Differential Expression Testing Research Articles

Resistance training in young men induces muscle transcriptome-wide changes associated with muscle structure and metabolism refining the response to exercise-induced stress.

Gene expression is an important process underpinning the acute and chronic adaptive response to resistance exercise (RE) training. To investigate the effect of training status on vastus lateralis muscle global transcriptome at rest and following acute RE. Muscle biopsies of nine young men (age: 26(2)years; body mass: 69(9)kg; height 172(6)cm) who undertook RE training for 10weeks were collected pre and 24h post-RE in the untrained (W1) and trained (W10) states and analysed using microarray. Tests of differential expression were conducted for rested and after RE contrasts in both training states. To control for false discovery rate (FDR), multiple testing correction was performed at a cut-off of FDR < 0.05. Unaccustomed RE (at W1) upregulated muscle gene transcripts related to stress (e.g., heat shock proteins), damage and inflammation, structural remodelling, protein turnover and increased translational capacity. Trained muscles (at W10) showed changes in the transcriptome signature regarding the regulation of energy metabolism, favouring a more oxidative one, upregulated antioxidant- and immune-related genes/terms, and gene transcripts related to the cytoskeleton and extracellular matrix, muscle contraction, development and growth. These results highlight that chronic repetition of RE changes muscle transcriptome response towards a more refined response to RE-induced stress.

Gene network variation and alternative paths to convergent evolution in turtles

Diversification of the turtle's shell comprises remarkable phenotypic transformations. For instance, two divergent species convergently evolved shell-closing systems with shoulder blade (scapula) segments that enable coordinated movements with the shell. We expected these unusual structures to originate via similar changes in underlying gene networks, as skeletal segment formation is an evolutionarily conserved developmental process. We tested this hypothesis by comparing transcriptomes of scapula tissue across three stages of embryonic development in three emydid turtles from natural populations. We found that alternative strategies for skeletal segmentation were associated with interspecific differences in gene co-expression networks. Notably, mesenchyme homeobox 2 (MEOX2) and HOXA3-5 were central hubs driving the activity of 2,806 genes in a candidate network for scapula segmentation, albeit in only one species. Even so, scapula muscle overgrowth corresponded to the activity of similar myogenic networks in both species. This and other derived developmental processes were not observed in the third species, which displayed the ancestral (unsegmented) scapula condition. Differential gene expression tests against this reference lineage supported histological and network analyses. Our findings illustrate that molecular underpinnings of convergent evolution, including during the diversification of the atypical turtle "body plan," are influenced by variation in underlying developmental processes.

BART: bioinformatics array research tool

BackgroundMicroarray experiments comprise more than half of all series in the Gene Expression Omnibus (GEO). However, downloading and analyzing raw or semi-processed microarray data from GEO is not intuitive and requires manual error-prone analysis and a bioinformatics background. This is due to a lack of standardization in array platform fabrication as well as the lack of a simple interactive tool for clustering, plotting, differential expression testing, and testing for functional enrichment.ResultsWe introduce the Bioinformatics Array Research Tool (BART), an R Shiny web application that automates the microarray download and analysis process across diverse microarray platforms. It provides an intuitive interface, automatically downloads and parses data from GEO, suggests groupings of samples for differential expression testing, performs batch effect correction, outputs quality control plots, converts probe IDs, generates full lists of differentially expressed genes, and performs functional enrichment analysis. We show that BART enables a more comprehensive analysis of a wider range of microarray datasets on GEO by comparing it to four leading online microarray analysis tools.ConclusionsBART allows a scientist with no bioinformatics background to extract knowledge from their own microarray data or microarray experiments available from GEO. BART is functional on more microarray experiments and provides more comprehensive analyses than extant microarray analysis tools. BART is hosted on bart.salk.edu, includes a user tutorial, and is available for download from https://bitbucket.org/Luisa_amaral/bart.

