Transcriptome and long noncoding RNA sequencing of three extracellular vesicle subtypes released from the human colon cancer LIM1863 cell line

Abstract

Previously we reported that LIM1863 colorectal cancer (CRC) cells secrete three distinct extracellular vesicle subtypes - two subpopulations of exosomes (apical EpCAM-Exos and basolateral A33-Exos) and shed microvesicles (sMVs) - with distinct protein and miRNA signatures. Here, we extend our omics approach to understand the fundamental role of LIM1863-derived EVs by performing a comprehensive analysis of their mRNAs and long non-coding RNAs (lncRNAs) using RNA-Seq. We show that 2,389 mRNAs, 317 pseudogene transcripts, 1,028 lncRNAs and 206 short non-coding RNAs selectively distributed to (i.e., are enriched in) LIM1863 EVs, relative to the parent cell. An Ensembl/UniProtKB analysis revealed 1,937 mRNAs encode canonical proteins, 348 isoforms (including splice-variant proteins), and 119 'missing proteins' (i.e., annotated in Ensembl but not UniProtKB). Further dissection of our protein/RNA data revealed that 6/151 observed RNA binding proteins have the potential to interact with ~75% of EV-enriched RNAs. Intriguingly, the co-existence of U1 and U2 ribonucleoproteins and their cognate snRNAs in LIM1863 EVs suggests a possible association of CRC EVs with recipient cell splicing events. Our data reveal several potential lncRNA CRC biomarkers and novel splicing/fusion genes that, collectively, will advance our understanding of EV biology in CRC and accelerate the development of EV-based diagnostics and therapeutics.

Figure 1. Gene expression profiling and number of RNA transcripts that preferentially distribute to extracellular vesicle subtypes released from colorectal cancer cell line LIM1863.

(A) TopHat2 and Cufflinks2 identified a total of 50,101 RNA transcripts (>1 FPKM) in LIM1863 cells and released EVs. Using the statistical criteria Log₂FC > 1, p-value < 0.05 and probability >0.7, a total of 3,940 RNA transcripts are enriched in EV subtypes (sMVs, A33-Exos, EpCAM-Exos) compared to CL. These transcripts are further annotated into four RNA classes: protein coding, pseudogene, short noncoding and long noncoding. (B) qRT-PCR validation of 7 RNA transcripts. (C) Number of short noncoding, long noncoding, pseudogene and protein coding RNA transcripts enriched in sMVs, A33-Exos and EpCAM-Exos. (D) Venn diagrams of protein coding, pseudogene, long noncoding and short noncoding RNA transcripts significantly enriched in sMVs, A33-Exos and EpCAM-Exos.

Figure 2. Annotation of EV-enriched protein coding RNA transcripts using UniProt and GO databases.

(A) Canonical transcripts. (B) Isoform transcripts. (C) Missing protein transcripts (i.e., protein annotations seen in Ensembl but not UniProtKB). Ensembl/GENCODE analysis further categorized EV-enriched mRNAs (relative to CL) into protein coding (left), and those predicted to be targets of nonsense-mediated decay (middle) and non-stop decay (right). DAVID Bioinformatics Resource was used to annotate biological process GO for EV-enriched transcripts common to all EV subtypes and mRNAs encoding (D) canonical and (E) isoform proteins selectively distributed into EV subtypes.

Figure 3. Noncoding RNA transcripts and RNA binding proteins enriched in LIM1863 cell-derived EVs.

(A) Distribution of the expression levels of EV-enriched pseudogene transcripts. (B) Venn diagram of EV-enriched short noncoding RNAs. (C) Venn diagram of RNA binding proteins identified in LIM1863 cell-derived EVs.

Figure 4. U1 and U2 complex related proteins and snRNAs identified in this study.

(A) X-ray structure (left) (reproduced under a CC-BY license agreement (https://creativecommons.org/licences/by/4.0/)) of U1 snRNP complex revealed Sm proteins (marked with red solid five-pointed stars), U1 snRNPs (marked with red solid five-pointed stars) from MS/MS data and U1 snRNAs (marked with hollow five-pointed star) from this study may form pre-complexes in EVs. Significant peptide spectra mapping to these proteins and normalized gene expression for U1 snRNA (RNU1-1) the table (right panel). (B) In U2 snRNP complex, and peptide spectra for proteins such as Sm, splicing factors and U2 snRNA (shown in the table in the right panel). *Major U1 snRNA transcript. †Major U2 snRNA transcript, also referred to RNU2 snRNA.