Abdallah M. Eteleeb

Genome-Wide Profiling of PARP1 Reveals an Interplay with Gene Regulatory Regions and DNA Methylation

Poly (ADP-ribose) polymerase-1 (PARP1) is a nuclear enzyme involved in DNA repair, chromatin remodeling and gene expression. PARP1 interactions with chromatin architectural multi-protein complexes (i.e. nucleosomes) alter chromatin structure resulting in changes in gene expression. Chromatin structure impacts gene regulatory processes including transcription, splicing, DNA repair, replication and recombination. It is important to delineate whether PARP1 randomly associates with nucleosomes or is present at specific nucleosome regions throughout the cell genome. We performed genome-wide association studies in breast cancer cell lines to address these questions. Our studies show that PARP1 associates with epigenetic regulatory elements genome-wide, such as active histone marks, CTCF and DNase hypersensitive sites. Additionally, the binding of PARP1 to chromatin genome-wide is mutually exclusive with DNA methylation pattern suggesting a functional interplay between PARP1 and DNA methylation. Indeed, inhibition of PARylation results in genome-wide changes in DNA methylation patterns. Our results suggest that PARP1 controls the fidelity of gene transcription and marks actively transcribed gene regions by selectively binding to transcriptionally active chromatin. These studies provide a platform for developing our understanding of PARP1’s role in gene regulation.

Involvement of PARP1 in the Regulation of Alternative Splicing

Specialized chromatin structures such as nucleosomes with specific histone modifications decorate exons in eukaryotic genomes, suggesting a functional connection between chromatin organization and the regulation of pre-mRNA splicing. Through profiling the functional location of Poly (ADP) ribose polymerase, we observed that it is associated with the nucleosomes at exon/intron boundaries of specific genes, suggestive of a role for this enzyme in alternative splicing. Poly (ADP) ribose polymerase has previously been implicated in the PARylation of splicing factors as well as regulation of the histone modification H3K4me3, a mark critical for co-transcriptional splicing. In light of these studies, we hypothesized that interaction of the chromatin-modifying factor, Poly (ADP) ribose polymerase with nucleosomal structures at exon–intron boundaries, might regulate pre-mRNA splicing. Using genome-wide approaches validated by gene-specific assays, we show that depletion of PARP1 or inhibition of its PARylation activity results in changes in alternative splicing of a specific subset of genes. Furthermore, we observed that PARP1 bound to RNA, splicing factors and chromatin, suggesting that Poly (ADP) ribose polymerase serves as a gene regulatory hub to facilitate co-transcriptional splicing. These studies add another function to the multi-functional protein, Poly (ADP) ribose polymerase, and provide a platform for further investigation of this protein’s function in organizing chromatin during gene regulatory processes.

UTR extension and alternate polyadenylation in neuroplasticity: an emerging paradigm?

Results Computational analyses of the novel UTR sequences, focusing on RNA-binding protein (RNAbp) interaction motifs revealed strongly over-represented RNAbps with known roles in nervous system pathologies. We consider the implications of 3’UTR transcript extension and protein interaction in the context of axonal plasticity and the consequences of mis-regulation of this process during neurological disease.

Multi-institutional Analysis Shows that Low PCAT-14 Expression Associates with Poor Outcomes in Prostate Cancer

