The global landscape of intron retentions in lung adenocarcinoma

Abstract

Background: The transcriptome complexity in an organism can be achieved by alternative splicing of precursor messenger RNAs. It has been revealed that alternations in mRNA splicing play an important role in a number of diseases including human cancers.

Methods: In this study, we exploited whole transcriptome sequencing data from five lung adenocarcinoma tissues and their matched normal tissues to interrogate intron retention, a less studied alternative splicing form which has profound structural and functional consequence by modifying open reading frame or inserting premature stop codons.

Results: Abundant intron retention events were found in both tumor and normal tissues, and 2,340 and 1,422 genes only contain tumor-specific retentions and normal-specific retentions, respectively. Combined with gene expression analysis, we showed that genes with tumor-specific retentions tend to be over-expressed in tumors, and the abundance of intron retention within genes is negatively related with gene expression, indicating the action of nonsense mediated decay. Further functional analysis demonstrated that genes with tumor-specific retentions include known lung cancer driver genes and are found enriched in pathways important in carcinogenesis.

Conclusions: We hypothesize that intron retentions and consequent nonsense mediated decay may collectively counteract the over-expression of genes promoting cancer development. Identification of genes with tumor-specific retentions may also help develop targeted therapies.

Figure 1

Genomic features of retained introns. A. The size distribution of introns in different classes. The bin size is 500 bp, and introns larger than 10,000 bp were included in one bin. The distribution showed tumor-specific retained introns are larger than other introns (see text). TSR: tumor-specific retained introns; NSR: normal-specific retained introns; Other: non-retained introns. B. The percentage of retained introns in different orders. Forward order is showed above the X-axis, and reverse order is showed under the X-axis. The histogram for introns after ten was truncated to fit the plot. Tumor and normal samples are labeled by blue and darkred.

Figure 2

Overrepresentation of genes with tumor-specific retention (TSR) in differentially expressed genes (DEGs) between tumor and normal samples. A: Venn diagram of DEG and genes with TSR. B: pie chart of the expression pattern in tumor and normal tissues for genes with TSR. A majority of DEGs are up-regulated in tumors. Non-DEG: genes without differentially expression in tumor and normal tissues; up-DEG: differentially expressed genes up-regulated in tumor tissues; down-DEG: differentially expressed genes up-regulated in normal tissues.