated with having few copies of an almost identical repeat at 10q26.3. Standard expression analyses have not given definitive answers as to the genes involved. To investigate the pathogenic effects of short D4Z4 arrays on gene expression in the very gene-poor 4q35.2 and to find chromatin landmarks there for transcription control, unannotated genes and chromatin structure, we mapped DNaseI-hypersensitive (DH) sites in FSHD and control myoblasts. Using custom tiling arrays (DNase-chip), we found unexpectedly many DH sites in the two large gene deserts in this 4-Mb region. One site was seen preferentially in FSHD myoblasts. Several others were mapped >0.7 Mb from genes known to be active in the muscle lineage and were also observed in cultured fibroblasts, but not in lymphoid, myeloid or hepatic cells. Their selective occurrence in cells derived from mesoderm suggests functionality. Our findings indicate that the gene desert regions of 4q35.2 may have functional significance, possibly also to FSHD, despite their paucity of known genes.

Both noncoding and protein-coding RNAs contribute to gene expression evolution in the primate brain

Despite striking differences in cognition and behavior between humans and our closest primate relatives, several studies have found little evidence for adaptive change in protein-coding regions of genes expressed primarily in the brain. Instead, changes in gene expression may underlie many cognitive and behavioral differences. Here, we used digital gene expression: tag profiling (here called Tag-Seq, also called DGE:tag profiling) to assess changes in global transcript abundance in the frontal cortex of the brains of 3 humans, 3 chimpanzees, and 3 rhesus macaques. A substantial fraction of transcripts we identified as differentially transcribed among species were not assayed in previous studies based on microarrays. Differentially expressed tags within coding regions are enriched for gene functions involved in synaptic transmission, transport, oxidative phosphorylation, and lipid metabolism. Importantly, because Tag-Seq technology provides strand-specific information about all polyadenlyated transcripts, we were able to assay expression in noncoding intragenic regions, including both sense and antisense noncoding transcripts (relative to nearby genes). We find that many noncoding transcripts are conserved in both location and expression level between species, suggesting a possible functional role. Lastly, we examined the overlap between differential gene expression and signatures of positive selection within putative promoter regions, a sign that these differences represent adaptations during human evolution. Comparative approaches may provide important insights into genes responsible for differences in cognitive functions between humans and nonhuman primates, as well as highlighting new candidate genes for studies investigating neurological disorders.

Abstract B44: Chromatin structure impacts androgen receptor transcriptional specificity in prostate cancer cell lines

Androgen receptor (AR) regulated gene expression and the phenotypic impact of AR activity varies significantly between different prostate cancer (PC) models. Specifically, in genetically engineered prostate epithelial cells, AR activation causes slowed growth and differentiation whereas in patient-derived AR-positive prostate cancer cell lines, AR activity increases proliferation. While the factors modulating AR transcriptional specificity remain poorly understood, recent reports have underscored AR co-regulators and AR-dependent enhancer sequences as modulators of AR activity. We hypothesized that differences in chromatin structure may also contribute significantly to AR transcriptional specificity and phenotypic effect.

We used genome-wide DNase I hypersensitivity (HS) analysis coupled with tiled microarrays (DNase-chip) and ultra-high throughput sequencing (DNase-seq) to identify regions of open chromatin that are markers for various gene regulatory elements. DNaseI HS analysis was conducted in both LNCaP cells, which increase their proliferation in response to androgen, and LHSR-AR cells, a genetically engineered prostate epithelial cell line that differentiates in response to androgen. Each cell line was processed using our established DNaseI HS protocol prior to and after androgen stimulation. For DNase-chip, hypersensitive sites were identified using a custom microarray that included proximal promoter and adjacent regions of the top 100 genes differentially regulated by AR activation in each cell line. DNase-seq analysis was conducted on the entire genome of each cell line. Parzen scores based upon the frequency of detected genomic sequence were generated and used as a relative measure of DNA accessibility.

