#apaperaday: Efficient Modulation of Exon Skipping via Antisense Circular RNAs

In today’s #apaperaday, Prof. Aartsma-Rus reads and comments on the paper titled: Efficient Modulation of Exon Skipping via Antisense Circular RNAs

Today’s pick is from Ren et al in the journal Research on the use of circular RNAs to induce exon skipping. DOI: 10.34133/research.0045. Exon skipping can be induced with antisense oligonucleotides (ASOs) or antisense genes (U7 snRNA).

Authors argue that for ASOs there are delivery problems (when considering muscle I would tend to agree). U7 snRNA antisense genes use the original U7 snRNA where they replace the original antisense sequence with an exon targeting sequence. U7 is delivered with gene therapy (AAV). Authors argue that the U7 snRNA system may produce by-products with unknown risks. From research I have been exposed to so far in model systems, there is no indication yet for risks, but maybe there is a risk. In either case authors offer an alternative: circular RNAs.

These circular RNAs are expressed from genes and are up to 50 nt long and target an exon. Authors started with a minigene system to target exon 51 in the dystrophin gene. The minigene was transfected into HEK293 cells along with the circular RNA plasmid.

In this very artificial system (minigene for exon 50-52 lacks most of the introns for intron 51 and 52 and HEK293 cells normally do not produce dystrophin) they observed exon 51 skipping as intended. Authors quantify their exon skipping levels found in RT-PCR. Won’t mention the numbers because they overestimate. Authors indicate that skipping levels were higher than with U7 snRNA – that makes sense as the plasmid for the circular RNA probably was smaller so delivery may have been easier…

Authors then went for slightly less artificial and used lentiviruses to deliver circular RNAs targeting exons in genes expressed in HEK293 cells. In 9/13 targeted exons, they observed exon skipping.

Then authors delivered the exon 51 circular RNAs with AAV into a mouse model with a deletion of exon 50 (note that the minigene contained mouse exon 50-52). In the mouse they observed exon skipping and dystrophin restoration mainly in heart.

Some notes about the pictures: the Western blot is not done properly. Only a single reference wild type control is used on each blot, rather than a linear range. This makes quantification unreliable. It is unclear why efficiency was so much higher in heart, and so poor in muscles.

Also a puzzling finding in the supplementary data. Authors generated a circular RNA for eteplirsen (exon 51 skipping PMO approved by FDA) and tested that in their system and they see good exon skipping with the ASO and the circular RNA equivalent. In a mouse system???

This is surprising as there are several mismatches between the target of eteplirsen (human exon 51 targeting) and mouse exon 51. So you would not expect this to work at all. Very unexpected finding. Note that the eteplirsen is a transcribed linear RNA rather than the PMO.

Authors conclude that they provided proof of concept that circular RNAs can induce exon skipping. I am left confused because they should not have induced exon 51 skipping with the eteplirsen sequence targeting mouse exon 51…unless they are very unspecific.

For now I conclude that much more work is needed and that authors need to elaborate on what exactly they did (e.g. mousify the eteplirsen sequence?). I am not yet convinced about this approach.