exoRBase 2.0: an atlas of mRNA, lncRNA and circRNA in extracellular vesicles from human biofluids

Abstract

Extracellular vesicles (EVs) are small membranous vesicles that contain an abundant cargo of different RNA species with specialized functions and clinical implications. Here, we introduce an updated online database (http://www.exoRBase.org), exoRBase 2.0, which is a repository of EV long RNAs (termed exLRs) derived from RNA-seq data analyses of diverse human body fluids. In exoRBase 2.0, the number of exLRs has increased to 19 643 messenger RNAs (mRNAs), 15 645 long non-coding RNAs (lncRNAs) and 79 084 circular RNAs (circRNAs) obtained from ∼1000 human blood, urine, cerebrospinal fluid (CSF) and bile samples. Importantly, exoRBase 2.0 not only integrates and compares exLR expression profiles but also visualizes the pathway-level functional changes and the heterogeneity of origins of circulating EVs in the context of different physiological and pathological conditions. Our database provides an attractive platform for the identification of novel exLR signatures from human biofluids that will aid in the discovery of new circulating biomarkers to improve disease diagnosis and therapy.

Figure 1.

A schematic overview of the exoRBase 2.0 core content and framework. ExoRBase 2.0 integrates RNA-seq data of EVs from human blood, urine, CSF and bile samples. The exLRs, including mRNAs, lncRNAs and circRNAs, were identified, annotated and quantified according to the Assembling Splice Junctions Analysis (ASJA) and CircRNA Identifier (CIRI2) bioinformatic tools. To interpret EV mRNA expression profiles, pathway enrichment analysis was performed on MSigDB gene sets using the ssGSEA method. The EV-origin approach was applied to predict the proportions of tissue/blood cell sources. ExoRBase 2.0 visualizes and compares exLR expression profiles as well as the enrichment levels of functional pathways and the origins of circulating EVs.

Figure 2.

The enhanced user interface of exoRBase 2.0. (A) The browse pages for mRNA, lncRNA, circRNA, pathway and tissue/cell origin. (B) The search section with four search pages. (C) The search result page with line and heat map charts and a result table. ‘Tumor mean’, ‘Benign mean’ and ‘Healthy mean’ represent the average expression/enrichment values of blood samples from tumor, benign and healthy cohorts, respectively. ‘Urine mean’, ‘CSF mean’ and ‘Bile mean’ indicate average values for urine, CSF and bile samples, respectively. The ‘Different group’ shows significantly divergent groups compared with healthy individuals. (D) The detail section of each target containing ‘Summary’, ‘Profile’ and ‘Comparison’ pages.

Figure 3.

Detailed information and expression profile of FGB. (A) The ‘Summary’ information of FGB gene in the ‘Detail’ section. (B) The expression landscape of FGB across different groups incorporated into exoRBase 2.0 on the ‘Profile’ page. (C, D) Differential expression analysis of FGB between HCC and healthy/benign cohorts on the ‘Comparison’ page. Abbreviations: CSF, cerebrospinal fluid; BRCA, breast cancer; CHD, coronary heart disease; CRC, colorectal cancer; GC, gastric cancer; KIRC, kidney cancer; HCC, hepatocellular carcinoma; ML, malignant lymphoma; OV, ovarian cancer; PAAD, pancreatic adenocarcinoma; SCLC, small cell lung cancer.

Figure 4.

The relative proportions of EV origins. (A) The ‘Origin’ tab at the top of each page and the ‘Tissue/Cell origin’ tab on the home page for accessing to the relative information obtained from EV-origin method. (B) The ‘Relative tissue/cell origin proportions’ page with selection of the ‘Healthy’ dataset. (C) The relative abundances of 23 types of blood cells in the top 20 healthy samples. (D) The relative abundances of 23 types of blood cells in the Healthy001 sample. In the two charts, the fractions of blood cells in each sample are indicated by different colors, and the lengths of bars indicate the enrichment levels of blood cell populations.