A Framework for Multi-Omic Prediction of Treatment Response to Biologic Therapy for Psoriasis

Biologic therapies have shown high efficacy in psoriasis, but individual response varies and is poorly understood. To inform biomarker discovery in the Psoriasis Stratification to Optimise Relevant Therapy (i.e., PSORT) study, we evaluated a comprehensive array of omics platforms across three time points and multiple tissues in a pilot investigation of 10 patients with severe psoriasis, treated with the tumor necrosis factor (TNF) inhibitor, etanercept. We used RNA sequencing to analyze mRNA and small RNA transcriptome in blood, lesional and nonlesional skin, and the SOMAscan platform to investigate the serum proteome. Using an integrative systems biology approach, we identified signals of treatment response in genes and pathways associated with TNF signaling, psoriasis pathology, and the major histocompatibility complex region. We found association between clinical response and TNF-regulated genes in blood and skin. Using a combination of differential expression testing, upstream regulator analysis, clustering techniques, and predictive modeling, we show that baseline samples are indicative of patient response to biologic therapies, including signals in blood, which have traditionally been considered unreliable for inference in dermatology. In conclusion, our pilot study provides both an analytical framework and empirical basis to estimate power for larger studies, specifically the ongoing PSORT study, which we show as powered for biomarker discovery and patient stratification.

Improved de novo genome assembly and analysis of the Chinese cucurbit Siraitia grosvenorii, also known as monk fruit or luo-han-guo.

BackgroundLuo-han-guo (Siraitia grosvenorii), also called monk fruit, is a member of the Cucurbitaceae family. Monk fruit has become an important area for research because of the pharmacological and economic potential of its noncaloric, extremely sweet components (mogrosides). It is also commonly used in traditional Chinese medicine for the treatment of lung congestion, sore throat, and constipation. Recently, a single reference genome became available for monk fruit, assembled from 36.9x genome coverage reads via Illumina sequencing platforms. This genome assembly has a relatively short (34.2 kb) contig N50 length and lacks integrated annotations. These drawbacks make it difficult to use as a reference in assembling transcriptomes and discovering novel functional genes.FindingsHere, we offer a new high-quality draft of the S. grosvenorii genome assembled using 31 Gb (∼73.8x) long single molecule real time sequencing reads and polished with ∼50 Gb Illumina paired-end reads. The final genome assembly is approximately 469.5 Mb, with a contig N50 length of 432,384 bp, representing a 12.6-fold improvement. We further annotated 237.3 Mb of repetitive sequence and 30,565 consensus protein coding genes with combined evidence. Phylogenetic analysis showed that S. grosvenorii diverged from members of the Cucurbitaceae family approximately 40.9 million years ago. With comprehensive transcriptomic analysis and differential expression testing, we identified 4,606 up-regulated genes in the early fruit compared to the leaf, a number of which were linked to metabolic pathways regulating fruit development and ripening.ConclusionsThe availability of this new monk fruit genome assembly, as well as the annotations, will facilitate the discovery of new functional genes and the genetic improvement of monk fruit.

Necklace: combining reference and assembled transcriptomes for more comprehensive RNA-Seq analysis.

BackgroundRNA sequencing (RNA-seq) analyses can benefit from performing a genome-guided and de novo assembly, in particular for species where the reference genome or the annotation is incomplete. However, tools for integrating an assembled transcriptome with reference annotation are lacking.FindingsNecklace is a software pipeline that runs genome-guided and de novo assembly and combines the resulting transcriptomes with reference genome annotations. Necklace constructs a compact but comprehensive superTranscriptome out of the assembled and reference data. Reads are subsequently aligned and counted in preparation for differential expression testing.ConclusionsNecklace allows a comprehensive transcriptome to be built from a combination of assembled and annotated transcripts, which results in a more comprehensive transcriptome for the majority of organisms. In addition RNA-seq data are mapped back to this newly created superTranscript reference to enable differential expression testing with standard methods.