BACKGROUND Long noncoding RNAs (lncRNAs) are an emerging class of relatively underexplored oncogenic molecules with biological and clinical significance. Current inadequacies for stratifying patients with aggressive disease presents a strong rationale to systematically identify lncRNAs as clinical predictors in localized prostate cancer. OBJECTIVE To identify RNA biomarkers associated with aggressive prostate cancer. DESIGN, SETTING, AND PARTICIPANTS Radical prostatectomy microarray and clinical data was obtained from 910 patients in three published institutional cohorts: Mayo Clinic I (N=545, median follow-up 13.8 yr), Mayo Clinic II (N=235, median follow-up 6.7 yr), and Thomas Jefferson University (N=130, median follow-up 9.6 yr). OUTCOME MEASUREMENTS AND STATISTICAL ANALYSIS The primary clinical endpoint was distant metastasis-free survival. Secondary endpoints include prostate cancer-specific survival and overall survival. Univariate and multivariate Cox regression were used to evaluate the association of lncRNA expression and these endpoints. RESULTS AND LIMITATIONS An integrative analysis revealed Prostate Cancer Associated Transcript-14 (PCAT-14) as the most prevalent lncRNA that is aberrantly expressed in prostate cancer patients. Down-regulation of PCAT-14 expression significantly associated with Gleason score and a greater probability of metastatic progression, overall survival, and prostate cancer-specific mortality across multiple independent datasets and ethnicities. Low PCAT-14 expression was implicated with genes involved in biological processes promoting aggressive disease. In-vitro analysis confirmed that low PCAT-14 expression increased migration while overexpressing PCAT-14 reduced cellular growth, migration, and invasion. CONCLUSIONS We discovered that androgen-regulated PCAT-14 is overexpressed in prostate cancer, suppresses invasive phenotypes, and lower expression is significantly prognostic for multiple clinical endpoints supporting its significance for predicting metastatic disease that could be used to improve patient management. PATIENT SUMMARY We discovered that aberrant prostate cancer associated transcript-14 expression during prostate cancer progression is prevalent across cancer patients. Prostate cancer associated transcript-14 is also prognostic for metastatic disease and survival highlighting its importance for stratifying patients that could benefit from treatment intensification.

A comparison of per sample global scaling and per gene normalization methods for differential expression analysis of RNA-seq data

Normalization is an essential step with considerable impact on high-throughput RNA sequencing (RNA-seq) data analysis. Although there are numerous methods for read count normalization, it remains a challenge to choose an optimal method due to multiple factors contributing to read count variability that affects the overall sensitivity and specificity. In order to properly determine the most appropriate normalization methods, it is critical to compare the performance and shortcomings of a representative set of normalization routines based on different dataset characteristics. Therefore, we set out to evaluate the performance of the commonly used methods (DESeq, TMM-edgeR, FPKM-CuffDiff, TC, Med UQ and FQ) and two new methods we propose: Med-pgQ2 and UQ-pgQ2 (per-gene normalization after per-sample median or upper-quartile global scaling). Our per-gene normalization approach allows for comparisons between conditions based on similar count levels. Using the benchmark Microarray Quality Control Project (MAQC) and simulated datasets, we performed differential gene expression analysis to evaluate these methods. When evaluating MAQC2 with two replicates, we observed that Med-pgQ2 and UQ-pgQ2 achieved a slightly higher area under the Receiver Operating Characteristic Curve (AUC), a specificity rate > 85%, the detection power > 92% and an actual false discovery rate (FDR) under 0.06 given the nominal FDR (≤0.05). Although the top commonly used methods (DESeq and TMM-edgeR) yield a higher power (>93%) for MAQC2 data, they trade off with a reduced specificity (<70%) and a slightly higher actual FDR than our proposed methods. In addition, the results from an analysis based on the qualitative characteristics of sample distribution for MAQC2 and human breast cancer datasets show that only our gene-wise normalization methods corrected data skewed towards lower read counts. However, when we evaluated MAQC3 with less variation in five replicates, all methods performed similarly. Thus, our proposed Med-pgQ2 and UQ-pgQ2 methods perform slightly better for differential gene analysis of RNA-seq data skewed towards lowly expressed read counts with high variation by improving specificity while maintaining a good detection power with a control of the nominal FDR level.

