Credit: Lara Crow/NPG

Post-transcriptional editing of RNA is a widespread, conserved mechanism by which genetic diversity can be introduced without any changes to the DNA, but its role in cancer has remained unclear despite the observation that the number of editing-induced RNA mutations is comparable to the number of somatic mutations in coding regions of DNA. Now, three studies demonstrate that adenosine-to-inosine (A-to-I) RNA editing (which is the most common form) introduces transcriptome diversity in tumours from a range of cancer types; diversity that is both functionally important and clinically relevant.

transcriptome diversity ... that is both functionally important and clinically relevant

Using RNA sequencing (RNA-seq) data from The Cancer Genome Atlas (TCGA), Han et al. and Paz-Yaacov et al. systematically analysed genome- and transcriptome-wide profiles of A-to-I RNA editing, respectively. They showed that levels of editing in tumour tissue differed significantly from those in matched normal tissue from the same individuals, with examples of both under-editing (for example, in chromophobe kidney cancer) and over-editing (for example, in breast tumours; a finding replicated by Fumagalli et al.).

Consistent with the known role of ADAR1 (adenosine deaminase acting on RNA), all three studies report a significant correlation between A-to-I RNA editing activity and ADAR1 expression in tumour tissues. Furthermore, Fumagalli et al. analysed TCGA RNA-seq data from patient samples and found that levels of both ADAR1 and A-to-I RNA editing in a range of cancer types correlate significantly with ADAR1 copy number and with expression of STAT1, which they used as a marker of type I interferon signalling. The significance of the STAT1 correlation increased for most cancer types studied when adjusting for ADAR1 copy number, suggesting that both factors independently regulate A-to-I RNA editing in these cancers. These molecular correlates provided insight into potential regulatory mechanisms, but what are the clinical correlates of A-to-I RNA editing?

Paz-Yaacov et al. reported a significant correlation between A-to-I RNA editing levels and survival in patients with certain cancers; similarly, Han et al. identified a proportion of A-to-I RNA editing sites that correlated significantly with established clinical parameters (disease stage, patient survival or tumour subtype). However, the potential contribution of interferon signalling to these clinical correlations was not reported in these studies.

Han et al. then focused on sites that showed nonsynonymous RNA editing, in which editing of the transcript could be translated into changes in the resulting protein sequence. For three such editing events that were found in two or more cancer types, functional assays revealed a significant impact of A-to-I RNA editing on viability in two different cell lines. This demonstration of in vitro 'driver-like' functional effects is consistent with the in silico predictions of Paz-Yaacov et al., which identified nonsynonymous A-to-I RNA editing events that are likely to translate into altered protein function.

Further supporting the functional importance of A-to-I RNA editing, Fumagalli et al. showed that reducing editing levels in several breast cancer cell lines (through ADAR1 knockdown) significantly decreased proliferation and increased apoptosis. Notably, Han et al. went on to identify functional effects of clinical relevance, demonstrating that certain nonsynonymous editing events influenced the sensitivity of a mouse leukaemia cell line to the cytotoxic effects of a range of compounds, suggesting a role for A-to-I RNA editing in drug sensitivity.

Together, these three studies provide evidence that A-to-I RNA editing is a functionally important and clinically relevant source of transcriptome diversity in cancer, and they highlight the need to study its potential roles as a pathogenic mechanism, a biomarker of disease and a therapeutic target.