Systematic characterization of autophagy-related genes during the adipocyte differentiation using public-access data

Autophagy contributes to reorganizing intracellular components and forming fat droplets during the adipocyte differentiation. Here, we systematically describe the role of autophagy-related genes and gene sets during the differentiation of adipocytes. We used a public dataset from the European Nucleotide Archive from an RNA-seq experiment in which 3T3-L1 cells were induced by a differentiation induction medium, total RNA was extracted and sequenced at four different time points. Raw reads were aligned to the UCSC mouse reference genome (mm10) using HISAT2, and aligned reads were summarized at the gene or exon level using HTSeq. DESeq2 and DEXSeq were used to model the gene and exon counts and test for differential expression and relative exon usage, respectively. After applying the appropriate transformation, gene counts were used to perform the gene set and pathway enrichment analysis. Data were obtained, processed and annotated using R and Bioconductor. Several autophagy-related genes and autophagy gene sets, as defined in the Gene Ontology, were actively regulated during the course of the adipocyte differentiation. We further characterized these gene sets by clustering their members to a few distinct temporal profiles. Other potential functionally related genes were identified using a machine learning procedure. In summary, we characterized the autophagy gene sets and their members to biologically meaningful groups and elected a number of genes to be functionally related based on their expression patterns, suggesting that autophagy plays a critical role in removal of some intracellular components and supply of energy sources for lipid biogenesis during adipogenesis.

SCANPY: large-scale single-cell gene expression data analysis

Scanpy is a scalable toolkit for analyzing single-cell gene expression data. It includes methods for preprocessing, visualization, clustering, pseudotime and trajectory inference, differential expression testing, and simulation of gene regulatory networks. Its Python-based implementation efficiently deals with data sets of more than one million cells (https://github.com/theislab/Scanpy). Along with Scanpy, we present AnnData, a generic class for handling annotated data matrices (https://github.com/theislab/anndata).

CORNAS: coverage-dependent RNA-Seq analysis of gene expression data without biological replicates

BackgroundIn current statistical methods for calling differentially expressed genes in RNA-Seq experiments, the assumption is that an adjusted observed gene count represents an unknown true gene count. This adjustment usually consists of a normalization step to account for heterogeneous sample library sizes, and then the resulting normalized gene counts are used as input for parametric or non-parametric differential gene expression tests. A distribution of true gene counts, each with a different probability, can result in the same observed gene count. Importantly, sequencing coverage information is currently not explicitly incorporated into any of the statistical models used for RNA-Seq analysis.ResultsWe developed a fast Bayesian method which uses the sequencing coverage information determined from the concentration of an RNA sample to estimate the posterior distribution of a true gene count. Our method has better or comparable performance compared to NOISeq and GFOLD, according to the results from simulations and experiments with real unreplicated data. We incorporated a previously unused sequencing coverage parameter into a procedure for differential gene expression analysis with RNA-Seq data.ConclusionsOur results suggest that our method can be used to overcome analytical bottlenecks in experiments with limited number of replicates and low sequencing coverage. The method is implemented in CORNAS (Coverage-dependent RNA-Seq), and is available at https://github.com/joel-lzb/CORNAS.

Skin Transcriptomes of common bottlenose dolphins (Tursiops truncatus) from the northern Gulf of Mexico and southeastern U.S. Atlantic coasts

Common bottlenose dolphins serve as sentinels for the health of their coastal environments as they are susceptible to health impacts from anthropogenic inputs through both direct exposure and food web magnification. Remote biopsy samples have been widely used to reveal contaminant burdens in free-ranging bottlenose dolphins, but do not address the health consequences of this exposure. To gain insight into whether remote biopsies can also identify health impacts associated with contaminant burdens, we employed RNA sequencing (RNA-seq) to interrogate the transcriptomes of remote skin biopsies from 116 bottlenose dolphins from the northern Gulf of Mexico and southeastern U.S. Atlantic coasts. Gene expression was analyzed using principal component analysis, differential expression testing, and gene co-expression networks, and the results correlated to season, location, and contaminant burden. Season had a significant impact, with over 60% of genes differentially expressed between spring/summer and winter months. Geographic location exhibited lesser effects on the transcriptome, with 23.5% of genes differentially expressed between the northern Gulf of Mexico and the southeastern U.S. Atlantic locations. Despite a large overlap between the seasonal and geographical gene sets, the pathways altered in the observed gene expression profiles were somewhat distinct. Co-regulated gene modules and differential expression analysis both identified epidermal development and cellular architecture pathways to be expressed at lower levels in animals from the northern Gulf of Mexico. Although contaminant burdens measured were not significantly different between regions, some correlation with contaminant loads in individuals was observed among co-expressed gene modules, but these did not include classical detoxification pathways. Instead, this study identified other, possibly downstream pathways, including those involved in cellular architecture, immune response, and oxidative stress, that may prove to be contaminant responsive markers in bottlenose dolphin skin.