Our results demonstrate significant differences in chromatin structure between our two androgen-responsive models. DNase-chip analysis of specific genes such as KLK3 displays both overlapping and cell-type specific regions of chromatin availability that correlate with AR-regulated expression. KLK3 is differentially transcribed only in LNCaP, not LHSR-AR, in response to stimulation. Accordingly, a HS site matching a known AR enhancer element is found four kilobases upstream of the KLK3 TSS in LNCaP, but not in LHSR-AR. While chromatin structure explains some of the differences between these cell lines, it is not sufficient to account for all the differences observed. DNase-seq analysis in LNCaP cells suggests that there little immediate impact of AR activity on chromatin structure (approximately 77% of peaks are shared between stimulated and un-stimulated samples). In the proximal promoter regions of AR-regulated genes, the Parzen score increases in response to stimulation (Wilcoxon p &lt; 0.001), but the degree of change is relatively modest and may not be biologically significant. In addition, the number of peaks within the 100kb surrounding the TSS of AR-regulated genes did not change with androgen stimulation (604 un-stimulated, 613 stimulated, 75% in common). Finally, we examined the overlap between HS peaks and previously published putative AR binding sites. Of 279 suspected binding sites on chromosomes 19–22, none corresponded to a HS peak in unstimulated LNCaP cells, but 147 overlapped a HS peak in stimulated LNCaP cells. Overall, our LNCaP DNase-seq data suggests that the promoter regions of AR-regulated genes are primed for transcription, but that AR activation causes key alterations to the distal chromatin landscape. Further work is required to understand the mechanism behind phenotype-specific differences in the chromatin landscape and its effects on AR transcriptional specificity.

Citation Information: Cancer Res 2009;69(23 Suppl):B44.

High-resolution mapping studies of chromatin and gene regulatory elements

Microarray and high-throughput sequencing technologies have enabled the development of comprehensive assays to identify locations of particular chromatin structures and regulatory elements. It is now possible to create genome-wide maps of DNA methylation, trans-factor binding sites, histone variants and histone tail modifications, nucleosome positions, regions of open chromatin, and chromosome locations and interactions. This review provides a summary of these new assays that are changing the way in which molecular biology research is being performed. While the generation of large amounts of data from these experiments is becoming increasingly easier, the development of corresponding analysis methods has progressed more slowly. It will likely be years before the full extent of the information contained in these data is fully appreciated.

A general integrative genomic feature transcription factor binding site prediction method applied to analysis of USF1 binding in cardiovascular disease

Transcription factors are key mediators of human complex disease processes. Identifying the target genes of transcription factors will increase our understanding of the biological network leading to disease risk. The prediction of transcription factor binding sites (TFBSs) is one method to identify these target genes; however, current prediction methods need improvement. We chose the transcription factor upstream stimulatory factor l (USF1) to evaluate the performance of our novel TFBS prediction method because of its known genetic association with coronary artery disease (CAD) and the recent availability of USF1 chromatin immunoprecipitation microarray (ChIP-chip) results. The specific goals of our study were to develop a novel and accurate genome-scale method for predicting USF1 binding sites and associated target genes to aid in the study of CAD. Previously published USF1 ChIP-chip data for 1 per cent of the genome were used to develop and evaluate several kernel logistic regression prediction models. A combination of genomic features (phylogenetic conservation, regulatory potential, presence of a CpG island and DNaseI hypersensitivity), as well as position weight matrix (PWM) scores, were used as variables for these models. Our most accurate predictor achieved an area under the receiver operator characteristic curve of 0.827 during cross-validation experiments, significantly outperforming standard PWM-based prediction methods. When applied to the whole human genome, we predicted 24,010 USF1 binding sites within 5 kilobases upstream of the transcription start site of 9,721 genes. These predictions included 16 of 20 genes with strong evidence of USF1 regulation. Finally, in the spirit of genomic convergence, we integrated independent experimental CAD data with these USF1 binding site prediction results to develop a prioritised set of candidate genes for future CAD studies. We have shown that our novel prediction method, which employs genomic features related to the presence of regulatory elements, enables more accurate and efficient prediction of USF1 binding sites. This method can be extended to other transcription factors identified in human disease studies to help further our understanding of the biology of complex disease.