Comprehensive Analysis of Long Non-Coding RNAs in Ovarian Cancer Reveals Global Patterns and Targeted DNA Amplification

Rozita Akrami,Anders Jacobsen,Chris Sander,Nikolaus Schultz,Jessica Hoell,Erik Larsson,Aedín C Culhane

doi:10.1371/journal.pone.0080306

Abstract

Long non-coding RNAs (lncRNAs) are emerging as potent regulators of cell physiology, and recent studies highlight their role in tumor development. However, while established protein-coding oncogenes and tumor suppressors often display striking patterns of focal DNA copy-number alteration in tumors, similar evidence is largely lacking for lncRNAs. Here, we report on a genomic analysis of GENCODE lncRNAs in high-grade serous ovarian adenocarcinoma, based on The Cancer Genome Atlas (TCGA) molecular profiles. Using genomic copy-number data and deep coverage transcriptome sequencing, we derived dual copy-number and expression data for 10,419 lncRNAs across 407 primary tumors. We describe global correlations between lncRNA copy-number and expression, and associate established expression subtypes with distinct lncRNA signatures. By examining regions of focal copy-number change that lack protein-coding targets, we identified an intergenic lncRNA on chromosome 1, OVAL, that shows narrow focal genomic amplification in a subset of tumors. While weakly expressed in most tumors, focal amplification coincided with strong OVAL transcriptional activation. Screening of 16 other cancer types revealed similar patterns in serous endometrial carcinomas. This shows that intergenic lncRNAs can be specifically targeted by somatic copy-number amplification, suggestive of functional involvement in tumor initiation or progression. Our analysis provides testable hypotheses and paves the way for further study of lncRNAs based on TCGA and other large-scale cancer genomics datasets.

Highlights

Recent transcriptomic studies in mammals have revealed an abundance of long non-coding RNAs that lie interspersed with coding genes in complex ways [1,2,3]
HOTAIR expression is high in metastatic breast cancer tumors, and its inhibition blocks metastasis in rodent models [8], MALAT1 expression correlates with metastases and survival in lung cancer [9], and polyA+ transcriptome sequencing (RNA-seq) recently identified PCAT-1 as a growth-promoting long non-coding RNAs (lncRNAs) in prostate cancer [10]
We found that the GENCODE lncRNA subset [16], which encompasses 10,419 manually annotated lncRNA genes, showed a high degree of polyadenylation as determined by normal tissue RNA sequencing (RNA-seq) data (Figure 1B)

Summary

Introduction

Recent transcriptomic studies in mammals have revealed an abundance of long non-coding RNAs (lncRNAs) that lie interspersed with coding genes in complex ways [1,2,3]. HOTAIR expression is high in metastatic breast cancer tumors, and its inhibition blocks metastasis in rodent models [8], MALAT1 expression correlates with metastases and survival in lung cancer [9], and polyA+ transcriptome sequencing (RNA-seq) recently identified PCAT-1 as a growth-promoting lncRNA in prostate cancer [10]. In large cancer genomics datasets, such as those produced by The Cancer Genome Atlas (TCGA) consortium, important cancer genes reveal themselves through striking patterns of recurrent DNA-level alteration, including focal copy-number amplification and deletion [12,13]. While lncRNAs and coding genes should in principle be susceptible to activation or deactivation through similar mechanisms, there is far little evidence that lncRNAs are targeted by copy-number alterations in cancer independently of proximal coding genes (recently reviewed in 6,14)

Methods

Results

Conclusion