Identification of transcript regulatory patterns in cell differentiation.

Studying transcript regulatory patterns in cell differentiation is critical in understanding its complex nature of the formation and function of different cell types. This is done usually by measuring gene expression at different stages of the cell differentiation. However, if the gene expression data available are only from the mature cells, we have some challenges in identifying transcript regulatory patterns that govern the cell differentiation. We propose to exploit the information of the lineage of cell differentiation in terms of correlation structure between cell types. We assume that two different cell types that are close in the lineage will exhibit many common genes that are co-expressed relative to those that are far in the lineage. Current analysis methods tend to ignore this correlation by testing for differential expression assuming some sort of independence between cell types. We employ a Bayesian approach to estimate the posterior distribution of the mean of expression in each cell type, by taking into account the cell formation path in the lineage. This enables us to infer genes that are specific in each cell type, indicating the genes are involved in directing the cell differentiation to that particular cell type. We illustrate the method using gene expression data from a study of haematopoiesis. R codes to perform the analysis are available in http://www1.maths.leeds.ac.uk/∼arief/R/CellDiff/. a.gusnanto@leeds.ac.uk. Supplementary data are available at Bioinformatics online.

Whole Transcriptome Analysis of Pre-invasive and Invasive Early Squamous Lung Carcinoma in Archival Laser Microdissected Samples

BackgroundPreinvasive squamous cell cancer (PSCC) are local transformations of bronchial epithelia that are frequently observed in current or former smokers. Their different grades and sizes suggest a continuum of dysplastic change with increasing severity, which may culminate in invasive squamous cell carcinoma (ISCC). As a consequence of the difficulty in isolating cancerous cells from biopsies, the molecular pathology that underlies their histological variability remains largely unknown.MethodTo address this issue, we have employed microdissection to isolate normal bronchial epithelia and cancerous cells from low- and high-grade PSCC and ISCC, from paraffin embedded (FFPE) biopsies and determined gene expression using Affymetric Human Exon 1.0 ST arrays. Tests for differential gene expression were performed using the Bioconductor package limma followed by functional analyses of differentially expressed genes in IPA.ResultsExamination of differential gene expression showed small differences between low- and high-grade PSCC but substantial changes between PSCC and ISCC samples (184 vs 1200 p-value <0.05, fc ±1.75). However, the majority of the differentially expressed PSCC genes (142 genes: 77%) were shared with those in ISCC samples. Pathway analysis showed that these shared genes are associated with DNA damage response, DNA/RNA metabolism and inflammation as major biological themes. Cluster analysis identified 12 distinct patterns of gene expression including progressive up or down-regulation across PSCC and ISCC. Pathway analysis of incrementally up-regulated genes revealed again significant enrichment of terms related to DNA damage response, DNA/RNA metabolism, inflammation, survival and proliferation. Altered expression of selected genes was confirmed using RT-PCR, as well as immunohistochemistry in an independent set of 45 ISCCs.ConclusionsGene expression profiles in PSCC and ISCC differ greatly in terms of numbers of genes with altered transcriptional activity. However, altered gene expression in PSCC affects canonical pathways and cellular and biological processes, such as inflammation and DNA damage response, which are highly consistent with hallmarks of cancer.

The involvement of RUNX2 and SPARC genes in the bacterial chondronecrosis with osteomyelitis in broilers