Differential expression analysis methods for ribonucleic acid-sequencing data

Introduction High-throughput mRNA sequencing (also known as RNA-Seq) promises to be the technique of choice for studying transcriptome profiles, offering several advantages over old techniques such as microarrays. This technique provides the ability to develop precise methodologies for a variety of RNA-Seq applications, including gene expression quantification, novel transcript and exon discovery, differential expression (DE) analysis and splice variant detection. With the introduction of this technique, there has been a significant effort in developing new methods and statistical models to accurately model RNA-Seq data and test for differences in gene expression between biological conditions. In this review, we examine some of the most recently and widely used methods for DE analysis. We provide a detailed review of these methods by looking at the following three main aspects: statistical methods for normalisation, statistical modelling of gene expression and statistical methods for DE testing. Conclusion No single DE method can be considered as the best among available methods. Some methods perform well in particular situations, but their performance is poor in others. Introduction With the advent of next generation sequencing (NGS) technologies, where a large volume of short deoxyribonucleic acid (DNA) sequences (reads) are generated, new methods and techniques have been developed for transcriptome analysis. Ribonucleic acid-sequencing (RNA-Seq) technology, which is based on the direct sequencing of complementary DNA (cDNA)1, provides the ability for the reconstruction of transcripts, estimation of mRNA abundances and testing for differential expression (DE) genes between two or more conditions. This technology has enabled researchers and scientists to study the transcriptome at an unprecedented rate and has lately become a common platform for transcriptome analysis. This technique offers several advantages over the old microarray technology2. For instance, whereas microarrays generate expression signal intensities, RNA-Seq data generates digital gene expression counts. Unlike microarrays, RNA-Seq has a low background noise with high resolution. While microarrays offer resolution at the probe length, RNA-Seq allows for a single base resolution. Such granularity allows for the detection of splice variants. The dynamic range for quantifying expression differences is limited to a few hundred folds in microarrays, and can be nearly 10,000 fold with RNA-Seq data. One key limitation for microarrays is that they rely on a reference genome while RNA-Seq can take advantage of such an annotation. It also offers the ability for de novo transcriptomics. An RNA-Seq experiment starts with the extraction of total RNA or a portion, such as polyadenylatedRNA2. The extracted RNA is then converted to a library of doublestranded cDNA and sheared into small fragments. In the next step, adapters are attached to one or both sides of each cDNA fragment. Using NGS platforms, such as Illumina’s HighSeq 2500, Roche 454 GS FLX Sequencer, Applied Biosystems SOLiD Sequencer, Helicos HeliScope, or Pacific Biosciences/RS sequencer (Table 1 shows more detailed information about the most recently NGS platforms), each cDNA fragment is sequenced and a short sequence (read) from one end of the fragment (single-end tag) or from both ends (paired-end tag) is obtained (Figure 1). The obtained reads are mapped to the reference genome or transcriptome to measure the abundance of each transcript. If the reference genome or transcriptome is not available, short sequences (reads) can be assembled de novo to identify the full set of transcripts, followed by abundance estimation. One of the primary applications in RNA-Seq is the study of gene expression profiling across experimental conditions. The number of reads that map to a gene is an approximation of its expression at the transcription level. Thus, the study of determining which genes have changed significantly in terms of their expression across biological samples is referred to as DE analysis. This step is essential in any RNA-Seq study. Identifying which genes are DEs between samples help researchers to understand the functions of genes in response to a given condition. In this review, we examine the most recently developed and widely used methods for DE analysis. We observe different statistical models that each method uses to test for DE. Because a large number of methods and tools have been developed in the last few years for DE analysis, not all DE methods * Corresponding authors Emails: ametel01@louisville.edu; eric. rouchka@louisville.edu Department of Computer Engineering and Computer Science, University of Louisville, Louisville, KY, USA Ge no m e Bi oi nf or m at ic s

An Island-Based Approach for Differential Expression Analysis

High-throughput mRNA sequencing (also known as RNA-Seq) promises to be the technique of choice for studying transcriptome profiles. This technique provides the ability to develop precise methodologies for transcript and gene expression quantification, novel transcript and exon discovery, and splice variant detection. One of the limitations of current RNA-Seq methods is the dependency on annotated biological features (e.g. exons, transcripts, genes) to detect expression differences across samples. This forces the identification of expression levels and the detection of significant changes to known genomic regions. Any significant changes that occur in unannotated regions will not be captured. To overcome this limitation, we developed a novel segmentation approach, Island-Based (IB), for analyzing differential expression in RNA-Seq and targeted sequencing (exome capture) data without specific knowledge of an isoform. The IB segmentation determines individual islands of expression based on windowed read counts that can be compared across experimental conditions to determine differential island expression. In order to detect differentially expressed genes, the significance of islands (p-values) are combined using Fishers method. We tested and evaluated the performance of our approach by comparing it to the existing differentially expressed gene (DEG) methods: CuffDiff, DESeq, and edgeR using two benchmark MAQC RNA-Seq datasets. The IB algorithm outperforms all three methods in both datasets as illustrated by an increased auROC.

RNASeq profiling of UTR expression during neuronal plasticity.

