Deep sequencing techniques and advanced data analysis methods recently enabled the characterization of thousands of circular RNA isoforms (circRNAs) from a number of tissues and organisms. There is emerging evidence that some circRNAs may have important biological functions or serve as diagnostic biomarkers in disease conditions. In order to analyze circRNA expression in the heart and its changes in different conditions researchers at the Technical University of Munich performed RNA-Seq analysis of ribosome-depleted libraries from rats (neonatal and adult), mice (sham or after transverse aortic constriction, TAC) and humans (failing, non-failing). All samples were sequenced after a treatment with exonuclease RNase R or a mock treatment and >9000 candidate circRNAs were detected for each species. Additionally, the researchers performed separate isolation of nuclear and cytoplasmic RNA and co-immunoprecipitated RNA interacting with endogenous argonaute 2 (Ago2) in primary cardiac myocytes.
They found circRNAs to be significantly enriched in the cytoplasm compared to linear transcripts and to have a similar level of association with Ago2. Notably in all three species they observed dozens of circRNAs arising from the titin (Ttn) gene, which is known to undergo highly complex alternative splicing during heart maturation. Correspondingly they observed extensive differential regulation of Ttn circRNAs between neonatal and adult rat hearts, suggesting that circRNA formation could be involved in the regulation of titin splicing. These researchers expect that their inventory of cardiac circRNAs, as well as the information on their conservation and differential expression will provide an important basis for further studies addressing their function and suitability as biomarkers.
Identification and characterization of cardiac circRNAs
(A) Processing of heart samples from rats (neonatal, NRH, and adult, ARH, n = 3 each), mouse (pressure overload, TAC, n = 2 and sham, n = 3) and humans (failing, HF, non-failing, HN, n = 2 each) for circRNA identification. Deep sequencing reads (shown in gray) spanning the backsplice allow identification of circRNA candidates in RNA-Seq data. (B) Venn diagram of identified backsplices and their conservation. (C) Density histogram of relative enrichment of backsplice-spanning reads from highly expressed rat circRNA candidates (detected with ≥3 reads in either all NRH or all ARH mock treated samples, n = 988) and regular splice junction spanning reads after RNase R digestion. Dashed line indicates the selected threshold of 21.1 (n = 767 at ≥21.1). (D) Relative enrichment of reads from the known nuclear lncRNAs Malat1 and Meg3 and cytoplasm-enriched Gapdh in the respective compartments confirms successful separation of the two fractions. (E) CircRNAs are highly enriched in the cytoplasm of NRCMs compared to linear transcripts. (F) Enrichment of the known microRNA targets H19 and Pten (compared to Gapdh) in the Ago2-RIP, which is much lower in the Hsp90-RIP, confirms successful enrichment of microRNA targets. (G) CircRNAs show a similar extent of enrichment in the Ago2-RIP in NRCMs as do linear transcripts. (All p-values were calculated using a Mann–Whitney U test).