Economic losses due to an increase of leg disorders in broilers have become a major concern of the poultry industry. Despite the efforts to reduce skeletal abnormalities in chickens, insufficient progress has been made. Bacterial chondronecrosis with osteomyelitis (BCO) is one of the main disorders that affect bone integrity in broilers. However, the genetic pathways and genes involved in most bone problems, including BCO, remains unclear. In this study, femoral samples from male broilers with 45 days of age affected or not with BCO were used to compare the relative expression with a reverse transcription real time PCR approach of 13 candidate genes: SPP1 (osteopontin), TNFRSF11B (osteoprotegerin), SPARC (osteonectin), CALB1 (calbidin 1), CALM (Calmodulin 2), IBSP (sialoprotein), COL1A2 (collagen, type I, α 2), BMP2 (bone morphogenetic protein 2), BMP3 (bone morphogenetic protein 3), RANKL (κ-B nuclear factor ligand), SMAD1 (SMAD family member 1), LEPR (leptin receptor) and RUNX2 (related transcription factor Runt 2). Differential expression test between affected and non-affected groups was performed using the REST software. The RUNX2 and SPARC genes were downregulated (P<0.05) in the affected group, with reduced expression of fourfold when compared with the non-affected group. This result indicates that the downregulation of RUNX2 and SPARC can contribute to an increased incidence of BCO in broilers.

RNA-seq analysis is easy as 1-2-3 with limma, Glimma and edgeR.

The ability to easily and efficiently analyse RNA-sequencing data is a key strength of the Bioconductor project. Starting with counts summarised at the gene-level, a typical analysis involves pre-processing, exploratory data analysis, differential expression testing and pathway analysis with the results obtained informing future experiments and validation studies. In this workflow article, we analyse RNA-sequencing data from the mouse mammary gland, demonstrating use of the popular edgeR package to import, organise, filter and normalise the data, followed by the limma package with its voom method, linear modelling and empirical Bayes moderation to assess differential expression and perform gene set testing. This pipeline is further enhanced by the Glimma package which enables interactive exploration of the results so that individual samples and genes can be examined by the user. The complete analysis offered by these three packages highlights the ease with which researchers can turn the raw counts from an RNA-sequencing experiment into biological insights using Bioconductor.

RNA-seq analysis is easy as 1-2-3 with limma, Glimma and edgeR

The ability to easily and efficiently analyse RNA-sequencing data is a key strength of the Bioconductor project. Starting with counts summarised at the gene-level, a typical analysis involves pre-processing, exploratory data analysis, differential expression testing and pathway analysis with the results obtained informing future experiments and validation studies. In this workflow article, we analyse RNA-sequencing data from the mouse mammary gland, demonstrating use of the popular edgeR package to import, organise, filter and normalise the data, followed by the limma package with its voom method, linear modelling and empirical Bayes moderation to assess differential expression and perform gene set testing. This pipeline is further enhanced by the Glimma package which enables interactive exploration of the results so that individual samples and genes can be examined by the user. The complete analysis offered by these three packages highlights the ease with which researchers can turn the raw counts from an RNA-sequencing experiment into biological insights using Bioconductor.

IMPROVED PERFORMANCE OF GENE SET ANALYSIS ON GENOME-WIDE TRANSCRIPTOMICS DATA WHEN USING GENE ACTIVITY STATE ESTIMATES.

Gene set analysis methods continue to be a popular and powerful method of evaluating genome-wide transcriptomics data. These approach require a priori grouping of genes into biologically meaningful sets, and then conducting downstream analyses at the set (instead of gene) level of analysis. Gene set analysis methods have been shown to yield more powerful statistical conclusions than single-gene analyses due to both reduced multiple testing penalties and potentially larger observed effects due to the aggregation of effects across multiple genes in the set. Traditionally, gene set analysis methods have been applied directly to normalized, log-transformed, transcriptomics data. Recently, efforts have been made to transform transcriptomics data to scales yielding more biologically interpretable results. For example, recently proposed models transform log-transformed transcriptomics data to a confidence metric (ranging between 0 and 100%) that a gene is active (roughly speaking, that the gene product is part of an active cellular mechanism). In this manuscript, we demonstrate, on both real and simulated transcriptomics data, that tests for differential expression between sets of genes using are typically more powerful when using gene activity state estimates as opposed to log-transformed gene expression data. Our analysis suggests further exploration of techniques to transform transcriptomics data to meaningful quantities for improved downstream inference.

TRACING CO-REGULATORY NETWORK DYNAMICS IN NOISY, SINGLE-CELL TRANSCRIPTOME TRAJECTORIES.