Background Neurons interact with, and are influenced by, tissues that are remote from the cell body. For example, sensory neuron cell bodies are located within peri-spinal ganglia but are connected to both the spinal cord and skin via their axons projecting through dorsal roots and peripheral nerves. Biochemical signals from anatomical compartments (spinal cord / root / ganglion / nerve / skin) modulate the molecular biology of neurons which can respond to signals from any/all of these remote regions. One mechanism by which neurons respond to these signals and interact with their targets is by actively transporting mRNA to that region. There, the mRNA is translated to produce protein at locally-determined positions and times. A growing body of evidence shows that untranslated regions (UTRs) of genes are important for this targeting. For example, 3’-UTRs contain 50nt “zip code” consensus binding sites for cis-acting zip code-binding proteins (ZPBs) that drive axonal targeting of mRNA [1]. We therefore hypothesized that gene expression during collateral sprouting, an axonal growth process that is highly responsive to target-derived factors, might involve differential regulation of UTR components.

Abstract 2547: Discovery and characterization of late-stage breast cancer estrogen receptor alpha 1 bound long non-coding RNAs

Breast cancer (BC) is the second most common newly diagnosed cancer and the second leading cause of cancer death among women in the United States. Despite the proven benefits of adjuvant endocrine therapy in women with hormone receptor positive BC, relapses still occur even after initial treatment with endocrine therapy for 5 years, referred to as late-stage relapse. Long non-coding RNAs (lncRNAs) have been shown to be dysregulated in breast cancer. Recent studies have also shown lncRNAs to function by interfacing with corresponding RNA binding proteins to play critical regulatory roles of diverse cellular processes. Therefore, we hypothesize that lncRNAs may interact with ER to regulate genes promoting late-stage relapse. To address this, we aimed to identify lncRNAs bound to the estrogen receptor alpha 1 protein (ESR1) that promote late-stage relapse breast cancer. We first used transcriptome sequencing to identify altered expression levels of lncRNAs between 72 primary tumors and 24 late-stage relapse breast cancer patients. We detected 1192 altered lncRNAs when comparing the metastatic to the primary samples (FDR Citation Format: Jessica Monique Silva-Fisher, Abdallah M. Eteleeb, Torsten Nielsen, Charles M. Perou, Jorge S. Reis-Filho, Mathew J. Ellis, Elaine R. Mardis, Christopher A. Maher. Discovery and characterization of late-stage breast cancer estrogen receptor alpha 1 bound long non-coding RNAs [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2017; 2017 Apr 1-5; Washington, DC. Philadelphia (PA): AACR; Cancer Res 2017;77(13 Suppl):Abstract nr 2547. doi:10.1158/1538-7445.AM2017-2547

Abstract 993: Identification of estrogen receptor alpha 1 bound lncRNAs in aggressive breast cancer

Breast cancer (BC) is the second most common newly diagnosed cancer and the second leading cause of cancer death among women in the United States. Around 70% of diagnosed BCs are estrogen receptor positive (ER+). Despite the proven benefits of adjuvant endocrine therapy in women with hormone receptor positive breast cancer, relapses still occur even after initial treatment with endocrine therapy for 5 years, referred to as late stage relapse. While existing studies have focused on the role of protein-coding genes, long non-coding RNAs (lncRNAs) are an emerging and under-characterized class of transcripts that have been shown to be dysregulated in breast cancer. Recently, lncRNAs have been shown to function by interfacing with corresponding RNA binding proteins to play critical regulatory roles in chromatin remodeling and diverse cellular processes by acting as decoys, guides, and scaffolds. As estrogen receptor expression is controlled mostly by epigenetic and post-transcriptional mechanisms, and very rarely at the genomic level, we hypothesize that lncRNAs may interact with ER to promote aggressive disease. To address this, we aimed to identify lncRNAs bound to the estrogen receptor alpha 1 (ESR1) that promote late stage breast cancer. To accomplish this, we first used transcriptome sequencing to identify altered expression levels of lncRNAs between primary tumors and late-stage relapse breast cancer patients. We detected 2086 altered lncRNAs when comparing the metastatic to the primary samples with an FDR Citation Format: Jessica M. Silva-Fisher, Abdallah M. Eteleeb, Torsten O. Nielsen, Charles M. Perou, Jorge S. Reis-Filho, Mathew J. Ellis, Elaine R. Mardis, Christopher A. Maher. Identification of estrogen receptor alpha 1 bound lncRNAs in aggressive breast cancer. [abstract]. In: Proceedings of the 107th Annual Meeting of the American Association for Cancer Research; 2016 Apr 16-20; New Orleans, LA. Philadelphia (PA): AACR; Cancer Res 2016;76(14 Suppl):Abstract nr 993.