The availability of gene expression data at the single cell level makes it possible to probe the molecular underpinnings of complex biological processes such as differentiation and oncogenesis. Promising new methods have emerged for reconstructing a progression 'trajectory' from static single-cell transcriptome measurements. However, it remains unclear how to adequately model the appreciable level of noise in these data to elucidate gene regulatory network rewiring. Here, we present a framework called Single Cell Inference of MorphIng Trajectories and their Associated Regulation (SCIMITAR) that infers progressions from static single-cell transcriptomes by employing a continuous parametrization of Gaussian mixtures in high-dimensional curves. SCIMITAR yields rich models from the data that highlight genes with expression and co-expression patterns that are associated with the inferred progression. Further, SCIMITAR extracts regulatory states from the implicated trajectory-evolvingco-expression networks. We benchmark the method on simulated data to show that it yields accurate cell ordering and gene network inferences. Applied to the interpretation of a single-cell human fetal neuron dataset, SCIMITAR finds progression-associated genes in cornerstone neural differentiation pathways missed by standard differential expression tests. Finally, by leveraging the rewiring of gene-gene co-expression relations across the progression, the method reveals the rise and fall of co-regulatory states and trajectory-dependent gene modules. These analyses implicate new transcription factors in neural differentiation including putative co-factors for the multi-functional NFAT pathway.

Conditional estimation of local pooled dispersion parameter in small-sample RNA-Seq data improves differential expression test.

High throughput sequencing technology in transcriptomics studies contribute to the understanding of gene regulation mechanism and its cellular function, but also increases a need for accurate statistical methods to assess quantitative differences between experiments. Many methods have been developed to account for the specifics of count data: non-normality, a dependence of the variance on the mean, and small sample size. Among them, the small number of samples in typical experiments is still a challenge. Here we present a method for differential analysis of count data, using conditional estimation of local pooled dispersion parameters. A comprehensive evaluation of our proposed method in the aspect of differential gene expression analysis using both simulated and real data sets shows that the proposed method is more powerful than other existing methods while controlling the false discovery rates. By introducing conditional estimation of local pooled dispersion parameters, we successfully overcome the limitation of small power and enable a powerful quantitative analysis focused on differential expression test with the small number of samples.

Abstract 3142: Circulating microRNAs can identify cancer-free women at risk for breast cancer

Abstract Introduction: MicroRNAs (miRNAs) are well documented to regulate cancer cell activity by modulating signaling pathways to promote disease onset and progression. Recent evidence correlates miRNAs measured in the serum of breast cancer (BCa) patients to expression in breast tumors. Thus, miRNAs are important biomarkers for stages of tumor development. However, no studies have evaluated the potential of circulating miRNAs (c-miRNAs) as risk biomarkers that predict BCa development years before tumor identification. Objective: Use global miRNA analysis of serum from at-risk but pre-cancerous women who have subsequently been diagnosed with BCa or remain cancer-free to 1) identify expression levels of miRNAs with known BCa association, 2) define cohorts of c-miRNAs differentiating women later developing cancer from cancer-free women with similar clinical risk, and 3) provide insight into c-miRNA mechanisms that promote BCa development. Methods: Subjects were identified from a cohort of 571 women enrolled in the High Risk Breast Program at the UVM Cancer Center; 35 diagnosed with BCa since enrollment. Enrollment criteria included no prior cancer and increased BCa risk due to family history, benign breast disease, prior irradiation for Hodgkin's disease, known pathogenic abnormality, and/or &amp;gt; 20% lifetime modeled BCa risk score. Clinical data and serum are collected at study enrollment and subsequent cancer-free 4-year interval visits. 28 BCa cases were matched to controls for risk factor and age (+/- 5 yrs) at serum sampling (DCIS, Stage IV BCa, BRCA1/2 excluded). RNA was isolated from 68 serum samples representing 52 individual women following a standardized protocol (Farina et al J Cell Biochem 2014) and over 2500 mature human miRNAs profiled on Affymetrix microRNA v4.0 microarrays. Principal component analyses (PCA), hierarchical clustering, differential expression testing, gene ontology, and biological pathway analyses were performed. Results: Our analyses show that women who developed BCa (largely hormone receptor positive, HER2 and lymph node negative) share similar patterns of miRNA expression that differ from women who remain cancer-free. PCA, based on a subset of dynamic miRNAs, segregates serum samples into 3 groups: cancer, cancer-free, and overlapped. Several miRNAs consistently cluster together when cases are contrasted against controls suggesting a functional relationship in providing an environment ideal for BCa development. Pathway analyses of miRNA gene targets reveal enriched PDGF, Wnt, Ras, and MAPK signaling pathways. Conclusion: These data suggest that a cohort of dynamic c-miRNAs discriminates at-risk women who will develop BCa from those who will remain cancer-free. While we consider these findings preliminary until a larger sample set is screened, we conclude that circulating miRNAs will become valuable biomarkers of BCa risk. Citation Format: Nicholas H. Farina, Jon E. Ramsey, Melissa E. Sands, Tiffany J. Rounds, Janet L. Stein, Gary S. Stein, Jane B. Lian, Marie E. Wood. Circulating microRNAs can identify cancer-free women at risk for 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 3142.

Abstract 2662: Interrogation of the landscape of long noncoding RNAs in breast cancer to identify an ER-regulated predictor of tamoxifen resistance

Abstract Background: We previously performed a bioinformatics-based analysis on 7,256 RNA sequencing libraries from tumors, normal tissues, and cell lines to delineate the landscape of long noncoding RNAs (lncRNAs) in the human transcriptome. This analysis identified 58,648 lncRNAs, including over 45,000 novel transcripts (Iyer and Niknafs et al, Nature Genetics, 2015). We now interrogate this lncRNA compendium to identify top candidate estrogen receptor (ER)-associated lncRNAs in breast cancer and characterize their association with disease progression. Methods: To identify and prioritize lncRNAs that were differentially expressed in cancer vs normal tissue, and in ER-positive vs ER-negative disease, we performed differential expression testing on over 1000 RNA sequencing libraries including breast cancer and normal tissue samples from the The Cancer Genome Atlas (TCGA) project. The effect of the top prioritized lncRNA on cancer phenotypes was studied via proliferation, colony formation, invasion and tamoxifen resistance assays in MCF7 and T47D cell lines, and via mouse xenograft studies. The mechanism by which this lncRNA promotes tumor progression was investigated by identifying its top protein interactors and its subdomains responsible for function, and then studying the effects of disrupting function of this lncRNA on cancer phenotypes. Results: Differential expression analyses nominated BRCAT431 as the top overexpressed ER-regulated lncRNA in breast cancer. In vitro experiments demonstrate that shRNA-mediated knockdown of BRCAT431 significantly decreased proliferation, colony formation, and invasion (by &amp;gt;50% in most assays). Tamoxifen resistance was associated with significantly increased BRCAT431 levels in both MCF7 and T47D cells, and knockdown of BRCAT431 reversed tamoxifen resistance. Xenograft studies demonstrate that knockdown of BRCAT431 also significantly decreased xenograft growth and tumor invasion by &amp;gt;50%. RNA pulldown followed by mass spectrometry identified the RNA binding protein hnRNPL as a key protein interacting with BRCAT431. Deletion studies identified a 27 base-pair region of BRCAT431 necessary for its interaction with hnRNPL, and loss of this region abrogated BRCAT431-induced invasion. Finally, guilt-by-association studies demonstrate a strong association between BRCAT431 overexpression and tumor grade, recurrence, and metastases. Conclusion: In this study, we develop the largest reported compendium of breast cancer lncRNAs. We prioritize BRCAT431 as the top lncRNA upregulated in ER-positive breast cancers, and demonstrate that it confers aggressive oncogenic phenotypes in vitro and in vivo. We identify a novel mechanism by which this lncRNA functions through interaction with hnRNPL. Our results suggest that by promoting tamoxifen resistance, this lncRNA increases the clinical risk of recurrence and metastases in breast cancer. Citation Format: Yashar Niknafs, Sumin Han, Teng Ma, James Rae, Arul Chinnaiyan, Felix Feng. Interrogation of the landscape of long noncoding RNAs in breast cancer to identify an ER-regulated predictor of tamoxifen resistance. [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 2662.